Citation
Psychophysiology and memory : the encoding and retrieval of affective text

Material Information

Title:
Psychophysiology and memory : the encoding and retrieval of affective text
Creator:
Spence, Ellen Louise
Publication Date:
Language:
English
Physical Description:
vi, 172 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Acceleration ( jstor )
Cognitive psychology ( jstor )
Emotional expression ( jstor )
Heart rate ( jstor )
Images ( jstor )
Memory ( jstor )
Memory retrieval ( jstor )
Psychophysiology ( jstor )
Reading tables ( jstor )
Verb valency ( jstor )
Department of Clinical and Health Psychology thesis Ph.D ( mesh )
Dissertations, Academic -- College of Health Related Professions -- Department of Clinical and Health Psychology -- UF ( mesh )
Emotions -- physiology ( mesh )
Memory -- physiology ( mesh )
Psychophysiology ( mesh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1992.
Bibliography:
Includes bibliographical references (leaves 166-170).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Ellen Louise Spence.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Ellen Louise Spence. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
27028999 ( OCLC )
027278340 ( ALEPH )

Downloads

This item has the following downloads:


Full Text




















PSYCHOPHYSIOLOGY AND MEMORY:
THE ENCODING AND RETRIEVAL OF AFFECTIVE TEXT














By

ELLEN LOUISE SPENCE















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1992
















ACKNOWLEDGEMENTS


The present research project was the culmination of several pilot studies, and involved the collaborative efforts of many people. I am especially grateful to Peter Lang, who generously supported this effort, both by making available the considerable resources of his psychophysiological laboratory and technical staff, and through his financial support of my graduate work at Florida. I would also like to express my deep appreciation to the members of my dissertation committee, Russell Bauer, Keith Berg, Dawn Bowers, Ira Fischler, and Edward Valenstein. They not only made valuable contributions to my dissertation research, but also provided continued support, encouragement, and intellectual guidance throughout my graduate training.

I would also like to thank Bruce Cuthbert, who provided training and support in the laboratory in too many ways to enumerate; and to Margaret Bradley for her literary, technical, and intellectual contributions to this project and the earlier pilot studies. I also benefitted from many discussions and sharing of ideas with my fellow graduate students, Scott Vrana and Mark Greenwald.

Finally, I would like to thank my husband, Ed Cook, for his untiring support and assistance in literally every aspect of this project.






ii










TABLE OF CONTENTS

ACKNOWLEDGEMENTS . . . . . . .. . . . . . ii

ABSTRACT . . . . . . . . . . . . . . . v

INTRODUCTION . . . . . . . . . . . 1
Theoretical Background .................. . 1
The Network Model ...... ............ . . 1
Imagery and Response Activation . . . . . . . 2
Bio-Information Theory ................. 3
Emotion and Memory ........... .... . . 4
The Research Problem. .. ............... . . 4
Encoding Hypotheses ................ .. .. .. 5
Retrieval Hypotheses. .................. ... 10

METHODS ................... . . . . 14
Subjects . . . . . . . . . . . . . . 14
Materials ................... . . ... 14
Texts .. .. . . . . . . . . ... .. . 14
Test Lists ....................... . . . . 17
Apparatus . . . . . . . . . . .. . 21
Procedures .................. . . .... 22
Data Reduction .................. ........ 25
Encoding Measures . . . . . . . . . . . 25
Retrieval Measures . . . . . . . . . . 26
Data Analysis . . . . . . . . . ..... . . 27
Imagery Ability . . . . . . . . . . .. . 27
Affective Classification of Texts .. ............ 27
Encoding Measures .... . . . . ..... .... 28
Retrieval Measures ................... .. 29

RESULTS. ................ . ...... .. 31
Encoding Measures . . . . . . . . . .. . 31
Read Time . . . . . . . . . .. . . 31
Heart Rate .. . . ......... . . .... . 33
Startle Reflex ........... . .. .. . . 35
Retrieval Measures ... . . . . . .... .. . 38
Free Recall . . . . . . . . . . . . 38
Episodic Recognition ... . . . . . .. . . . 38

DISCUSSION . . . . . . . . . . . . . 46
Encoding Measures ... . . . . . . .. . . 47
Retrieval Measures . . . . . . . . .. . . 51
Overall Conclusions ........... . . . . . 54

APPENDIX A QUESTIONNAIRES, INSTRUCTIONS AND DEBRIEFING STATEMENT 55

APPENDIX B TEXT MATERIALS AND TEST LIST WORDS . . . . . 65

APPENDIX C IMAGERY ABILITY SCORES FOR INSTRUCTION GROUPS ... . 102



iii










APPENDIX D TABLES OF MEANS FOR DEPENDENT MEASURES ....... 103 APPENDIX E AFFECTIVE RATINGS FOR PARAGRAPHS . . . . . 156

APPENDIX F SELF-ASSESSMENT MANIKIN (SAM) .... ........ 161

APPENDIX G SUMMARY OF RESULTS FOR PILOT PRIMING STUDIES .. . 164 REFERENCES . . . . . . . . . . . ... 166

BIOGRAPHICAL SKETCH . . . . . . ... . .... 171



















































iv
















Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


PSYCHOPHYSIOLOGY AND MEMORY:
THE ENCODING AND RETRIEVAL OF AFFECTIVE TEXT By

Ellen Louise Spence

May, 1992


Chairman: Peter J. Lang, Ph.D.
Major Department: Clinical and Health Psychology

The conceptual framework for this research was derived from Lang's bio-informational theory of emotion which states that affective memory representations contain information about physiological mobilization and action. Imaginal processing of emotional text content was posited to further increase the representation of physiological activation in memory. The resulting affective information structure contains more information than representations of neutral content, and supports faster and more probable retrieval from memory.

Subjects were trained in a verbal or imaginal encoding strategy. They then read brief narratives that depicted pleasant, neutral or unpleasant experiences. Verbal instructions emphasized attention to the exact words used in the text, while imagery instructions focussed on the events described as a vivid personal experience. Read time was employed as a measure of the time required to construct the text schema, while



v









heart rate and the startle reflex verified the presence of emotional response during encoding of the paragraphs. Affective content and imaginal processing were predicted to increase encoding time, to produce physiological activation during construction of the text schema, and to facilitate memory retrieval.

The results showed that affectively unpleasant content was read

more slowly than pleasant material. Unpleasant paragraphs also prompted greater heart rate decrease, and larger and faster probe startle responses, compared to pleasant contents. These content effects were all more pronounced for the later sentences in the paragraph, after the affective content was established. The read time and heart rate valence effects tended to be more reliable for poor imagers and for subjects under verbal instructions. The startle/valence effect was more reliable under imagery instructions.

Retrieval measures included free recall and episodic recognition reaction time. Emotional content facilitated free recall performance relative to neutral content. Affective content, good imagery ability, and imagery instructions facilitated the speed and accuracy of recognition memory for study word targets. Episodic priming occurred for neutral texts, but was not reliable for affective texts.

These results support the hypotheses that both the encoding and retrieval of narrative text vary significantly with affective versus neutral content. In addition, emotional content interacted with encoding strategy (imaginal or verbal) to further modulate these processes.





vi
















INTRODUCTION


The purpose of this research is to examine how narrative text is processed into story-images, and how the information structure of such images differs from representations of the specifically verbal content of text. The experimental procedure employed here is designed to test a set of propositions that were conceptually derived from the bioinformational theory of imagery (Lang, 1977, 1979, 1984, 1985), and from associated supporting research (e.g., Cook, Melamed, Cuthbert, McNeil, & Lang, 1988; Lang, Levin, Miller, Kozak, 1983; Miller, Levin, Kozak, Cook, & McLean, & Lang, 1987; Vrana & Lang, 1990).


Theoretical Background

The Network Model

In general, cognitive psychologists hold that meaningful text is parsed by the brain into concept units (e.g., Anderson, 1976; Kintsch, 1974; Schank, 1970). Following Anderson and Bower (1973), these concepts are presumed to be inter-linked by association (perhaps, with a propositional sub-organization), forming an information network that represents an individual's knowledge of the text content. Subsequently, if a newly presented stimulus matches a concept in the network (e.g., a word from the previously seen text is presented on a CRT screen) retrieval is initiated. This associative retrieval process eventually activates the entire network structure, and is reported by the subject



1









2

as his or her verbal memory for the text. Several parameters determine network activation and the resulting memory retrieval. These include the number and type of network concepts represented in the activating percept, and the associative strength between concepts in memory. Measures of recognition memory and free-recall of text content have been used to assess their influences (e.g., Anderson & Bower, 1973; Carroll & Kirsner, 1982; Galambos & Rips, 1982; Kintsch & van Dijk, 1978; McKoon & Ratcliff, 1980; Ratcliff & McKoon, 1978).


Imagery and Response Activation

Variants of the above model have been widely employed in the

semantic analysis of text processing. In 1973, Pylyshyn suggested that a similar analysis might also account for imaginal memory. In his view, images are internally constructed propositional descriptions of objects. Kieras (1978) later elaborated this idea, suggesting that imagery differed from other information structures, specifically, in the amount of modality-specific information that was in the network. Thus, vivid imagers, more than poor imagers, would give elaborate sensory descriptions of previously seen percepts, or readily detect errors of detail in a picture memory task. While this seems plausible, the experimental evidence for this view is not compelling (see Strosahl & Ascough, 1981). Furthermore, this sensory information emphasis neglects a vast psychophysiological literature, which has reliably found imagery effects, not in afference, but in the efferent system (e.g., Jacobson, 1930; Lusebrink & McGuigan, 1989; McGuigan & Schoonover, 1973; Miller et al., 1987). That is, when subjects imagine, remember, or re-experience









3
past events, this mental imagery is reliably accompanied by somatic and visceral changes which are similar in pattern to those occurring in the referent perceptual contexts. An analysis of imagery based on the activation of efferent information was developed in a neurophysiological framework by Hebb (1968), and elaborated in cognitive terms by Lang's (1979) bio-informational theory of emotional imagery.


Bio-Information Theory

The bio-informational model of memory imagery assumes the

essentials of an associative network model. It differs from other views in that response information is included as a concept category. That is to say, an image network would include sensory concepts (e.g., color, object representations, spatial locations) and perhaps also meaning concepts (e.g., logical relationships, categorical or other semantic information), as in other network theories. In this view, however, images would also include response concepts (e.g., overt behavior patterns and covert physiological events). This latter information would be in the same associative net, and its activation would have similar consequences for overall network firing. Furthermore, there is a natural neural connection between response concepts in memory (which might be activated, for example, by text content) and the actual efferent events to which they refer. Thus, while imagery is in principle a state in which overt behavior is inhibited, "efferent leakage" --a sub-overt activation of visceral organs and somatic muscle intrinsic to the represented behavior--is a normal part of the image process. In point of fact, Lang (1979) argues that in the intact organism image processing is defined by its efferent consequences.












Emotion and Memory

The view taken here (see Lang, 1985; Lang, Bradley & Cuthbert, 1990) is that emotions are action dispositions, which are either appetitive--involving approach and consummatory-related behaviors, or they are aversive--involving defensive, avoidant, and protective behavioral strategies. Both the appetitive and aversive emotional states vary in intensity or arousal, which is associated with increased efferent activity and a greater overt action potential.

The responses in emotion include expressive language, overt

behavioral acts, and a supporting somatic and visceral physiology (Lang, 1964; Lang, 1978). Therefore, the network of an emotional memory image is held to include, in addition to response concepts relevant to the remembered perceptual context (e.g., eye movements, postural adjustments), response representations of remembered emotional reactions. Furthermore, while overt manifestations may be inhibited by the context of recall, the memory image should generate an affective physiological pattern, aversive or appetitive, moderate or intense, depending on the emotional context being recalled (Lang et al., 1983).


The Research Problem

The present experiment examines the encoding and subsequent memory retrieval of affective text. Subjects initially completed self report measures of imagery ability and were then told they would be asked to learn a series of narrative paragraphs. Half of the subjects received imagery instructions. They were told to imagine the content of the paragraphs as a vivid personal experience, clearly imaging the events










5

and their specific action. The other half received verbal instructions: they were told to focus on the exact words used in the text. Instructions to both groups were presented as optimal learning strategies that would facilitate performance on the word recognition task following each text.

Subjects read the study paragraphs at their own pace, sentence by sentence, on a CRT screen. The time taken to read each sentence was recorded, and heart rate was monitored continuously. For a subset of subjects, an acoustic startle stimulus occurred during some of the texts. After each paragraph, a test list of words and control symbols was presented. Subjects held reaction time buttons, one in each hand; they indicated with a right hand button press if a target word had appeared in the preceding text, and with a left hand button press if the word was not in the text. The words in the test lists were all preceded by a very brief (300 ms) priming word or symbol. Some of the priming words were also from the preceding text, and were expected to facilitate recognition speed for the target words. After this part of the experiment was completed, subjects were asked to write down the gist of as many of the paragraphs as they could recall. Subjects provided affective ratings of the texts at the end of the experimental session.

Predictions, derived from bio-informational theory were tested for both encoding and retrieval phases of the study.


Encoding Hypotheses

During the encoding process, narrative text activates representations in memory which constitute an image of the events described. As









6

suggested earlier, cognitive representation of an emotion includes information on visceral and motor activation, inherent in emotional experience, but not included in neutral, non-action representations. Thus, during encoding, more units of memorial information are contacted by affective than by neutral text. This additional information and the associated efferent expression may modulate encoding time. This action content from memory will be evident in the pattern of physiological activation instigated in the subject. Furthermore, network processing occurs over time as significant narrative elements are encoded and a story "schema" is constructed (Kintsch & van Dijk, 1978). Thus, the above effects on physiological response and processing time will more probably occur late in an encoding task.

Attention to a given aspect of a stimulus array determines the content of memorial representations and affects memory performance (Jacoby, 1982). It is suggested here that encoding and retrieval of affective text are also modulated by the subject's processing style, i.e., whether she focusses on the specific words in the text, or focusses on the developing narrative imagery. This variable is examined here in two ways: (1) by instructions--to either image or attend to the specific words of the texts; and (2) by examining individual differences in the normal prevalence and vividness of imagery, i.e., by comparing "good" and "poor" imagers.

The present research examines the following general hypotheses:

1. The action content in emotional memory is evident in the pattern of physiological activation during the encoding of affective text.









7

2. Affective and neutral cognitive representations differ in type and quantity of associated information, which differentially affects the time it takes to encode these contents.

3. A story schema develops over time during text processing, and thus emotional activation is more pronounced late in the encoding process.

4. Instructions to image and good imagery ability increase the probability of physiological activation at encoding.

A long tradition of psychophysiological research on attention and affect has employed heart rate as a dependent variable. Many investigators have found that heart rate deceleration accompanies an attentional set (Graham and Clifton, 1966; Obrist, 1981; Lang, bhman, & Simons, 1978). Lacey and Lacey (1974) described heart rate deceleration as part of an "information intake" orientation in which the organism shows a generalized inhibition of motor activity accompanied by an attentional focus on the stimulus of interest. Cardiac acceleration, on the other hand, typically accompanies tasks in which the subject focusses on internal mental processes and inhibits environmental intake, such as performance of mental arithmetic (Lacey & Lacey, 1974) and generation of mental images (Miller et al., 1987).

Studies of emotion have revealed divergent patterns of heart rate response to affective content, depending on whether the subject is engaged in information intake or internal processing. For example, in perceptual situations such as slide-viewing, heart rate is generally found to decrease more when normal subjects are exposed to affectively unpleasant pictures compared to neutral or positive stimuli (Klorman,











Weisberg, & Wiesenfeld, 1977). This may reflect greater attention to aversive stimuli. On the other hand, heart rate increases have been observed during imagery of aversive events, relative to neutral mentation (Lang, Kozak, Miller, Levin, & McLean, 1980), perhaps indexing a more intensive internal processing. Perception and imagery are not, of course, wholly independent. In order to be processed, any stimulus, text or picture, must initially be an object of orientation and attention. However, if the material subsequently contacts relevant representations in memory, the cognitive work of image construction may begin. Presuming that heart rate is a reliable covariate of these processes, we might expect (1) that the response to a stimulus would change over time, and (2) that subjects who quickly develop images, because of the particular pertinence of the stimulus or their general disposition to image, are more likely to show the heart rate pattern of imagery than that of attention. An example of the latter phenomenon has been observed in phobic subjects (Hamm, Globisch, Cuthbert, & Vaitl, 1991; Klorman, Wiesenfeld, & Austin, 1975), who readily access a vivid fear image, and show a strong accelerative response (rather than the usual deceleration in aversion) to slides of their own primary fear objects.

In the present study heart rate response should vary with both encoding strategy and activation of an affective "action set." The above considerations suggest the following working hypotheses:

1. It is predicted that heart rate will decrease during text

encoding, consistent with the experimental task of information intake.









9
2. Greater heart rate deceleration is expected for subjects receiving verbal encoding instructions and subjects who report poor imagery ability. As these subjects are primarily "attending" and are less likely to be engaging in imaginal processing, greater heart rate deceleration is predicted for unpleasant text content.

3. Imagery instructions and self-reported vivid imagery are predicted to result in heart rate acceleration during the encoding process, reflecting the internal cognitive effort of image generation. The pattern of heart rate change associated with emotional imagery (greater heart rate increase for more aversive content) is predicted for these subjects.

4. Consistent with text processing theory, heart rate change

associated with the development of affective schemata is more probable late in the encoding process.

The startle reflex is a generalized flexor response elicited by

sudden, intense stimuli. The response complex entails a large number of protective elements, such as eyeblink, forward and downward head movements, raising of the shoulders, contraction of the abdomen, and flexion of the limbs (Davis, 1984). Latency and magnitude of the eyeblink component of this reflex are modulated by both the attentional and the emotional state of the organism (Anthony, 1985; Graham, 1979; Lang, Bradley, & Cuthbert, 1990). Inhibition of the reflex response to an acoustic startle probe occurs during visual attention to pleasant, interesting stimuli (Simons & Zelson, 1985). On the other hand, startle probe responses are exaggerated in aversive emotional states (Berg & Davis, 1984; Bradley, Cuthbert, & Lang, 1990; Cook, Hawk, Davis, &









10

Stevenson, 1991; Cook, Davis, Hawk, Spence, & Gautier, in press; Vrana & Lang, 1990; Vrana, Spence, & Lang, 1988; see Lang, Bradley, Cuthbert, & Patrick, in press, for a review). In the present study, startle probes are presented unpredictably while subjects read the study paragraphs. The blink component of the startle response is recorded. It is expected that variations in its latency and magnitude will track the development of the affective schema during the encoding process. Specifically, it is predicted that:

1. Potentiation of the startle reflex (greater magnitude and faster latency) is expected as a result of activation of an aversive emotional schema; relative inhibition is expected for pleasant schemata.

2. These effects will be enhanced late in the encoding process, after the affective schema is well developed.

3. Imagery instructions and good imagery ability will result in greater affective modulation of the startle response, relative to verbal instructions and self-reported poor imagery ability.


Retrieval Hypotheses

Bio-information theory states that emotional images have an extra set of concepts in the network--namely those involving response information. Network theory states that greater elaboration of representations in memory enhances retrieval (Anderson, 1976). This mechanism may underlie the corollary finding that moderately arousing affective events are better and more accurately remembered than less arousing events (e.g., Bradley, Greenwald, Petry, & Lang, 1991; Heuer & Reisberg, 1990). These findings are also consistent with the Yerkes-











Dodson Law (1908) in which performance on a variety of learning discrimination tasks is positively related to increases in physiological arousal, within the moderate range.

Thus, the present research tests the following general hypothesis regarding the effects of emotional content on measures of memory retrieval:

1. More affectively arousing text, (i.e., elaborate information structures in memory) will result in enhanced performance for both recognition and free recall measures of retrieval.

Arousal is more probable when subjects employ an imaginal encoding strategy, given that imaginal processing emphasizes the inclusion of sensory and visceral/motor activation codes in the resulting memory representation. Conversely, instructions to attend exclusively on the surface characteristics of the words (verbal instructions condition), would be expected to reduce the inclusion of such information in memory. The instructions in this experiment may also be construed as an affective levels of processing manipulation (Craik & Lockart, 1972) in which imaginal encoding instructions involved "deep" processing, and verbal instructions prompted "surface" processing of the paragraphs. These considerations motivate the following hypothesis:

2. Imaginal encoding strategies, resulting from instructions or subject characteristics, emphasize the representation of visceral/motor activation codes in memory for affective text relative to verbal encoding strategies, thereby enhancing memory performance.

3. The above effects are specific to the reactivation of the

study paragraph's information network by the related test word, and do









12

not represent a simple persistence of the affective response in memory. That is, the above results will not be found for nonstudy words.

Priming can be defined as facilitation of the response to a test item by an immediately preceding cue (Ratcliff & McKoon, 1978). This paradigm has proved a useful tool in the development of theory on how knowledge is represented in memory, such as the network theory developed by Anderson (1983). According to network theory, when a concept is processed cognitively it is activated in memory. This activation automatically spreads through the information network to other associated concepts. Episodic priming results from temporary logical relationships established among concepts when subjects read a particular text, and has been employed by researchers to demonstrate that memory for text is represented in logical form as meaning propositions, and that activation of this associative network follows the logical structure of these associations. Priming effects are typically determined by comparing the reaction time for memory recognition of a target stimulus preceded by a highly associated word, or "prime", relative to that for a neutral or control condition.

In Lang's original presentation of bio-informational theory (1977; 1979), he followed Pylyshyn (1973) in suggesting that there is a general, propositional sub-organization in imaginal/affective information structures. To examine this assumption, an episodic recognition priming paradigm was adapted from the work of Ratcliff and McKoon (1978, 1981; McKoon & Ratcliff, 1980) for the present study. Their research explored the dynamics of spreading activation in logical information structures in neutral narrative text. They observed that









13

when a word in a test list, taken from a study paragraph, was primed (i.e., immediately preceded) by another word from the same paragraph, recognition time was faster than if the prime was an unrelated word or symbol. The effect was greater, the closer the logical relationship between prime word and target word. Earlier pilot studies (Bradley, Spence, & Lang, 1992; Appendix G) in which episodic priming was compared for affectively unpleasant and neutral texts suggested that emotional content altered the priming phenomenon reported by Ratcliff and McKoon. The basic priming phenomenon is tested here for affective and neutral text, i.e., do related text words facilitate recognition of text study words. The following specific questions are addressed:

1. Does episodic priming occur with affective narratives, as has been shown for neutral text, consistent with a propositional memorial sub-structure?

2. Are there differences in priming for affectively pleasant and affectively unpleasant text?

3. Is priming enhanced by imaginal processing instructions or by the subjects' disposition to process imaginally?
















METHODS


Subjects

Fifty-seven female undergraduates at the University of Florida

served as subjects in this experiment. Subjects were recruited through introductory psychology classes, at summer registration, or through signs posted at sorority houses on campus. Subjects either received class credit or were paid $10 for participation in the study.

The research sample included a subset of 25 subjects to whom acoustic startle probes were presented during text encoding. Probe administration had almost no effect on the other dependent variables Thus, for all measures except the startle probe, all 57 subjects are included in the analyses, with the few sample differences noted. The analysis of the startle probe includes, of course, only the subset of 25 subjects.


Materials

Texts

The study materials used in this research consisted of two sets of 36 6-line paragraphs. Each set contained 18 unpleasant, 9 neutral, and

9 pleasant exemplars. The paragraphs were classified into affective categories based on a validation study in which college students read and imaged a larger set of 108 texts (Spence et al., 1987). The validation data were psychophysiological measures of heart rate, skin


14









15

conductance, and electromyographic (EMG) activity from the corrugator and zygomatic facial muscles, recorded during a 30-second imagery period. Subjects also provided ratings of the paragraphs, using the automated SAM system (Lang, 1980; Hodes, Cook, & Lang, 1985) after the imagery task. Texts were selected for this research on the basis of:

(1) extreme pleasant or unpleasant valence ratings, and (2) the presence of the appropriate psychophysiological activity. For unpleasant texts, the criterion physiological responses during imagery included increased heart rate and skin conductance level (associated with arousal), as well as high corrugator (frown) muscle tension (associated with negative emotional valence). The selected affectively pleasant texts showed high zygomatic (smile) muscle activity accompanied by low corrugator muscle activity, in addition to SAM ratings of positive emotional valence and high arousal. Neutral texts were selected on the basis of neutral valence and low arousal SAM ratings, and low levels of psychophysiological activity on all four measures. The number of exemplars contributing to each affective category (18 unpleasant, 9 neutral, and

9 pleasant) reflected the primary focus on the comparison between effects on cognitive processes of unpleasant versus neutral/pleasant affect in prior pilot research. A summary of the validation study including the selected texts and the corresponding data can be found in Appendix E.

Paragraph topic and contents of the test list were matched between affective categories. Every unpleasant text in Set 1 had a "matched" text that was neutral or pleasant in Set 2; likewise, every neutral/ pleasant text in Set 1 had a "matched" unpleasant text in Set 2.









16

"Matched" means that the texts were similar in topic and sentence construction, but varied in affective content. For example, a text describing a beach picnic was matched with a similar text describing a shark attack at the beach. In all cases, each pair of matched texts was followed by the same test list. To illustrate with the previous example, the study word "fins" appeared in the test list following the matched texts pertaining to a visit to the beach. However, in the affectively unpleasant text, "fins" referred to shark fins spotted prior to a shark attack; in the pleasant text, the word referred to swim fins that were lying on the beach.

The use of matched texts and test lists was employed to control for inadvertent effects on reaction time (RT) that might be considered material-specific. That is, it could be argued that the words in test lists derived from affective texts might speed or slow recognition RT because of a number of factors, including concreteness versus abstractness of the nouns, differential frequency characteristics, or affective valence of the test words themselves. The text matching strategy controlled for this confound. See Appendix B for a listing of study text pairs and their associated test list words.

These texts were also rated by subjects in the present study at the end of the experimental session, using the booklet form of the SAM ratings scale--see Appendix F for an illustration. Like the automated SAM system, the booklet form included the affective dimensions of valence, arousal, and dominance. Each of these dimensions was set to a 9-point scale. A rating of 1 represented the lowest possible value for the valence (i.e., unpleasant), arousal, or dominance scale; and a









17

rating of 9 represented the highest value on these scales. Texts receiving a valence rating of 1 to 3 were classified as affectively unpleasant; texts rated from 4 through 6 were classified as neutral; and texts with a rating of 7 or above were classified as pleasant. Statistical analyses for the present study were based on affective categories derived from these subject ratings.


Test Lists

After the subject had finished reading the study text, an episodic recognition task began immediately. The test list consisted of 12 prime-target pairs. The prime appeared on the monitor screen for 300 ms prior to the onset of the target. The target remained on the screen until the subject responded with a right or left button press. In the case of verbal targets, the button press indicated whether or not the target had appeared in the study text; whereas for non-verbal targets, the button press was determined by the target itself and was not related to text content. Subjects were instructed to respond only on the basis of the target stimuli, without regard to the prime. The lists were structured so that subjects responded with their right hand ("yes" response) to half of the items, and with their left hand ("no" response) to the other half. This prevented uneven response probabilities from influencing the RT data. Verbal primes and targets were always nouns, although they varied as to whether they were concrete or abstract in content. No test word appeared more than once in the experiment.









18

For the subset of subjects for whom a startle probe occurred on some occasions while they read the text, care was taken that test words and startle probes did not occur in the same sentence.

Targets. Half the targets (6) were verbal stimuli (Study or New words) and half were non-verbal stimuli (*****, =====). Verbal targets were also evenly divided between Study (3) and New words (3). Study words were always nouns that occurred in the preceding text. New words were nouns that did not occur in the study text, and had no semantic relationship to the topic of the study text or to the other words in the test list. Statistical analyses focussed on reaction time and accuracy for (1) the prime-target pairs for which the target was a study word or a new word, and (2) study word targets that were preceded by either a study word or new word prime.

Non-verbal targets were also equally divided, half (3) requiring a right-hand button press (******), and half a left-hand button press (======). These stimuli differed from the verbal targets in that they required pattern recognition for a choice reaction time task rather than verbal decoding for a recognition memory task. In addition, the nonverbal targets were repeated in every test list while verbal test list items appeared only once in the experiment. Thus, the non-verbal items involved minimal cognitive effort and responses to them became automatic over trials. Predictably, reaction time for the non-verbal targets was much faster (by over 200 ms) and less variable than for verbal targets. These data were thus not comparable to the verbal target conditions and were excluded from further analysis.









19

Primes. All four classes of targets (Study word, New word,

******, and ======) were primed by three alternatives: A Study word, a New word, or a series of X's. The New and XXXXX primes served as alternative control conditions. The priming condition in which a study word was primed by a string of X's was eliminated as a control, or neutral, prime for reasons detailed by Jonides and Mack (1984), in which important criteria for the characteristics of neutral primes are outlined. For example, they argue that non-word primes should not be compared with word primes because of the differential processing requirements and possible expectation sets that accompany these two types of priming stimuli. In addition, they caution against the use of a prime which is repeated on each trial (such as the XXXXX primes) as a comparison condition for a prime (such as a new or study word), which is not repeated. Again, differential processing requirements as well as possible differences in the alerting, or warning, functions of these two type of prime were considered to vary significantly. Because new word primes satisfied the above criteria and XXXXX primes did not, the analyses of priming effects were restricted to the comparison of recognition reaction time for New-Study with Study-Study prime-target pairs.

An exemplar of a study paragraph and its associated test list are provided in Figure 1. Note that in the presentation of this text on the subject monitor during the experiment, each sentence was presented in a single line of text, so that the paragraph appeared as six lines of text. The format of the present document prevents this presentation.









20



STUDY TEXT EXEMPLAR

You're having a great time on a picnic by the ocean with

some good friends. Some are out swimming, and you look twice

when you think you spot sharks. Your feet pound through the

surf as you run calling, the fins now obvious. Unaware, your

friends don't hear you, and you are gripped with horror. Your heart pounding in your chest, you run across the sand shouting

and waving. You scream in desperation as someone is jerked

under, and the water turns red.


Prime Target Pair Type

scissors degree New Word-New Word

calf picnic New Word-Study Word friends tent Study Word-New Word

charm New Word-Left Hand Response

XXXXX XXXXX-Left Hand Response

pint ******* New Word-Right Hand Response

fins Study Word-Left Hand Response feet ******* Study Word-Right Hand Response

XXXXX wagon XXXXX-New Word

XXXXX ocean XXXXX-Study heart sand Study-Study





Figure 1. Exemplar of affectively unpleasant study paragraph and its associated test list.









21

Apparatus

The experiment was controlled by two computers working in concert: Terak and Datax LSI-11 computers. The Terak controlled the timing of events in the experiments, presented the texts, and measured and recorded read times and reaction times. For the subset of subjects receiving startle probes, the Datax controlled presentation of the acoustic startle stimulus, and recorded EKG and blink response.

Beckman standard silver-silver chloride electrodes were placed in a modified Lead 2 configuration to measure the electrocardiogram (EKG). The signal was filtered and amplified by Coulbourn modules. A Schmitt trigger detected the R-waves and interrupted the computer clock to record inter-beat intervals with a resolution of 1 ms. EKG was measured continuously during presentation of each text, beginning 3 seconds prior to the beginning and continuing during the time the subject read each sentence of the text.

For the startle subset of the sample, EMG was measured with Med Associates miniature silver-silver chloride electrodes, using Beckman electrode electrolyte as the contact medium. The sensors were placed about 1 cm beneath the left eye. One electrode was centered beneath the eye, and the other was placed beneath the outer canthus. The amplification for the signal was set at 6 (X 10,000), and frequencies below 90 Hz and above 1000 Hz were filtered out by a Coulbourn HiGain Bioamplifier. The signal was rectified and integrated by a Coulbourn Contour Following Integrator (200 msec time constant), and then fed through an Analog-to-Digital converter. EMG was collected at 1000 Hz for 250 msec following the onset of the startle stimulus. The acoustic









22

startle stimulus consisted of a 50 msec burst of 100 dB (A) white noise with an instantaneous rise time, that was gated by a relay and presented binaurally through Pioneer (SE-250) stereo headphones from a tape that ran continuously during the experimental session.

Texts were displayed on an Amdek CRT monitor in the subject room. The monitor was placed approximately 2 feet in front of the subject, and a transparent plastic screen separated the subject from the monitor itself. Subjects responded using hand-held RT buttons, mounted on 6inch cylinders, which were activated by pressing down with the thumb. The subject held a cylinder in each hand throughout the experimental session. She controlled presentation of each successive sentence with a button press, and a final button press removed the entire text from the screen. Read time consisted of the time between button presses, as the subject read through the texts at her own pace. The subject also pressed either the right or left button to respond during the episodic recognition task. RTs and reading times were recorded to the nearest msec.

An intercom permitted communication between the subject and the experimenter throughout the experimental session.


Procedures

Upon arrival at the laboratory, subjects signed an informed consent form and completed imagery questionnaires and a vocabulary measure. The subject was then seated in a comfortable chair in the experimental chamber and sensors were attached in order to obtain the psychophysiological measures.









23

Instructions were read to each subject for verbal or imagery processing, and the subject was given practice in reading texts and responding to test lists. For startle subjects, headphones were placed on the subject's head and instructions were given to ignore the brief bursts of white noise that would occasionally occur while she was reading the study texts. The consent form, imagery questionnaire (QMI) and the exact instructions read to each subject are included in Appendix A.

After the experimenter had insured that the subject understood the instructions and had verified that the physiological measures were being recorded accurately, the subject was left alone in the experimental chamber to complete the experiment independently. The duration of this part of the experiment was approximately 45 minutes.

Prior to the onset of each trial, a 3-second pre-trial baseline period occurred. At the end of this period, the following message appeared on the subject's monitor screen: "Please press button to begin." This message signalled the subject to bring the first sentence of the study text onto the monitor screen by pressing one or both buttons. Each successive button press presented an additional sentence of the text, until the entire paragraph was displayed on the screen. For a subset of subjects, a brief burst of white noise was presented over headphones at either 2.0, 2.25, or 2.50 seconds after onset of one of the first two sentences (12 texts) or last two sentences (12 texts) of the texts. When the subject had read the entire text, a final button press cleared the screen. After a 200 ms pause, the test list began. Initially a "+" appeared on the screen for 150 ms, centered on the line









24

where the prime was presented. After the "+" was removed, the prime appeared immediately and remained on the screen for 300 ms. After the prime was removed from the screen, the target was presented on the line immediately below where the prime had appeared, and remained on the screen until the subject responded. When the subject gave an incorrect response, the word "ERROR" appeared on the screen for 2 seconds. After the subject response there was another 200 ms pause, after which the "+" appeared again for 150 ms, followed by the prime and target stimuli, as described for the previous test pair. This sequence continued until all 12 prime-target pairs were presented. A variable inter-trial interval occurred between text presentations ranging from 15 to 25 seconds in length. After the last text had disappeared from the screen, the following message appeared on the screen: "The session has been completed. Thank you for your participation."

After the subject had completed this part of the study, the

physiological sensors were removed, and she was taken to another room where she was asked to recall and write down a brief phrase characterizing the content of as many of the previously presented texts as possible. This task began approximately 10 minutes after the end of the experiment. Subjects generally required about 5 or 10 minutes to complete this task, although they were given as much time as needed.

After the free recall task was completed, the subject rated the texts for affective valence, arousal, and dominance. Ratings were made using a paper-and-pencil version of the Self-Assessment Manikin (SAM; Hodes, Cook, & Lang, 1985; Lang, 1980; see also Appendix F). Typed versions of the texts were interleaved in the SAM ratings booklet; the









25

subject was asked to refer to the typed version in order to recall the text, and then to mark her rating of how she felt while reading the text during the experiment. Finally, the subject was debriefed, and either given course credit or signed a voucher to be paid $10 for her participation in the study. The total experimental session lasted approximately 2-1/2 hours.


Data Reduction

Encoding Measures

Read time. Read time for each sentence was measured by computer with a resolution of one msec. The read times for sentences 1 and 2 were then averaged to provide a mean read time for the "early" sentence pair. Read times were also averaged over sentences 3 and 4, and sentences 5 and 6, and constituted the "middle" and "late" sentence pair read times, respectively. The final data set included three mean read times (for the early, middle, and late sentence pairs) for each of the study paragraphs, within each affective category. For subjects receiving startle probes, averages were calculated for each sentence at each level of the Startle Position factor (early, late, no startle), within each of the three affective categories.

Heart rate. Inter-beat intervals were converted into heart rate values using the algorithm of Graham (1979). Because subjects read the paragraphs at their own rate, the sampling periods for heart rate were highly variable among sentences. Thus, a real time analysis of the cardiac wave form across sentences was not possible. Rather, the average heart rate for each sentence of each text was calculated and









26

deviated from the pre-trial baseline for that text. Mean heart rate change scores were then calculated for affectively pleasant, neutral, and unpleasant texts. For subjects receiving startle probes, separate averages were calculated for each of the three levels of the Startle Position factor (Early, Late, No Startle) within each affective category.

Blink reflex. Integrated blink EMG data were scored with the

computer program provided by Balaban, Losito, Simons, and Graham (1986). Average startle baseline values and magnitude (in arbitrary A-D units) were calculated for startle probes presented early and late in the study texts (i.e., during sentences 1 or 2 versus 5 or 6), for each of the affective categories. The onset latency of the startle reflex was also scored. Trials in which the startle magnitude value was zero were omitted from the calculation of latency averages, which were calculated for probes presented early and late in the text and for each of the affective categories.


Retrieval Measures

Free recall. For each subject, the list of short descriptive

phrases was matched to a master list of paragraph topics in the study. A paragraph was considered to have been recalled if the descriptive phrase matched the topic of one of the paragraphs she read during the experimental session. Paragraphs identified by the subject in this manner were counted as having been recalled. The recall measure was calculated for each subject: For each affective category (pleasant, neutral, unpleasant), the number of texts listed by the subject was









27

divided by the total number of paragraphs she had rated in that affective category.

Reaction time. Reaction time for each of the 12 prime-target pairs was recorded to the nearest msec, for correct responses only. Average RT and error rates were calculated for each prime-target pair, for each affective category.

Accuracy. Error rate was calculated for each of the 12 primetarget pairs by dividing the number of errors within an affective category by the total number of paragraphs in that category, for each subject. This number was then multiplied by 100 so that error rate was expressed as a percentage.


Data Analysis

Imagery Ability

Subjects were classified according to imagery ability, based on a median split of 80 on the Questionnaire upon Mental Imagery (QMI; Sheehan, 1967). The range and distribution of QMI scores obtained for subjects in this study conformed to those obtained in other studies in the Lang laboratory, and are documented in Appendix C. Groups of subjects assigned to receive imagery and verbal processing instructions were balanced for QMI score to control for the effects of imagery ability and its possible interaction with the instructional manipulation.


Affective Classification of Texts

In the design of this experiment, a text's affective category was determined by ratings and psychophysiological measures obtained in an









28

earlier validation study (Spence et al., 1985). Subjects in the present study, however, also provided ratings of the texts after the experiment. This permitted a post-hoc recategorization of texts, based on this more proximal, group-specific assessment. The resulting redistribution of paragraphs produced an approximately equal number of affectively pleasant and unpleasant texts, but fewer neutral texts. For each subject, unequal numbers of texts contributed to the cell averages. Three subjects were not included; two of them failed to rate any paragraph in the neutral range, and another subject rated only one text as affectively neutral. The ratings distributions for the validation study, the present study, and a contingency table illustrating the stability of ratings classifications are provided in Appendix E.

Statistical analyses were performed, using both categorization systems, with similar overall results. Because they more closely represent this specific sample, and to avoid redundancy, only results based on the subjects' own ratings are presented here. Tests of valence involve orthogonal contrasts of pleasant versus unpleasant (linear effect), and affective (pleasant and unpleasant) versus neutral content (quadratic effect).


Encoding Measures

Read time and heart rate. These data were analyzed in a 2 X 2 X 2 X 3 X 3 factorial design. Between-subject variables were Imagery Ability (Good, Poor), Instructions (Imagery, Verbal), and Group (Startle Probe, No Startle Probe); within-subject variables were Affective Content (Pleasant, Neutral, Unpleasant) and Sentence Pair (1&2, 3&4,









29

5&6). Read times were subjected to ANOVA, and heart rate data were analyzed in an ANCOVA, the 3-second pre-trial heart rate average serving as the covariate for heart rate change scores.

Recall that a subset of 25 subjects in this study received startle probes while they read the study paragraphs. For these subjects, separate analyses of read time and heart rate included the additional factor of startle position (Early, Late, No Startle) because presentation of startle probes might be expected to affect encoding processes for the study paragraphs. All data analyses of the complete sample included group (Startle Probe vs. No Startle Probe) as a factor, however, in order to detect any differences related to this variable.

Startle magnitude and latency. A 2 X 2 X 3 X 2 ANOVA was carried out for each of these dependent measures. The between-subject factors were Imagery Ability and Instructions; the within-subject factors were Affective Content and Startle Position (Early, Late). Retrieval Measures

Free recall. These data were subjected to a 2 X 2 X 2 X 3 ANOVA with the between-subject factors of Imagery Ability, Instructions, and Probe Group; the within-subject factor was Affective Content.

Episodic RecoQnition for Target Words. A 2 X 2 X 2 X 3 X 2 factorial design included the between-subject factors of Imagery Ability, Instructions, and Group. The within-subjects variables consisted of Affective Content and Target Type (study word, new word). Non-word targets were not included in these analyses because of the widely different characteristics between the non-verbal and verbal









30

targets. As explained earlier (page 18), each of the non-verbal targets appeared three times in each test list while target words appeared only once in the experiment. Subject recognition for the non-verbal targets thus became automatic over trials, resulting in much faster and less variable reaction times than for verbal targets. These data were thus not comparable to the verbal target conditions and were excluded from further analysis.

Episodic Priming Analyses. A 2 X 2 X 2 X 3 X 2 factorial design again included the between-subject factors of Imagery Ability, Instructions, and Startle Probe Group. The within-subjects variables consisted of Affective Content, and Prime Type (Study Word, New Word). As stated earlier (page 19), XXXXX-Study pairs were excluded from analyses of priming condition. Important differences between the XXXXX and study word primes (verbal vs. non-verbal stimuli; repetition both within and between test lists; different levels of processing requirements for decoding by subjects) served as the rationale for their elimination as a control condition.

Recognition accuracy. The effects of the experimental factors on accuracy in the recognition RT task were analyzed in order to address the question of speed-accuracy trade-off as a possible factor in recognition performance. The accuracy measure was calculated as percent errors; analyses of these data paralleled those for recognition RT.
















RESULTS


Encoding Measures

Read Time

Table 1 presents mean read times for the early, middle, and late sentence pairs of paragraphs with affectively pleasant, neutral, or unpleasant content. Consistent with findings in the text processing literature, subjects read the later sentences in the paragraphs more rapidly than earlier sentences, F (2,96) = 59.71, 2 < .01 Furthermore, paragraphs with pleasant content were generally read more quickly than unpleasant texts, yielding a linear relationship between read time and pleasant versus unpleasant affective content, Valence F (2,92) = 2.17, p = .12; Valence (linear) F (1,46) = 9.46, p < .01. Finally, on the average, good imagers tended to read all sentences of the paragraphs faster (mean per sentence = 4.7 secs) than poor imagers (mean per sentence = 5.6 secs), F (1,46) = 3.55, R = .06.

Affective content and sentence interacted significantly,

F (4,184) = 3.97, p < .01 in the omnibus test. As can be seen in Table 1, read times for early sentences followed a quadratic pattern with affective content, Valence (quadratic) X Sentence F (2,92) = 5.86,



1 Where appropriate, reported p values reflect application of the Greenhouse-Geisser correction.





31









32

Table 1

Mean read times (in seconds) for early, middle, and late sentence pairs
of texts with affectively pleasant, neutral, and unpleasant content


Pleasant Neutral Unpleasant

Sentence Pair Mean

Early 5.49 5.35 5.59 5.48 Middle 5.08 5.27 5.30 5.22 Late 4.63 4.72 4.79 4.71

Mean 5.07 5.12 5.23


P < .01. That is, for these first two sentences only, mean reading times for both emotional contents (pleasant and unpleasant) were longer than for neutral paragraphs. The overall linear valence effect-shortest read times for pleasant paragraphs--developed in the later sentences (Valence (linear) Fs (1,46) = 1.57, 9.64, and 7.07; ps < .22, .01, and .02; for early, middle and late sentence pairs, respectively).

In subsequent tests, the Valence X Sentence interaction was also found to vary with subgroup: Thus, good imagers showed the early quadratic valence effect more strongly than poor imagers, Valence (quadratic) X Imagery Ability X Sentence F (2,92) = 3.47, 2 < .01. For poor imagers the linear valence effect (with the fastest read times for pleasant content) was the dominant response across all trials (see Table 7, Appendix D).

No reliable main effects or interactions involving instructions were observed for read time.









33


Heart Rate

Baseline. A preliminary 3 X 2 X 2 X 2 ANOVA (Valence X Imagery Ability X Instructions X Probe Group) of heart rate baseline values yielded no significant main effects or interactions of heart rate change with the independent variables. Thus, baseline heart rate was not systematically related to any of the factors in this study.

Encoding. Heart rate decreased from baseline during paragraph encoding, F (2,52) = 52.17, R < .01. As can be seen in Table 2, this deceleration was less during encoding of pleasant than unpleasant content, Valence F (2,51) = 2.57, R = .11 in the omnibus test; Valence (linear) F (1,25) = 15.52, p < .01. Examination of the Table also suggests that the above effect was stronger late in the paragraph. The Valence X Sentence interaction did not, however, reach statistical significance in the omnibus test (F (4,104) = 2.14, R = .12).


Table 2

Mean heart rate deceleration for early, middle, and late sentence pairs
of texts with affectively unpleasant, neutral, and pleasant content


Pleasant Neutral Unpleasant

Sentence Pair Mean

Early -1.02 -1.88 -1.45 -1.45 Middle -2.74 -3.11 -3.49 -3.11 Late -3.02 -3.34 -3.69 -3.35

Mean -2.26 -2.78 -2.88









34

Nevertheless, based on prior hypotheses, separate tests were undertaken for each sentence pair. These tests revealed that while the effect of valence on heart rate change was not significant early in the paragraph, Valence (linear) F (1,25) = 2.58, 2 = .12, pleasant contents resulted in reliably less heart rate deceleration than unpleasant contents for middle and late sentence pairs, Valence (linear) Fs (1,25) = 20.65 and

8.24, respectively; both 2s < .01.

Table 3 suggests that subjects receiving verbal instructions

showed greater heart rate deceleration compared to subjects receiving imagery instructions; however, the statistical test only approached significance, F (1,25) = 3.89, 2 = .06. On the other hand, instructions significantly modulated the affect/heart rate relationship, yielding a Valence (linear) X Instructions interaction, F (1,25) = 7.01, 2 < .02. The greater relative deceleration for unpleasant than pleasant texts, found overall, was more pronounced in subjects receiving verbal instructions.


Table 3

Mean heart rate change (beats per minute) for texts with
pleasant, neutral and unpleasant content, for subjects receiving verbal and imagery instructions


Pleasant Neutral Unpleasant

Instructions Mean

Verbal -2.92 -3.88 -3.91 -3.57 Imagery -1.61 -1.67 -1.85 -1.71

Mean -2.26 -2.78 -2.88









35

The stronger valence effect for verbal instructed subjects was greater for those who received startle probes, resulting in a Valence (linear) X Instructions X Probe Group interaction, F (1,25) = 5.63, P < .03 in the omnibus test (see Table 19, Appendix D). Furthermore, startle probes potentiated a stronger linear valence effect in poor imagers, Valence (linear) X Imagery Ability X Startle Group interaction, E (1,25) = 5.87, R < .03 in the omnibus test (see Table 18, Appendix D). To summarize the effects of these moderator variables on heart rate and affect, the subjects who showed the strongest linear valence pattern (least deceleration to pleasant content and greatest to unpleasant) were poor imagers, under verbal instructions, who also received startle probes.


Startle Reflex

Baseline EMG level. Analysis of baseline values for the startle response yielded no significant main effects or interactions of the independent variables in this study.

Startle magnitude. The omnibus analysis indicated that startle responses did not vary significantly with text valence overall, F < 1. The Startle Position X Valence (linear) interaction, however, approached significance, F (1,15) = 4.27, p = .057. As can be seen in Table 4, mean blink magnitudes increased linearly from pleasant to unpleasant paragraphs for late probes, valence (linear) F (1,15) = 4.81, p < .05. Neither a significant linear nor quadratic valence effect was found for early probes, F < 1.









36

No main effects or interactions involving instructions or imagery ability occurred for this measure.


Table 4

Mean startle magnitude (arbitrary A-D units) for probes measured
early and late in affectively pleasant, neutral and unpleasant texts

Pleasant Neutral Unpleasant

Startle Position Mean

Early 114.6 133.9 111.2 119.9 Late 134.0 149.8 165.1 149.6

Mean 124.3 141.9 138.2


Onset latency. Compared to the magnitude data, startle latency showed a closer--but still not statistically significant--overall relationship with affective content (Pleasant = 55.2 msec, Neutral = 52.9 msec, Unpleasant = 52.7 msec), F (2,28) = 1.63, p = .21 in the omnibus test; Valence (linear) F = 2.11, 2 = .17. Emotional content affected blink latency differently, however, for subjects receiving verbal versus imagery processing instructions, and particularly, depending on whether the startle probe occurred early or late in the paragraph (see Figure 2), Valence X Instructions X Startle Position interaction, F (2,28) = 3.72, 2 = .053 in the omnibus test; Valence (linear) X Instructions X Startle Position, F (1,14) = 5.99, < .03.

As can be seen in the left panel of Figure 2, instructions

interacted with emotional content to affect startle latency early in the encoding process, Valence (linear) X Instructions F (1,14) = 5.41,







Imagery SO' 0 Verbal E 6o


C
a- 55


50 -,

45
Pleasant Neutral Unpleasant Pleasant Neutral Unpleasant EARLY LATE

Figure 2. Startle latency for verbal versus imagery instructions, for probes presented
early and late in affectively pleasant, neutral and unpleasant texts.









38

p < .04. While subjects receiving imagery instructions showed the expected linear valence relationship, those receiving verbal instructions showed a deviant pattern. For late probes, both instruction groups startled faster to unpleasant content, Valence (linear) F (1,14) = 5.61, p < .04.

Imagery ability did not modulate startle latency for pleasant and unpleasant content. The only effect of imagery ability was a generally faster early probe startle for good imagers, compared to poor imagers--a difference that disappeared with the late probes, Startle Position F (1,14) = 11.88, 2 < .02 (see Table 22, Appendix D).


Retrieval Measures

Free Recall

Consistent with study predictions, paragraphs with emotionally evocative content, both pleasant and unpleasant, were recalled in greater proportion compared to affectively neutral texts (pleasant = .41, neutral = .29, unpleasant = .42), Valence F (2,100) = 11.20, p < .01 in the omnibus analysis; Valence (quadratic) F (1,50) = 18.20, p < .01. Instructions, imagery ability, and presence or absence of startle probes did not affect free recall. Episodic Recognition

As described earlier, analyses of recognition reaction time data were accomplished only for word targets and primes. The effects of the experimental factors (Valence, Instructions, Imagery Ability, Probe Group) were tested as they interacted with target type (study words versus new words) and priming condition (study-study, new-study).









39

Target recognition latency. Study word targets were words that

appeared in the preceding text and required a "yes" response. New word targets were non-text-related words which required a "no" response. Recognition reaction time for study and new word targets was not significantly different, F < 1. However, emotional content, imagery ability, and instructions interacted with target type, F (2,92) = 4.36, p < .03 in the omnibus test, Valence (quadratic) X Instructions X Imagery Ability X Target Type F (1,46) = 4.62, R < .04. Thus, separate follow-up analyses were done for new word and study word targets. As expected, no significant effects of the experimental factors were found for new word targets, all Fs < 1.9, all ps > .17.

Recognition memory for study word targets followed a pattern over affective contents similar to free recall (pleasant = 857 msec, neutral = 881 msec, unpleasant = 853 msec). In this case, however, the statistical test was not significant, Valence F (2,92) = 1.90, p = .17; Valence (quadratic) F (1,46) = 2.37, p = .13. Nevertheless, latencies did interact significantly with all of the main experimental factors, Valence X Instructions X Imagery Ability F (2,92) = 4.24, p < .03, Valence (quadratic) X Instructions X Imagery Ability F (1,46) = 5.30, p < .03. Examination of Table 4 reveals that target recognition was clearly faster for emotional than neutral content only for good imagers receiving imagery instructions, valence (quadratic), F (1,46) = 5.94, p < .02). Subsequent separate tests of affective content differences for other groups were not significant, F (1,46) = 2.04, 2 = .16 for poor imagers receiving verbal instructions,; all other Fs < 1.









40

Target recognition errors. Subjects made a larger percentage of errors for study word compared to new word targets (study = 8.40 percent; new = 3.93 percent), F (1,46) = 59.39, p < .01. Target type also interacted with valence and instructions to affect error rate, Valence X Instructions X Target Type F (2,92) = 4.02, p < .04 in the omnibus test; Valence (quadratic) X Instructions X Target Type, F = 5.22, p < .03. Thus, separate follow-up analyses were carried out for new word and study word targets.


Table 4

Mean recognition reaction time (msec) for study word targets
for affectively pleasant, neutral, and unpleasant texts
for good and poor imagers receiving imagery and verbal instructions


Imagery Instructions

Pleasant Neutral Unpleasant Mean Imagery Ability

Good 805 890 819 838 Poor 854 851 861 856

Mean 830 871 840


Verbal Instructions

Pleasant Neutral Unpleasant Mean Imagery Ability

Good 856 823 829 836 Poor 912 961 902 925

Mean 884 892 865









41
Subjects made more errors in rejecting new word targets with

imagery compared to verbal instructions (imagery = 4.71 percent, verbal = 3.16 percent), F (1,46) = 4.37, R < .05. Other tests were not significant: Errors for new word targets did not vary with emotional content of the paragraphs, presence or absence of startle probes, or imagery ability of the subjects.

Overall, study word errors did not vary significantly with valence (F < 1). Examination of Table 5 does suggest that subjects receiving verbal instructions made more errors in recognition of study word targets from affectively neutral paragraphs compared to emotional (both pleasant and unpleasant) texts. Statistical support for this interpretation was, however, marginal, Valence X Instructions interaction, F (2,92) = 3.27, 2 = .06; Valence (quadratic) X Instructions F (1,46) = 3.43, 2 = .07.

Errors in recognition of study word targets did not vary reliably with imagery ability, or presence or absence of startle probes. There is no evidence in these data that the observed recognition latency differences in affective contents can be attributed to a speed-accuracy trade-off.

Priming condition and target recognition latency. A subset of the reaction times to study words were re-examined, considering the effect of the preceding prime. Specifically, the analysis compared recognition latency for study word targets primed by another study word (SS) with recognition latency to study words primed by a new word (NS). The same experimental factors (affective valence, instructions, imagery ability, presence or absence of startle probes) were assessed, as previously.









42

Table 5

Mean percent errors for recognition of study word targets from pleasant, neutral and unpleasant texts, for subjects receiving verbal and imagery instructions

Pleasant Neutral Unpleasant

Instructions Mean

Imagery 8.65 6.79 8.00 7.81 Verbal 7.05 11.19 8.70 8.98

Mean 7.85 9.00 8.45


In the overall analysis, emotional content interacted with priming condition to affect reaction time (see Figure 3), Valence X Priming Condition F (2,92) = 8.52, 2 < .01; Valence (linear) X Priming Condition, F (1,46) = 6.56, 2 < .02; Valence (quadratic) X Priming Condition, F (1,46) = 9.52, p < .01. These effects were followed up by separate tests for each priming condition.

Study-Study. Recognition reaction time was fastest for pleasant and slowest for unpleasant content when the prime-target pair consisted of study words (SS) (pleasant = 832 msec, neutral = 849 msec, unpleasant = 859 msec), Valence F (2,92) = 1.14, p = .32 in the omnibus analysis; Valence (linear) F (1,46) = 4.71, p < .04. Recognition memory for SS prime-target pairs did not vary with imagery ability, instructions, or startle probe condition.

New-Study. Reaction time for recognition of study word targets primed by new words (NS) was affected by emotional content in a different manner--reaction time was relatively slower for neutral compared to emotional (either pleasant or unpleasant) content (pleasant = 845 msec, neutral = 907, unpleasant = 825 msec),







43








950-
0 New-Study L 2s 925 Study-Study
(I)
0 900 S875

0 850 825



PLEASANT NEUTRAL UNPLEASANT

AFFECTIVE CATEGORY










Figure 3. Mean reaction times for Study-Study and New-Study prime-target pairs following affectively pleasant, neutral and unpleasant paragraphs.









44

Valence F (2,92) = 6.08, p < .01 in the omnibus test; Valence (quadratic) F (1,46) = 6.97, 2 < .01. Other experimental factors did not reliably affect reaction time for NS pairs.

Recognition errors were least when a study word target was primed by another study word; performance declined somewhat with new word primes, (SS = 6.23 percent, NS = 8.38), Priming Condition F (1,46) =

5.66, 2 < .03. While error rate was unaffected by imagery ability, or the presence or absence of startle probes, error rates for SS and NS prime-target pairs were differentially affected by emotional content and instructions, Valence X Instructions X Priming Condition interaction E (2,92) = 3.12, p = .06; Valence (quadratic) X Instructions X Priming Condition F (1,46) = 4.28, 2 < .05. This interaction was followed up by separate analyses of each priming condition.

Study-Study. Recognition of study words primed by other study

words was not affected by emotional content of the paragraphs, Valence F < 1 in the omnibus test. However, instructions interacted with imagery ability to affect subject performance, F (1,46) = 6.05, p < .02. Good imagers receiving imagery instructions made significantly more errors than poor imagers (good = 8.69 percent, poor = 3.98 percent), F (1,26) =

7.95, p < .01. Under the verbal instructions condition, however, imagery ability did not affect error rates (good imagers = 5.44 percent, poor imagers = 6.82 percent), F < 1; (see Table 48, Appendix D).

Comparison of the error data results with the reaction time

analyses does not support a speed-accuracy trade-off explanation for the valence effects on recognition reaction time described earlier.









45

New-Study. For study words primed by new words, no main effects or interactions of the experimental factors on error rate were observed.


Episodic Priming

Facilitation of recognition RT for SS relative to NS prime-target pairs was not reliable when averaged across experimental conditions, (SS = 847 msec; NS = 859 msec), F(1,46) = 1.65, R > .20. As noted previously (p. 44), however, priming did vary with emotional content of the paragraphs, Valence X Priming Condition F (2,92) = 8.52, p < .01 (see also Figure 3). Thus, individual tests were conducted for each affective category. Consistent with findings in the episodic priming literature, reliable priming occurred for neutral texts, (56 ms), E (1,46) = 10.79, R = .01. However, for affectively pleasant texts, there was no significant priming (F < 1); and for affectively unpleasant texts, RT latency was actually 34 ms longer for study-study compared to new-study prime-target pairs, F (1,46) = 5.17, o < .03.
















DISCUSSION


The research provided broad support for the hypotheses that both the encoding and retrieval of narrative text vary significantly with the emotionality of story content. Thus at the learning stage, subjects tended to read pleasant texts faster, and unpleasant texts slower, than they did the neutral narratives. Furthermore, this effect of emotional valence was more pronounced later in encoding, even though reading times were then faster overall. Slower reading times for unpleasant paragraphs were accompanied by strong decelerative heart rate responses, and in response to startle probes, faster and larger blink reflexes. Again, these effects were reliably stated only late in the subjects' reading of the paragraph, when the story schema was already well developed.

Memory measures also showed the effects of emotional content.

Thus, pleasant and unpleasant texts were better remembered than neutral paragraphs at free recall. A similar pattern over affective contents was found for recognition reaction time, which was most evident for imagery instructed good imagers. Examination of primed recognition suggests that non-text primes distracted subjects, producing a performance deficit for neutral content, but did not interfere with the recognition of words from affective paragraphs.






46









47

Encoding Measures

According to text processing theory (Kintsch & van Dijk, 1978), the basic structure of a text schema is constructed during the initial stages of encoding. In later sentences, the processing task shifts to elaboration of the story framework. The greater cognitive processing demands of schema development have been associated with longer reading times for earlier sentences of logical narratives, with later sentences being read more quickly. Consistent with study predictions and the text processing literature, earlier sentences of the study paragraphs were read more slowly than later sentences.

Emotional valence of the paragraphs was also expected to modulate reading time in this study. It was hypothesized that affective text contains information on visceral and motor activation, resulting in more elaborate and complex representations, compared to memory for neutral content. However, since this additional information does not increase the logical complexity of the text--a factor known to increase encoding time in text processing--it was not clear how emotional content might modulate encoding of textual materials.

The results of this study suggested that emotional content

affected reading time for early sentences differently than for later sentences. The tendency to read affectively unpleasant paragraphs more slowly than pleasant texts was significant overall, although this effect only developed later in the encoding process. In addition, this overall effect was strongest for subjects with self-reported poor imagery ability, who showed the linear effect of valence consistently throughout the narrative.









48

On the other hand, early in the encoding process, good imagers

read sentences from emotional (both pleasant and unpleasant) paragraphs more slowly than those with neutral content. This effect was replaced with the overall linear effect of valence in later sentences. Because the early quadratic pattern of valence effects (unpleasant and pleasant slower than neutral) occurred during the initial sentences of the paragraphs for subjects reporting good but not poor imagery ability, it may be argued that these results reflect individual differences in schema development. It was suggested earlier that representations of affective experience include additional information consisting of the visceral/motor activation inherent in emotion (Lang, 1979). Hence, construction of the schema for emotionally evocative text may result in longer processing time (as reflected by read time) for affective compared to neutral text, but only for subjects with a predisposition to include the "action set" of emotion in the encoding process.

Heart rate decreased reliably during text encoding, consistent with theories that deceleration is associated with information intake (Lacey & Lacey, 1974). In addition, heart rate interacted with valence, resulting in greater heart rate deceleration for affectively unpleasant and less deceleration for pleasant contents. Consistent with read time results, this effect emerged later in the encoding process.

Study predictions and prior research predicted that imaginal

versus verbal encoding strategies would result in different patterns of heart rate change over sentences. Heart rate decrease was predicted for verbal instructions, reflecting an information intake orientation. This result is consistent with studies of heart rate response to presentation









49

of affective picture slides (Greenwald et al., 1989). However, research also exists in which subjects listened to affective and neutral narratives over headphones (Cook et al., 1988) and showed heart rate acceleration in response to unpleasant content, despite the fact that they were engaged in a task requiring information intake. The visual aspect of the experimental task is common to both studies showing heart rate deceleration to unpleasant stimuli, suggesting that the sensory aspect of the experimental task may modulate the direction of heart rate change in response to affective content.

Heart rate was expected to increase late in the encoding process as a result of the internal processing demands in the imagery instructions condition, consistent with prior research on imaginal processing of emotional content (Lang et al., 1980; Vrana & Lang, 1990). While study results confirmed the predicted heart rate decrease in the verbal instructions condition, heart rate did not increase for the imagery instructions condition. Nevertheless, the lack of heart rate deceleration in the imagery instructions condition might be construed as evidence for acceleratory effects cancelling out the deceleration component.

While the read time and heart rate results showed consistent

effects of emotional content on subject responses, these measures are nevertheless limited as direct measures of emotion because they are inherently susceptible to a variety of other factors, as noted above. Read time and heart rate are measures of perceptual processing and stimulus intake, and thus vary with instructions, characteristics of the experimental materials, and individual predisposition of the subjects.









50

The startle reflex, on the other hand, has been shown in recent studies to be a robust measure of emotional response that is invariant across a variety of experimental tasks and sensory modalities.

Lang has conceptualized startle as a protective reflex that is primed or facilitated in negative affective contexts (Lang et al., 1990). That is, the response to a noxious stimulus (the startle probe) would be expected to be larger when the organism is in an affectively negative state, and smaller during an affectively positive state. Thus, the startle response served as a manipulation check in the present study, to assess the development of affect in the subject during the encoding process.

Consistent with the hypothesis that schema development occurs over time, emotional content only affected startles presented late in the encoding process. Startle magnitude showed the expected pattern of greater potentiation for unpleasant relative to pleasant content. However, encoding strategy did not affect this aspect of the startle response.

Consistent with the startle magnitude results, startle latency was faster for probes presented late in affectively unpleasant compared to pleasant paragraphs. These effects were evidenced more for subjects receiving imagery instructions, who showed this linear effect for early probe startles as well. Subject receiving verbal instructions, on the other hand, only developed this effect late in the encoding process. These results suggest that imagery instructions may result in earlier contact with the emotional content of text than instructions that focus on the verbal characteristics of the narrative.









51

In summary, development of the affective information structure, or text schema, was hypothesized to occur over time during encoding. Late in the encoding process, a consistent pattern of results emerged in which affectively unpleasant content resulted in longer read times, greater heart rate decrease, and larger and faster startle responses, compared to paragraphs rated as pleasant. For read time and heart rate, the negative versus positive valence effect was most evident for subjects reporting poor imagery ability and those receiving verbal instructions, while imagery processing strategies tended to minimize these effects. Instructions and individual differences variables did not interact with startle magnitude effects, but for startle latency, good imagers showed the linear effect of emotional content earlier in the processing task.


Retrieval Measures

While measures of encoding of affective texts appeared to vary with the negative versus positive valence of the materials, retrieval measures varied more according to whether or not paragraphs were emotional in content. Both recognition and free recall measures showed greater facilitation for affective compared to neutral paragraphs.

Recognition of study word targets varied with presence or absence of emotional content as well as imagery ability and instructions. Good imagers receiving imagery instructions showed faster and more accurate recognition for words from emotionally evocative compared to neutral paragraphs. These results are consistent with the idea that imaginal encoding strategies are more likely to activate arousing emotional









52

response information. Consistent with the bio-informational hypotheses, and with other recent studies, more arousing materials were remembered faster and more reliably (Bradley et al., in press; Heuer & Reisberg, 1990).

In addition, recognition RT for new word targets showed no relationship to either affective content or encoding strategy, suggesting that the representation of text in memory must be activated (by processing of a study word target) in order for the experimental factors to affect performance. It may then be reasoned that the recognition RT measure does in fact reflect the characteristics of the information network represented in memory rather than nonspecific, generalized effects of emotional arousal.

Analyses of the priming data yielded significant facilitation of reaction time for study-study compared to new-study prime-target pairs for neutral texts, but priming did not occur for affective texts. However, Figure 2 clearly shows that the lack of priming for affective materials resulted from a change in RT for New-Study pairs. That is, recognition RT was faster for New-Study pairs from affective than from neutral paragraphs. It appears from these data that a new word prime has a greater inhibitory effect on recognition of a study word target from an affectively neutral text, relative to when the study word target is part of an affective information structure. Lang has suggested that the associative structure of affective information, particularly when processed in an imagery mode, contains a greater amount and different types of information (motor and visceral activation codes) than the materials typically employed in priming experiments. It appears that









53

one characteristic of such affective information structures is that they facilitate both probability and speed of memory retrieval.

However, one important measure of episodic priming deviated from this general pattern of valence effects. When study words primed by other study words were considered independently, reaction times were slower for affectively unpleasant content relative to pleasant content. This result does not conform to the study predictions for memory performance. This linear effect of valence (slower for unpleasant and faster for pleasant content) study-study reaction times is more consistent with the pattern of valence effects observed for the encoding measures during the later sentences. For reading time and heart rate, this linear valence effect was associated with poor imagery ability and verbal instructions, and was less pronounced for good imagers and imagery instructions. These results suggest that a perceptual orientation to the text processing task may moderate the valence effects on a variety of dependent measures. Given that the perceptual aspect of the episodic recognition task is inherently enhanced when the exact information is presented at retrieval that was present at encoding, the study-study reaction time data may reflect processes similar to those acting on read time and heart rate late in the encoding process.

Accuracy data from this experiment showed a similar pattern of results to the recognition RT data. Recognition was both faster and more accurate when study words had previously appeared in an emotionally evocative text, and when subjects employed imagery during encoding.









54


Overall Conclusions

In summary, the results of this study show that while encoding measures varied along the valence dimension, retrieval from memory appeared to be largely mediated by the presence or absence of emotional content in the texts. These results were consistent for all measures of encoding, including read time, heart rate, and blink magnitude and latency. In addition, recognition and recall measures of memory retrieval were influenced in a similar fashion by the arousal dimension of the study texts. An important exception to this general trend was study word targets primed by study word primes. This research suggests that variation in emotional content and manipulation of information processing strategies (e.g., imaginal versus verbal encoding instructions) work separately and in interaction to affect both the encoding and retrieval for textual materials. Further study of the effects of these variables on verbal learning and memory promises to extend our understanding of these cognitive processes in important ways.
















APPENDIX A
QUESTIONNAIRES, INSTRUCTIONS AND DEBRIEFING STATEMENT


OMI Vividness of Imagery Scale

The aim of this test is to determine the vividness of your

imagery. The items of the test will bring certain images to your mind. You are to rate the vividness of each image by reference to the accompanying rating scale which is shown at the bottom of the page. For example, if your image is "vague and dim" you give it a rating of 5. Record your answer in the brackets provided after each item. Just write the appropriate number after each item. Before you turn to the items on the next page, familiarize yourself with the different categories on the rating scale. Throughout the test, refer to the rating scale when judging the vividness of each image.

* Rating 1 Perfectly clear and as vivid as the actual experience *

* Rating 2 Very clear and comparable in vividness to the actual *

* experience *

* Rating 3 Moderately clear and vivid *

* Rating 4 Not clear or vivid, but recognizable *

* Rating 5 Vague and dim

* Rating 6 So vague and dim as to be hardly discernable *

* Rating 7 No image present at all, you're only "knowing that" you *

* are thinking of the object *





55









56

An example of an item on the test would be one which asked you to consider an image which comes to your mind's eye of a red apple. If your visual image was moderately clear and vivid you would check the rating scale and mark "3" in the brackets as follows: Item Rating

5. A red apple (3)

Now turn to the next page when you have understood these instructions and begin the test.

Think of some relative or friend whom you frequently see,

considering carefully the picture that rises before your mind's eye. Classify the imagers suggested by each of the following questions as indicated by the degree of clearness and vividness specified on the Rating Scale.

1. The exact contour of face, head, shoulders, and body (

2. Characteristic poses of head, attitude of body, etc.

3. The precise carriage, length of step, etc., in walking

4. The different colors in some familiar costume

Think of seeing the following, considering carefully the picture which comes before your mind's eye, and classify the image suggested by the following question as indicated by the degrees of clearness and vividness specified on the Rating Scale.

5. The sun as it is sinking below the horizon (

Think of each of the following sounds, considering carefully the image which comes to your mind's eye, and classify the images suggested by each of the following questions as indicated by the degree of clearness and vividness specified on the Rating Scale.









57

6. The whistle of a locomotive

7. The honk of an automobile

8. The mewing of a cat

9. The sound of escaping steam 10. The clapping of hands in applause (



Think of "feeling" or touching each of the following, considering carefully the image which comes to your mind's touch, and classify the images suggested by each of the following questions as indicated by the degrees of clearness and vividness specified on the Rating Scale.

Rating

11. Sand ( 12. Linen ( ) 13. Fur

14. The prick of a pin ( 15. The warmth of a tepid bath ( )



Think of performing each of the following acts, considering

carefully the image which comes to your mind's arms, legs, lips, etc., and classify the images suggested as indicated by the degree of clearness and vividness specified on the Rating Scale. 16. Running upstairs

17. Springing across a gutter ( 18. Drawing a circle on paper ( ) 19. Reaching up to a high shelf ( 20. Kicking something out of your way (









58



Think of tasting each of the following, considering carefully the image which comes to your mind's mouth, and classify the images suggested by each of the following as indicated by the degree of clearness and vividness specified on the Rating Scale. 21. Salt

22. Granulated (white) sugar 23. Oranges

24. Jelly

25. Your favorite soup



Think of smelling each of the following, considering carefully the image which comes to your mind's nose, and classify the images suggested by each of the following questions as indicated by the degree of clearness and vividness specified on the Rating Scale. 26. An ill-ventilated room 27. Cooking cabbage

28. Roast beef

29. Fresh paint

30. New leather



Think of each of the following sensations, considering carefully the image which comes before your mind, and classify the images suggested as indicated by the degree of clearness and vividness specified on the Rating Scale. 31. Fatigue (










59

32. Hunger ( 33. A sore throat ( 34. Drowsiness ( 35. Feeling full after a large meal (










60

Instructions to Subjects

In this study, you will be reading paragraphs that will appear on this monitor screen, and then respond yes or no to a list of words, depending on whether or not they were in those paragraphs.

The exact procedure will be as follows: You will hold these two

buttons, one in each hand (get subject to hold buttons), and use them to indicate yes or no to the test words as they appear. You should press the right button if the word is in the paragraph and the left button if the word is not in the paragraph. You should press a button as fast as possible while still making accurate decisions. You will also use the buttons to present the paragraphs line by line.

After you press both buttons to start, the first line of a

paragraph will appear on the monitor screen. Read through the sentence at a comfortable pace, and then press both buttons again to move to the next sentence.

IMAGERY INSTRUCTIONS:

As you read the sentences, please make a special effort to form as vivid an image of the events described as you can. Place yourself in the image as a participant in the story--as if you were actually experiencing it all personally. The more you can experience the actions and feelings, the more vivid your image will be. Read through the lines at your own pace, constructing your image as you read. The more vivid your image of the events described in the paragraph, the more it is a personal experience, the better you'll be able to remember it for the subsequent test.









61

VERBAL INSTRUCTIONS:

As you read the sentences, you should concentrate on the specific words used in each sentence. Read through the lines of the paragraph at your own pace. Try to remember each sentence in the paragraph and the specific vocabulary employed. As you read, try to retain the precise wording of each sentence. The better your memory of the exact language used, the better your performance will be on the word test that comes after.

MAIN INSTRUCTIONS:

After you press the buttons to remove the last sentence, a word pair will be presented on the screen. The first word will appear very briefly, so you must stay alert and pay close attention. Read this first word, but do not press a button. Wait for the second word to appear. When the second word appears, your task is to press one of the buttons to indicate whether it was in the paragraph you just studied. Press the button in your right hand for yes, and the button in your left hand for no. After you respond, another pair of words will appear. If you are wrong, an error message will appear on the screen, and then the next test pair will appear as normal. The test lists move very fast, so it's important for you to stay alert and to respond as quickly and accurately as possible. After the test list there will be a short pause, and then a message will appear on the screen instructing you to press both buttons to start the next trial. At the end of the session a message will come on the screen to let you know you have finished, and I'll come in shortly after that. In order to make your reaction times more precise, please hold your thumbs lightly on the button throughout









62

the trials--so that you don't have to reach for them when you want to make a response.

We will go through two practice trials now to be sure you understand how the experiment goes.

Do you have any questions?

Present practice paragraphs and continue specialized instructions: IMAGERY INSTRUCTIONS: Tell me about the image you formed. Were you aware of knocking on the door? Movement? Feeling of knocking? Could you see the faces? The candles on the cake? Were you aware of feeling surprised? Happy? REINFORCE AND INSTRUCT AS PER IMAGERY 18 VERBAL INSTRUCTIONS:

Repeat the first sentence in the paragraph as precisely as you can. What age was mentioned in the paragraph? What had you agreed to do that evening? Did you smile with happiness or delight? REINFORCE AND INSTRUCT FOR EXACT WORDS AND IGNORE ANY IMAGERY BEHAVIOR OR RESPONSES.

Remind subject about baseline periods, sitting quietly not to disturb physiological recordings, and tell them you'll be right next door.









63

Debriefing Statement

This study is concerned with determining whether people form

different mental representations of emotional vs. non-emotional texts. You may have noticed that some of the paragraphs that you read were pleasant, some were unpleasant, and some were relatively neutral. We were measuring your reaction time to target words taken from those paragraphs. Some target words were preceded (or primed) by words taken from the same sentence, some target words were primed by words taken from another sentence in the paragraph, and some targets were primed by words that did not appear in the paragraph at all. Usually reaction time is faster when a word is primed by a word that is closely related (when the prime and target were from the same sentence in this case), and measuring the changes in reaction time based on the logical distance between the prime and the target can provide information about the structure of the mental representation you have constructed from the paragraph.

We think that mental representations of emotional paragraphs may be more complex and extensive than those of neutral paragraphs, because emotional stimuli more often include representations of motor activity (jumping away from the palmetto bug) and of visceral activation (your heart is pounding from the scare). This more extensive information network would result in longer reaction times for affective paragraphs, which is what we have found in an earlier study.

We also think that people who form vivid images may form

particularly extensive information structures, and so we would expect those persons to have longer reaction times also.









64

This research is part of a larger program of experiments which concern the study of modes of information processing and emotion in persons who vary in imagery ability. We are working to develop a model of how affective memories are represented, stored and retrieved in the brain.

It is important that subjects participating in these studies not have preconceived ideas about the experimenter's hypotheses and expectations, in order to obtain valid experimental data. Therefore, please do not discuss this experiment with other psychology students until after the term is over.

I have read this form and had any questions concerning the study answered to my satisfaction.








SUBJECT SIGNATURE
















APPENDIX B
TEXT MATERIALS AND TEST LIST WORDS


Instructions and Practice Paragraph Experimenter Instructions

This is a warm-up exercise to teach you special symbols used in this study. This one (press right button to display *******) signals you to make a "yes" response. This one (press left button to display

) signals you to make a "no" response. Try it a few times:

******** ======== ******** ======== These symbols will appear in the test lists which will follow each paragraph. What does

signal you to do? What does ******* signal you to do? Practice Paragraph

Today you turned twenty, and you wonder if any of your friends will notice. You go out running after class and go over to study at a friend's house later. Knocking at her door, you wonder if she has any idea that it's your birthday. As you enter, people seem to jump out from everywhere shouting "SURPRISE!" You gasp with delight as you see a big cake with candles and balloons all over. You smile with affection as they laugh gleefully that you suspected nothing. STUDY WORDS: NOTICE, BIRTHDAY, IDEA, PEOPLE, DELIGHT, CAKE, NOTHING NEW WORDS: POSTER, ERASER, BOLT, PERFUME, FLINCH, HYDROGEN, BLOOMS







65









66

Text Materials

1. Evening Walk

Unpleasant. One night you and your roommate go for a walk through a bad neighborhood. You notice several dark figures slip quietly between two buildings just ahead. You stop and listen closely; beads of sweat break out on your forehead. Your breath catches as you hear soft footsteps nearby, and you both panic. Echoes of your footsteps break the night's stillness as you run for your car. At the car you're fumbling with your keys when two guys grab your roommate.

Neutral. One night you and your roommate take a relaxed walk

through the neighborhood. You've been studying all evening and it feels good to get up and walk around. There is a light sweat on your forehead and the cool breeze feels refreshing. Quiet and preoccupied with your studies, you have no particular desire to talk. A lone car drives by momentarily creating echoes in the night's stillness. You walk back home, thinking maybe you'll read one more chapter before bed.



STUDY WORDS: ROOMMATE, WALK, SWEAT, FOREHEAD, STILLNESS, ECHOES,

NEIGHBORHOOD

NEW WORDS: RECORD, WINNERS, PARROT, IMAGINATION, BARREL, GEESE,

PLASTER









67

2. Bicycle Ride

Unpleasant. You are riding your bicycle to school early in the

morning, in a rush. You are late for an exam, but you can't seem to get free of all the cars. You're stuck on a narrow street in a line of traffic moving slowly up the hill. Exasperated, you bolt for the other sidewalk between cars, at top speed. You look around to a loud horn, but the wind blows hair across your face. Suddenly you are struck hard from the side, slamming you onto the pavement.

Neutral. You are riding your bicycle to school, relaxed and in no particular rush. Pedaling smoothly along, you mentally review your schedule for that morning. You go along steadily in a line of bikes moving with the traffic up a hill. Picking up speed as you reach a wide sidewalk on campus, you feel more relaxed. You coast downhill, lifting your head to let the wind blow through your hair. Arriving at the library you climb off your bike and lock it up in the rack.



STUDY WORDS: BICYCLE, SCHOOL, RUSH, TRAFFIC, HILL, SPEED, SIDEWALK NEW WORDS: NEPHEW, THIRST, MINK, CLAMP, HAM, SPARROW, COAL









68

3. Hot Summer Night

Unpleasant. Lying in the dark naked on a night too hot for covers, you can't get any sleep. You have the ceiling fan on high trying to get comfortable in the humid heat. Suddenly you feel a huge palmetto bug drop from the ceiling onto your chest. You can feel the insect crawl rapidly over your chest and up onto your neck. You leap onto the carpet in panic and disgust, quickly switching on the lamp. Scanning the bedclothes, you hesitantly lift the sheets looking for the insect.

Neutral. You are in bed, drifting off to sleep on a summer night too warm for covers. The ceiling fan cools off the summer heat, and the ice cream helped too. You lie there relaxing, watching occasional car lights play on the ceiling. The weekend starts tomorrow, and you think about things you'd like to do. The phone rings--you switch on the lamp and walk across the carpet to get it. It's a good friend calling long distance, and you sit down to have a talk.



STUDY WORDS: SLEEP, COVERS, FAN, HEAT, LAMP, CARPET, CEILING NEW WORDS: EAGLE, STEAK, VICTORY, WALNUT, LAKE, BOWL, TREE









69

4. Dental Appointment

Unpleasant. Your anxiety is raised by a call confirming your appointment with the dentist. Walking in, you are instantly aware of the whine of the drill from a back room. As you climb into the chair you glimpse the pointed instruments on the tray. You wince as he drags the sharp steel probe along your sensitive gumline. Hands wet with sweat, you feel the long needle pierce the back of your mouth. As he approaches you with the whining drill you grip the chair hard.

Neutral. You get a phone call that your appointment has been

cancelled with the dentist. Now you have some extra time and you decide to have a leisurely breakfast. You settle into a comfortable chair, reaching for a bagel on the tray. You turn on the morning news to get the latest developments on the Iran probe. Putting the last bite into your mouth you wipe your hands on your napkin. You look through the paper and read the funnies before leaving for class.



STUDY WORDS: APPOINTMENT, DENTIST, CHAIR, TRAY, HANDS, MOUTH, PROBE NEW WORDS: SOCKS, PUPPY, KNOB, RAILWAY, BOXES, DINNER, PLANET









70

5. Airplane Trip

Unpleasant. You startle as the plane lurches violently in

midflight on your way home. Confused, the passengers stir restlessly as the cabin lights go out. Suddenly there's an odd noise--the engines' sputtering silences all talk. Grabbing the chair arms hard, you feel the hull shudder violently. The hostess is thrown sideways, spilling the drinks she was carrying. Screams pierce the dark as the plane tilts sideways and starts to fall.

Pleasant. You are flushed with excitement as the plane soars

high, taking you home. Soon you will be with your boyfriend again after a long separation. The noise of small talk around you is of no interest, you are so happy. Smiling with a rush of pleasure, you imagine him in your arms again. When the hostess comes by with drinks, you decide on champagne to celebrate. As you approach your destination you smile happily, full of anticipation.



STUDY WORDS: PLANE, HOME, NOISE, TALK, DRINKS, HOSTESS, ARMS NEW WORDS: REMAINDER, COUNTER, PEPPER, APPLES, GRAPEFRUIT, FRONTIER,

CAPTIVE









71

6. Football Game

Unpleasant. The football game is a few minutes past halftime inside the packed stadium. You have a bad headache and the hot sun makes you feel worse by the minute. The crowd roars and tension is high--they are only yards away from a touchdown. You're sick and uncomfortable on the hard seat as play continues on the field. Your stomach churns painfully as the stands pulsate with the roaring cheers. Your head throbs and you're sick to your stomach as the game finally ends.

Pleasant. The football game is in the last few minutes inside the packed stadium. You sweat freely in the hot sun as your team moves into playing position. Tension is high--you are one point behind, and thirty yards from a touchdown. You jump from your seat cheering wildly as the play moves down the field. The stands pulsate with the crowd's cheers as the halfback gets the ball. Touchdown! You thrust your arms overhead as the crowd roars with excitement.



STUDY WORDS: MINUTES, STADIUM, TENSION, YARDS, SEAT, FIELD, PLAY NEW WORDS: SUGAR, MONSTER, ECONOMY, MAYOR, PRISON, HAT, SPINACH









72

7. First Date

Unpleasant. Out on a first date your shyness around this new

person is painfully obvious. Sitting stiffly in your seat, there is no conversation before the movie starts. You are aware of your sweaty palms and feel a slow blush covering your face. The silence drags on endlessly as you nervously search for something to say. You finally manage to make a joke and fall silent after the brief response. You sit there silent and miserable, wishing the evening were over.

Pleasant. Out on a date, you're having a wonderful time with a very special person. Conversation seems effortless as you laugh and talk before the movie starts. You feel a blush of excitement and attraction as you make plans for later on. You agree to go dancing later in the evening as silence falls over the theater. You lean over close with a last little joke--he squeezes your hand in response. You sit there, warm and excited, wishing the evening would go on forever.



STUDY WORDS: DATE, PERSON, CONVERSATION, MOVIE, JOKE, RESPONSE, BLUSH NEW WORDS: BUSHEL, STRAND, SCARF, CEREAL, SWEATER, SALUTE, REMINDER









73

8. Beach Picnic

Unpleasant. You're having a great time on a picnic by the ocean with some good friends. Some are out swimming, and you look twice when you think you spot sharks. Your feet pound through the surf as you run calling, the fins now obvious. Unaware, your friends don't hear you, and you are gripped with horror. Your heart pounding in your chest, you run across the sand shouting and waving. You scream in desperation as someone is jerked under, and the water turns red.

Pleasant. You're having a great time on a picnic by the ocean

with some good friends. Some are out swimming, and you wave to them as you walk along the surf. Walking along, you stumble as your feet trip over a pair of lost swim fins. You look up to see a frisbee sailing in your direction, and you go after it. Your heart pounds in your chest as you run hard across the sand to catch it. You laugh and shout to your friends as you grab it, full of high spirits.



STUDY WORDS: PICNIC, OCEAN, FEET, FINS, HEART, SAND, FRIENDS NEW WORDS: DEGREE, WAGON, PINT, CHARM, TENT, CALF, SCISSORS









74

9. End of School Term

Neutral. It's the end of another school year and you are going home for the summer. You have been packing, and just finished finals for your courses this term. You start carrying plants and clothes and suitcases out to your parents' car. Everything is finally loaded up and you drive away, waving to your roommate. Leafing through a magazine in the back seat you think about a summer job. Thinking of the long trip ahead, you settle down in the comfortable car.

Unpleasant. It's the end of the school year and you have to go home for the summer. You feel down, having just told your parents you failed two finals this term. You wearily drag plants and clothes and suitcases out to their car in silence. Your father says that maybe you should quit school and you feel more depressed. Sitting in the car, you think about the boring job you'll have to start soon. You sigh and settle down in the back seat, dreading the long trip ahead of you.



STUDY WORDS: YEAR, SUMMER, FINALS, TERM, PLANTS, CLOTHES, SUITCASES NEW WORDS: CHIMNEY, LANTERN, GLUE, MELON, RECEIPT, CANAL, SILVER









75

10. Elevator

Neutral. You are part of a crowd of shoppers, running errands in a new shopping mall. Walking around, you discover the mall has a glass elevator and a nice fountain. Your errands are run, and you notice you feel slightly on edge from hunger. On your way out, you stop and treat yourself to a delicious cold milkshake. It's a relief to get out into the cool darkness after the busy confusion there. As you drive away, you start to relax and think about watching a movie at home.

Unpleasant. You are part of a crowd of shoppers, running errands in a large shopping mall. The elevator in the mall is full, but you're in a hurry and squeeze in anyway. People are pressed in around you; you feel uncomfortably hot and on edge. Suddenly the elevator jerks to a stop; you startle as a loud alarm goes off. There is no relief as panic spreads in the darkness in the tight enclosure. You clench your teeth as a baby screams; you feel yourself sweating heavily.



STUDY WORDS: SHOPPERS, ERRANDS, MALL, ELEVATOR, EDGE, RELIEF, DARKNESS NEW WORDS: SUNSET, RANCH, BALLET, VINES, LIQUID, CARAMEL, EGGS









76

11. Father Waits Up

Neutral. Coming home from racquetball and a movie with a friend, it's gotten late. Although it's midnight your father is still up, so you introduce your friend. As you close the front door you offer a vague explanation for your late return. You groan as you sit on the couch suddenly aware of sore muscles from the game. Your father has a sleepy look on his face as he smiles and says goodnight. Pretty tired yourselves, your friend soon leaves and you head off for bed.

Unpleasant. Coming back from a late movie with a girlfriend, it's gotten very late. Her father has been waiting up, and you realize that it's hours after midnight. He slams the door shut after you come in, and loudly demands an explanation. Your muscles tense as he rudely orders you in and sits you both on the couch. You feel nervous and a little scared from the furious look on his face. You feel yourself sweating and your face getting red as he goes on yelling.



STUDY WORDS: FATHER, MIDNIGHT, EXPLANATION, DOOR, MUSCLES, COUCH, FACE NEW WORDS: SALMON, PURSE, PASTURE, TOAST, MARE, PRUNE, BRICK









77

12. Parents' Pool Party

Neutral. The evening of your parents' pool party the guests are not very exciting. Finally tired of hearing the old familiar stories, you head for your room. You take off your watch, your shoes and clothes, and put on your bathing suit. Things have quieted down outside by the pool, and you dive in for a quick swim. You swim vigorously through the cool refreshing water using your full strength. You come out of the pool feeling much better and head off for the refreshments.

Unpleasant. It's late and some of the guests at the pool party are getting very drunk. You ask them to quiet down but they laugh, grabbing your arms and both legs. You panic--you're wearing expensive shoes and a watch that isn't waterproof. As they drag you towards the pool you yell that your watch will be ruined. Angrily, you fight with all your strength as they swing you out over the water. You plunge into the cold water and come up choking, completely furious.



STUDY WORDS: EVENING, GUESTS, WATCH, SHOES, POOL, STRENGTH, WATER NEW WORDS: SPACE, DWARF, CREAM, MIRROR, ATMOSPHERE, ONION, SONG









78

13. Canoe Ride

Neutral. You lean forward as the canoe glides smoothly along

among the water lilies. Slowing the boat, you scan the surface of the water close to the shore. You spot a duck nesting on her eggs over to the side, hidden among the reeds. You hurriedly check the camera--the timing is critical in this situation. You watch closely, adjusting the focus as the full nest comes into view. Delighted with your luck, you get several excellent shots of this rare find.

Unpleasant. You slip out of the canoe into the brackish water among the water lilies. Unnoticed, a huge alligator glides off the shore, sliding under the surface. You jump as you spot something big moving in the reeds over to the side of you. Panic shoots through you as you suddenly know you're in a dangerous situation. Your heart stops as you see the alligator now gliding towards you in full view. Desperate to escape, you run hard, crashing through the reeds and lilies.



STUDY WORDS: CANOE, LILIES, SURFACE, SHORE, SIDE, REEDS, SITUATION NEW WORDS: DESK, KITCHEN, ORANGES, PORCH, STEAM, TEMPLE, CRUST









79

14. Tennis Game

Unpleasant. A cute guy you met invited you to play tennis but you're not in good form. As the game begins you make clumsy mistakes and feel more and more embarrassed. You can't seem to get the feel of the game as the ball lands in the net again. You breathe hard, trying to catch your wind between volleys, your face flushed. As you lose the match, you shout hollow congratulations and manage a smile. You redden as you see people watching you; you just want to get off the court.

Pleasant. You are playing tennis with a good friend, feeling energetic and in top form. Running hard across the court, every move feels right--you're playing well. You delight in the solid feel of each shot as you slam the ball across the net. You breathe hard, catching your wind between volleys, your face flushed. Your partner shouts congratulations with a big smile as you win another match. You walk off the court feeling great, ready for a shower and a cold drink.



STUDY WORDS: TENNIS, FORM, FEEL, NET, SMILE, MATCH, VOLLEYS NEW WORDS: JUDGE, PILLOW, ROBIN, POUND, JUNGLE, BRACELET, CUPS









80

15. Favorite Rock Star

Unpleasant. You are over visiting with friends, while the radio plays in the background. Suddenly the DJ says that your favorite musician was just shot and killed. You sit frozen and numb; some of the others sit quietly crying in the room. You can't believe this is true. Your spirit sinks as you realize how empty this makes the future. The news is so upsetting that you don't know what to do next.

Pleasant. You are visiting with friends, while the radio plays in the background. Suddenly the DJ announces that you won a trip to meet your favorite musician. You sit thrilled and amazed as the others scream and dance around the room. You can't believe it! Your spirit soars as you realize the future holds this great meeting. The news is so exciting that you don't know what to do next.



STUDY WORDS: RADIO, BACKGROUND, OTHERS, ROOM, SPIRIT, FUTURE, MUSICIAN NEW WORDS: DIRT, TRIFLE, ABILITY, COTTAGE, POTATO, CANDY, BOTTLE









81

16. Driving in Traffic

Unpleasant. You are driving your car across town in heavy traffic, on a hot summer day. The windows are down, and you are sweating heavily in the hot summer sun. A wasp blows in striking you in the face, and you accidentally hit the horn. You jerk the wheel, swerving dangerously out of your lane, and just miss a car. You wave around your head and neck, anxiously aware of the buzzing behind you. You feel a sharp sting on your neck and strike at it, trying to get it off.

Pleasant. You are out for a drive heading across town in your

beautiful new sports car. The windows and top are down; you bask in the sun and some admiring glances. You notice an old friend walking along the sidewalk and you blow your horn. His head turns towards you and you notice the buzzing interest among his group. Turning the wheel, you move neatly into the curb lane, and drive up beside him. The day is now perfect--he climbs in and you head to the country for a ride.



STUDY WORDS: TOWN, CAR, HEAD, BUZZING, WHEEL, LANE, HORN NEW WORDS: BOOT, OVEN, STAGE, MACKEREL, BLEACH, SOLDIER, BUBBLE









82

17. Hotel Party

Unpleasant. You are at a large exclusive party which packs several floors of an old hotel. Out a window you suddenly see fire exploding into the sky up on the hotel roof. Someone runs in yelling that the hotel is on fire from the fireworks up there. Incredibly fast, the halls fill with smoke and panic spreads through the crowd. In disbelief, you stare at windows shattering in the repeated explosions. Your lungs and eyes burn as you struggle towards a balcony, gasping for air.

Pleasant. At the hotel, you two leave the party and wander up to the rooftop garden. It's a beautiful clear night; you stand on the roof gazing up at the sky. Under the full moon, soft kisses along your throat fill you with warmth. The noisy crowds down in the halls filled with smoke seem very far away now. Back inside, you lose yourselves in soft caresses and explosions of desire. His long, deep kisses send thrills of pleasure all through your body.



STUDY WORDS: PARTY, HOTEL, ROOF, SKY, HALLS, SMOKE, EXPLOSIONS NEW WORDS: CLIP, PASTRY, BASEBALL, WEED, TODDLER, NOVEL, GRASS









83
18. Ski Slopes

Unpleasant. It is the dead of winter, a windy day with dark skies and freezing temperature. You glumly climb into the chair on the ski lift, behind your best friend. Far below you great clouds of snow are being blown up off the deserted slopes. Jerking up the hill, a strong gust of icy wind hits the ski lift hard. You watch helplessly as your friend's chair swings violently on the cable. You scream, staring in disbelief, as your friend is thrown from the lift.

Pleasant. It is winter and a beautiful day with clear blue skies and mild temperature. Full of energy, you gaily hop onto the ski lift with your best friend. You watch expert skiers far below raise up clouds of snow on the long slopes. As you glide up on the lift a sudden gust of wind knocks your friend's hat off. The chair swings on the cable as your friend reaches out to grab it, laughing. Soon the hat is only a small bright spot on the sparkling white hill below.



STUDY WORDS: WINTER, TEMPERATURE, CLOUDS, SNOW, GUST, LIFT, SLOPES NEW WORDS: HAMBURGER, CAKE, NATION, BEAR, CUBICLE, MERCHANT, COMBS










84

19. Mexican Vacation

Neutral. You are on vacation on the Mexican coast, arranging some deep sea fishing. Standing on the deck, you are aware of the constant rocking motion of the boat. The guide explains that big tuna inhabit the deep shadows around the shoals. You reflect that speaking Spanish all the time will probably get to be a pain. You mentally check your supplies, adding the pills that prevent sea sickness. Hopefully the weather will be good, but you'll pack a raincoat just in case.

Unpleasant. You are on vacation on the Mexican coast, out doing some deep sea fishing. Standing on the deck, the constant motion of the boat makes you lightheaded. Nausea creeps over you as you search for tuna in the shadows under the waves. Suddenly another tourist makes a bad cast and a hook sinks into your cheek. You scream with pain and several Mexicans start speaking rapid Spanish at you. Blood drips down onto your shirt and you feel like you're going to be sick.



STUDY WORDS: VACATION, COAST, DECK, MOTION, TUNA, SHADOWS, PAIN NEW WORDS: BLADE, VENTURE, CHIMNEY, PORTIONS, HAWK, GERMS, CAVE









85
20. Shower

Pleasant. It is early morning as you reach for a towel to dry off after your shower. The fluffy towel feels pleasantly soft as you dry yourself with long strokes. You rub your hair lightly and dry your face and neck, feeling more awake now. Stepping out of the shower stall you catch the fragrant odors in the bathroom. Finally, you're done; your skin feels smooth, your body feels fresh and clean. Refreshed and ready for the day, you put on your robe and go back to your room.

Unpleasant. You dry off from your shower with a thin cotton towel in the old locker room. The towel turns into a small soggy lump after a few half-hearted strokes. You are vaguely aware of something crawling on your hair or maybe your neck. The narrow stall is dark, humid, and smells of mold and stale body odors. You jump and your skin crawls as you feel a bug scuttle down your naked body. Your heart pounds as you nervously scan the walls and floor in the dark stall.



STUDY WORDS: SHOWER, TOWEL, STALL, ODORS, SKIN, BODY, STROKES NEW WORDS: LOG, HINGES, SATIN, DRILL, VEST, CHALK, KIDNEY









86

21. Moving into Apartment/ Rat and Baby

Neutral. The apartment is on a busy, tree-lined avenue in a

pleasant part of the city. The guys are gradually unpacking the truck parked in front along the curb. They carry in the crib, the beds, and chairs, while you go out for food. It is near dinner time and everyone is ready to relax for a minute and eat. You give the kids some toys and a cookie, and see that the baby is asleep. As you come back with dinner you see the truck is empty and everyone's inside.

Unpleasant. The apartment is on a crowded, dirty avenue in the slums of the city. It is late, but no adults are home in the dark, empty apartment. A large rat climbs up into the baby's crib, sniffing around for food. The baby draws back and cries anxiously as the ugly animal sniffs around. Roaches crawl over a broken cookie on the floor amid scattered toys. The rat approaches the frightened, crying child in the dark room.



STUDY WORDS: APARTMENT, AVENUE, CRIB, FOOD, COOKIE, TOYS, BABY NEW WORDS: MONKEY, TORNADO, THEME, BUTTER, BASEMENT, FUNGUS, GARMENT









87

22. Radio Station/Blood Donation

Neutral. You are touring the radio station on campus, fascinated by all the equipment. They show you everything from the phonograph needle to the cot for naps. You are chosen as a volunteer to play a record and blood rushes to your cheeks. You manage to get it started and playing with only a little difficulty. As you pull the lever to lift the needle off, the announcer gives you a smile. You stand back and relax, watching the next person play another record.

Unpleasant. You are sitting in the blood donor station, noticing all the medical equipment. The sheer size of the long needles makes you sick as you sit waiting on the cot. You stare at the dark blood filling the syringe of the volunteer next to you. It is your turn now, and you smile with difficulty as she tightens the tourniquet. You feel your hands sweat as she prepares the syringe; it's taking her forever. A sharp pain shoots through your arm as the needle slowly pierces your skin.



STUDY WORDS: STATION, EQUIPMENT, NEEDLE, COT, BLOOD, VOLUNTEER,

DIFFICULTY

NEW WORDS: CRACK, SQUIRREL, CAMEL, INVASION, MEADOW, SALAD, DISPUTE










88

23. Graduation

Pleasant. Coming down the stairs the morning of your graduation you're happy and excited. As you sit down to breakfast your parents are smiling at each other a lot. Looking down, your eyes fix on a set of car keys sitting in your plate. You jump up from your chair, grabbing them, and run to the front door. Sitting by the curb is an incredible present--a beautiful, shining new car. You just can't believe it! You are so excited you hardly know what to do.

Unpleasant. You move slowly down the stairs the morning of your graduation feeling sick. Determined to attend the ceremony that afternoon, you sit down to breakfast. Noise from the television set hurts your head as you stare at your plate. You feel miserable as you realize you're too sick to even eat your toast. Looking out you see a friend stepping up on the curb with a present for you. Your head throbs and you feel terrible--you are going to miss your graduation.



STUDY WORDS: STAIRS, MORNING, SET, PLATE, CURB, PRESENT, BREAKFAST NEW WORDS: BONE, CABBAGE, SKATER, SHIRT, STATUE, TILE, CABINET









89

24. Party/Nighttime Intruder

Pleasant. Arriving home at night, you hear sounds of a party from across your back yard. You just finished your last final; you are up and really ready to celebrate. Your hand moves fast to pick up the telephone by your bed when it rings. You hear laughter and music behind the voice--it's the cute neighbor you like. Your heart beats with joy as you rush out, the screen door slamming behind you. You walk into the apartment full of music and dancing, and your mood soars.

Unpleasant. Alone and lying awake at night you hear sounds of something out in the yard. You become instantly alert and scarcely breathe as you lie rigid, listening. You hand is lightly shaking as you reach for the telephone sitting by the bed. Your voice dies in your throat as you think you see a shadow moving outside. Your heart pounds in your chest as you hear someone cutting the screen door. The police are finally on the line, and you struggle to give your address.



STUDY WORDS: SOUNDS, YARD, TELEPHONE, BED, HEART, SCREEN, VOICE NEW WORDS: DOLLAR, CHEMIST, HUNGER, MINT, GOOSE, PLOW, FRUIT









90
25. Sailboat

Unpleasant. You duck your head as waves crash over the bow, drenching the whole crew. The wind tears powerfully at the sails, pulling you over in a sideways motion. Holding the rail, you brace yourself as the boat lurches unsteadily forward. The day is cold, the sea dark and churning with dirty caps of gray foam. You take a sharp breath; the ice cold spray cuts hard against your face. You clutch the mast with all your strength--you are going under.

Pleasant. You laugh as the waves race past the bow, spraying most of the crew. The wind pulls powerfully at the sails, increasing the boat's forward motion. Holding to the mast, you steady yourself as your boat surges into the lead. The day is perfect, the sea shining bright blue with flecks of foam. You take a deep breath; the spray feels invigorating, fresh against your face. Around you everyone is shouting and cheering--you are winning the race!



STUDY WORDS: WAVES, BOW, DAY, FOAM, BREATH, SPRAY, CREW NEW WORDS: TOPIC, CORNER, FEVER, MAP, HELMET, RUBY, OWL









91

26. Walk Across Campus

Pleasant. Walking across campus in a hurry, you thread your way among crowds of students. With a shock you spot the attractive guy you met at the library last night. Your attention is drawn to him as he walks the across the lawn towards you. He comes closer and looks up, and as your eyes meet your heart skips a beat. Your pace slows quickly, and you finally come to almost a complete stop. You thrill with delight as he flashes you a smile and stops to say hello.

Unpleasant. Walking across campus in a hurry, you thread your way amid crowds of students. With a shock, you spot your boyfriend and one of your friends holding hands. In disbelief, your attention fixes on them talking closely on the lawn. Your heart pounds as you walk up to them, and you feel upset and agitated. Your pace slows as they look over, startled, and their talk comes to a stop. Taking a deep breath, you fight back your feelings and force yourself to speak.



STUDY WORDS: CAMPUS, STUDENTS, ATTENTION, LAWN, PACE, STOP, SHOCK NEW WORDS: CELERY, CRANE, MUD, SPOON, PICKLE, LEMONADE, MAGAZINES









92

27. Coming Home from Library

Neutral. Coming back from a good study session at the library, it's early evening. Your neck and shoulders feel tight since you've been reading for two hours. You stretch with relief as you put down your books--you're ready to relax. Turning on the radio, you sit down in your favorite chair and open your mail. Later on you test yourself, reviewing parts of the information you covered. Most of it comes back to you, but you decide to review your notes once again.

Unpleasant. Coming back from a long cramming session at the

library, it's late at night. Your shoulders and neck feel stiff from hunching over your work for hours. You put down your books, aware of a dull tension headache behind both eyes. As you sit down to open your mail, the fatigue and worry are overwhelming. You try to rehearse parts of the information you covered, but you can't think. You just can't seem to stop worrying about the big test and your bad grades.



STUDY WORDS: SESSION, LIBRARY, SHOULDERS, HOURS, PARTS, INFORMATION,

MAIL

NEW WORDS: CONQUEST, STONE, DREAM, COMPUTER, GREASE, LUMBER, THOUGHT









93

28. Store Clerk

Neutral. You are opening the register as you start your shift in a nice clothing store. Only a few customers are in your area of the store, so you can take your time. You turn off the alarm, find the key to the drawer lock, and get out a pen. You are all set up now, and you wander over to look through the new blouses. There a beautiful white silk one with a front pocket that you especially like. A customer is browsing through the skirts nearby and you go over to help her.

Unpleasant. You are closing the register after the late shift at a convenience store. The store is in a secluded area, and you don't like being there alone at 2 a.m. At last it's time to close; you turn the lock on the door and set the alarm. Suddenly you look up to see a man's face watching you through the window. You freeze with fear, staring, as he pulls a gun out of his jacket pocket. You jump as he shouts something at you, motioning with a jerk towards the door.



STUDY WORDS: REGISTER, SHIFT, STORE, AREA, LOCK, ALARM, POCKET NEW WORDS: TACKLE, SINK, DIME, ORGAN, RIBBON, BRIDE, BUMPS









94

29. First Day of Class/Final Exam

Neutral. You walk into the classroom on the first day of class, full of confidence. Sitting down, you put your books and papers under your chair and sit back. A booklet is handed out, and a few homework questions for the first assignment. You take out pencil and paper as the organization of the course is outlined. Your brain is full of information as you note down the date of the first test. Walking out after class you reflect that this looks like an interesting course.

Unpleasant. You walk into the classroom to take your final with no confidence at all. Behind in the course, you crammed all night, but there was too much to learn. Opening the test booklet you pale with apprehension as the questions look hard. You find you don't even have a pencil and, embarrassed, have to borrow one. Your brain is filled with confusion as you read the test; you just can't think. Avoiding your professor's eyes, you turn in the half empty exam and leave.



STUDY WORDS: CLASSROOM, CONFIDENCE, BOOKLET, QUESTIONS, BRAIN, TEST,

PENCIL

NEW WORDS: VENDOR, RELUCTANCE, PEANUTS, MEDICINE, LUNCH, MOTEL,

SPINACH




Full Text
123
Table 21
Mean startle magnitude (arbitrary A-D units) for probes presented
early and late in affectively pleasant, neutral, and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
114.6
133.9
111.2
119.9
Late
134.0
149.8
165.1
149.6
Mean
124.3
141.9
138.2
134.8
Table 22
Mean startle
magnitude (arbitrary A-D
units) for probes presented
early and late
in
affectively pleasant.
neutral, and unpleasant texts
for
good and
poor imagery
ability subjects
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
82.0
62.8
102.8
82.5
Late
101.1
125.2
133.3
119.9
Mean
91 .6
94.0
118.1
101.2
Pleasant
Startle Position
Early 147.2
Late 166.9
157.1
Poor Imagers
Neutral
Unpleasant
Mean
204.9
119.6
157.2
174.4
196.9
179.4
189.6
158.3
168.3
Mean


54
Overall Conclusions
In summary, the results of this study show that while encoding
measures varied along the valence dimension, retrieval from memory
appeared to be largely mediated by the presence or absence of emotional
content in the texts. These results were consistent for all measures of
encoding, including read time, heart rate, and blink magnitude and
latency. In addition, recognition and recall measures of memory
retrieval were influenced in a similar fashion by the arousal dimension
of the study texts. An important exception to this general trend was
study word targets primed by study word primes. This research suggests
that variation in emotional content and manipulation of information
processing strategies (e.g., imaginal versus verbal encoding
instructions) work separately and in interaction to affect both the
encoding and retrieval for textual materials. Further study of the
effects of these variables on verbal learning and memory promises to
extend our understanding of these cognitive processes in important ways.


152
Table 49
Mean percent errors for study words primed by a study word, a new word,
or XXXXX for affectively pleasant, neutral, or unpleasant texts.
with and without startle probes, for good and poor imagers
NO STARTLE SUBJECTS
Good Imagers
Pleasant Neutral Unpleasant Mean
Prime Type
Study Word
6.34
7.66
7.67
7.22
New Word
6.95
9.19
9.52
8.56
XXXXX
7.39
5.88
8.16
7.14
Mean
6.89
7.58
8.45
7.64
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
4.38
3.26
3.17
3.60
New Word
11.35
8.90
7.53
9.30
XXXXX
6.95
11.35
11.25
9.85
Mean
7.60
7.84
7.32
7.58


101
36. Tour of New York City
Pleasant. You are on a trip to New York City, touring the sights
for the first time. Your tour route ends on the deck at the Lincoln
Center overlooking the city. You are breathless in admiration at the
spectacular skyscrapers all around. Another special building is a few
streets away and you decide to go there too. Soon you start thinking
about the evening's entertainment and buy a newspaper. You're ready for
another night of exploring great restaurants and nightclubs.
Unpleasant. You are on a trip to New York City, touring the
sights for the first time. Your tour route ends on the deck at the
Lincoln Center overlooking the city. It's very windy up there, and you
feel uncomfortable as the building sways. You feel dizzy, sensing the
building move as you look far down at the streets. You pale at the
sight of the city so far below, and start to lose your balance. Your
breathing is shallow and rapid, your face feels cold and sweaty.
STUDY WORDS: SIGHTS, TIME, ROUTE, CITY, BUILDING, STREETS, TRIP
NEW WORDS: PRIZE, GLASSES, SPONGE, CARROT, VEST, PRAIRIE, LAWYER


97
32. Staff Meeting/Partv
Pleasant. The staff drifts in and the party gets started, now
that the meeting is over. Your mouth waters as you open the pizza and
get out the ice cold cokes. With no hesitation, you put away your
project notes and grab a cold drink. You feel very hungry as you
breathe in the delicious smell of the hot pizza. The clock hits seven,
and silence falls over the room as you all start to eat. The ice cold
coke perfectly complements the pizza's spicy sauce and cheese.
Unpleasant. You feel tense as the staff gathers for a meeting in
the conference room. Your stomach churns and your mouth is dry as you
nervously wait to speak. With some hesitation you look over your notes
again, your hands shaking. You cringe with embarrassment as you stumble
badly over the introduction. They all sit in silence staring at you;
the clock ticks loudly in the silence. Your heart pounds and your face
is burning as you start over once more.
STUDY WORDS: STAFF, MEETING, HESITATION, NOTES, CLOCK, SILENCE, MOUTH
NEW WORDS:
MUSEUM, YARN, WIDOW, RULER, COLISEUM, STOCKINGS, VIOLENCE


107
Table 10continued
VERBAL INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.09
4.60
5.11
4.93
Middle
4.43
4.40
4.58
4.47
Late
4.14
4.23
4.06
4.14
Mean
4.55
4.41
4.59
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
6.09
6.08
6.08
6.08
Middle
5.65
6.27
5.85
5.93
Late
4.99
5.13
5.26
5.12
Mean
5.58
5.83
5.73
5.58


81
16. Driving in Traffic
Unpleasant. You are driving your car across town in heavy
traffic, on a hot summer day. The windows are down, and you are
sweating heavily in the hot summer sun. A wasp blows in striking you in
the face, and you accidentally hit the horn. You jerk the wheel,
swerving dangerously out of your lane, and just miss a car. You wave
around your head and neck, anxiously aware of the buzzing behind you.
You feel a sharp sting on your neck and strike at it, trying to get it
off.
Pleasant. You are out for a drive heading across town in your
beautiful new sports car. The windows and top are down; you bask in the
sun and some admiring glances. You notice an old friend walking along
the sidewalk and you blow your horn. His head turns towards you and you
notice the buzzing interest among his group. Turning the wheel, you
move neatly into the curb lane, and drive up beside him. The day is now
perfecthe climbs in and you head to the country for a ride.
STUDY WORDS: TOWN, CAR, HEAD, BUZZING, WHEEL, LANE, HORN
NEW WORDS: BOOT, OVEN, STAGE, MACKEREL, BLEACH, SOLDIER, BUBBLE


120
Table 19continued
STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1.49
-0.81
-1 .47
-1 .26
Middle
-2.61
-2.08
-3.20
-2.63
Late
-3.36
-2.69
-3.06
-3.04
Mean
-2.48 -1.86 -2.58
Verbal Instructions
-2.31
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.35
-2.58
-1 .84
-1 .59
Middle
-2.97
-4.15
-4.29
-3.81
Late
-2.90
-4.42
-4.56
-3.96
Mean
-2.08
-3.72
-3.57
-3.12


135
Table 35
Mean recognition reaction time (msec) for study words primed by
a study word, a new word, or XXXXX for affectively pleasant, neutral.
or unpleasant texts, with and without startle probes
Type
No
Pleasant
Startle
Neutral
Subjects
Unpleasant
Mean
Study Word
844
876
870
863
New Word
838
933
842
871
XXXXX
906
908
896
903
N-S Priming
- 6
+ 57
-28
+ 17
X-S Priming
+62
+ 32
+ 26
+ 40
Startle Subjects
Pleasant Neutral Unpleasant Mean
Prime Type
Study Word
821
821
849
830
New Word
852
881
807
847
XXXXX
881
868
852
867
N-S Priming
+ 31
+ 60
-42
+ 17
X-S Priming
+ 47
+61
+ 3
+ 37


25
subject was asked to refer to the typed version in order to recall the
text, and then to mark her rating of how she felt while reading the text
during the experiment. Finally, the subject was debriefed, and either
given course credit or signed a voucher to be paid $10 for her
participation in the study. The total experimental session lasted
approximately 2-1/2 hours.
Data Reduction
Encoding Measures
Read time. Read time for each sentence was measured by computer
with a resolution of one msec. The read times for sentences 1 and 2
were then averaged to provide a mean read time for the "early" sentence
pair. Read times were also averaged over sentences 3 and 4, and
sentences 5 and 6, and constituted the "middle" and "late" sentence pair
read times, respectively. The final data set included three mean read
times (for the early, middle, and late sentence pairs) for each of the
study paragraphs, within each affective category. For subjects
receiving startle probes, averages were calculated for each sentence at
each level of the Startle Position factor (early, late, no startle),
within each of the three affective categories.
Heart rate. Inter-beat intervals were converted into heart rate
values using the algorithm of Graham (1979). Because subjects read the
paragraphs at their own rate, the sampling periods for heart rate were
highly variable among sentences. Thus, a real time analysis of the
cardiac wave form across sentences was not possible. Rather, the
average heart rate for each sentence of each text was calculated and


APPENDIX C
IMAGERY ABILITY SCORES FOR INSTRUCTION GROUPS
Imagery Instructions
Verbal
Instructions
Good Imagers
Poor Imagers
Good Imagers
Poor Imagers
40
81
37
84
45
86
43
85
48
87
49
85
58
93
50
92
58
94
51
95
59
95
56
96
62
97
57
98
63
101
60
104
67
105
61
113
69
105
65
121
76
108
76
123
78
111
76
125
79
114
76
79
117
77
147
79
102


PSYCHOPHYSIOLOGY AND MEMORY:
THE ENCODING AND RETRIEVAL OF AFFECTIVE TEXT
By
ELLEN LOUISE SPENCE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1992


12
not represent a simple persistence of the affective response in memory.
That is, the above results will not be found for nonstudy words.
Priming can be defined as facilitation of the response to a test
item by an immediately preceding cue (Ratcliff & McKoon, 1978). This
paradigm has proved a useful tool in the development of theory on how
knowledge is represented in memory, such as the network theory developed
by Anderson (1983). According to network theory, when a concept is
processed cognitively it is activated in memory. This activation
automatically spreads through the information network to other
associated concepts. Episodic priming results from temporary logical
relationships established among concepts when subjects read a particular
text, and has been employed by researchers to demonstrate that memory
for text is represented in logical form as meaning propositions, and
that activation of this associative network follows the logical
structure of these associations. Priming effects are typically
determined by comparing the reaction time for memory recognition of a
target stimulus preceded by a highly associated word, or "prime",
relative to that for a neutral or control condition.
In Lang's original presentation of bio-informational theory (1977;
1979), he followed Pylyshyn (1973) in suggesting that there is a
general, propositional sub-organization in imaginal/affective
information structures. To examine this assumption, an episodic
recognition priming paradigm was adapted from the work of Ratcliff and
McKoon (1978, 1981; McKoon & Ratcliff, 1980) for the present study.
Their research explored the dynamics of spreading activation in logical
information structures in neutral narrative text. They observed that


51
In summary, development of the affective information structure, or
text schema, was hypothesized to occur over time during encoding. Late
in the encoding process, a consistent pattern of results emerged in
which affectively unpleasant content resulted in longer read times,
greater heart rate decrease, and larger and faster startle responses,
compared to paragraphs rated as pleasant. For read time and heart rate,
the negative versus positive valence effect was most evident for
subjects reporting poor imagery ability and those receiving verbal
instructions, while imagery processing strategies tended to minimize
these effects. Instructions and individual differences variables did
not interact with startle magnitude effects, but for startle latency,
good imagers showed the linear effect of emotional content earlier in
the processing task.
Retrieval Measures
While measures of encoding of affective texts appeared to vary
with the negative versus positive valence of the materials, retrieval
measures varied more according to whether or not paragraphs were
emotional in content. Both recognition and free recall measures showed
greater facilitation for affective compared to neutral paragraphs.
Recognition of study word targets varied with presence or absence
of emotional content as well as imagery ability and instructions. Good
imagers receiving imagery instructions showed faster and more accurate
recognition for words from emotionally evocative compared to neutral
paragraphs. These results are consistent with the idea that imaginal
encoding strategies are more likely to activate arousing emotional


128
Table 29
Mean recognition reaction time (msec) for study word and new
word targets for affectively pleasant, neutral, and unpleasant texts
for subjects receiving imagery and verbal instructions
Study Word Targets
Pleasant Neutral Unpleasant Mean
Instructions
Imagery
830
871
840
847
Verbal
884
892
865
880
Mean
857
882
852
864
New Word Targets
Pleasant
Neutral
Unpleasant
Mean
Instructions
Imagery
840
859
845
848
Verbal
895
914
909
906
Mean
867
887
877
877


39
Target recognition latency. Study word targets were words that
appeared in the preceding text and required a "yes" response. New word
targets were non-text-related words which required a "no" response.
Recognition reaction time for study and new word targets was not
significantly different, F < 1. However, emotional content, imagery
ability, and instructions interacted with target type, F (2,92) = 4.36,
p < .03 in the omnibus test, Valence (quadratic) X Instructions X
Imagery Ability X Target Type F (1,46) = 4.62, p < .04. Thus, separate
follow-up analyses were done for new word and study word targets. As
expected, no significant effects of the experimental factors were found
for new word targets, all Fs < 1.9, all ps > .17.
Recognition memory for study word targets followed a pattern over
affective contents similar to free recall (pleasant = 857 msec, neutral
= 881 msec, unpleasant = 853 msec). In this case, however, the
statistical test was not significant, Valence F (2,92) = 1.90, p = .17;
Valence (quadratic) F (1,46) = 2.37, p = .13. Nevertheless, latencies
did interact significantly with all of the main experimental factors,
Valence X Instructions X Imagery Ability F (2,92) = 4.24, p < .03,
Valence (quadratic) X Instructions X Imagery Ability F (1,46) = 5.30,
p < .03. Examination of Table 4 reveals that target recognition was
clearly faster for emotional than neutral content only for good imagers
receiving imagery instructions, valence (quadratic), F (1,46) = 5.94,
p < .02). Subsequent separate tests of affective content differences
for other groups were not significant, F (1,46) = 2.04, p = .16 for poor
imagers receiving verbal instructions,; all other Fs < 1.


43
AFFECTIVE CATEGORY
Figure 3. Mean reaction times for Study-Study and New-Study
prime-target pairs following affectively pleasant, neutral and
unpleasant paragraphs.


121
Table 20
Mean heart rate change (beats per minute) for affectively pleasant,
neutral and unpleasant texts, with and without startle probes.
for good and poor imagers receiving verbal and imagery instructions
NO STARTLE SUBJECTS
Imagery Instructions
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
-0.28
1
u>
o
CD
-0.99
-1 .45
Poor
-1.18
+ 0.13
-1.24
-0.76
Mean
-0.73
-1 .47
-1.12
-1.10
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
-3.76
-3.75
-4.24
-3.92
Poor
-3.74
-4.33
-4.25
-4.11
Mean
-3.76
-4.04
-4.25
-4.01


o
0)
CO
E
65
60
55
50
# Imagery
O Verbal
Pleasant Neutral Unpleasant
Pleasant Neutral Unpleasant
EARLY LATE
Figure 2. Startle latency for verbal versus imagery instructions, for probes presented
early and late in affectively pleasant, neutral and unpleasant texts.
LO
^3


APPENDIX D TABLES OF MEANS FOR DEPENDENT MEASURES 103
APPENDIX E AFFECTIVE RATINGS FOR PARAGRAPHS 156
APPENDIX F SELF-ASSESSMENT MANIKIN (SAM) 161
APPENDIX G SUMMARY OF RESULTS FOR PILOT PRIMING STUDIES 164
REFERENCES 166
BIOGRAPHICAL SKETCH 171
IV


49
of affective picture slides (Greenwald et al., 1989). However,
research also exists in which subjects listened to affective and neutral
narratives over headphones (Cook et al., 1988) and showed heart rate
acceleration in response to unpleasant content, despite the fact that
they were engaged in a task requiring information intake. The visual
aspect of the experimental task is common to both studies showing heart
rate deceleration to unpleasant stimuli, suggesting that the sensory
aspect of the experimental task may modulate the direction of heart rate
change in response to affective content.
Heart rate was expected to increase late in the encoding process
as a result of the internal processing demands in the imagery
instructions condition, consistent with prior research on imaginal
processing of emotional content (Lang et al., 1980; Vrana & Lang, 1990).
While study results confirmed the predicted heart rate decrease in the
verbal instructions condition, heart rate did not increase for the
imagery instructions condition. Nevertheless, the lack of heart rate
deceleration in the imagery instructions condition might be construed as
evidence for acceleratory effects cancelling out the deceleration
component.
While the read time and heart rate results showed consistent
effects of emotional content on subject responses, these measures are
nevertheless limited as direct measures of emotion because they are
inherently susceptible to a variety of other factors, as noted above.
Read time and heart rate are measures of perceptual processing and
stimulus intake, and thus vary with instructions, characteristics of the
experimental materials, and individual predisposition of the subjects.


APPENDIX A
QUESTIONNAIRES, INSTRUCTIONS AND DEBRIEFING STATEMENT
OMI Vividness of Imagery Scale
The aim of this test is to determine the vividness of your
imagery. The items of the test will bring certain images to your mind.
You are to rate the vividness of each image by reference to the
accompanying rating scale which is shown at the bottom of the page. For
example, if your image is "vague and dim" you give it a rating of 5.
Record your answer in the brackets provided after each item. Just write
the appropriate number after each item. Before you turn to the items on
the next page, familiarize yourself with the different categories on the
rating scale. Throughout the test, refer to the rating scale when
judging the vividness of each image.
* Rating 1 Perfectly clear and as vivid as the actual experience *
* Rating 2 Very clear and comparable in vividness to the actual *
* experience *
* Rating 3 Moderately clear and vivid *
* Rating 4 Not clear or vivid, but recognizable *
* Rating 5 Vague and dim *
* Rating 6 So vague and dim as to be hardly discernable *
* Rating 7 No image present at all, you're only "knowing that" you *
* are thinking of the object *
55


I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Lra Fischler
Professor of Psychology
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
This dissertation was submitted to the Graduate Faculty of the
College of Health Related Professions and to the Graduate School and was
accepted as partial fulfillment of the requirements for the degree of
Doctor of Philosophy.
May, 1992
P-
Dean, College of Health Related Professions
Dean, Graduate School


68
3. Hot Summer Night
Unpleasant. Lying in the dark naked on a night too hot for
covers, you can't get any sleep. You have the ceiling fan on high
trying to get comfortable in the humid heat. Suddenly you feel a huge
palmetto bug drop from the ceiling onto your chest. You can feel the
insect crawl rapidly over your chest and up onto your neck. You leap
onto the carpet in panic and disgust, quickly switching on the lamp.
Scanning the bedclothes, you hesitantly lift the sheets looking for the
insect.
Neutral. You are in bed, drifting off to sleep on a summer night
too warm for covers. The ceiling fan cools off the summer heat, and the
ice cream helped too. You lie there relaxing, watching occasional car
lights play on the ceiling. The weekend starts tomorrow, and you think
about things you'd like to do. The phone ringsyou switch on the lamp
and walk across the carpet to get it. It's a good friend calling long
distance, and you sit down to have a talk.
STUDY WORDS: SLEEP, COVERS, FAN, HEAT, LAMP, CARPET, CEILING
NEW WORDS:
EAGLE, STEAK, VICTORY, WALNUT, LAKE, BOWL, TREE


69
4. Dental Appointment
Unpleasant. Your anxiety is raised by a call confirming your
appointment with the dentist. Walking in, you are instantly aware of
the whine of the drill from a back room. As you climb into the chair
you glimpse the pointed instruments on the tray. You wince as he drags
the sharp steel probe along your sensitive gumline. Hands wet with
sweat, you feel the long needle pierce the back of your mouth. As he
approaches you with the whining drill you grip the chair hard.
Neutral. You get a phone call that your appointment has been
cancelled with the dentist. Now you have some extra time and you decide
to have a leisurely breakfast. You settle into a comfortable chair,
reaching for a bagel on the tray. You turn on the morning news to get
the latest developments on the Iran probe. Putting the last bite into
your mouth you wipe your hands on your napkin. You look through the
paper and read the funnies before leaving for class.
STUDY WORDS: APPOINTMENT, DENTIST, CHAIR, TRAY, HANDS, MOUTH, PROBE
NEW WORDS: SOCKS, PUPPY, KNOB, RAILWAY, BOXES, DINNER, PLANET


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EDM1POWJ0_UZV7AP INGEST_TIME 2015-03-31T17:52:59Z PACKAGE AA00029754_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


159
TABLE 55
Mean dominance ratings for pleasant, neutral, and unpleasant texts
for subjects with good and poor imagery ability
receiving verbal and imagery instructions
Pleasant Neutral Unpleasant
Instructions
Good
Poor
Good
Poor
Good
Poor
Imagery
7.4
6.7
4.9
5.2
2.5
2.6
Verbal
7.1
6.6
5.5
4.3
2.2
2.4
TABLE 56
Contingency table comparing affective classification of texts
based on a priori (pilot-based) and experimental subject rating data
A PRIORI RATINGS
SUBJECT RATINGS
Pleasant
Neutral
Unpleasant
Total
Pleasant
474
267
36
777
Neutral
33
213
144
390
Unpleasant
9
18
858
885
Total
1038
498
516
2052
Numbers of texts: (57 subjects X 36 texts = 2052 total ratings)
The number of texts that remained in the a priori affective
categories according to ratings by the present sample appears on the
diagonal of this table. The number of texts that changed categories for
the present sample appears in the rows. For example, of the texts
classified as affectively unpleasant in the validation study, 858


REFERENCES
Anderson, J. R. (1976). Language, memory, and thought. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Anderson, J. R. (1983). The architecture of cognition. Cambridge, MA:
Harvard University Press.
Anderson, J. R., & Bower, G. H. (1973). Human associative memory.
Washington, D.C.: Hemisphere Press.
Anthony, B. J. (1985). In the blink of an eye: Implications of reflex
modification for information processing. In P. K. Ackels, J. R.
Jennings, & M. G. H. Coles (Eds.), Advances in psychophysiology
(Vol. 1, pp. 167-218). Greenwich, CT: JAI Press.
Balaban, M. T., Losito, B., Simons, R. F., & Graham, F. K. (1986).
Offline latency and amplitude scoring of the human reflex eyeblink
with Fortran IV (computer program abstract). Psychophysiology.
23, 612.
Berg, W. K., & Davis, M. (1984). Diazepam blocks fear-enhanced startle
elicted electrically from the brainstem. Physiology and Behavior.
32, 333-336.
Bower, G. H., Gilligan, S. G., & Monteiro, K. P. (1981). Selectivity
of learning caused by affective states. Journal of Experimental
Psychology: General. 110. 451-473.
Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1990). Startle reflex
modification: Emotion or attention? Psychophysiology, 27, 513
522.
Bradley, M. M., Greenwald, M. K., Petry, M. C., & Lang, P. j. (1991).
Remembering pictures: Pleasure and arousal in memory.
Unpublished manuscript, University of Florida.
Bradley, M. M., Spence, E. L., & Lang, P. J. (1992). Retrieval effects
for affective texts. Unpublished manuscript, University of
Florida.
Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition. 5, 73-
99.
166


74
9. End of School Term
Neutral. It's the end of another school year and you are going
home for the summer. You have been packing, and just finished finals
for your courses this term. You start carrying plants and clothes and
suitcases out to your parents' car. Everything is finally loaded up and
you drive away, waving to your roommate. Leafing through a magazine in
the back seat you think about a summer job. Thinking of the long trip
ahead, you settle down in the comfortable car.
Unpleasant. It's the end of the school year and you have to go
home for the summer. You feel down, having just told your parents you
failed two finals this term. You wearily drag plants and clothes and
suitcases out to their car in silence. Your father says that maybe you
should quit school and you feel more depressed. Sitting in the car, you
think about the boring job you'll have to start soon. You sigh and
settle down in the back seat, dreading the long trip ahead of you.
STUDY WORDS: YEAR, SUMMER, FINALS, TERM, PLANTS, CLOTHES, SUITCASES
NEW WORDS: CHIMNEY, LANTERN, GLUE, MELON, RECEIPT, CANAL, SILVER


APPENDIX F
SELF-ASSESSMENT MANIKIN (SAM)


APPENDIX G
SUMMARY OF RESULTS FOR PILOT PRIMING STUDIES
The two pilot studies completed prior to the present experiment
suggested that affective content altered the episodic priming effects
demonstrated by Ratcliff and McKoon. The design and results of the two
studies are briefly summarized below:
Study 1
The initial study was adapted from a priming paradigm utilized by
Ratcliff and McKoon (1981), in which subjects read sets of two three-
line paragraphs followed by a test list of prime-target pairs. Subjects
responded to the target word in prime-target pairs, and recognition RT
was compared for study words primed by a word from the same paragraph,
compared to RT for study words primed by a word from the other
paragraph.
Ratcliff and McKoon's paradigm was replicated as closely as
possible. A copy of the texts and test lists was obtained from those
authors and their methods of presentation for these materials were
followed as closely as possible, with the exception that in some cases
emotionally evocative paragraphs were substituted for their texts.
Study 2
Another paradigm developed by Ratcliff and McKoon (1978) was
adapted for the second study. In this paradigm, episodic priming was
164


158
TABLE 54
Mean arousal ratings for pleasant, neutral, and unpleasant texts
for subjects reporting good and poor imagery ability
Pleasant
Neutral
Unpleasant
Arousal Ratings
Good
6.7
4.6
7.8
Poor
5.8
4.5
7.2
Dominance Ratings
Mean dominance ratings are summarized in Table 55 for affectively
pleasant, neutral, and unpleasant texts. Dominance ratings were highest
for pleasant texts and decreased for texts rated as more unpleasant,
yielding a significant overall effect of valence, F (2,94) = 347.39,
p < .0001. While imagery ability did not affect dominance ratings for
affectively pleasant or unpleasant texts, good imagers gave higher
dominance ratings than poor imagers for neutral paragraphs in the verbal
instructions condition (Good = 5.5; Poor = 4.3). However, in the
imagery instructions condition, poor imagers gave higher dominance
ratings than good imagers for neutral texts (Good = 4.9; Poor = 5.2),
resulting in a significant Valence X Instructions X Imagery Ability
interaction: F (2,94) = 3.63; p < .04.
Stability of text ratings between a priori and subject classification:
A contingency table was calculated in order to determine how many
texts changed affective classification, and which affective categories
showed greater change, comparing the previous validation study to the
present experiment (see Table 56).


117
Table 18
Mean heart rate change (beats per minute) for early, middle, and late
sentence pairs of affectively pleasant, neutral and unpleasant
texts, with and without startle probes, for good and poor imagers
NO STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1.20
-3.06
-1 .24
-1 .83
Middle
-2.38
-3.37
-3.12
-2.96
Late
-2.48
-3.83
-3.49
-3.27
Mean
-2.02
-3.42
-2.62
-2.69
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1 .03
-1.06
-1.25
-1.11
Middle
-3.01
-2.82
-3.35
-3.06
Late
-3.35
-2.41
-3.64
-3.13
Mean
-2.46
-2.10
-2.75
-2.46


42
Table 5
Mean percent errors for recognition of study word targets from
pleasant, neutral and unpleasant texts, for subjects receiving
verbal and imagery instructions
Pleasant
Neutral
Unpleasant
Instructions
Mean
Imagery
8.65
6.79
8.00
7.81
Verbal
7.05
11.19
8.70
8.98
Mean
7.85
9.00
8.45
In the overall analysis, emotional content interacted with priming
condition to affect reaction time (see Figure 3), Valence X Priming
Condition F (2,92) = 8.52, p < .01; Valence (linear) X Priming
Condition, F (1,46) = 6.56, p < .02; Valence (quadratic) X Priming
Condition, F (1,46) = 9.52, p < .01. These effects were followed up by
separate tests for each priming condition.
Study-Study. Recognition reaction time was fastest for pleasant
and slowest for unpleasant content when the prime-target pair consisted
of study words (SS) (pleasant = 832 msec, neutral = 849 msec, unpleasant
= 859 msec), Valence F (2,92) = 1.14, p = .32 in the omnibus analysis;
Valence (linear) F (1,46) = 4.71, p < .04. Recognition memory for SS
prime-target pairs did not vary with imagery ability, instructions, or
startle probe condition.
New-Studv. Reaction time for recognition of study word targets
primed by new words (NS) was affected by emotional content in a
different manner--reaction time was relatively slower for neutral
compared to emotional (either pleasant or unpleasant) content
(pleasant = 845 msec, neutral = 907, unpleasant = 825 msec),


87
22. Radio Station/Blood Donation
Neutral. You are touring the radio station on campus, fascinated
by all the equipment. They show you everything from the phonograph
needle to the cot for naps. You are chosen as a volunteer to play a
record and blood rushes to your cheeks. You manage to get it started
and playing with only a little difficulty. As you pull the lever to
lift the needle off, the announcer gives you a smile. You stand back
and relax, watching the next person play another record.
Unpleasant. You are sitting in the blood donor station, noticing
all the medical equipment. The sheer size of the long needles makes you
sick as you sit waiting on the cot. You stare at the dark blood filling
the syringe of the volunteer next to you. It is your turn now, and you
smile with difficulty as she tightens the tourniquet. You feel your
hands sweat as she prepares the syringe; it's taking her forever. A
sharp pain shoots through your arm as the needle slowly pierces your
skin.
STUDY WORDS: STATION, EQUIPMENT, NEEDLE, COT, BLOOD, VOLUNTEER,
DIFFICULTY
NEW WORDS: CRACK, SQUIRREL, CAMEL, INVASION, MEADOW, SALAD, DISPUTE


19
Primes. All four classes of targets (Study word, New word,
******t ancj ======) were primed by three alternatives: A Study word, a
New word, or a series of X's. The New and XXXXX primes served as
alternative control conditions. The priming condition in which a study
word was primed by a string of X's was eliminated as a control, or
neutral, prime for reasons detailed by Jonides and Mack (1984), in which
important criteria for the characteristics of neutral primes are
outlined. For example, they argue that non-word primes should not be
compared with word primes because of the differential processing
requirements and possible expectation sets that accompany these two
types of priming stimuli. In addition, they caution against the use of
a prime which is repeated on each trial (such as the XXXXX primes) as a
comparison condition for a prime (such as a new or study word), which is
not repeated. Again, differential processing requirements as well as
possible differences in the alerting, or warning, functions of these two
type of prime were considered to vary significantly. Because new word
primes satisfied the above criteria and XXXXX primes did not, the
analyses of priming effects were restricted to the comparison of
recognition reaction time for New-Study with Study-Study prime-target
pairs.
An exemplar of a study paragraph and its associated test list are
provided in Figure 1. Note that in the presentation of this text on the
subject monitor during the experiment, each sentence was presented in a
single line of text, so that the paragraph appeared as six lines of
text. The format of the present document prevents this presentation.


88
23. Graduation
Pleasant. Coming down the stairs the morning of your graduation
you're happy and excited. As you sit down to breakfast your parents are
smiling at each other a lot. Looking down, your eyes fix on a set of
car keys sitting in your plate. You jump up from your chair, grabbing
them, and run to the front door. Sitting by the curb is an incredible
presenta beautiful, shining new car. You just can't believe it! You
are so excited you hardly know what to do.
Unpleasant. You move slowly down the stairs the morning of your
graduation feeling sick. Determined to attend the ceremony that
afternoon, you sit down to breakfast. Noise from the television set
hurts your head as you stare at your plate. You feel miserable as you
realize you're too sick to even eat your toast. Looking out you see a
friend stepping up on the curb with a present for you. Your head throbs
and you feel terribleyou are going to miss your graduation.
STUDY WORDS: STAIRS, MORNING, SET, PLATE, CURB, PRESENT, BREAKFAST
NEW WORDS: BONE, CABBAGE, SKATER, SHIRT, STATUE, TILE, CABINET


89
24. Partv/Nighttime Intruder
Pleasant. Arriving home at night, you hear sounds of a party from
across your back yard. You just finished your last final; you are up
and really ready to celebrate. Your hand moves fast to pick up the
telephone by your bed when it rings. You hear laughter and music behind
the voiceit's the cute neighbor you like. Your heart beats with joy
as you rush out, the screen door slamming behind you. You walk into the
apartment full of music and dancing, and your mood soars.
Unpleasant. Alone and lying awake at night you hear sounds of
something out in the yard. You become instantly alert and scarcely
breathe as you lie rigid, listening. You hand is lightly shaking as you
reach for the telephone sitting by the bed. Your voice dies in your
throat as you think you see a shadow moving outside. Your heart pounds
in your chest as you hear someone cutting the screen door. The police
are finally on the line, and you struggle to give your address.
STUDY WORDS: SOUNDS, YARD, TELEPHONE, BED, HEART, SCREEN, VOICE
NEW WORDS: DOLLAR, CHEMIST, HUNGER, MINT, GOOSE, PLOW, FRUIT


66
Text Materials
1. Evening Walk
Unpleasant. One night you and your roommate go for a walk through
a bad neighborhood. You notice several dark figures slip quietly
between two buildings just ahead. You stop and listen closely; beads of
sweat break out on your forehead. Your breath catches as you hear soft
footsteps nearby, and you both panic. Echoes of your footsteps break
the night's stillness as you run for your car. At the car you're
fumbling with your keys when two guys grab your roommate.
Neutral. One night you and your roommate take a relaxed walk
through the neighborhood. You've been studying all evening and it feels
good to get up and walk around. There is a light sweat on your forehead
and the cool breeze feels refreshing. Quiet and preoccupied with your
studies, you have no particular desire to talk. A lone car drives by
momentarily creating echoes in the night's stillness. You walk back
home, thinking maybe you'll read one more chapter before bed.
STUDY WORDS: ROOMMATE, WALK, SWEAT, FOREHEAD, STILLNESS, ECHOES,
NEIGHBORHOOD
NEW WORDS: RECORD, WINNERS, PARROT, IMAGINATION, BARREL, GEESE,
PLASTER


147
Table 44
Mean percent errors for studv
words primed bv a studv
word, a new word.
or XXXXX for affectively pleasant.
neutral, or unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Prime Type
Study Word
6.81
5.05
6.83
6.23
New Word
7.92
9.00
8.22
8.38
XXXXX
10.32
12.91
8.49
10.58
Mean
8.35
8.99
7.85
8.40
Table
45
Mean percent errors for studv
words primed bv a studv
word, a new word.
or XXXXX for
affectively pleasant.
neutral, or unpleasant texts,
for qood and poor imaqers
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Prime Type
Study Word
7.44
5.28
8.47
7.06
New Word
8.04
7.41
8.17
7.87
XXXXX
9.06
13.78
9.86
10.90
Mean
8.18
8.82
8.83
8.61
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Prime Type
Study Word
6.22
4.82
5.15
5.40
New Word
8.41
10.59
7.67
8.89
XXXXX
7.92
12.04
10.79
10.25
Mean
7.52
9.15
7.87
8.18


APPENDIX D
TABLES OF MEANS FOR DEPENDENT MEASURES
Table 6
Mean reading time (sec) for early, middle, and late sentence pairs
of affectively pleasant, neutral and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.49
5.35
5.59
5.48
Middle
5.08
5.27
5.30
5.22
Late
4.63
4.72
4.79
4.71
Mean
5.07
5.12
5.23
5.14
Table 7
Mean reading
time (sec)
for early, middle, and late sentence
of affectively pleasant, neutral
and unpleasant
texts
for qood and
poor imagery ability subjects
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.10
4.77
5.16
5.01
Middle
4.63
4.65
4.84
4.70
Late
4.35
4.44
4.39
4.39
Mean
4.69
4.62
4.80
4.70
103


29
5&6). Read times were subjected to ANOVA, and heart rate data were
analyzed in an ANCOVA, the 3-second pre-trial heart rate average serving
as the covariate for heart rate change scores.
Recall that a subset of 25 subjects in this study received startle
probes while they read the study paragraphs. For these subjects,
separate analyses of read time and heart rate included the additional
factor of startle position (Early, Late, No Startle) because
presentation of startle probes might be expected to affect encoding
processes for the study paragraphs. All data analyses of the complete
sample included group (Startle Probe vs. No Startle Probe) as a factor,
however, in order to detect any differences related to this variable.
Startle magnitude and latency. A2X2X3X2 ANOVA was carried
out for each of these dependent measures. The between-subject factors
were Imagery Ability and Instructions; the within-subject factors were
Affective Content and Startle Position (Early, Late).
Retrieval Measures
Free recall. These data were subjected to a 2 X 2 X 2 X 3 ANOVA
with the between-subject factors of Imagery Ability, Instructions, and
Probe Group; the within-subject factor was Affective Content.
Episodic Recognition for Target Words. A2X2X2X3X2
factorial design included the between-subject factors of Imagery
Ability, Instructions, and Group. The within-subjects variables
consisted of Affective Content and Target Type (study word, new word).
Non-word targets were not included in these analyses because of the
widely different characteristics between the non-verbal and verbal


154
Table 50
Mean percent errors for study words primed by a study word, a new word,
or XXXXX for affectively pleasant, neutral, or unpleasant texts,
with and without startle probes, for imagery and verbal instructions
NO STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
3.36
6.17
4.32
4.62
New Word
10.89
4.67
9.67
8.41
XXXXX
5.51
5.82
8.29
6.54
Mean
6.59
5.55
7.43
6.52
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
7.36
4.75
6.52
6.21
New Word
7.52
13.42
7.38
9.44
XXXXX
8.83
11.41
11.12
10.45
Mean
7.91
9.86
8.34
8.70


133
Table 33
primed bv a
study word.
a new word.
or XXXXX for affectively
leasant, neutral, or unpleasant texts
; for good and
poor imaqers
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
808
813
830
817
New Word
808
894
799
834
xxxxx
876
862
842
860
N-S Priming
0
+81
-31
+ 17
X-S Priming
+ 68
+ 49
+ 12
+ 43
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
857
884
889
877
New Word
882
920
850
884
XXXXX
911
914
906
910
N-S Priming
+25
+ 36
-39
+ 7
X-S Priming
+ 54
+ 30
+ 17
+ 34


62
the trials--so that you don't have to reach for them when you want to
make a response.
We will go through two practice trials now to be sure you
understand how the experiment goes.
Do you have any questions?
Present practice paragraphs and continue specialized instructions:
IMAGERY INSTRUCTIONS:
Tell me about the image you formed.
Were you aware of knocking on the door? Movement? Feeling of knocking?
Could you see the faces? The candles on the cake? Were you aware of
feeling surprised? Happy? REINFORCE AND INSTRUCT AS PER IMAGERY 18
VERBAL INSTRUCTIONS:
Repeat the first sentence in the paragraph as precisely as you can.
What age was mentioned in the paragraph?
What had you agreed to do that evening?
Did you smile with happiness or delight?
REINFORCE AND INSTRUCT FOR EXACT WORDS AND IGNORE ANY IMAGERY BEHAVIOR
OR RESPONSES.
Remind subject about baseline periods, sitting quietly not to disturb
physiological recordings, and tell them you'll be right next door.


138
Table 37
Mean recognition RT (msec) for study words primed by a study word, a new
word, or XXXXX for affectively pleasant, neutral, or unpleasant texts,
with and without startle probes, for imagery and verbal instructions
NO STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
827
911
858
865
New Word
821
936
808
855
XXXXX
877
872
908
886
N-S Priming
- 6
+25
-50
-10
X-S Priming
+ 50
-39
+ 50
+21
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
862
841
882
862
New Word
856
930
876
887
XXXXX
934
944
883
920
N-S Priming
- 6
+89
- 6
+ 25
X-S Priming
+ 72
+ 103
+ 1
+ 59


145
Table 43
Mean percent errors for new word and study word targets.
for affectively pleasant, neutral, and unpleasant texts,
for good and poor imagers, receiving imagery and verbal instructions
NEW WORD TARGETS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
5.38
5.96
4.43
5.25
Poor
3.85
4.20
4.43
4.16
Mean
4.61 5.08 4.43
Verbal Instructions
4.71
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
3.00
2.29
5.46
3.58
Poor
2.51
2.87
2.84
2.74
Mean
2.75
2.58
4.15
3.16


95
30. Motorist Survev/Traffic Ticket
Neutral. You have an assignment to survey motorist opinions of
the 55 mph speed limit. With two classmates, you flag down cars on a
street corner under the lights. You stop a policeman for fun and laugh
with surprise when he says he'll answer. He talks about how speeding is
a big mistake, even when the roads are empty. You listen with interest
as he quotes a lot of statistics about accidents. As he drives away,
you think what a great addition to your survey this will be.
Unpleasant. Driving along, a motorist swerves past you going well
over the speed limit. Looking up, you see a police car close behind you
with flashing red lights. You feel a shock of surprise and nervousness
as the policeman pulls you over. Unaware of his mistake, he flatly
accuses you of speeding and reckless driving. As you try to explain, he
angrily cuts you off and tells you to speak in court. You fight to
control your angry feelings as he silently writes you a ticket.
STUDY WORDS: MOTORIST, LIMIT, SURPRISE, POLICEMEN, SPEEDING, MISTAKE,
LIGHTS
NEW WORDS: MOUNTAIN, STRAWBERRY, OXYGEN, KITCHEN, MIXTURE, GRADE,
FLOCKS


78
13. Canoe Ride
Neutral. You lean forward as the canoe glides smoothly along
among the water lilies. Slowing the boat, you scan the surface of the
water close to the shore. You spot a duck nesting on her eggs over to
the side, hidden among the reeds. You hurriedly check the camera--the
timing is critical in this situation. You watch closely, adjusting the
focus as the full nest comes into view. Delighted with your luck, you
get several excellent shots of this rare find.
Unpleasant. You slip out of the canoe into the brackish water
among the water lilies. Unnoticed, a huge alligator glides off the
shore, sliding under the surface. You jump as you spot something big
moving in the reeds over to the side of you. Panic shoots through you
as you suddenly know you're in a dangerous situation. Your heart stops
as you see the alligator now gliding towards you in full view.
Desperate to escape, you run hard, crashing through the reeds and
lilies.
STUDY WORDS: CANOE, LILIES, SURFACE, SHORE, SIDE, REEDS, SITUATION
NEW WORDS: DESK, KITCHEN, ORANGES, PORCH, STEAM, TEMPLE, CRUST


33
Heart Rate
Baseline. A preliminary 3 X 2 X 2 X 2 ANOVA (Valence X Imagery
Ability X Instructions X Probe Group) of heart rate baseline values
yielded no significant main effects or interactions of heart rate change
with the independent variables. Thus, baseline heart rate was not
systematically related to any of the factors in this study.
Encoding. Heart rate decreased from baseline during paragraph
encoding, F (2,52) = 52.17, p < .01. As can be seen in Table 2, this
deceleration was less during encoding of pleasant than unpleasant
content, Valence F (2,51) = 2.57, p= .11 in the omnibus test; Valence
(linear) F (1,25) = 15.52, p < .01. Examination of the Table also
suggests that the above effect was stronger late in the paragraph. The
Valence X Sentence interaction did not, however, reach statistical
significance in the omnibus test (F (4,104) = 2.14, p = .12).
Table 2
Mean heart rate deceleration for early, middle, and late sentence pairs
of texts with affectively unpleasant, neutral, and pleasant content
Pleasant
Neutral
Unpleasant
ice Pair
Mean
Early
-1.02
-1.88
-1.45
-1 .45
Middle
-2.74
-3.11
-3.49
-3.11
Late
-3.02
-3.34
-3.69
-3.35
-2.26
-2.78
-2.88
Mean


7
2. Affective and neutral cognitive representations differ in type
and quantity of associated information, which differentially affects the
time it takes to encode these contents.
3. A story schema develops over time during text processing, and
thus emotional activation is more pronounced late in the encoding
process.
4. Instructions to image and good imagery ability increase the
probability of physiological activation at encoding.
A long tradition of psychophysiological research on attention and
affect has employed heart rate as a dependent variable. Many
investigators have found that heart rate deceleration accompanies an
attentional set (Graham and Clifton, 1966; Obrist, 1981; Lang, Ohman, &
Simons, 1978). Lacey and Lacey (1974) described heart rate deceleration
as part of an "information intake" orientation in which the organism
shows a generalized inhibition of motor activity accompanied by an
attentional focus on the stimulus of interest. Cardiac acceleration, on
the other hand, typically accompanies tasks in which the subject
focusses on internal mental processes and inhibits environmental intake,
such as performance of mental arithmetic (Lacey & Lacey, 1974) and
generation of mental images (Miller et al., 1987).
Studies of emotion have revealed divergent patterns of heart rate
response to affective content, depending on whether the subject is
engaged in information intake or internal processing. For example, in
perceptual situations such as slide-viewing, heart rate is generally
found to decrease more when normal subjects are exposed to affectively
unpleasant pictures compared to neutral or positive stimuli (Klorman,


71
6. Football Game
Unpleasant. The football game is a few minutes past halftime
inside the packed stadium. You have a bad headache and the hot sun
makes you feel worse by the minute. The crowd roars and tension is
high--they are only yards away from a touchdown. You're sick and
uncomfortable on the hard seat as play continues on the field. Your
stomach churns painfully as the stands pulsate with the roaring cheers.
Your head throbs and you're sick to your stomach as the game finally
ends.
Pleasant. The football game is in the last few minutes inside the
packed stadium. You sweat freely in the hot sun as your team moves into
playing position. Tension is high--you are one point behind, and thirty
yards from a touchdown. You jump from your seat cheering wildly as the
play moves down the field. The stands pulsate with the crowd's cheers
as the halfback gets the ball. Touchdown! You thrust your arms
overhead as the crowd roars with excitement.
STUDY WORDS: MINUTES, STADIUM, TENSION, YARDS, SEAT, FIELD, PLAY
NEW WORDS: SUGAR, MONSTER, ECONOMY, MAYOR, PRISON, HAT, SPINACH


163


150
Table 48
Mean percent errors for study words primed by a study word, a new word,
or XXXXX for affectively pleasant, neutral, or unpleasant texts,
for good and poor imagers receiving verbal and imagery instructions
IMAGERY INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
9.51
8.18
8.37
8.69
New Word
9.29
7.20
11.01
9.17
XXXXX
10.84
5.62
8.15
8.20
Mean
9.88
7.00
9.18
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
4.78
4.50
2.66
3.98
New Word
10.63
6.50
7.60
8.24
XXXXX
6.84
8.74
10.22
8.60
Mean
7.42
6.58
6.82


61
VERBAL INSTRUCTIONS:
As you read the sentences, you should concentrate on the specific
words used in each sentence. Read through the lines of the paragraph at
your own pace. Try to remember each sentence in the paragraph and the
specific vocabulary employed. As you read, try to retain the precise
wording of each sentence. The better your memory of the exact language
used, the better your performance will be on the word test that comes
after.
MAIN INSTRUCTIONS:
After you press the buttons to remove the last sentence, a word
pair will be presented on the screen. The first word will appear very
briefly, so you must stay alert and pay close attention. Read this
first word, but do not press a button. Wait for the second word to
appear. When the second word appears, your task is to press one of the
buttons to indicate whether it was in the paragraph you just studied.
Press the button in your right hand for yes, and the button in your left
hand for no. After you respond, another pair of words will appear. If
you are wrong, an error message will appear on the screen, and then the
next test pair will appear as normal. The test lists move very fast, so
it's important for you to stay alert and to respond as quickly and
accurately as possible. After the test list there will be a short pause,
and then a message will appear on the screen instructing you to press
both buttons to start the next trial. At the end of the session a
message will come on the screen to let you know you have finished, and
I'll come in shortly after that. In order to make your reaction times
more precise, please hold your thumbs lightly on the button throughout


110
Table 12
Mean reading time (sec) for early, middle, and late sentence pairs
of affectively pleasant, neutral and unpleasant texts, with and without
startle probes, for subjects receiving verbal and imagery instructions
NO STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.74
5.49
5.73
5.65
Middle
5.50
5.24
5.56
5.43
Late
4.87
4.82
5.11
4.93
Mean
5.37 5.18 5.47
Verbal Instructions
5.34
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.07
5.18
4.98
5.08
Middle
4.67
4.89
4.76
4.78
Late
4.23
4.46
4.40
4.36
Mean
4.66
4.84
4.72
4.74


72
7. First Date
Unpleasant. Out on a first date your shyness around this new
person is painfully obvious. Sitting stiffly in your seat, there is no
conversation before the movie starts. You are aware of your sweaty
palms and feel a slow blush covering your face. The silence drags on
endlessly as you nervously search for something to say. You finally
manage to make a joke and fall silent after the brief response. You sit
there silent and miserable, wishing the evening were over.
Pleasant. Out on a date, you're having a wonderful time with a
very special person. Conversation seems effortless as you laugh and
talk before the movie starts. You feel a blush of excitement and
attraction as you make plans for later on. You agree to go dancing
later in the evening as silence falls over the theater. You lean over
close with a last little jokehe squeezes your hand in response. You
sit there, warm and excited, wishing the evening would go on forever.
STUDY WORDS: DATE, PERSON, CONVERSATION, MOVIE, JOKE, RESPONSE, BLUSH
NEW WORDS:
BUSHEL, STRAND, SCARF, CEREAL, SWEATER, SALUTE, REMINDER


96
31. Driving on Highway/Accident
Neutral. Your comfortable automobile drives on steadily after the
brief summer storm. The rain has stopped and you reach over to turn off
the windshield wipers. Your hands lightly steady the wheel as the road
follows a wide smooth curve. Keeping an eye on the curve, you turn the
radio on to your favorite station. Driving without effort, you notice
the wild flowers in the ditch by the road. You yawn and stretch as you
check the clock againjust one more hour to go.
Unpleasant. You tense up as the small automobile swerves
dangerously in the heavy storm. The windshield wipers are useless as
rain blows in sheets across the road. Your hands grip the wheel as you
peer through the dark and see a curve ahead. You turn hard to make the
curve--you can feel the car skidding sideways. In a desperate effort to
stop the car, you run off the road into the ditch. Your head slams hard
against the window as you feel the car pitch sideways.
STUDY WORDS: AUTOMOBILE, STORM, RAIN, WIPERS, EFFORT, DITCH, ROAD
NEW WORDS:
PIPE, FROST, PEARS, CHUNK, CLOTH, SUIT, SNAIL


INTRODUCTION
The purpose of this research is to examine how narrative text is
processed into story-images, and how the information structure of such
images differs from representations of the specifically verbal content
of text. The experimental procedure employed here is designed to test a
set of propositions that were conceptually derived from the bio-
informational theory of imagery (Lang, 1977, 1979, 1984, 1985), and from
associated supporting research (e.g., Cook, Melamed, Cuthbert, McNeil, &
Lang, 1988; Lang, Levin, Miller, Kozak, 1983; Miller, Levin, Kozak,
Cook, & McLean, & Lang, 1987; Vrana & Lang, 1990).
Theoretical Background
The Network Model
In general, cognitive psychologists hold that meaningful text is
parsed by the brain into concept units (e.g., Anderson, 1976; Kintsch,
1974; Schank, 1970). Following Anderson and Bower (1973), these
concepts are presumed to be inter-linked by association (perhaps, with a
propositional sub-organization), forming an information network that
represents an individual's knowledge of the text content. Subsequently,
if a newly presented stimulus matches a concept in the network (e.g., a
word from the previously seen text is presented on a CRT screen)
retrieval is initiated. This associative retrieval process eventually
activates the entire network structure, and is reported by the subject
1


105
Table 9
Mean reading time (sec) for early, middle, and late sentence
pairs
of affectively pleasant, neutral and unpleasant texts
with and without startle probes
No Startle Subjects
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.41
5.34
5.36
5.37
Middle
5.09
5.06
5.16
5.10
Late
4.55
4.64
4.76
4.65
Mean
5.01
5.01 5.10
Startle Subjects
5.04
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.57
5.37
5.82
5.59
Middle
5.08
5.48
5.44
5.33
Late
4.71
4.80
4.82
4.78
Mean
5.12
5.22
5.36
5.23


91
26. Walk Across Campus
Pleasant. Walking across campus in a hurry, you thread your way
among crowds of students. With a shock you spot the attractive guy you
met at the library last night. Your attention is drawn to him as he
walks the across the lawn towards you. He comes closer and looks up,
and as your eyes meet your heart skips a beat. Your pace slows quickly,
and you finally come to almost a complete stop. You thrill with delight
as he flashes you a smile and stops to say hello.
Unpleasant. Walking across campus in a hurry, you thread your way
amid crowds of students. With a shock, you spot your boyfriend and one
of your friends holding hands. In disbelief, your attention fixes on
them talking closely on the lawn. Your heart pounds as you walk up to
them, and you feel upset and agitated. Your pace slows as they look
over, startled, and their talk comes to a stop. Taking a deep breath,
you fight back your feelings and force yourself to speak.
STUDY WORDS: CAMPUS, STUDENTS, ATTENTION, LAWN, PACE, STOP, SHOCK
NEW WORDS: CELERY, CRANE, MUD, SPOON, PICKLE, LEMONADE, MAGAZINES


50
The startle reflex, on the other hand, has been shown in recent studies
to be a robust measure of emotional response that is invariant across a
variety of experimental tasks and sensory modalities.
Lang has conceptualized startle as a protective reflex that is
primed or facilitated in negative affective contexts (Lang et al.,
1990). That is, the response to a noxious stimulus (the startle probe)
would be expected to be larger when the organism is in an affectively
negative state, and smaller during an affectively positive state. Thus,
the startle response served as a manipulation check in the present
study, to assess the development of affect in the subject during the
encoding process.
Consistent with the hypothesis that schema development occurs over
time, emotional content only affected startles presented late in the
encoding process. Startle magnitude showed the expected pattern of
greater potentiation for unpleasant relative to pleasant content.
However, encoding strategy did not affect this aspect of the startle
response.
Consistent with the startle magnitude results, startle latency was
faster for probes presented late in affectively unpleasant compared to
pleasant paragraphs. These effects were evidenced more for subjects
receiving imagery instructions, who showed this linear effect for early
probe startles as well. Subject receiving verbal instructions, on the
other hand, only developed this effect late in the encoding process.
These results suggest that imagery instructions may result in earlier
contact with the emotional content of text than instructions that focus
on the verbal characteristics of the narrative.


27
divided by the total number of paragraphs she had rated in that
affective category.
Reaction time. Reaction time for each of the 12 prime-target
pairs was recorded to the nearest msec, for correct responses only.
Average RT and error rates were calculated for each prime-target pair,
for each affective category.
Accuracy. Error rate was calculated for each of the 12 prime-
target pairs by dividing the number of errors within an affective
category by the total number of paragraphs in that category, for each
subject. This number was then multiplied by 100 so that error rate was
expressed as a percentage.
Data Analysis
Imagery Ability
Subjects were classified according to imagery ability, based on a
median split of 80 on the Questionnaire upon Mental Imagery (QMI;
Sheehan, 1967). The range and distribution of QMI scores obtained for
subjects in this study conformed to those obtained in other studies in
the Lang laboratory, and are documented in Appendix C. Groups of
subjects assigned to receive imagery and verbal processing instructions
were balanced for QMI score to control for the effects of imagery
ability and its possible interaction with the instructional
manipulation.
Affective Classification of Texts
In the design of this experiment, a text's affective category was
determined by ratings and psychophysiological measures obtained in an


104
Table 7continued
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.87
5.93
6.01
5.94
Middle
5.53
5.90
5.77
5.73
Late
4.90
5.00
5.19
5.03
Mean
5.44
5.61
5.66
5.57
Table
8
Mean reading
time (sec) for
early, i
middle, and late
sentence
of affectively pleasant, neutral and unpleasant texts
for subiects receiving verbal
and imagery instructions
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.39
5.36
5.58
5.44
Middle
5.12
5.21
5.39
5.24
Late
4.69
4.76
4.92
4.79
Mean
5.07
5.11
5.30
5.16
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.59
5.34
5.60
5.51
Middle
5.04
5.34
5.22
5.20
Late
4.56
4.68
4.66
4.63
Mean
5.06
5.12
5.16
5.11


26
deviated from the pre-trial baseline for that text. Mean heart rate
change scores were then calculated for affectively pleasant, neutral,
and unpleasant texts. For subjects receiving startle probes, separate
averages were calculated for each of the three levels of the Startle
Position factor (Early, Late, No Startle) within each affective
category.
Blink reflex. Integrated blink EMG data were scored with the
computer program provided by Balaban, Losito, Simons, and Graham (1986).
Average startle baseline values and magnitude (in arbitrary A-D units)
were calculated for startle probes presented early and late in the study
texts (i.e., during sentences 1 or 2 versus 5 or 6), for each of the
affective categories. The onset latency of the startle reflex was also
scored. Trials in which the startle magnitude value was zero were
omitted from the calculation of latency averages, which were calculated
for probes presented early and late in the text and for each of the
affective categories.
Retrieval Measures
Free recall. For each subject, the list of short descriptive
phrases was matched to a master list of paragraph topics in the study.
A paragraph was considered to have been recalled if the descriptive
phrase matched the topic of one of the paragraphs she read during the
experimental session. Paragraphs identified by the subject in this
manner were counted as having been recalled. The recall measure was
calculated for each subject: For each affective category (pleasant,
neutral, unpleasant), the number of texts listed by the subject was


136
Table 36
Mean recognition RT (msec) for study words primed by a study word, a new
word, or XXXXX for affectively pleasant, neutral, or unpleasant texts.
with and without startle probes, for good and poor imagers
NO STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
828
863
859
850
New Word
818
941
838
866
XXXXX
942
909
899
917
N-S Priming
-10
+ 78
-21
+ 16
X-S Priming
+ 114
+ 46
+ 40
+ 67
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
861
890
880
877
New Word
858
924
846
876
XXXXX
869
907
893
890
N-S Priming
- 3
+ 34
-34
- 1
X-S Priming
+ 8
+ 17
+ 13
+ 13


67
2. Bicycle Ride
Unpleasant. You are riding your bicycle to school early in the
morning, in a rush. You are late for an exam, but you can't seem to get
free of all the cars. You're stuck on a narrow street in a line of
traffic moving slowly up the hill. Exasperated, you bolt for the other
sidewalk between cars, at top speed. You look around to a loud horn,
but the wind blows hair across your face. Suddenly you are struck hard
from the side, slamming you onto the pavement.
Neutral. You are riding your bicycle to school, relaxed and in no
particular rush. Pedaling smoothly along, you mentally review your
schedule for that morning. You go along steadily in a line of bikes
moving with the traffic up a hill. Picking up speed as you reach a wide
sidewalk on campus, you feel more relaxed. You coast downhill, lifting
your head to let the wind blow through your hair. Arriving at the
library you climb off your bike and lock it up in the rack.
STUDY WORDS: BICYCLE, SCHOOL, RUSH, TRAFFIC, HILL, SPEED, SIDEWALK
NEW WORDS:
NEPHEW, THIRST, MINK, CLAMP, HAM, SPARROW, COAL


143
Table 41
Mean percent errors for new word and study word targets,
for affectively pleasant, neutral, and unpleasant texts.
for subjects receiving imagery and verbal instructions
New Word Targets
Pleasant Neutral Unpleasant Mean
Instructions
Imagery
4.61
5.08
4.43
4.71
Verbal
2.75
2.58
4.15
3.16
Mean
3.68
3.83
4.29
3.93
Study Word
Targets
Pleasant
Neutral
Unpleasant
Mean
Instructions
Imagery
8.65
6.79
8.00
7.81
Verbal
7.05
11.19
8.70
8.98
Mean
7.85
9.00
8.45
8.40


44
Valence F (2,92) = 6.08, g < .01 in the omnibus test; Valence
(quadratic) F (1,46) = 6.97, g < .01. Other experimental factors did
not reliably affect reaction time for NS pairs.
Recognition errors were least when a study word target was primed
by another study word; performance declined somewhat with new word
primes, (SS = 6.23 percent, NS = 8.38), Priming Condition F (1,46) =
5.66, g < .03. While error rate was unaffected by imagery ability, or
the presence or absence of startle probes, error rates for SS and NS
prime-target pairs were differentially affected by emotional content and
instructions, Valence X Instructions X Priming Condition interaction
F (2,92) = 3.12, g = .06; Valence (quadratic) X Instructions X Priming
Condition F (1,46) = 4.28, g < .05. This interaction was followed up by
separate analyses of each priming condition.
Study-Study. Recognition of study words primed by other study
words was not affected by emotional content of the paragraphs, Valence F
< 1 in the omnibus test. However, instructions interacted with imagery
ability to affect subject performance, F (1,46) = 6.05, g < .02. Good
imagers receiving imagery instructions made significantly more errors
than poor imagers (good = 8.69 percent, poor = 3.98 percent), F (1,26) =
7.95, g < .01. Under the verbal instructions condition, however,
imagery ability did not affect error rates (good imagers = 5.44 percent,
poor imagers = 6.82 percent), F < 1; (see Table 48, Appendix D) .
Comparison of the error data results with the reaction time
analyses does not support a speed-accuracy trade-off explanation for the
valence effects on recognition reaction time described earlier.


130
Table 31
Mean recognition reaction time (msec) for study word and new word
targets for affectively pleasant, neutral, and unpleasant texts
for good and poor imagers receiving imagery and verbal instructions
NEW WORD TARGETS
Imagery Instructions
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
848
858
839
848
Poor
833
860
851
848
Mean
840
859
845
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
864
894
891
883
Poor
925
935
927
929
895 914 909
Mean


Table 12continued
111
STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.03
5.24
5.43
5.23
Middle
4.75
5.19
5.21
5.05
Late
4.52
4.71
4.73
4.65
Mean
4.77
5.04
5.12
4.98
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
6.11
5.49
6.21
5.94
Middle
5.41
5.78
5.67
5.62
Late
4.90
4.90
4.92
4.90
Mean
5.47
5.39
5.60
5.49


119
Table 19
Mean heart rate change (beats per minute) for early, middle, and late
sentence pairs of affectively pleasant, neutral and unpleasant texts.
with and without startle probes, for verbal and imagery instructions
NO STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
+ 0.66
-0.54
+0.26
+0.12
Middle
-1 .35
-2.22
-1.68
-1 .75
Late
-1 .50
-1 .66
-1 .92
-1.69
Mean
-0.73 -1.47 -1.12
Verbal Instructions
-1.10
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-2.88
-3.58
-2.75
-3.07
Middle
-4.04
-3.96
-4.79
-4.26
Late
-4.34
-4.58
-5.20
-4.71
Mean
-3.76
-4.04
-4.25
-4.01


160
ratings remained in this category for the present sample, while in 144
instances the texts were rated as neutral, and a priori unpleasant texts
were rated as pleasant in 36 instances by the present sample.
Examination of this table reveals that while the affective classifi
cations of text content were relatively more stable across samples for
texts rated as pleasant or unpleasant, ratings for neutral texts were
less stable and tended to move to the affectively pleasant category.


134
Table 34
Mean recognition
reaction
time (msec) for studv word
. targets primed bv
a studv word, a
new word
, or XXXXX for
affectively
pleasant, neutral.
or unpleasant texts for imagery
and verbal instructions
Imagery Instructions
Pleasant Neutral
Unpleasant
Mean
Prime Type
Study Word
810
863
850
841
New Word
830
897
791
839
xxxxx
849
852
879
860
N-S Priming
+20
+ 34
-59
- 2
X-S Priming
+ 39
-11
+29
+ 20
Verbal Instructions
Pleasant Neutral
Unpleasant
Mean
Prime Type
Study Word
855
834
869
853
New Word
860
917
859
878
XXXXX
938
924
869
911
N-S Priming
+ 5
+83
-10
+26
X-S Priming
+83
+90
0
+ 57


115
Table 17
Mean heart rate change (beats per minute) for early, middle, and late
sentence pairs of affectively pleasant, neutral and unpleasant texts
for good and poor imagers receiving verbal and imagery instructions
IMAGERY INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.26
-1.18
-0.19
-0.54
Middle
-1.27
-2.47
-1 .88
-1 .87
Late
-1.59
-2.47
-1.65
-1.90
Mean
-1 .04
-2.04
-1 .24
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.57
-0.17
-1 .03
-0.59
Middle
-2.69
-1 .84
-3.00
-2.51
Late
-3.26
-1 .88
-3.34
-2.83
Mean
-2.18
-1 .30
-2.45


15
conductance, and electromyographic (EMG) activity from the corrugator
and zygomatic facial muscles, recorded during a 30-second imagery
period. Subjects also provided ratings of the paragraphs, using the
automated SAM system (Lang, 1980; Hodes, Cook, & Lang, 1985) after the
imagery task. Texts were selected for this research on the basis of:
(1) extreme pleasant or unpleasant valence ratings, and (2) the presence
of the appropriate psychophysiological activity. For unpleasant texts,
the criterion physiological responses during imagery included increased
heart rate and skin conductance level (associated with arousal), as well
as high corrugator (frown) muscle tension (associated with negative
emotional valence). The selected affectively pleasant texts showed high
zygomatic (smile) muscle activity accompanied by low corrugator muscle
activity, in addition to SAM ratings of positive emotional valence and
high arousal. Neutral texts were selected on the basis of neutral
valence and low arousal SAM ratings, and low levels of psycho-
physiological activity on all four measures. The number of exemplars
contributing to each affective category (18 unpleasant, 9 neutral, and
9 pleasant) reflected the primary focus on the comparison between
effects on cognitive processes of unpleasant versus neutral/pleasant
affect in prior pilot research. A summary of the validation study
including the selected texts and the corresponding data can be found in
Appendix E.
Paragraph topic and contents of the test list were matched between
affective categories. Every unpleasant text in Set 1 had a "matched"
text that was neutral or pleasant in Set 2; likewise, every neutral/
pleasant text in Set 1 had a "matched" unpleasant text in Set 2.


34
Nevertheless, based on prior hypotheses, separate tests were undertaken
for each sentence pair. These tests revealed that while the effect of
valence on heart rate change was not significant early in the paragraph,
Valence (linear) F (1,25) = 2.58, p = .12, pleasant contents resulted in
reliably less heart rate deceleration than unpleasant contents for
middle and late sentence pairs, Valence (linear) Fs (1,25) = 20.65 and
8.24, respectively; both ps < .01.
Table 3 suggests that subjects receiving verbal instructions
showed greater heart rate deceleration compared to subjects receiving
imagery instructions; however, the statistical test only approached
significance, F (1,25) = 3.89, p = .06. On the other hand, instructions
significantly modulated the affect/heart rate relationship, yielding a
Valence (linear) X Instructions interaction, F (1,25) = 7.01, p < .02.
The greater relative deceleration for unpleasant than pleasant texts,
found overall, was more pronounced in subjects receiving verbal
instructions.
Table 3
Mean heart rate change (beats per minute) for texts with
pleasant, neutral and unpleasant content, for subjects receiving
verbal and imagery instructions
Instructions
Verbal
Imagery
Pleasant
Neutral
Unpleasant
Mean
-2.92
CO
CO
ro
1
-3.91
-3.57
-1.61
-1.67
-1 .85
-1 .71
-2.26
-2.78
-2.88
Mean


Dodson Law (1908) in which performance on a variety of learning
discrimination tasks is positively related to increases in physiological
arousal, within the moderate range.
Thus, the present research tests the following general hypothesis
regarding the effects of emotional content on measures of memory
retrieval:
1. More affectively arousing text, (i.e., elaborate information
structures in memory) will result in enhanced performance for both
recognition and free recall measures of retrieval.
Arousal is more probable when subjects employ an imaginal encoding
strategy, given that imaginal processing emphasizes the inclusion of
sensory and visceral/motor activation codes in the resulting memory
representation. Conversely, instructions to attend exclusively on the
surface characteristics of the words (verbal instructions condition),
would be expected to reduce the inclusion of such information in memory.
The instructions in this experiment may also be construed as an
affective levels of processing manipulation (Craik & Lockart, 1972) in
which imaginal encoding instructions involved "deep" processing, and
verbal instructions prompted "surface" processing of the paragraphs.
These considerations motivate the following hypothesis:
2. Imaginal encoding strategies, resulting from instructions or
subject characteristics, emphasize the representation of visceral/motor
activation codes in memory for affective text relative to verbal
encoding strategies, thereby enhancing memory performance.
3. The above effects are specific to the reactivation of the
study paragraph's information network by the related test word, and do


2
as his or her verbal memory for the text. Several parameters determine
network activation and the resulting memory retrieval. These include
the number and type of network concepts represented in the activating
percept, and the associative strength between concepts in memory.
Measures of recognition memory and free-recall of text content have been
used to assess their influences (e.g., Anderson & Bower, 1973; Carroll &
Kirsner, 1982; Galambos & Rips, 1982; Kintsch & van Dijk, 1978; McKoon &
Ratcliff, 1980; Ratcliff & McKoon, 1978).
Imagery and Response Activation
Variants of the above model have been widely employed in the
semantic analysis of text processing. In 1973, Pylyshyn suggested that
a similar analysis might also account for imaginal memory. In his view,
images are internally constructed propositional descriptions of objects.
Kieras (1978) later elaborated this idea, suggesting that imagery
differed from other information structures, specifically, in the amount
of modality-specific information that was in the network. Thus, vivid
imagers, more than poor imagers, would give elaborate sensory
descriptions of previously seen percepts, or readily detect errors of
detail in a picture memory task. While this seems plausible, the
experimental evidence for this view is not compelling (see Strosahl &
Ascough, 1981). Furthermore, this sensory information emphasis neglects
a vast psychophysiological literature, which has reliably found imagery
effects, not in afference, but in the efferent system (e.g., Jacobson,
1930; Lusebrink & McGuigan, 1989; McGuigan & Schoonover, 1973; Miller et
al., 1987). That is, when subjects imagine, remember, or re-experience


Figure 4. Self-Assessment Manikin (SAM) used to obtain affective ratings of texts.


5
and their specific action. The other half received verbal instructions:
they were told to focus on the exact words used in the text.
Instructions to both groups were presented as optimal learning
strategies that would facilitate performance on the word recognition
task following each text.
Subjects read the study paragraphs at their own pace, sentence by
sentence, on a CRT screen. The time taken to read each sentence was
recorded, and heart rate was monitored continuously. For a subset of
subjects, an acoustic startle stimulus occurred during some of the
texts. After each paragraph, a test list of words and control symbols
was presented. Subjects held reaction time buttons, one in each hand;
they indicated with a right hand button press if a target word had
appeared in the preceding text, and with a left hand button press if the
word was not in the text. The words in the test lists were all preceded
by a very brief (300 ms) priming word or symbol. Some of the priming
words were also from the preceding text, and were expected to facilitate
recognition speed for the target words. After this part of the
experiment was completed, subjects were asked to write down the gist of
as many of the paragraphs as they could recall. Subjects provided
affective ratings of the texts at the end of the experimental session.
Predictions, derived from bio-informational theory were tested for
both encoding and retrieval phases of the study.
Encoding Hypotheses
During the encoding process, narrative text activates representa
tions in memory which constitute an image of the events described. As


82
17. Hotel Party
Unpleasant. You are at a large exclusive party which packs
several floors of an old hotel. Out a window you suddenly see fire
exploding into the sky up on the hotel roof. Someone runs in yelling
that the hotel is on fire from the fireworks up there. Incredibly fast,
the halls fill with smoke and panic spreads through the crowd. In
disbelief, you stare at windows shattering in the repeated explosions.
Your lungs and eyes burn as you struggle towards a balcony, gasping for
air.
Pleasant. At the hotel, you two leave the party and wander up to
the rooftop garden. It's a beautiful clear night; you stand on the roof
gazing up at the sky. Under the full moon, soft kisses along your
throat fill you with warmth. The noisy crowds down in the halls filled
with smoke seem very far away now. Back inside, you lose yourselves in
soft caresses and explosions of desire. His long, deep kisses send
thrills of pleasure all through your body.
STUDY WORDS: PARTY, HOTEL, ROOF, SKY, HALLS, SMOKE, EXPLOSIONS
NEW WORDS: CLIP, PASTRY, BASEBALL, WEED, TODDLER, NOVEL, GRASS


METHODS
Subjects
Fifty-seven female undergraduates at the University of Florida
served as subjects in this experiment. Subjects were recruited through
introductory psychology classes, at summer registration, or through
signs posted at sorority houses on campus. Subjects either received
class credit or were paid $10 for participation in the study.
The research sample included a subset of 25 subjects to whom
acoustic startle probes were presented during text encoding. Probe
administration had almost no effect on the other dependent variables
Thus, for all measures except the startle probe, all 57 subjects are
included in the analyses, with the few sample differences noted. The
analysis of the startle probe includes, of course, only the subset of 25
subjects.
Materials
Texts
The study materials used in this research consisted of two sets of
36 6-line paragraphs. Each set contained 18 unpleasant, 9 neutral, and
9 pleasant exemplars. The paragraphs were classified into affective
categories based on a validation study in which college students read
and imaged a larger set of 108 texts (Spence et al., 1987). The
validation data were psychophysiological measures of heart rate, skin
14


20
STUDY TEXT EXEMPLAR
You're having a great time on a picnic by the ocean with
some good friends. Some are out swimming, and you look twice
when you think you spot sharks. Your feet pound through the
surf as you run calling, the fins now obvious. Unaware, your
friends don't hear you, and you are gripped with horror. Your
heart pounding in your chest, you run across the sand shouting
and waving.
You scream in
desperation as someone is jerked
under, and
the water turns
red.
Prime
Tarqet
Pair Type
scissors
degree
New Word-New Word
calf
picnic
New Word-Study Word
friends
tent
Study Word-New Word
charm
New Word-Left Hand Response
xxxxx
XXXXX-Left Hand Response
pint
*******
New Word-Right Hand Response
fins
Study Word-Left Hand Response
feet
*******
Study Word-Right Hand Response
XXXXX
wagon
XXXXX-New Word
xxxxx
ocean
XXXXX-Study
heart
sand
Study-Study
Figure 1. Exemplar of affectively unpleasant study paragraph and
its associated test list.


92
27. Coming Home from Library
Neutral. Coming back from a good study session at the library,
it's early evening. Your neck and shoulders feel tight since you've
been reading for two hours. You stretch with relief as you put down
your booksyou're ready to relax. Turning on the radio, you sit down
in your favorite chair and open your mail. Later on you test yourself,
reviewing parts of the information you covered. Most of it comes back
to you, but you decide to review your notes once again.
Unpleasant. Coming back from a long cramming session at the
library, it's late at night. Your shoulders and neck feel stiff from
hunching over your work for hours. You put down your books, aware of a
dull tension headache behind both eyes. As you sit down to open your
mail, the fatigue and worry are overwhelming. You try to rehearse parts
of the information you covered, but you can't think. You just can't
seem to stop worrying about the big test and your bad grades.
STUDY WORDS: SESSION, LIBRARY, SHOULDERS, HOURS, PARTS, INFORMATION,
MAIL
NEW WORDS:
CONQUEST, STONE, DREAM, COMPUTER, GREASE, LUMBER, THOUGHT


86
21. Moving into Apartment/ Rat and Babv
Neutral. The apartment is on a busy, tree-lined avenue in a
pleasant part of the city. The guys are gradually unpacking the truck
parked in front along the curb. They carry in the crib, the beds, and
chairs, while you go out for food. It is near dinner time and everyone
is ready to relax for a minute and eat. You give the kids some toys and
a cookie, and see that the baby is asleep. As you come back with dinner
you see the truck is empty and everyone's inside.
Unpleasant. The apartment is on a crowded, dirty avenue in the
slums of the city. It is late, but no adults are home in the dark,
empty apartment. A large rat climbs up into the baby's crib, sniffing
around for food. The baby draws back and cries anxiously as the ugly
animal sniffs around. Roaches crawl over a broken cookie on the floor
amid scattered toys. The rat approaches the frightened, crying child in
the dark room.
STUDY WORDS: APARTMENT, AVENUE, CRIB, FOOD, COOKIE, TOYS, BABY
NEW WORDS:
MONKEY, TORNADO, THEME, BUTTER, BASEMENT, FUNGUS, GARMENT


22
startle stimulus consisted of a 50 msec burst of 100 dB (A) white noise
with an instantaneous rise time, that was gated by a relay and presented
binaurally through Pioneer (SE-250) stereo headphones from a tape that
ran continuously during the experimental session.
Texts were displayed on an Amdek CRT monitor in the subject room.
The monitor was placed approximately 2 feet in front of the subject, and
a transparent plastic screen separated the subject from the monitor
itself. Subjects responded using hand-held RT buttons, mounted on 6-
inch cylinders, which were activated by pressing down with the thumb.
The subject held a cylinder in each hand throughout the experimental
session. She controlled presentation of each successive sentence with a
button press, and a final button press removed the entire text from the
screen. Read time consisted of the time between button presses, as the
subject read through the texts at her own pace. The subject also
pressed either the right or left button to respond during the episodic
recognition task. RTs and reading times were recorded to the nearest
msec.
An intercom permitted communication between the subject and the
experimenter throughout the experimental session.
Procedures
Upon arrival at the laboratory, subjects signed an informed
consent form and completed imagery questionnaires and a vocabulary
measure. The subject was then seated in a comfortable chair in the
experimental chamber and sensors were attached in order to obtain the
psychophysiological measures.


106
Table 10
Mean reading time (sec) for early, middle, and late sentence pairs
of affectively pleasant, neutral and unpleasant texts, for good and poor
imagery ability subjects receiving verbal and imagery instructions
IMAGERY INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.11
4.95
5.22
5.09
Middle
4.83
4.90
5.09
4.94
Late
4.56
4.64
4.72
4.64
Mean
4.84
4.83
5.01
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.66
5.78
5.95
5.79
Middle
5.42
5.53
5.68
5.54
Late
4.82
4.88
5.12
4.94
Mean
5.30
5.40
5.58


21
Apparatus
The experiment was controlled by two computers working in concert:
Terak and Datax LSI-11 computers. The Terak controlled the timing of
events in the experiments, presented the texts, and measured and
recorded read times and reaction times. For the subset of subjects
receiving startle probes, the Datax controlled presentation of the
acoustic startle stimulus, and recorded EKG and blink response.
Beckman standard silver-silver chloride electrodes were placed in
a modified Lead 2 configuration to measure the electrocardiogram (EKG).
The signal was filtered and amplified by Coulbourn modules. A Schmitt
trigger detected the R-waves and interrupted the computer clock to
record inter-beat intervals with a resolution of 1 ms. EKG was measured
continuously during presentation of each text, beginning 3 seconds prior
to the beginning and continuing during the time the subject read each
sentence of the text.
For the startle subset of the sample, EMG was measured with Med
Associates miniature silver-silver chloride electrodes, using Beckman
electrode electrolyte as the contact medium. The sensors were placed
about 1 cm beneath the left eye. One electrode was centered beneath the
eye, and the other was placed beneath the outer canthus. The
amplification for the signal was set at 6 (X 10,000), and frequencies
below 90 Hz and above 1000 Hz were filtered out by a Coulbourn HiGain
Bioamplifier. The signal was rectified and integrated by a Coulbourn
Contour Following Integrator (200 msec time constant), and then fed
through an Analog-to-Digital converter. EMG was collected at 1000 Hz
for 250 msec following the onset of the startle stimulus. The acoustic


140
Table 38
Mean recognition RT (msec) for study words primed by a study word, a new
word, or XXXXX for affectively pleasant, neutral, or unpleasant texts.
for good and poor imagers receiving verbal and imagery instructions
IMAGERY INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
778
863
827
823
New Word
791
939
782
837
XXXXX
845
868
847
853
Mean
805
890
819
838
N-S Priming
+ 13
+76
-45
X-S Priming
+ 67
+ 5
+ 20
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
842
863
872
859
New Word
869
855
800
841
XXXXX
852
836
912
867
Mean
854
851
861
856
N-S Priming
+ 27
- 8
-72
X-S Priming
+ 10
-27
+ 40


53
one characteristic of such affective information structures is that they
facilitate both probability and speed of memory retrieval.
However, one important measure of episodic priming deviated from
this general pattern of valence effects. When study words primed by
other study words were considered independently, reaction times were
slower for affectively unpleasant content relative to pleasant content.
This result does not conform to the study predictions for memory
performance. This linear effect of valence (slower for unpleasant and
faster for pleasant content) study-study reaction times is more
consistent with the pattern of valence effects observed for the encoding
measures during the later sentences. For reading time and heart rate,
this linear valence effect was associated with poor imagery ability and
verbal instructions, and was less pronounced for good imagers and
imagery instructions. These results suggest that a perceptual
orientation to the text processing task may moderate the valence effects
on a variety of dependent measures. Given that the perceptual aspect of
the episodic recognition task is inherently enhanced when the exact
information is presented at retrieval that was present at encoding, the
study-study reaction time data may reflect processes similar to those
acting on read time and heart rate late in the encoding process.
Accuracy data from this experiment showed a similar pattern of
results to the recognition RT data. Recognition was both faster and
more accurate when study words had previously appeared in an emotionally
evocative text, and when subjects employed imagery during encoding.


heart rate and the startle reflex verified the presence of emotional
response during encoding of the paragraphs. Affective content and
imaginal processing were predicted to increase encoding time, to produce
physiological activation during construction of the text schema, and to
facilitate memory retrieval.
The results showed that affectively unpleasant content was read
more slowly than pleasant material. Unpleasant paragraphs also prompted
greater heart rate decrease, and larger and faster probe startle
responses, compared to pleasant contents. These content effects were
all more pronounced for the later sentences in the paragraph, after the
affective content was established. The read time and heart rate valence
effects tended to be more reliable for poor imagers and for subjects
under verbal instructions. The startle/valence effect was more reliable
under imagery instructions.
Retrieval measures included free recall and episodic recognition
reaction time. Emotional content facilitated free recall performance
relative to neutral content. Affective content, good imagery ability,
and imagery instructions facilitated the speed and accuracy of
recognition memory for study word targets. Episodic priming occurred
for neutral texts, but was not reliable for affective texts.
These results support the hypotheses that both the encoding and
retrieval of narrative text vary significantly with affective versus
neutral content. In addition, emotional content interacted with
encoding strategy (imaginal or verbal) to further modulate these
processes.
vi


48
On the other hand, early in the encoding process, good imagers
read sentences from emotional (both pleasant and unpleasant) paragraphs
more slowly than those with neutral content. This effect was replaced
with the overall linear effect of valence in later sentences. Because
the early quadratic pattern of valence effects (unpleasant and pleasant
slower than neutral) occurred during the initial sentences of the
paragraphs for subjects reporting good but not poor imagery ability, it
may be argued that these results reflect individual differences in
schema development. It was suggested earlier that representations of
affective experience include additional information consisting of the
visceral/motor activation inherent in emotion (Lang, 1979). Hence,
construction of the schema for emotionally evocative text may result in
longer processing time (as reflected by read time) for affective
compared to neutral text, but only for subjects with a predisposition to
include the "action set" of emotion in the encoding process.
Heart rate decreased reliably during text encoding, consistent
with theories that deceleration is associated with information intake
(Lacey & Lacey, 1974). In addition, heart rate interacted with valence,
resulting in greater heart rate deceleration for affectively unpleasant
and less deceleration for pleasant contents. Consistent with read time
results, this effect emerged later in the encoding process.
Study predictions and prior research predicted that imaginal
versus verbal encoding strategies would result in different patterns of
heart rate change over sentences. Heart rate decrease was predicted for
verbal instructions, reflecting an information intake orientation. This
result is consistent with studies of heart rate response to presentation


168
Heuer, F., & Reisberg, D. (1990). Vivid memories of emotional events:
The accuracy of remembered minutiae. Memory and Cognition. 18.
496-506.
Jacobson, E. (1930). Psychology and the integrative action of the
nervous system. Acta Symbolica. 1, 31-35.
Jacoby, L. L. (1982). Knowing and remembering: Some parallels in the
behavior of Korsakoff patients and normals. In L. S. Cermak
(Ed.), Human memory and amnesia (pp. 97-122). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Jonides, J., & Mack, R. (1984). On the cost and benefit of cost and
benefit. Psychological Bulletin. 96, 29-44.
Kieras, D. (1978). Beyond pictures and words: Alternate information
processing models for imagery effects in verbal memory.
Psychological Bulletin. 85, 532-554.
Kintsch, W. (1974). The representation of meaning in memory.
Hillsdale, NJ: Lawrence Erlbaum Associates.
Kintsch, W., & van Dijk, T. A. (1978). Toward a model of text
comprehension and production. Psychological Review. 85, 363-394.
Klorman, R., Weisberg, R. P., & Wiesenfeld, A. R. (1977). Individual
differences in fear and autonomic reactions to affective
stimulation. Psychophysiology. 14. 45-51.
Klorman, R., Wiesenfeld, A. R., & Austin, M. L. (1975). Autonomic
responses to affective visual stimuli. Psychophysiology. 12. 553-
560.
Lacey, J. I., & Lacey, B. C. (1974). Studies of heart rate and other
bodily processes in sensorimotor behavior. In P. A. Obrist, A. H.
Black, J. Brener, & L. V. DiCara (Eds.), Cardiovascular psycho
physiology (pp. 538-564). Chicago: Aldine.
Lang, P. J. (1964). Experimental studies of desensitization in
psychotherapy. In J. Wolpe (Ed.), The conditioning therapies.
New York: Holt, Rinehart, & Winston.
Lang, P. J. (1977). Psychophysiological assessment of anxiety and
fear. In J. D. Cone & R. P. Hawkins (Eds.), Behavioral
assessment: New directions in clinical psychology. New York:
Bruner-Mazel.
Lang, P. J. (1979). A bio-informational theory of emotional imagery.
Psychophysiology. 16. 495-512.


17
rating of 9 represented the highest value on these scales. Texts
receiving a valence rating of 1 to 3 were classified as affectively
unpleasant; texts rated from 4 through 6 were classified as neutral; and
texts with a rating of 7 or above were classified as pleasant.
Statistical analyses for the present study were based on affective
categories derived from these subject ratings.
Test Lists
After the subject had finished reading the study text, an episodic
recognition task began immediately. The test list consisted of 12
prime-target pairs. The prime appeared on the monitor screen for 300 ms
prior to the onset of the target. The target remained on the screen
until the subject responded with a right or left button press. In the
case of verbal targets, the button press indicated whether or not the
target had appeared in the study text; whereas for non-verbal targets,
the button press was determined by the target itself and was not related
to text content. Subjects were instructed to respond only on the basis
of the target stimuli, without regard to the prime. The lists were
structured so that subjects responded with their right hand ("yes"
response) to half of the items, and with their left hand ("no" response)
to the other half. This prevented uneven response probabilities from
influencing the RT data. Verbal primes and targets were always nouns,
although they varied as to whether they were concrete or abstract in
content. No test word appeared more than once in the experiment.


90
25. Sailboat
Unpleasant. You duck your head as waves crash over the bow,
drenching the whole crew. The wind tears powerfully at the sails,
pulling you over in a sideways motion. Holding the rail, you brace
yourself as the boat lurches unsteadily forward. The day is cold, the
sea dark and churning with dirty caps of gray foam. You take a sharp
breath; the ice cold spray cuts hard against your face. You clutch the
mast with all your strengthyou are going under.
Pleasant. You laugh as the waves race past the bow, spraying most
of the crew. The wind pulls powerfully at the sails, increasing the
boat's forward motion. Holding to the mast, you steady yourself as your
boat surges into the lead. The day is perfect, the sea shining bright
blue with flecks of foam. You take a deep breath; the spray feels
invigorating, fresh against your face. Around you everyone is shouting
and cheeringyou are winning the race!
STUDY WORDS: WAVES, BOW, DAY, FOAM, BREATH, SPRAY, CREW
NEW WORDS:
TOPIC, CORNER, FEVER, MAP, HELMET, RUBY, OWL


98
33. Rollercoaster
Neutral. You slam the door shut for your little cousins at the
rollercoaster ride. You are having an enjoyable day at an amusement
park with your relatives. Looking up from below, you hear faint laughs
and shrieks as they top the crest. Your stomach growls with hunger as
you follow their course down the incline. While they are still zipping
around curves on the ride, you go get a snack. With hot dog and coke in
hand, you greet them as they walk down the ramp.
Unpleasant. You and your cousins are climbing into your seats on
the rollercoaster ride. You are feeling very anxious and nervous, but
try to hide it from the others. You sit pale among the laughs and
shrieks, the car jerking up to the crest. Your stomach drops as the car
plunges, roaring fast down the steep incline. You whip hard around two
sharp curves, feeling more and more dizzy and sick. Your hands are
white as you grip the bar hard, trying to take a deep breath.
STUDY WORDS: COUSINS, RIDE, LAUGHS, SHRIEKS, STOMACH, INCLINE, CREST
NEW WORDS: BUTTONS, ENVELOPE, FLAME, CARTON, MAID, PAINT, OLIVE


131
Table 31continued
STUDY WORD TARGETS
Imagery Instructions
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
805
890
819
838
Poor
854
851
861
856
Mean
830
871
840
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
856
823
829
836
Poor
912
961
902
925
Mean
884
892
865


TABLE OF CONTENTS
ACKNOWLEDGEMENTS ii
ABSTRACT v
INTRODUCTION 1
Theoretical Background 1
The Network Model 1
Imagery and Response Activation 2
Bio-Information Theory 3
Emotion and Memory 4
The Research Problem 4
Encoding Hypotheses 5
Retrieval Hypotheses 10
METHODS 14
Subjects 14
Materials 14
Texts 14
Test Lists 17
Apparatus 21
Procedures 22
Data Reduction 25
Encoding Measures 25
Retrieval Measures 26
Data Analysis 27
Imagery Ability 27
Affective Classification of Texts 27
Encoding Measures 28
Retrieval Measures 29
RESULTS 31
Encoding Measures 31
Read Time 31
Heart Rate 33
Startle Reflex 35
Retrieval Measures 38
Free Recall 38
Episodic Recognition 38
DISCUSSION 46
Encoding Measures 47
Retrieval Measures 51
Overall Conclusions 54
APPENDIX A QUESTIONNAIRES, INSTRUCTIONS AND DEBRIEFING STATEMENT 55
APPENDIX B TEXT MATERIALS AND TEST LIST WORDS 65
APPENDIX C IMAGERY ABILITY SCORES FOR INSTRUCTION GROUPS 102
iii


6
suggested earlier, cognitive representation of an emotion includes
information on visceral and motor activation, inherent in emotional
experience, but not included in neutral, non-action representations.
Thus, during encoding, more units of memorial information are contacted
by affective than by neutral text. This additional information and the
associated efferent expression may modulate encoding time. This action
content from memory will be evident in the pattern of physiological
activation instigated in the subject. Furthermore, network processing
occurs over time as significant narrative elements are encoded and a
story "schema" is constructed (Kintsch & van Dijk, 1978). Thus, the
above effects on physiological response and processing time will more
probably occur late in an encoding task.
Attention to a given aspect of a stimulus array determines the
content of memorial representations and affects memory performance
(Jacoby, 1982). It is suggested here that encoding and retrieval of
affective text are also modulated by the subject's processing style,
i.e., whether she focusses on the specific words in the text, or
focusses on the developing narrative imagery. This variable is examined
here in two ways: (1) by instructionsto either image or attend to the
specific words of the texts; and (2) by examining individual differences
in the normal prevalence and vividness of imagery, i.e., by comparing
"good" and "poor" imagers.
The present research examines the following general hypotheses:
1. The action content in emotional memory is evident in the
pattern of physiological activation during the encoding of affective
text.


113
Table 15
Mean heart rate change (beats per minute) for early, middle, and late
sentence pairs of affectively pleasant, neutral and unpleasant texts
for subjects receiving verbal and imagery instructions
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.41
CO
V£>
O
-0.61
-0.57
Middle
-1.98
-2.15
-2.44
-2.19
Late
-2.43
-2.17
-2.49
-2.36
Mean
-1.61 -1.67 -1.85
Verbal Instructions
-1 .71
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1 .62
-3.08
-2.30
-2.33
Middle
-3.51
-4.06
-4.54
-4.04
Late
-3.62
-4.50
-4.88
-4.33
Mean
-2.92
-3.88
-3.91
-3.57


38
2 < .04. While subjects receiving imagery instructions showed the
expected linear valence relationship, those receiving verbal
instructions showed a deviant pattern. For late probes, both
instruction groups startled faster to unpleasant content, Valence
(linear) F (1,14) = 5.61, p < .04.
Imagery ability did not modulate startle latency for pleasant and
unpleasant content. The only effect of imagery ability was a generally
faster early probe startle for good imagers, compared to poor imagersa
difference that disappeared with the late probes, Startle Position
F (1,14) = 11.88, p < .02 (see Table 22, Appendix D).
Retrieval Measures
Free Recall
Consistent with study predictions, paragraphs with emotionally
evocative content, both pleasant and unpleasant, were recalled in
greater proportion compared to affectively neutral texts (pleasant =
.41, neutral = .29, unpleasant = .42), Valence F (2,100) = 11.20,
p < .01 in the omnibus analysis; Valence (quadratic) F (1,50) = 18.20,
p < .01. Instructions, imagery ability, and presence or absence of
startle probes did not affect free recall.
Episodic Recognition
As described earlier, analyses of recognition reaction time data
were accomplished only for word targets and primes. The effects of the
experimental factors (Valence, Instructions, Imagery Ability, Probe
Group) were tested as they interacted with target type (study words
versus new words) and priming condition (study-study, new-study).


32
Table 1
Mean read times (in
seconds) for early, middle
, and late
sentence pairs
of texts with affectively pleasant
, neutral,
and unpleasant content
Pleasant
Neutral
Unpleasant
Sentence Pair
Mean
Early
5.49
5.35
5.59
5.48
Middle
5.08
5.27
5.30
5.22
Late
4.63
4.72
4.79
4.71
Mean
5.07
5.12
5.23
2 < .01. That is, for these first two sentences only, mean reading
times for both emotional contents (pleasant and unpleasant) were longer
than for neutral paragraphs. The overall linear valence effect--
shortest read times for pleasant paragraphs--developed in the later
sentences (Valence (linear) Fs (1,46) = 1.57, 9.64, and 7.07; ps < .22,
.01, and .02; for early, middle and late sentence pairs, respectively).
In subsequent tests, the Valence X Sentence interaction was also
found to vary with subgroup: Thus, good imagers showed the early
quadratic valence effect more strongly than poor imagers, Valence
(quadratic) X Imagery Ability X Sentence F (2,92) = 3.47, p < .01. For
poor imagers the linear valence effect (with the fastest read times for
pleasant content) was the dominant response across all trials (see
Table 7, Appendix D).
No reliable main effects or interactions involving instructions
were observed for read time.


76
11. Father Waits Up
Neutral. Coining home from racquetball and a movie with a friend,
it's gotten late. Although it's midnight your father is still up, so
you introduce your friend. As you close the front door you offer a
vague explanation for your late return. You groan as you sit on the
couch suddenly aware of sore muscles from the game. Your father has a
sleepy look on his face as he smiles and says goodnight. Pretty tired
yourselves, your friend soon leaves and you head off for bed.
Unpleasant. Coming back from a late movie with a girlfriend, it's
gotten very late. Her father has been waiting up, and you realize that
it's hours after midnight. He slams the door shut after you come in,
and loudly demands an explanation. Your muscles tense as he rudely
orders you in and sits you both on the couch. You feel nervous and a
little scared from the furious look on his face. You feel yourself
sweating and your face getting red as he goes on yelling.
STUDY WORDS: FATHER, MIDNIGHT, EXPLANATION, DOOR, MUSCLES, COUCH, FACE
NEW WORDS: SALMON, PURSE, PASTURE, TOAST, MARE, PRUNE, BRICK


129
Table 30
Mean recognition reaction time (msec) for study word and new word
targets for affectively pleasant, neutral, and unpleasant texts
with and without startle probes
New Word Targets
Pleasant
Neutral
Unpleasant
Mean
Startle Group
No Startle
857
883
868
869
Startle
878
891
886
885
Mean
868
887
877
877
Pleasant
Study Word Targets
Neutral Unpleasant
Mean
Startle Group
No Startle
863
906
869
879
Startle
851
857
836
848
Mean
857
882
852
864


141
Table 38continued
VERBAL INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
837
763
832
811
New Word
824
849
817
830
XXXXX
907
856
837
867
Mean
856
823
829
836
N-S Priming
-13
+86
-15
X-S Priming
+ 70
+93
+ 5
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
872
905
905
894
New Word
896
984
900
927
XXXXX
969
992
901
954
Mean
912
960
902
925
N-S Priming
+24
+ 79
- 5
X-S Priming
+ 97
+87
- 4


56
An example of an item on the test would be one which asked you to
consider an image which comes to your mind's eye of a red apple. If
your visual image was moderately clear and vivid you would check the
rating scale and mark "3" in the brackets as follows:
Item Rating
5. A red apple (3)
Now turn to the next page when you have understood these
instructions and begin the test.
Think of some relative or friend whom you frequently see,
considering carefully the picture that rises before your mind's eye.
Classify the imagers suggested by each of the following questions as
indicated by the degree of clearness and vividness specified on the
Rating Scale.
1 The exact contour of face, head, shoulders, and body ( )
2. Characteristic poses of head, attitude of body, etc. ( )
3. The precise carriage, length of step, etc., in walking ( )
4. The different colors in some familiar costume ( )
Think of seeing the following, considering carefully the picture
which comes before your mind's eye, and classify the image suggested by
the following question as indicated by the degrees of clearness and
vividness specified on the Rating Scale.
5.The sun as it is sinking below the horizon ( )
Think of each of the following sounds, considering carefully the
image which comes to your mind's eye, and classify the images suggested
by each of the following questions as indicated by the degree of
clearness and vividness specified on the Rating Scale.


13
when a word in a test list, taken from a study paragraph, was primed
(i.e., immediately preceded) by another word from the same paragraph,
recognition time was faster than if the prime was an unrelated word or
symbol. The effect was greater, the closer the logical relationship
between prime word and target word. Earlier pilot studies (Bradley,
Spence, & Lang, 1992; Appendix G) in which episodic priming was compared
for affectively unpleasant and neutral texts suggested that emotional
content altered the priming phenomenon reported by Ratcliff and McKoon.
The basic priming phenomenon is tested here for affective and neutral
text, i.e., do related text words facilitate recognition of text study
words. The following specific questions are addressed:
1. Does episodic priming occur with affective narratives, as has
been shown for neutral text, consistent with a propositional memorial
sub-structure?
2. Are there differences in priming for affectively pleasant and
affectively unpleasant text?
3. Is priming enhanced by imaginal processing instructions or by
the subjects' disposition to process imaginally?


100
35. Sauna
Pleasant. You are walking out, having just finished your favorite
class at your gym. The muscles in your abdomen and legs feel tight and
you head for the sauna. Stepping into the hot steam of the sauna, you
can feel yourself relaxing. Only a few other people are there, and you
stretch out comfortably on a bench. The heat penetrating your tired
limbs is a wonderfully soothing feeling. You step out into the cool air
again, feeling refreshed and energized.
Unpleasant. You have just finished a strenuous class at a rundown
old gym in town. Your legs are tired and shaking; you have a sick
feeling in your abdomen. Stepping into the hot steam of the sauna you
sink onto the bench, exhausted. You lean back against the wall, head
pounding, realizing you've overdone it. The overwhelming heat gives you
a suffocated feeling; you need some fresh air. Walking unsteadily
towards the door, you lean against it but it won't open.
STUDY WORDS: CLASS, GYM, ABDOMEN, LEGS, STEAM, BENCH, FEELING
NEW WORDS: BOOTS, THEME, DELIGHT, BARN, SLIPPER, ISLAND, ANGLE


35
The stronger valence effect for verbal instructed subjects was
greater for those who received startle probes, resulting in a Valence
(linear) X Instructions X Probe Group interaction, F (1,25) = 5.63,
p < .03 in the omnibus test (see Table 19, Appendix D). Furthermore,
startle probes potentiated a stronger linear valence effect in poor
imagers, Valence (linear) X Imagery Ability X Startle Group interaction,
F (1,25) = 5.87, e < -03 in the omnibus test (see Table 18, Appendix D).
To summarize the effects of these moderator variables on heart rate and
affect, the subjects who showed the strongest linear valence pattern
(least deceleration to pleasant content and greatest to unpleasant) were
poor imagers, under verbal instructions, who also received startle
probes.
Startle Reflex
Baseline EMG level. Analysis of baseline values for the startle
response yielded no significant main effects or interactions of the
independent variables in this study.
Startle magnitude. The omnibus analysis indicated that startle
responses did not vary significantly with text valence overall, F < 1 .
The Startle Position X Valence (linear) interaction, however, approached
significance, F (1,15) = 4.27, p = -057. As can be seen in Table 4,
mean blink magnitudes increased linearly from pleasant to unpleasant
paragraphs for late probes, valence (linear) F (1,15) = 4.81, p < -05.
Neither a significant linear nor quadratic valence effect was found for
early probes, F < 1.


172
specialization in child clinical psychology. In future research, I plan
to combine psychophysiological, cognitive, and behavioral measures to
explore the interaction of affective and attentional factors in
attentional deficits and learning disorders in pediatric populations.


DISCUSSION
The research provided broad support for the hypotheses that both
the encoding and retrieval of narrative text vary significantly with the
emotionality of story content. Thus at the learning stage, subjects
tended to read pleasant texts faster, and unpleasant texts slower, than
they did the neutral narratives. Furthermore, this effect of emotional
valence was more pronounced later in encoding, even though reading times
were then faster overall. Slower reading times for unpleasant
paragraphs were accompanied by strong decelerative heart rate responses,
and in response to startle probes, faster and larger blink reflexes.
Again, these effects were reliably stated only late in the subjects'
reading of the paragraph, when the story schema was already well
developed.
Memory measures also showed the effects of emotional content.
Thus, pleasant and unpleasant texts were better remembered than neutral
paragraphs at free recall. A similar pattern over affective contents
was found for recognition reaction time, which was most evident for
imagery instructed good imagers. Examination of primed recognition
suggests that non-text primes distracted subjects, producing a
performance deficit for neutral content, but did not interfere with the
recognition of words from affective paragraphs.
46


30
targets. As explained earlier (page 18), each of the non-verbal targets
appeared three times in each test list while target words appeared only
once in the experiment. Subject recognition for the non-verbal targets
thus became automatic over trials, resulting in much faster and less
variable reaction times than for verbal targets. These data were thus
not comparable to the verbal target conditions and were excluded from
further analysis.
Episodic Priming Analyses. A2X2X2X3X2 factorial design
again included the between-subject factors of Imagery Ability,
Instructions, and Startle Probe Group. The within-subjects variables
consisted of Affective Content, and Prime Type (Study Word, New Word).
As stated earlier (page 19), XXXXX-Study pairs were excluded from
analyses of priming condition. Important differences between the XXXXX
and study word primes (verbal vs. non-verbal stimuli; repetition both
within and between test lists; different levels of processing
requirements for decoding by subjects) served as the rationale for their
elimination as a control condition.
Recognition accuracy. The effects of the experimental factors on
accuracy in the recognition RT task were analyzed in order to address
the question of speed-accuracy trade-off as a possible factor in
recognition performance. The accuracy measure was calculated as percent
errors; analyses of these data paralleled those for recognition RT.


94
29. First Day of Class/Final Exam
Neutral. You walk into the classroom on the first day of class,
full of confidence. Sitting down, you put your books and papers under
your chair and sit back. A booklet is handed out, and a few homework
questions for the first assignment. You take out pencil and paper as
the organization of the course is outlined. Your brain is full of
information as you note down the date of the first test. Walking out
after class you reflect that this looks like an interesting course.
Unpleasant. You walk into the classroom to take your final with
no confidence at all. Behind in the course, you crammed all night, but
there was too much to learn. Opening the test booklet you pale with
apprehension as the questions look hard. You find you don't even have a
pencil and, embarrassed, have to borrow one. Your brain is filled with
confusion as you read the test; you just can't think. Avoiding your
professor's eyes, you turn in the half empty exam and leave.
STUDY WORDS: CLASSROOM, CONFIDENCE, BOOKLET, QUESTIONS, BRAIN, TEST,
PENCIL
NEW WORDS: VENDOR, RELUCTANCE, PEANUTS, MEDICINE, LUNCH, MOTEL,
SPINACH


BIOGRAPHICAL SKETCH
I completed my undergraduate education at the Ohio State
University in 1968, where my major field of study was English
Literature, my minor was French Language and Literature. I was admitted
to the masters program in general/experimental psychology at the
California State University, Northridge, in 1978. My masters research
was a psychophysiological study of the relationship between subjects'
electrodermal lability and their ability to control their heart rate
with biofeedback training.
I was admitted to the graduate program in the Department of
Clinical and Health Psychology at the University of Florida in 1982.
During my graduate training I served as research assistant in Dr. Lang's
psychophysiology laboratory. In this setting I gained experience in the
design, execution, and analysis of a series of experiments which were
part of a larger program of research involving the study of various
aspects of affective cognition. The present dissertation represents the
culmination of this pilot work which validated the affective properties
of the experimental materials employed in the present research, and
provided early replications of several of the phenomena reported in this
work.
I completed my clinical internship at the University of Alabama at
Birmingham Psychology Training Consortium where I developed a
171


125
Table 25
Mean startle
latency (msec) for probes
presented
early and late
in affectively pleasant, neutral.
and unpleasant texts
for qood and
poor imaqerv ability subiects
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
59.2
53.5
56.5
56.4
Late
55.0
51.6
52.0
52.9
Mean
57.1
52.6
54.2
54.6
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
51 .4
48.9
54.0
51 .4
Late
55.2
57.6
48.5
53.8
Mean
53.3
53.2
51.3
52.6
Table 26
Mean startle
latency (msec) for probes
presented early and late
in affectively pleasant, neutral.
and unpleasant texts
for subiects receivinq verbal and
imaqery instructions
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
54.6
53.1
49.3
52.3
Late
50.6
49.6
47.0
49.1
Mean
52.6
51 .4
48.2
50.7


99
34. Hike through Woods
Pleasant. It's late spring, and you and your group have hiked
miles into the woods. Suddenly, a beautiful baby fawn steps uncertainly
onto the path just ahead. It gazes at you in silence with deep brown
eyes, munching on some leaves. You glance sideways to meet the look of
pure delight on the others' faces. Caught with this wonderful sight,
you stand completely still on the path. You marvel at this small,
spotted brown fawn standing trustingly before you.
Unpleasant. It's late spring, and you and your group have hiked
miles into the woods. You jump violently off the path realizing you've
just stepped on a rattlesnake. It coils back among some leaves fixing
you in silence with expressionless eyes. You glance sideways at the
looks of fear and helplessness on the others' faces. Frozen at the
sight of the snake's coiled body, you are too afraid to move. Striking
with lightning speed, you feel its fangs sink deep into your calf.
STUDY WORDS: PATH, EYES, SPRING, WOODS, LEAVES, MILES, GROUP
NEW WORDS: CORN, FRAUD, FLOOD, BELLS, PARSLEY, PEBBLE, CORK


170
Osgood, C., Suci, G., & Tannenbaum, P. (1957). The measurement of
meaning. Urbana, IL: University of Illinois Press.
Paivio, A. (1971). Imagery and verbal processes. New York: Holt,
Rinehart, & Winston.
Pylyshyn, Z. (1973). What the mind's eye tells the mind's brain: A
critique of mental imagery. Psychological Bulletin. 80, 1-22.
Ratcliff, R., & McKoon, G. (1978). Priming in item recognition:
Evidence for the propositional structure of sentences. Journal of
Verbal Learning and Verbal Behavior. 17. 403-417.
Ratcliff, R., & McKoon, G. (1981). Automatic and strategic priming in
recognition. Journal of Verbal Learning and Verbal Behavior, 20.
204-215.
Schank, R. C. (1970). A conceptual dependency representation for a
computer-oriented semantics. (Doctoral dissertation, University
of Texas, 1970). Dissertation Abstracts International. 30,
2994A-2995A.
Sheehan, P. W. (1967). A shortened form of Betts' questionnaire upon
mental imagery. Journal of Clinical Psychology. 223, 380-389.
Simons, R. F., & Zelson, M. F. (1985). Engaging visual stimuli and
reflex blink modification. Psychophysiology, 22, 44-49.
Sokolov, Y. N. (1963). Perception and the conditioned reflex (S. W.
Waydenfeld, Trans.). New York: Macmillan. (Original work
published 1958).
Spence, E. L., Bradley, M., & Lang, P. J. (1985). Imagery ability and
psychophysiological response to affective texts. Psycho
physiology. 22, 613. Abstract.
Strosahl, K. D. & Ascough, J. C. (1981) Clinical uses of mental
imagery: Experimental foundations, theoretical misconceptions,
and research issues. Psychological Bulletin, 89, 422-438.
Vrana, S. R., & Lang, P. J. (1990). Fear imagery and the startle-probe
reflex. Journal of Abnormal Psychology. 99, 189-197.
Vrana, S. R., Spence, E. L., & Lang, P. J. (1988). The startle probe
response: A new measure of emotion? Journal of Abnormal
Psychology. 97. 487-491.
Yerkes, R. M., & Dodson, J. D. (1908). The relation of strength of
stimulus to rapidity of habit-formation. Journal of Comparative
and Neurology and Psychology. 18. 459-482.


93
28. Store Clerk
Neutral. You are opening the register as you start your shift in
a nice clothing store. Only a few customers are in your area of the
store, so you can take your time. You turn off the alarm, find the key
to the drawer lock, and get out a pen. You are all set up now, and you
wander over to look through the new blouses. There a beautiful white
silk one with a front pocket that you especially like. A customer is
browsing through the skirts nearby and you go over to help her.
Unpleasant. You are closing the register after the late shift at
a convenience store. The store is in a secluded area, and you don't
like being there alone at 2 a.m. At last it's time to close; you turn
the lock on the door and set the alarm. Suddenly you look up to see a
man's face watching you through the window. You freeze with fear,
staring, as he pulls a gun out of his jacket pocket. You jump as he
shouts something at you, motioning with a jerk towards the door.
STUDY WORDS: REGISTER, SHIFT, STORE, AREA, LOCK, ALARM, POCKET
NEW WORDS:
TACKLE, SINK, DIME, ORGAN, RIBBON, BRIDE, BUMPS


139
Table 37continued
STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
793
815
842
816
New Word
840
859
774
824
XXXXX
820
831
850
834
N-S Priming
+ 47
+ 44
-68
+ 7
X-S Priming
+ 27
+ 16
+ 8
+ 17
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
848
827
856
844
New Word
864
904
841
870
XXXXX
942
905
855
901
N-S Priming
+ 16
+ 77
-15
+ 26
X-S Priming
+ 94
+ 78
- 1
+ 57


144
Table 42
Mean percent errors for new word and study word targets,
for affectively pleasant, neutral, and unpleasant texts.
with and without presentation of startle probes
New Word Targets
Pleasant
Neutral
Unpleasant
Mean
Startle Group
No Startle
3.47
3.56
3.47
3.50
Startle
3.90
4.10
5.11
4.37
Mean
3.68
3.83
4.29
3.93
Pleasant
Study Word
Neutral
Targets
Unpleasant
Mean
Startle Group
No Startle
7.25
7.71
7.88
7.61
Startle
8.45
10.27
8.82
9.18
Mean
7.85
9.00
8.35
8.40


80
15. Favorite Rock Star
Unpleasant. You are over visiting with friends, while the radio
plays in the background. Suddenly the DJ says that your favorite
musician was just shot and killed. You sit frozen and numb; some of the
others sit quietly crying in the room. You can't believe this is true.
Your spirit sinks as you realize how empty this makes the future. The
news is so upsetting that you don't know what to do next.
Pleasant. You are visiting with friends, while the radio plays in
the background. Suddenly the DJ announces that you won a trip to meet
your favorite musician. You sit thrilled and amazed as the others
scream and dance around the room. You can't believe it! Your spirit
soars as you realize the future holds this great meeting. The news is
so exciting that you don't know what to do next.
STUDY WORDS: RADIO, BACKGROUND, OTHERS, ROOM, SPIRIT, FUTURE, MUSICIAN
NEW WORDS: DIRT, TRIFLE, ABILITY, COTTAGE, POTATO, CANDY, BOTTLE


RESULTS
Encoding Measures
Read Time
Table 1 presents mean read times for the early, middle, and late
sentence pairs of paragraphs with affectively pleasant, neutral, or
unpleasant content. Consistent with findings in the text processing
literature, subjects read the later sentences in the paragraphs more
rapidly than earlier sentences, F (2,96) = 59.71, p < 011.
Furthermore, paragraphs with pleasant content were generally read more
quickly than unpleasant texts, yielding a linear relationship between
read time and pleasant versus unpleasant affective content, Valence
F (2,92) = 2.17, p = .12; Valence (linear) F (1,46) = 9.46, p < .01.
Finally, on the average, good imagers tended to read all sentences of
the paragraphs faster (mean per sentence = 4.7 secs) than poor imagers
(mean per sentence = 5.6 secs), F (1,46) = 3.55, p = .06.
Affective content and sentence interacted significantly,
F (4,184) = 3.97, p < .01 in the omnibus test. As can be seen in
Table 1, read times for early sentences followed a quadratic pattern
with affective content, Valence (quadratic) X Sentence F (2,92) = 5.86,
1 Where appropriate, reported p values reflect application of the
Greenhouse-Geisser correction.
31


ACKNOWLEDGEMENTS
The present research project was the culmination of several pilot
studies, and involved the collaborative efforts of many people. I am
especially grateful to Peter Lang, who generously supported this effort,
both by making available the considerable resources of his psychophysio-
logical laboratory and technical staff, and through his financial
support of my graduate work at Florida. I would also like to express my
deep appreciation to the members of my dissertation committee, Russell
Bauer, Keith Berg, Dawn Bowers, Ira Fischler, and Edward Valenstein.
They not only made valuable contributions to my dissertation research,
but also provided continued support, encouragement, and intellectual
guidance throughout my graduate training.
I would also like to thank Bruce Cuthbert, who provided training
and support in the laboratory in too many ways to enumerate; and to
Margaret Bradley for her literary, technical, and intellectual
contributions to this project and the earlier pilot studies. I also
benefitted from many discussions and sharing of ideas with my fellow
graduate students, Scott Vrana and Mark Greenwald.
Finally, I would like to thank my husband, Ed Cook, for his
untiring support and assistance in literally every aspect of this
project.
11


137
Table 36continued
STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
787
763
800
784
New Word
797
847
760
801
XXXXX
810
815
785
803
N-S Priming
+ 10
+84
-40
+ 18
X-S Priming
+ 23
+52
-15
+20
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
854
878
897
876
New Word
907
916
855
892
XXXXX
952
921
920
931
N-S Priming
+ 53
+ 38
-42
+ 17
X-S Priming
+98
+ 43
+ 23
+ 55


75
10. Elevator
Neutral. You are part of a crowd of shoppers, running errands in
a new shopping mall. Walking around, you discover the mall has a glass
elevator and a nice fountain. Your errands are run, and you notice you
feel slightly on edge from hunger. On your way out, you stop and treat
yourself to a delicious cold milkshake. It's a relief to get out into
the cool darkness after the busy confusion there. As you drive away,
you start to relax and think about watching a movie at home.
Unpleasant. You are part of a crowd of shoppers, running errands
in a large shopping mall. The elevator in the mall is full, but you're
in a hurry and squeeze in anyway. People are pressed in around you; you
feel uncomfortably hot and on edge. Suddenly the elevator jerks to a
stop; you startle as a loud alarm goes off. There is no relief as panic
spreads in the darkness in the tight enclosure. You clench your teeth
as a baby screams; you feel yourself sweating heavily.
STUDY WORDS: SHOPPERS, ERRANDS, MALL, ELEVATOR, EDGE, RELIEF, DARKNESS
NEW WORDS: SUNSET, RANCH, BALLET, VINES, LIQUID, CARAMEL, EGGS


73
8. Beach Picnic
Unpleasant. You're having a great time on a picnic by the ocean
with some good friends. Some are out swimming, and you look twice when
you think you spot sharks. Your feet pound through the surf as you run
calling, the fins now obvious. Unaware, your friends don't hear you,
and you are gripped with horror. Your heart pounding in your chest, you
run across the sand shouting and waving. You scream in desperation as
someone is jerked under, and the water turns red.
Pleasant. You're having a great time on a picnic by the ocean
with some good friends. Some are out swimming, and you wave to them as
you walk along the surf. Walking along, you stumble as your feet trip
over a pair of lost swim fins. You look up to see a frisbee sailing in
your direction, and you go after it. Your heart pounds in your chest as
you run hard across the sand to catch it. You laugh and shout to your
friends as you grab it, full of high spirits.
STUDY WORDS: PICNIC, OCEAN, FEET, FINS, HEART, SAND, FRIENDS
NEW WORDS: DEGREE, WAGON, PINT, CHARM, TENT, CALF, SCISSORS


146
Table 43continued
STUDY WORD TARGETS
Imagery Instructions
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
9.88
7.00
9.18
8.69
Poor
7.42
6.58
6.82
6.94
Mean
8.65
6.79
8.00
7.81
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
6.48
10.64
8.49
8.54
Poor
7.62
11.73
8.92
9.42
Mean
7.05
11.19
8.70
8.98


36
No main effects or interactions involving instructions or imagery
ability occurred for this measure.
Table 4
Mean startle magnitude (arbitrary A-D units) for probes measured
early and late in affectively pleasant, neutral and unpleasant texts
Pleasant Neutral Unpleasant
Startle Position Mean
Early
114.6
133.9
111.2
119.9
Late
134.0
149.8
165.1
149.6
Mean
124.3
141 .9
138.2
Onset latency. Compared to the magnitude data, startle latency
showed a closerbut still not statistically significant--overall
relationship with affective content (Pleasant = 55.2 msec,
Neutral = 52.9 msec, Unpleasant = 52.7 msec), F (2,28) = 1.63, p = .21
in the omnibus test; Valence (linear) F = 2.11, p = .17. Emotional
content affected blink latency differently, however, for subjects
receiving verbal versus imagery processing instructions, and
particularly, depending on whether the startle probe occurred early or
late in the paragraph (see Figure 2), Valence X Instructions X Startle
Position interaction, F (2,28) = 3.72, p = .053 in the omnibus test;
Valence (linear) X Instructions X Startle Position, F (1,14) = 5.99,
p < .03.
As can be seen in the left panel of Figure 2, instructions
interacted with emotional content to affect startle latency early in the
encoding process, Valence (linear) X Instructions F (1,14) = 5.41,


148
Table 46
Mean percent errors for study words primed by a study word, a new word,
or XXXXX for affectively pleasant, neutral, or unpleasant texts for
subjects receiving imagery and verbal instructions
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
7.15
6.34
5.52
6.33
New Word
9.96
6.85
9.31
8.70
XXXXX
8.84
7.18
9.18
8.40
Mean
8.65
6.79
8.00
7.81
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
6.51
3.76
8.11
6.13
New Word
6.49
11.15
6.53
8.06
XXXXX
8.14
18.64
11.47
12.75
Mean
7.05
11.19
8.70
8.98


45
New-Studv. For study words primed by new words, no main effects
or interactions of the experimental factors on error rate were observed.
Episodic Priming
Facilitation of recognition RT for SS relative to NS prime-target
pairs was not reliable when averaged across experimental conditions,
(SS = 847 msec; NS = 859 msec), F(1,46) = 1.65, p > .20. As noted
previously (p. 44), however, priming did vary with emotional content of
the paragraphs, Valence X Priming Condition F (2,92) = 8.52, p < .01
(see also Figure 3). Thus, individual tests were conducted for each
affective category. Consistent with findings in the episodic priming
literature, reliable priming occurred for neutral texts, (56 ms), F
(1,46) = 10.79, p = .01. However, for affectively pleasant texts, there
was no significant priming (F < 1); and for affectively unpleasant
texts, RT latency was actually 34 ms longer for study-study compared to
new-study prime-target pairs, F (1,46) = 5.17, p < .03.


151
Table 48continued
VERBAL INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
5.36
2.38
8.56
5.44
New Word
6.80
7.62
5.33
6.58
xxxxx
7.28
21 .93
11.57
13.60
Mean
6.48
10.64
8.49
Poor Imagers
Pleasant Neutral Unpleasant Mean
Prime Type
Study Word
7.66
5.15
7.65
6.82
New Word
6.18
14.69
7.74
9.54
XXXXX
9.01
15.35
11.36
11 .90
Mean
7.62
11.73
8.92


UNIVERSITY OF FLORIDA
3 1262 08554 8138


60
Instructions to Subjects
In this study, you will be reading paragraphs that will appear on
this monitor screen, and then respond yes or no to a list of words,
depending on whether or not they were in those paragraphs.
The exact procedure will be as follows: You will hold these two
buttons, one in each hand (get subject to hold buttons), and use them to
indicate yes or no to the test words as they appear. You should press
the right button if the word is in the paragraph and the left button if
the word is not in the paragraph. You should press a button as fast as
possible while still making accurate decisions. You will also use the
buttons to present the paragraphs line by line.
After you press both buttons to start, the first line of a
paragraph will appear on the monitor screen. Read through the sentence
at a comfortable pace, and then press both buttons again to move to the
next sentence.
IMAGERY INSTRUCTIONS:
As you read the sentences, please make a special effort to form as
vivid an image of the events described as you can. Place yourself in
the image as a participant in the storyas if you were actually
experiencing it all personally. The more you can experience the actions
and feelings, the more vivid your image will be. Read through the lines
at your own pace, constructing your image as you read. The more vivid
your image of the events described in the paragraph, the more it is a
personal experience, the better you'll be able to remember it for the
subsequent test.


18
For the subset of subjects for whom a startle probe occurred on
some occasions while they read the text, care was taken that test words
and startle probes did not occur in the same sentence.
Targets. Half the targets (6) were verbal stimuli (Study or New
words) and half were non-verbal stimuli (*****, =====). Verbal targets
were also evenly divided between Study (3) and New words (3). Study
words were always nouns that occurred in the preceding text. New words
were nouns that did not occur in the study text, and had no semantic
relationship to the topic of the study text or to the other words in the
test list. Statistical analyses focussed on reaction time and accuracy
for (1 ) the prime-target pairs for which the target was a study word or
a new word, and (2) study word targets that were preceded by either a
study word or new word prime.
Non-verbal targets were also equally divided, half (3) requiring a
right-hand button press (******), and half a left-hand button press
(======). These stimuli differed from the verbal targets in that they
required pattern recognition for a choice reaction time task rather than
verbal decoding for a recognition memory task. In addition, the non
verbal targets were repeated in every test list while verbal test list
items appeared only once in the experiment. Thus, the non-verbal items
involved minimal cognitive effort and responses to them became automatic
over trials. Predictably, reaction time for the non-verbal targets was
much faster (by over 200 ms) and less variable than for verbal targets.
These data were thus not comparable to the verbal target conditions and
were excluded from further analysis.


APPENDIX B
TEXT MATERIALS AND TEST LIST WORDS
Instructions and Practice Paragraph
Experimenter Instructions
This is a warm-up exercise to teach you special symbols used in
this study. This one (press right button to display *******) signals
you to make a "yes" response. This one (press left button to display
=======) signals you to make a "no" response. Try it a few times:
******** ======== ******** ========. These symbols will
appear in the test lists which will follow each paragraph. What does
======= signal you to do? What does ******* signal you to do?
Practice Paragraph
Today you turned twenty, and you wonder if any of your friends
will notice. You go out running after class and go over to study at a
friend's house later. Knocking at her door, you wonder if she has any
idea that it's your birthday. As you enter, people seem to jump out
from everywhere shouting "SURPRISE!" You gasp with delight as you see a
big cake with candles and balloons all over. You smile with affection
as they laugh gleefully that you suspected nothing.
STUDY WORDS: NOTICE, BIRTHDAY, IDEA, PEOPLE, DELIGHT, CAKE, NOTHING
NEW WORDS: POSTER, ERASER, BOLT, PERFUME, FLINCH, HYDROGEN, BLOOMS
65


157
Subjects with poor imagery ability rated unpleasant paragraphs as
less unpleasant and pleasant paragraphs as less pleasant than subjects
with good imagery ability, yielding an Imagery Ability X Valence
interaction: F (1,47) = 8.51, p < .01 (See Table 53).
TABLE 53
Mean valence ratings for pleasant, neutral, and unpleasant texts
for subjects reporting good and poor imagery ability
Pleasant
Neutral
Unpleasant
Valence Ratings
Good
8.4
5.0
1 .7
Poor
8.2
4.9
2.0
Arousal Ratings
Arousal ratings also differed reliably among affective categories
F(2,94) = 98.49, p < .0001. Affectively unpleasant texts were
significantly more arousing than pleasant texts, F (1,47) = 88.40,
p < .0001, and pleasant and unpleasant texts received higher arousal
ratings than neutral texts, F (1,47) = 100.99, p < .0001.
Imagery ability also affected arousal ratings. Table 54
summarizes mean arousal ratings for affectively negative, neutral, and
positive texts for subjects reporting good and poor imagery ability.
Good imagers rated texts as more arousing than poor imagers, F (1,47) =
5.35, p < .03.


112
Table 13
Mean heart rate change (beats per minute) for early, middle, and late
sentence pairs of affectively pleasant, neutral and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1 .02
-1 .88
-1 .45
-1.45
Middle
-2.74
-3.11
-3.49
-3.11
Late
-3.02
-3.34
-3.69
-3.35
Mean
-2.26
-2.78
-2.88
Table 14
Mean heart rate
chancre (beats
per minute)
for earlv.
middle, and late
sentence pairs of affectively
pleasant, neutral and unpleasant texts
for
subjects with
qood and poor imaqerv
ability
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1 .61
-2.80
-1 .80
-2.07
Middle
-3.08
-3.73
-3.71
-3.51
Late
-3.37
-4.08
-3.89
-3.78
Mean
-2.69
-3.54
-3.13
-3.12
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.43
-0.95
-1.10
-0.83
Middle
-2.40
-2.48
-3.27
-2.72
Late
-2.67
-2.60
-3.48
-2.92
Mean
-1 .83
-2.01
-2.62
-2.16


126
Table 26--continued
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
56.0
49.3
61 .2
55.5
Late
59.6
59.6
53.4
57.6
Mean
57.8
54.4
57.3
56.5
Table 27
Mean recognition
reaction time
(msec) for
study word and
new word
targets for affectively oleasant. neutral.
and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Target Type
Study
857
881
853
864
New
868
887
887
877
Mean
862
884
865
870


165
used to demonstrate logical distance between propositions in text.
Subjects read 6-line paragraphs and then responded to a test list of
prime-target pairs. Recognition RT for study words primed by another
study word from the same proposition was compared to RT for study words
primed by another study word taken from a different proposition in the
paragraph. Ratcliff and McKoon's design and methods were again closely
replicated in our adaptation of their study. However, rather than
combine their textual materials with affective texts developed in our
laboratory, paragraphs were employed that had been developed in a
previous validation study (Spence, Bradley, & Lang, 1985). These
materials were the same as those employed in the present experiment.
Consistent with the results of Study 1, recognition RT was slower
for study-study prime-target pairs from affectively negative texts
compared to affectively positive and neutral texts. No differential
priming was found between near and far propositions. In a free recall
task similar to that carried out in the dissertation experiment,
subjects were more likely to recall affectively negative texts than
positive or neutral texts.
The results of the present dissertation experiment are consistent
with these earlier studies in that recognition RT was significantly
slower for study-study prime-target pairs from affectively negative
compared to positive and neutral texts.


28
earlier validation study (Spence et al., 1985). Subjects in the present
study, however, also provided ratings of the texts after the experiment.
This permitted a post-hoc recategorization of texts, based on this more
proximal, group-specific assessment. The resulting redistribution of
paragraphs produced an approximately equal number of affectively
pleasant and unpleasant texts, but fewer neutral texts. For each
subject, unequal numbers of texts contributed to the cell averages.
Three subjects were not included; two of them failed to rate any
paragraph in the neutral range, and another subject rated only one text
as affectively neutral. The ratings distributions for the validation
study, the present study, and a contingency table illustrating the
stability of ratings classifications are provided in Appendix E.
Statistical analyses were performed, using both categorization
systems, with similar overall results. Because they more closely
represent this specific sample, and to avoid redundancy, only results
based on the subjects' own ratings are presented here. Tests of valence
involve orthogonal contrasts of pleasant versus unpleasant (linear
effect), and affective (pleasant and unpleasant) versus neutral content
(quadratic effect).
Encoding Measures
Read time and heart rate. These data were analyzed in a 2 X 2 X 2
X 3 X 3 factorial design. Between-subject variables were Imagery
Ability (Good, Poor), Instructions (Imagery, Verbal), and Group (Startle
Probe, No Startle Probe); within-subject variables were Affective
Content (Pleasant, Neutral, Unpleasant) and Sentence Pair (1&2, 3&4,


124
Table 23
Mean startle magnitude (arbitrary A-D units) for probes presented
early and late in affectively pleasant, neutral, and unpleasant texts
for good and poor imagery ability subjects
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
158.7
190.6
142.0
163.8
Late
167.2
195.7
200.7
187.9
Mean
163.0
193.2
171.4
175.9
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
70.5
77.2
80.4
76.0
Late
100.7
103.9
129.6
111.4
Mean
85.6
90.6
105.0
93.7
Table 24
Mean startle
latency (msec)
for probes
: presented early and
in affectively pleasant, neutral.
and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
55.3
51 .2
55.3
53.9
Late
55.1
54.6
50.2
53.3
Mean
55.2
52.9
52.7
53.6


167
Carroll, M., & Kirsner, K. (1982). Context and repetition effects in
lexical decision and recognition memory. Journal of Verbal
Learning and Verbal Behavior. 21. 55-69.
Christianson, S.-, & Loftus, E. F. (1987). Memory for traumatic
events. Applied cognitive psychology. 1, 225-239.
Cirillo, R. K., & Foss, D. J. (1980). Text structure and reading time
for sentences. Journal of Verbal Learning and Verbal Behavior.
19, 96-109.
Cook, E. W., Ill, Hawk, L. W., Davis, T. D., & Stevenson, V. E. (1991).
Affective individual differences and startle reflex modulation.
Journal of Abnormal Psychology, 100. 5-13.
Cook, E. W., Ill, Melamed, B. G., Cuthbert, B. N., McNeil, D. W., Lang,
P. J. (1988). Emotional imagery and the differential diagnosis
of anxiety. Journal of Consulting and Clinical Psychology. 56.
734-740.
Cook, E. W., Ill, Davis, T. L., Hawk, L. W., Jr., Spence, E. L., &
Gautier, C. H. (in press). Fearfulness and startle potentiation
during aversive visual stimuli. Psychophysiology.
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A
framework for memory research. Journal of Verbal Learning and
Verbal Behavior. 11. 671-684.
Davis, M. (1984). The mammalian startle response. In R. C. Eaton
(Ed.) Neural mechanisms of startle behavior (pp. 287-351). New
York: Plenum Press.
Galambos, J. A., & Rips, L. J. (1982). Memory for routines. Journal
of Verbal Learning and Verbal Memory, 21. 260-281 .
Graham, F. K. (1979). Distinguishing among orienting, defense, and
startle reflexes. In H. D. Kimmel, E. H. van Olst, & J. F
Orlebeke (Eds.). The orienting reflex in humans (pp. 137-167).
Hillsdale, NJ: Erlbaum.
Graham, F. K., & Clifton, R. K. (1966). Heart-rate change as a
component of the orienting response. Psychological Bulletin. 65,
305-320.
Hamm, A. 0., Globisch, J., Cuthbert, B. N., & Vaitl, D. (1991). Startle
reflex modulation in simple phobics and normals. Psycho
physiology. 27, S37.
Hebb, D. 0. (1968). Concerning imagery. Psychological Review. 75.
466-477.


77
12. Parents' Pool Party
Neutral. The evening of your parents' pool party the guests are
not very exciting. Finally tired of hearing the old familiar stories,
you head for your room. You take off your watch, your shoes and
clothes, and put on your bathing suit. Things have quieted down outside
by the pool, and you dive in for a quick swim. You swim vigorously
through the cool refreshing water using your full strength. You come
out of the pool feeling much better and head off for the refreshments.
Unpleasant. It's late and some of the guests at the pool party
are getting very drunk. You ask them to quiet down but they laugh,
grabbing your arms and both legs. You panicyou're wearing expensive
shoes and a watch that isn't waterproof. As they drag you towards the
pool you yell that your watch will be ruined. Angrily, you fight with
all your strength as they swing you out over the water. You plunge into
the cold water and come up choking, completely furious.
STUDY WORDS: EVENING, GUESTS, WATCH, SHOES, POOL, STRENGTH, WATER
NEW WORDS: SPACE, DWARF, CREAM, MIRROR, ATMOSPHERE, ONION, SONG


70
5. Airplane Trip
Unpleasant. You startle as the plane lurches violently in
midflight on your way home. Confused, the passengers stir restlessly as
the cabin lights go out. Suddenly there's an odd noisethe engines'
sputtering silences all talk. Grabbing the chair arms hard, you feel
the hull shudder violently. The hostess is thrown sideways, spilling
the drinks she was carrying. Screams pierce the dark as the plane tilts
sideways and starts to fall.
Pleasant. You are flushed with excitement as the plane soars
high, taking you home. Soon you will be with your boyfriend again after
a long separation. The noise of small talk around you is of no
interest, you are so happy. Smiling with a rush of pleasure, you
imagine him in your arms again. When the hostess comes by with drinks,
you decide on champagne to celebrate. As you approach your destination
you smile happily, full of anticipation.
STUDY WORDS: PLANE, HOME, NOISE, TALK, DRINKS, HOSTESS, ARMS
NEW WORDS: REMAINDER, COUNTER, PEPPER, APPLES, GRAPEFRUIT, FRONTIER,
CAPTIVE


PSYCHOPHYSIOLOGY AND MEMORY:
THE ENCODING AND RETRIEVAL OF AFFECTIVE TEXT
By
ELLEN LOUISE SPENCE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1992

ACKNOWLEDGEMENTS
The present research project was the culmination of several pilot
studies, and involved the collaborative efforts of many people. I am
especially grateful to Peter Lang, who generously supported this effort,
both by making available the considerable resources of his psychophysio-
logical laboratory and technical staff, and through his financial
support of my graduate work at Florida. I would also like to express my
deep appreciation to the members of my dissertation committee, Russell
Bauer, Keith Berg, Dawn Bowers, Ira Fischler, and Edward Valenstein.
They not only made valuable contributions to my dissertation research,
but also provided continued support, encouragement, and intellectual
guidance throughout my graduate training.
I would also like to thank Bruce Cuthbert, who provided training
and support in the laboratory in too many ways to enumerate; and to
Margaret Bradley for her literary, technical, and intellectual
contributions to this project and the earlier pilot studies. I also
benefitted from many discussions and sharing of ideas with my fellow
graduate students, Scott Vrana and Mark Greenwald.
Finally, I would like to thank my husband, Ed Cook, for his
untiring support and assistance in literally every aspect of this
project.
11

TABLE OF CONTENTS
ACKNOWLEDGEMENTS ii
ABSTRACT v
INTRODUCTION 1
Theoretical Background 1
The Network Model 1
Imagery and Response Activation 2
Bio-Information Theory 3
Emotion and Memory 4
The Research Problem 4
Encoding Hypotheses 5
Retrieval Hypotheses 10
METHODS 14
Subjects 14
Materials 14
Texts 14
Test Lists 17
Apparatus 21
Procedures 22
Data Reduction 25
Encoding Measures 25
Retrieval Measures 26
Data Analysis 27
Imagery Ability 27
Affective Classification of Texts 27
Encoding Measures 28
Retrieval Measures 29
RESULTS 31
Encoding Measures 31
Read Time 31
Heart Rate 33
Startle Reflex 35
Retrieval Measures 38
Free Recall 38
Episodic Recognition 38
DISCUSSION 46
Encoding Measures 47
Retrieval Measures 51
Overall Conclusions 54
APPENDIX A QUESTIONNAIRES, INSTRUCTIONS AND DEBRIEFING STATEMENT 55
APPENDIX B TEXT MATERIALS AND TEST LIST WORDS 65
APPENDIX C IMAGERY ABILITY SCORES FOR INSTRUCTION GROUPS 102
iii

APPENDIX D TABLES OF MEANS FOR DEPENDENT MEASURES 103
APPENDIX E AFFECTIVE RATINGS FOR PARAGRAPHS 156
APPENDIX F SELF-ASSESSMENT MANIKIN (SAM) 161
APPENDIX G SUMMARY OF RESULTS FOR PILOT PRIMING STUDIES 164
REFERENCES 166
BIOGRAPHICAL SKETCH 171
IV

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
PSYCHOPHYSIOLOGY AND MEMORY:
THE ENCODING AND RETRIEVAL OF AFFECTIVE TEXT
By
Ellen Louise Spence
May, 1992
Chairman: Peter J. Lang, Ph.D.
Major Department: Clinical and Health Psychology
The conceptual framework for this research was derived from Lang's
bio-informational theory of emotion which states that affective memory
representations contain information about physiological mobilization and
action. Imaginal processing of emotional text content was posited to
further increase the representation of physiological activation in
memory. The resulting affective information structure contains more
information than representations of neutral content, and supports faster
and more probable retrieval from memory.
Subjects were trained in a verbal or imaginal encoding strategy.
They then read brief narratives that depicted pleasant, neutral or
unpleasant experiences. Verbal instructions emphasized attention to the
exact words used in the text, while imagery instructions focussed on the
events described as a vivid personal experience. Read time was employed
as a measure of the time required to construct the text schema, while
v

heart rate and the startle reflex verified the presence of emotional
response during encoding of the paragraphs. Affective content and
imaginal processing were predicted to increase encoding time, to produce
physiological activation during construction of the text schema, and to
facilitate memory retrieval.
The results showed that affectively unpleasant content was read
more slowly than pleasant material. Unpleasant paragraphs also prompted
greater heart rate decrease, and larger and faster probe startle
responses, compared to pleasant contents. These content effects were
all more pronounced for the later sentences in the paragraph, after the
affective content was established. The read time and heart rate valence
effects tended to be more reliable for poor imagers and for subjects
under verbal instructions. The startle/valence effect was more reliable
under imagery instructions.
Retrieval measures included free recall and episodic recognition
reaction time. Emotional content facilitated free recall performance
relative to neutral content. Affective content, good imagery ability,
and imagery instructions facilitated the speed and accuracy of
recognition memory for study word targets. Episodic priming occurred
for neutral texts, but was not reliable for affective texts.
These results support the hypotheses that both the encoding and
retrieval of narrative text vary significantly with affective versus
neutral content. In addition, emotional content interacted with
encoding strategy (imaginal or verbal) to further modulate these
processes.
vi

INTRODUCTION
The purpose of this research is to examine how narrative text is
processed into story-images, and how the information structure of such
images differs from representations of the specifically verbal content
of text. The experimental procedure employed here is designed to test a
set of propositions that were conceptually derived from the bio-
informational theory of imagery (Lang, 1977, 1979, 1984, 1985), and from
associated supporting research (e.g., Cook, Melamed, Cuthbert, McNeil, &
Lang, 1988; Lang, Levin, Miller, Kozak, 1983; Miller, Levin, Kozak,
Cook, & McLean, & Lang, 1987; Vrana & Lang, 1990).
Theoretical Background
The Network Model
In general, cognitive psychologists hold that meaningful text is
parsed by the brain into concept units (e.g., Anderson, 1976; Kintsch,
1974; Schank, 1970). Following Anderson and Bower (1973), these
concepts are presumed to be inter-linked by association (perhaps, with a
propositional sub-organization), forming an information network that
represents an individual's knowledge of the text content. Subsequently,
if a newly presented stimulus matches a concept in the network (e.g., a
word from the previously seen text is presented on a CRT screen)
retrieval is initiated. This associative retrieval process eventually
activates the entire network structure, and is reported by the subject
1

2
as his or her verbal memory for the text. Several parameters determine
network activation and the resulting memory retrieval. These include
the number and type of network concepts represented in the activating
percept, and the associative strength between concepts in memory.
Measures of recognition memory and free-recall of text content have been
used to assess their influences (e.g., Anderson & Bower, 1973; Carroll &
Kirsner, 1982; Galambos & Rips, 1982; Kintsch & van Dijk, 1978; McKoon &
Ratcliff, 1980; Ratcliff & McKoon, 1978).
Imagery and Response Activation
Variants of the above model have been widely employed in the
semantic analysis of text processing. In 1973, Pylyshyn suggested that
a similar analysis might also account for imaginal memory. In his view,
images are internally constructed propositional descriptions of objects.
Kieras (1978) later elaborated this idea, suggesting that imagery
differed from other information structures, specifically, in the amount
of modality-specific information that was in the network. Thus, vivid
imagers, more than poor imagers, would give elaborate sensory
descriptions of previously seen percepts, or readily detect errors of
detail in a picture memory task. While this seems plausible, the
experimental evidence for this view is not compelling (see Strosahl &
Ascough, 1981). Furthermore, this sensory information emphasis neglects
a vast psychophysiological literature, which has reliably found imagery
effects, not in afference, but in the efferent system (e.g., Jacobson,
1930; Lusebrink & McGuigan, 1989; McGuigan & Schoonover, 1973; Miller et
al., 1987). That is, when subjects imagine, remember, or re-experience

3
past events, this mental imagery is reliably accompanied by somatic and
visceral changes which are similar in pattern to those occurring in the
referent perceptual contexts. An analysis of imagery based on the
activation of efferent information was developed in a neurophysiological
framework by Hebb (1968), and elaborated in cognitive terms by Lang's
(1979) bio-informational theory of emotional imagery.
Bio-Information Theory
The bio-informational model of memory imagery assumes the
essentials of an associative network model. It differs from other views
in that response information is included as a concept category. That is
to say, an image network would include sensory concepts (e.g., color,
object representations, spatial locations) and perhaps also meaning
concepts (e.g., logical relationships, categorical or other semantic
information), as in other network theories. In this view, however,
images would also include response concepts (e.g., overt behavior
patterns and covert physiological events). This latter information
would be in the same associative net, and its activation would have
similar consequences for overall network firing. Furthermore, there is
a natural neural connection between response concepts in memory (which
might be activated, for example, by text content) and the actual
efferent events to which they refer. Thus, while imagery is in
principle a state in which overt behavior is inhibited, "efferent
leakage" --a sub-overt activation of visceral organs and somatic muscle
intrinsic to the represented behavioris a normal part of the image
process. In point of fact, Lang (1979) argues that in the intact
organism image processing is defined by its efferent consequences.

4
Emotion and Memory
The view taken here (see Lang, 1985; Lang, Bradley & Cuthbert,
1990) is that emotions are action dispositions, which are either
appetitive--involving approach and consummatory-related behaviors, or
they are aversiveinvolving defensive, avoidant, and protective
behavioral strategies. Both the appetitive and aversive emotional
states vary in intensity or arousal, which is associated with increased
efferent activity and a greater overt action potential.
The responses in emotion include expressive language, overt
behavioral acts, and a supporting somatic and visceral physiology (Lang,
1964; Lang, 1978). Therefore, the network of an emotional memory image
is held to include, in addition to response concepts relevant to the
remembered perceptual context (e.g., eye movements, postural
adjustments), response representations of remembered emotional
reactions. Furthermore, while overt manifestations may be inhibited by
the context of recall, the memory image should generate an affective
physiological pattern, aversive or appetitive, moderate or intense,
depending on the emotional context being recalled (Lang et al., 1983).
The Research Problem
The present experiment examines the encoding and subsequent memory
retrieval of affective text. Subjects initially completed self report
measures of imagery ability and were then told they would be asked to
learn a series of narrative paragraphs. Half of the subjects received
imagery instructions. They were told to imagine the content of the
paragraphs as a vivid personal experience, clearly imaging the events

5
and their specific action. The other half received verbal instructions:
they were told to focus on the exact words used in the text.
Instructions to both groups were presented as optimal learning
strategies that would facilitate performance on the word recognition
task following each text.
Subjects read the study paragraphs at their own pace, sentence by
sentence, on a CRT screen. The time taken to read each sentence was
recorded, and heart rate was monitored continuously. For a subset of
subjects, an acoustic startle stimulus occurred during some of the
texts. After each paragraph, a test list of words and control symbols
was presented. Subjects held reaction time buttons, one in each hand;
they indicated with a right hand button press if a target word had
appeared in the preceding text, and with a left hand button press if the
word was not in the text. The words in the test lists were all preceded
by a very brief (300 ms) priming word or symbol. Some of the priming
words were also from the preceding text, and were expected to facilitate
recognition speed for the target words. After this part of the
experiment was completed, subjects were asked to write down the gist of
as many of the paragraphs as they could recall. Subjects provided
affective ratings of the texts at the end of the experimental session.
Predictions, derived from bio-informational theory were tested for
both encoding and retrieval phases of the study.
Encoding Hypotheses
During the encoding process, narrative text activates representa
tions in memory which constitute an image of the events described. As

6
suggested earlier, cognitive representation of an emotion includes
information on visceral and motor activation, inherent in emotional
experience, but not included in neutral, non-action representations.
Thus, during encoding, more units of memorial information are contacted
by affective than by neutral text. This additional information and the
associated efferent expression may modulate encoding time. This action
content from memory will be evident in the pattern of physiological
activation instigated in the subject. Furthermore, network processing
occurs over time as significant narrative elements are encoded and a
story "schema" is constructed (Kintsch & van Dijk, 1978). Thus, the
above effects on physiological response and processing time will more
probably occur late in an encoding task.
Attention to a given aspect of a stimulus array determines the
content of memorial representations and affects memory performance
(Jacoby, 1982). It is suggested here that encoding and retrieval of
affective text are also modulated by the subject's processing style,
i.e., whether she focusses on the specific words in the text, or
focusses on the developing narrative imagery. This variable is examined
here in two ways: (1) by instructionsto either image or attend to the
specific words of the texts; and (2) by examining individual differences
in the normal prevalence and vividness of imagery, i.e., by comparing
"good" and "poor" imagers.
The present research examines the following general hypotheses:
1. The action content in emotional memory is evident in the
pattern of physiological activation during the encoding of affective
text.

7
2. Affective and neutral cognitive representations differ in type
and quantity of associated information, which differentially affects the
time it takes to encode these contents.
3. A story schema develops over time during text processing, and
thus emotional activation is more pronounced late in the encoding
process.
4. Instructions to image and good imagery ability increase the
probability of physiological activation at encoding.
A long tradition of psychophysiological research on attention and
affect has employed heart rate as a dependent variable. Many
investigators have found that heart rate deceleration accompanies an
attentional set (Graham and Clifton, 1966; Obrist, 1981; Lang, Ohman, &
Simons, 1978). Lacey and Lacey (1974) described heart rate deceleration
as part of an "information intake" orientation in which the organism
shows a generalized inhibition of motor activity accompanied by an
attentional focus on the stimulus of interest. Cardiac acceleration, on
the other hand, typically accompanies tasks in which the subject
focusses on internal mental processes and inhibits environmental intake,
such as performance of mental arithmetic (Lacey & Lacey, 1974) and
generation of mental images (Miller et al., 1987).
Studies of emotion have revealed divergent patterns of heart rate
response to affective content, depending on whether the subject is
engaged in information intake or internal processing. For example, in
perceptual situations such as slide-viewing, heart rate is generally
found to decrease more when normal subjects are exposed to affectively
unpleasant pictures compared to neutral or positive stimuli (Klorman,

8
Weisberg, & Wiesenfeld, 1977). This may reflect greater attention to
aversive stimuli. On the other hand, heart rate increases have been
observed during imagery of aversive events, relative to neutral
mentation (Lang, Kozak, Miller, Levin, & McLean, 1980), perhaps indexing
a more intensive internal processing. Perception and imagery are not,
of course, wholly independent. In order to be processed, any stimulus,
text or picture, must initially be an object of orientation and
attention. However, if the material subsequently contacts relevant
representations in memory, the cognitive work of image construction may
begin. Presuming that heart rate is a reliable covariate of these
processes, we might expect (1) that the response to a stimulus would
change over time, and (2) that subjects who quickly develop images,
because of the particular pertinence of the stimulus or their general
disposition to image, are more likely to show the heart rate pattern of
imagery than that of attention. An example of the latter phenomenon has
been observed in phobic subjects (Hamm, Globisch, Cuthbert, & Vaitl,
1991; Klorman, Wiesenfeld, & Austin, 1975), who readily access a vivid
fear image, and show a strong accelerative response (rather than the
usual deceleration in aversion) to slides of their own primary fear
objects.
In the present study heart rate response should vary with both
encoding strategy and activation of an affective "action set." The
above considerations suggest the following working hypotheses:
1. It is predicted that heart rate will decrease during text
encoding, consistent with the experimental task of information intake.

9
2. Greater heart rate deceleration is expected for subjects
receiving verbal encoding instructions and subjects who report poor
imagery ability. As these subjects are primarily "attending" and are
less likely to be engaging in imaginal processing, greater heart rate
deceleration is predicted for unpleasant text content.
3. Imagery instructions and self-reported vivid imagery are
predicted to result in heart rate acceleration during the encoding
process, reflecting the internal cognitive effort of image generation.
The pattern of heart rate change associated with emotional imagery
(greater heart rate increase for more aversive content) is predicted for
these subjects.
4. Consistent with text processing theory, heart rate change
associated with the development of affective schemata is more probable
late in the encoding process.
The startle reflex is a generalized flexor response elicited by
sudden, intense stimuli. The response complex entails a large number of
protective elements, such as eyeblink, forward and downward head
movements, raising of the shoulders, contraction of the abdomen, and
flexion of the limbs (Davis, 1984). Latency and magnitude of the
eyeblink component of this reflex are modulated by both the attentional
and the emotional state of the organism (Anthony, 1985; Graham, 1979;
Lang, Bradley, & Cuthbert, 1990). Inhibition of the reflex response to
an acoustic startle probe occurs during visual attention to pleasant,
interesting stimuli (Simons & Zelson, 1985). On the other hand, startle
probe responses are exaggerated in aversive emotional states (Berg &
Davis, 1984; Bradley, Cuthbert, & Lang, 1990; Cook, Hawk, Davis, &

10
Stevenson, 1991; Cook, Davis, Hawk, Spence, & Gautier, in press; Vrana &
Lang, 1990; Vrana, Spence, & Lang, 1988; see Lang, Bradley, Cuthbert, &
Patrick, in press, for a review). In the present study, startle probes
are presented unpredictably while subjects read the study paragraphs.
The blink component of the startle response is recorded. It is expected
that variations in its latency and magnitude will track the development
of the affective schema during the encoding process. Specifically, it
is predicted that:
1. Potentiation of the startle reflex (greater magnitude and
faster latency) is expected as a result of activation of an aversive
emotional schema; relative inhibition is expected for pleasant schemata.
2. These effects will be enhanced late in the encoding process,
after the affective schema is well developed.
3. Imagery instructions and good imagery ability will result in
greater affective modulation of the startle response, relative to verbal
instructions and self-reported poor imagery ability.
Retrieval Hypotheses
Bio-information theory states that emotional images have an extra
set of concepts in the networknamely those involving response
information. Network theory states that greater elaboration of
representations in memory enhances retrieval (Anderson, 1976). This
mechanism may underlie the corollary finding that moderately arousing
affective events are better and more accurately remembered than less
arousing events (e.g., Bradley, Greenwald, Petry, & Lang, 1991; Heuer &
Reisberg, 1990). These findings are also consistent with the Yerkes-

Dodson Law (1908) in which performance on a variety of learning
discrimination tasks is positively related to increases in physiological
arousal, within the moderate range.
Thus, the present research tests the following general hypothesis
regarding the effects of emotional content on measures of memory
retrieval:
1. More affectively arousing text, (i.e., elaborate information
structures in memory) will result in enhanced performance for both
recognition and free recall measures of retrieval.
Arousal is more probable when subjects employ an imaginal encoding
strategy, given that imaginal processing emphasizes the inclusion of
sensory and visceral/motor activation codes in the resulting memory
representation. Conversely, instructions to attend exclusively on the
surface characteristics of the words (verbal instructions condition),
would be expected to reduce the inclusion of such information in memory.
The instructions in this experiment may also be construed as an
affective levels of processing manipulation (Craik & Lockart, 1972) in
which imaginal encoding instructions involved "deep" processing, and
verbal instructions prompted "surface" processing of the paragraphs.
These considerations motivate the following hypothesis:
2. Imaginal encoding strategies, resulting from instructions or
subject characteristics, emphasize the representation of visceral/motor
activation codes in memory for affective text relative to verbal
encoding strategies, thereby enhancing memory performance.
3. The above effects are specific to the reactivation of the
study paragraph's information network by the related test word, and do

12
not represent a simple persistence of the affective response in memory.
That is, the above results will not be found for nonstudy words.
Priming can be defined as facilitation of the response to a test
item by an immediately preceding cue (Ratcliff & McKoon, 1978). This
paradigm has proved a useful tool in the development of theory on how
knowledge is represented in memory, such as the network theory developed
by Anderson (1983). According to network theory, when a concept is
processed cognitively it is activated in memory. This activation
automatically spreads through the information network to other
associated concepts. Episodic priming results from temporary logical
relationships established among concepts when subjects read a particular
text, and has been employed by researchers to demonstrate that memory
for text is represented in logical form as meaning propositions, and
that activation of this associative network follows the logical
structure of these associations. Priming effects are typically
determined by comparing the reaction time for memory recognition of a
target stimulus preceded by a highly associated word, or "prime",
relative to that for a neutral or control condition.
In Lang's original presentation of bio-informational theory (1977;
1979), he followed Pylyshyn (1973) in suggesting that there is a
general, propositional sub-organization in imaginal/affective
information structures. To examine this assumption, an episodic
recognition priming paradigm was adapted from the work of Ratcliff and
McKoon (1978, 1981; McKoon & Ratcliff, 1980) for the present study.
Their research explored the dynamics of spreading activation in logical
information structures in neutral narrative text. They observed that

13
when a word in a test list, taken from a study paragraph, was primed
(i.e., immediately preceded) by another word from the same paragraph,
recognition time was faster than if the prime was an unrelated word or
symbol. The effect was greater, the closer the logical relationship
between prime word and target word. Earlier pilot studies (Bradley,
Spence, & Lang, 1992; Appendix G) in which episodic priming was compared
for affectively unpleasant and neutral texts suggested that emotional
content altered the priming phenomenon reported by Ratcliff and McKoon.
The basic priming phenomenon is tested here for affective and neutral
text, i.e., do related text words facilitate recognition of text study
words. The following specific questions are addressed:
1. Does episodic priming occur with affective narratives, as has
been shown for neutral text, consistent with a propositional memorial
sub-structure?
2. Are there differences in priming for affectively pleasant and
affectively unpleasant text?
3. Is priming enhanced by imaginal processing instructions or by
the subjects' disposition to process imaginally?

METHODS
Subjects
Fifty-seven female undergraduates at the University of Florida
served as subjects in this experiment. Subjects were recruited through
introductory psychology classes, at summer registration, or through
signs posted at sorority houses on campus. Subjects either received
class credit or were paid $10 for participation in the study.
The research sample included a subset of 25 subjects to whom
acoustic startle probes were presented during text encoding. Probe
administration had almost no effect on the other dependent variables
Thus, for all measures except the startle probe, all 57 subjects are
included in the analyses, with the few sample differences noted. The
analysis of the startle probe includes, of course, only the subset of 25
subjects.
Materials
Texts
The study materials used in this research consisted of two sets of
36 6-line paragraphs. Each set contained 18 unpleasant, 9 neutral, and
9 pleasant exemplars. The paragraphs were classified into affective
categories based on a validation study in which college students read
and imaged a larger set of 108 texts (Spence et al., 1987). The
validation data were psychophysiological measures of heart rate, skin
14

15
conductance, and electromyographic (EMG) activity from the corrugator
and zygomatic facial muscles, recorded during a 30-second imagery
period. Subjects also provided ratings of the paragraphs, using the
automated SAM system (Lang, 1980; Hodes, Cook, & Lang, 1985) after the
imagery task. Texts were selected for this research on the basis of:
(1) extreme pleasant or unpleasant valence ratings, and (2) the presence
of the appropriate psychophysiological activity. For unpleasant texts,
the criterion physiological responses during imagery included increased
heart rate and skin conductance level (associated with arousal), as well
as high corrugator (frown) muscle tension (associated with negative
emotional valence). The selected affectively pleasant texts showed high
zygomatic (smile) muscle activity accompanied by low corrugator muscle
activity, in addition to SAM ratings of positive emotional valence and
high arousal. Neutral texts were selected on the basis of neutral
valence and low arousal SAM ratings, and low levels of psycho-
physiological activity on all four measures. The number of exemplars
contributing to each affective category (18 unpleasant, 9 neutral, and
9 pleasant) reflected the primary focus on the comparison between
effects on cognitive processes of unpleasant versus neutral/pleasant
affect in prior pilot research. A summary of the validation study
including the selected texts and the corresponding data can be found in
Appendix E.
Paragraph topic and contents of the test list were matched between
affective categories. Every unpleasant text in Set 1 had a "matched"
text that was neutral or pleasant in Set 2; likewise, every neutral/
pleasant text in Set 1 had a "matched" unpleasant text in Set 2.

16
"Matched" means that the texts were similar in topic and sentence
construction, but varied in affective content. For example, a text
describing a beach picnic was matched with a similar text describing a
shark attack at the beach. In all cases, each pair of matched texts was
followed by the same test list. To illustrate with the previous
example, the study word "fins" appeared in the test list following the
matched texts pertaining to a visit to the beach. However, in the
affectively unpleasant text, "fins" referred to shark fins spotted prior
to a shark attack; in the pleasant text, the word referred to swim fins
that were lying on the beach.
The use of matched texts and test lists was employed to control
for inadvertent effects on reaction time (RT) that might be considered
material-specific. That is, it could be argued that the words in test
lists derived from affective texts might speed or slow recognition RT
because of a number of factors, including concreteness versus
abstractness of the nouns, differential frequency characteristics, or
affective valence of the test words themselves. The text matching
strategy controlled for this confound. See Appendix B for a listing of
study text pairs and their associated test list words.
These texts were also rated by subjects in the present study at
the end of the experimental session, using the booklet form of the SAM
ratings scale--see Appendix F for an illustration. Line the automated
SAM system, the booklet form included the affective dimensions of
valence, arousal, and dominance. Each of these dimensions was set to a
9-point scale. A rating of 1 represented the lowest possible value for
the valence (i.e., unpleasant), arousal, or dominance scale; and a

17
rating of 9 represented the highest value on these scales. Texts
receiving a valence rating of 1 to 3 were classified as affectively
unpleasant; texts rated from 4 through 6 were classified as neutral; and
texts with a rating of 7 or above were classified as pleasant.
Statistical analyses for the present study were based on affective
categories derived from these subject ratings.
Test Lists
After the subject had finished reading the study text, an episodic
recognition task began immediately. The test list consisted of 12
prime-target pairs. The prime appeared on the monitor screen for 300 ms
prior to the onset of the target. The target remained on the screen
until the subject responded with a right or left button press. In the
case of verbal targets, the button press indicated whether or not the
target had appeared in the study text; whereas for non-verbal targets,
the button press was determined by the target itself and was not related
to text content. Subjects were instructed to respond only on the basis
of the target stimuli, without regard to the prime. The lists were
structured so that subjects responded with their right hand ("yes"
response) to half of the items, and with their left hand ("no" response)
to the other half. This prevented uneven response probabilities from
influencing the RT data. Verbal primes and targets were always nouns,
although they varied as to whether they were concrete or abstract in
content. No test word appeared more than once in the experiment.

18
For the subset of subjects for whom a startle probe occurred on
some occasions while they read the text, care was taken that test words
and startle probes did not occur in the same sentence.
Targets. Half the targets (6) were verbal stimuli (Study or New
words) and half were non-verbal stimuli (*****, =====). Verbal targets
were also evenly divided between Study (3) and New words (3). Study
words were always nouns that occurred in the preceding text. New words
were nouns that did not occur in the study text, and had no semantic
relationship to the topic of the study text or to the other words in the
test list. Statistical analyses focussed on reaction time and accuracy
for (1 ) the prime-target pairs for which the target was a study word or
a new word, and (2) study word targets that were preceded by either a
study word or new word prime.
Non-verbal targets were also equally divided, half (3) requiring a
right-hand button press (******), and half a left-hand button press
(======). These stimuli differed from the verbal targets in that they
required pattern recognition for a choice reaction time task rather than
verbal decoding for a recognition memory task. In addition, the non
verbal targets were repeated in every test list while verbal test list
items appeared only once in the experiment. Thus, the non-verbal items
involved minimal cognitive effort and responses to them became automatic
over trials. Predictably, reaction time for the non-verbal targets was
much faster (by over 200 ms) and less variable than for verbal targets.
These data were thus not comparable to the verbal target conditions and
were excluded from further analysis.

19
Primes. All four classes of targets (Study word, New word,
******t ancj ======) were primed by three alternatives: A Study word, a
New word, or a series of X's. The New and XXXXX primes served as
alternative control conditions. The priming condition in which a study
word was primed by a string of X's was eliminated as a control, or
neutral, prime for reasons detailed by Jonides and Mack (1984), in which
important criteria for the characteristics of neutral primes are
outlined. For example, they argue that non-word primes should not be
compared with word primes because of the differential processing
requirements and possible expectation sets that accompany these two
types of priming stimuli. In addition, they caution against the use of
a prime which is repeated on each trial (such as the XXXXX primes) as a
comparison condition for a prime (such as a new or study word), which is
not repeated. Again, differential processing requirements as well as
possible differences in the alerting, or warning, functions of these two
type of prime were considered to vary significantly. Because new word
primes satisfied the above criteria and XXXXX primes did not, the
analyses of priming effects were restricted to the comparison of
recognition reaction time for New-Study with Study-Study prime-target
pairs.
An exemplar of a study paragraph and its associated test list are
provided in Figure 1. Note that in the presentation of this text on the
subject monitor during the experiment, each sentence was presented in a
single line of text, so that the paragraph appeared as six lines of
text. The format of the present document prevents this presentation.

20
STUDY TEXT EXEMPLAR
You're having a great time on a picnic by the ocean with
some good friends. Some are out swimming, and you look twice
when you think you spot sharks. Your feet pound through the
surf as you run calling, the fins now obvious. Unaware, your
friends don't hear you, and you are gripped with horror. Your
heart pounding in your chest, you run across the sand shouting
and waving.
You scream in
desperation as someone is jerked
under, and
the water turns
red.
Prime
Tarqet
Pair Type
scissors
degree
New Word-New Word
calf
picnic
New Word-Study Word
friends
tent
Study Word-New Word
charm
New Word-Left Hand Response
xxxxx
XXXXX-Left Hand Response
pint
*******
New Word-Right Hand Response
fins
Study Word-Left Hand Response
feet
*******
Study Word-Right Hand Response
XXXXX
wagon
XXXXX-New Word
xxxxx
ocean
XXXXX-Study
heart
sand
Study-Study
Figure 1. Exemplar of affectively unpleasant study paragraph and
its associated test list.

21
Apparatus
The experiment was controlled by two computers working in concert:
Terak and Datax LSI-11 computers. The Terak controlled the timing of
events in the experiments, presented the texts, and measured and
recorded read times and reaction times. For the subset of subjects
receiving startle probes, the Datax controlled presentation of the
acoustic startle stimulus, and recorded EKG and blink response.
Beckman standard silver-silver chloride electrodes were placed in
a modified Lead 2 configuration to measure the electrocardiogram (EKG).
The signal was filtered and amplified by Coulbourn modules. A Schmitt
trigger detected the R-waves and interrupted the computer clock to
record inter-beat intervals with a resolution of 1 ms. EKG was measured
continuously during presentation of each text, beginning 3 seconds prior
to the beginning and continuing during the time the subject read each
sentence of the text.
For the startle subset of the sample, EMG was measured with Med
Associates miniature silver-silver chloride electrodes, using Beckman
electrode electrolyte as the contact medium. The sensors were placed
about 1 cm beneath the left eye. One electrode was centered beneath the
eye, and the other was placed beneath the outer canthus. The
amplification for the signal was set at 6 (X 10,000), and frequencies
below 90 Hz and above 1000 Hz were filtered out by a Coulbourn HiGain
Bioamplifier. The signal was rectified and integrated by a Coulbourn
Contour Following Integrator (200 msec time constant), and then fed
through an Analog-to-Digital converter. EMG was collected at 1000 Hz
for 250 msec following the onset of the startle stimulus. The acoustic

22
startle stimulus consisted of a 50 msec burst of 100 dB (A) white noise
with an instantaneous rise time, that was gated by a relay and presented
binaurally through Pioneer (SE-250) stereo headphones from a tape that
ran continuously during the experimental session.
Texts were displayed on an Amdek CRT monitor in the subject room.
The monitor was placed approximately 2 feet in front of the subject, and
a transparent plastic screen separated the subject from the monitor
itself. Subjects responded using hand-held RT buttons, mounted on 6-
inch cylinders, which were activated by pressing down with the thumb.
The subject held a cylinder in each hand throughout the experimental
session. She controlled presentation of each successive sentence with a
button press, and a final button press removed the entire text from the
screen. Read time consisted of the time between button presses, as the
subject read through the texts at her own pace. The subject also
pressed either the right or left button to respond during the episodic
recognition task. RTs and reading times were recorded to the nearest
msec.
An intercom permitted communication between the subject and the
experimenter throughout the experimental session.
Procedures
Upon arrival at the laboratory, subjects signed an informed
consent form and completed imagery questionnaires and a vocabulary
measure. The subject was then seated in a comfortable chair in the
experimental chamber and sensors were attached in order to obtain the
psychophysiological measures.

23
Instructions were read to each subject for verbal or imagery
processing, and the subject was given practice in reading texts and
responding to test lists. For startle subjects, headphones were placed
on the subject's head and instructions were given to ignore the brief
bursts of white noise that would occasionally occur while she was
reading the study texts. The consent form, imagery questionnaire (QMI)
and the exact instructions read to each subject are included in
Appendix A.
After the experimenter had insured that the subject understood the
instructions and had verified that the physiological measures were being
recorded accurately, the subject was left alone in the experimental
chamber to complete the experiment independently. The duration of this
part of the experiment was approximately 45 minutes.
Prior to the onset of each trial, a 3-second pre-trial baseline
period occurred. At the end of this period, the following message
appeared on the subject's monitor screen: "Please press button to
begin." This message signalled the subject to bring the first sentence
of the study text onto the monitor screen by pressing one or both
buttons. Each successive button press presented an additional sentence
of the text, until the entire paragraph was displayed on the screen.
For a subset of subjects, a brief burst of white noise was presented
over headphones at either 2.0, 2.25, or 2.50 seconds after onset of one
of the first two sentences (12 texts) or last two sentences (12 texts)
of the texts. When the subject had read the entire text, a final button
press cleared the screen. After a 200 ms pause, the test list began.
Initially a "+" appeared on the screen for 150 ms, centered on the line

24
where the prime was presented. After the "+" was removed, the prime
appeared immediately and remained on the screen for 300 ms. After the
prime was removed from the screen, the target was presented on the line
immediately below where the prime had appeared, and remained on the
screen until the subject responded. When the subject gave an incorrect
response, the word "ERROR" appeared on the screen for 2 seconds. After
the subject response there was another 200 ms pause, after which the "+"
appeared again for 150 ms, followed by the prime and target stimuli, as
described for the previous test pair. This sequence continued until all
12 prime-target pairs were presented. A variable inter-trial interval
occurred between text presentations ranging from 15 to 25 seconds in
length. After the last text had disappeared from the screen, the
following message appeared on the screen: "The session has been
completed. Thank you for your participation."
After the subject had completed this part of the study, the
physiological sensors were removed, and she was taken to another room
where she was asked to recall and write down a brief phrase
characterizing the content of as many of the previously presented texts
as possible. This task began approximately 10 minutes after the end of
the experiment. Subjects generally required about 5 or 10 minutes to
complete this task, although they were given as much time as needed.
After the free recall task was completed, the subject rated the
texts for affective valence, arousal, and dominance. Ratings were made
using a paper-and-pencil version of the Self-Assessment Manikin (SAM;
Hodes, Cook, & Lang, 1985; Lang, 1980; see also Appendix F). Typed
versions of the texts were interleaved in the SAM ratings booklet; the

25
subject was asked to refer to the typed version in order to recall the
text, and then to mark her rating of how she felt while reading the text
during the experiment. Finally, the subject was debriefed, and either
given course credit or signed a voucher to be paid $10 for her
participation in the study. The total experimental session lasted
approximately 2-1/2 hours.
Data Reduction
Encoding Measures
Read time. Read time for each sentence was measured by computer
with a resolution of one msec. The read times for sentences 1 and 2
were then averaged to provide a mean read time for the "early" sentence
pair. Read times were also averaged over sentences 3 and 4, and
sentences 5 and 6, and constituted the "middle" and "late" sentence pair
read times, respectively. The final data set included three mean read
times (for the early, middle, and late sentence pairs) for each of the
study paragraphs, within each affective category. For subjects
receiving startle probes, averages were calculated for each sentence at
each level of the Startle Position factor (early, late, no startle),
within each of the three affective categories.
Heart rate. Inter-beat intervals were converted into heart rate
values using the algorithm of Graham (1979). Because subjects read the
paragraphs at their own rate, the sampling periods for heart rate were
highly variable among sentences. Thus, a real time analysis of the
cardiac wave form across sentences was not possible. Rather, the
average heart rate for each sentence of each text was calculated and

26
deviated from the pre-trial baseline for that text. Mean heart rate
change scores were then calculated for affectively pleasant, neutral,
and unpleasant texts. For subjects receiving startle probes, separate
averages were calculated for each of the three levels of the Startle
Position factor (Early, Late, No Startle) within each affective
category.
Blink reflex. Integrated blink EMG data were scored with the
computer program provided by Balaban, Losito, Simons, and Graham (1986).
Average startle baseline values and magnitude (in arbitrary A-D units)
were calculated for startle probes presented early and late in the study
texts (i.e., during sentences 1 or 2 versus 5 or 6), for each of the
affective categories. The onset latency of the startle reflex was also
scored. Trials in which the startle magnitude value was zero were
omitted from the calculation of latency averages, which were calculated
for probes presented early and late in the text and for each of the
affective categories.
Retrieval Measures
Free recall. For each subject, the list of short descriptive
phrases was matched to a master list of paragraph topics in the study.
A paragraph was considered to have been recalled if the descriptive
phrase matched the topic of one of the paragraphs she read during the
experimental session. Paragraphs identified by the subject in this
manner were counted as having been recalled. The recall measure was
calculated for each subject: For each affective category (pleasant,
neutral, unpleasant), the number of texts listed by the subject was

27
divided by the total number of paragraphs she had rated in that
affective category.
Reaction time. Reaction time for each of the 12 prime-target
pairs was recorded to the nearest msec, for correct responses only.
Average RT and error rates were calculated for each prime-target pair,
for each affective category.
Accuracy. Error rate was calculated for each of the 12 prime-
target pairs by dividing the number of errors within an affective
category by the total number of paragraphs in that category, for each
subject. This number was then multiplied by 100 so that error rate was
expressed as a percentage.
Data Analysis
Imagery Ability
Subjects were classified according to imagery ability, based on a
median split of 80 on the Questionnaire upon Mental Imagery (QMI;
Sheehan, 1967). The range and distribution of QMI scores obtained for
subjects in this study conformed to those obtained in other studies in
the Lang laboratory, and are documented in Appendix C. Groups of
subjects assigned to receive imagery and verbal processing instructions
were balanced for QMI score to control for the effects of imagery
ability and its possible interaction with the instructional
manipulation.
Affective Classification of Texts
In the design of this experiment, a text's affective category was
determined by ratings and psychophysiological measures obtained in an

28
earlier validation study (Spence et al., 1985). Subjects in the present
study, however, also provided ratings of the texts after the experiment.
This permitted a post-hoc recategorization of texts, based on this more
proximal, group-specific assessment. The resulting redistribution of
paragraphs produced an approximately equal number of affectively
pleasant and unpleasant texts, but fewer neutral texts. For each
subject, unequal numbers of texts contributed to the cell averages.
Three subjects were not included; two of them failed to rate any
paragraph in the neutral range, and another subject rated only one text
as affectively neutral. The ratings distributions for the validation
study, the present study, and a contingency table illustrating the
stability of ratings classifications are provided in Appendix E.
Statistical analyses were performed, using both categorization
systems, with similar overall results. Because they more closely
represent this specific sample, and to avoid redundancy, only results
based on the subjects' own ratings are presented here. Tests of valence
involve orthogonal contrasts of pleasant versus unpleasant (linear
effect), and affective (pleasant and unpleasant) versus neutral content
(quadratic effect).
Encoding Measures
Read time and heart rate. These data were analyzed in a 2 X 2 X 2
X 3 X 3 factorial design. Between-subject variables were Imagery
Ability (Good, Poor), Instructions (Imagery, Verbal), and Group (Startle
Probe, No Startle Probe); within-subject variables were Affective
Content (Pleasant, Neutral, Unpleasant) and Sentence Pair (1&2, 3&4,

29
5&6). Read times were subjected to ANOVA, and heart rate data were
analyzed in an ANCOVA, the 3-second pre-trial heart rate average serving
as the covariate for heart rate change scores.
Recall that a subset of 25 subjects in this study received startle
probes while they read the study paragraphs. For these subjects,
separate analyses of read time and heart rate included the additional
factor of startle position (Early, Late, No Startle) because
presentation of startle probes might be expected to affect encoding
processes for the study paragraphs. All data analyses of the complete
sample included group (Startle Probe vs. No Startle Probe) as a factor,
however, in order to detect any differences related to this variable.
Startle magnitude and latency. A2X2X3X2 ANOVA was carried
out for each of these dependent measures. The between-subject factors
were Imagery Ability and Instructions; the within-subject factors were
Affective Content and Startle Position (Early, Late).
Retrieval Measures
Free recall. These data were subjected to a 2 X 2 X 2 X 3 ANOVA
with the between-subject factors of Imagery Ability, Instructions, and
Probe Group; the within-subject factor was Affective Content.
Episodic Recognition for Target Words. A2X2X2X3X2
factorial design included the between-subject factors of Imagery
Ability, Instructions, and Group. The within-subjects variables
consisted of Affective Content and Target Type (study word, new word).
Non-word targets were not included in these analyses because of the
widely different characteristics between the non-verbal and verbal

30
targets. As explained earlier (page 18), each of the non-verbal targets
appeared three times in each test list while target words appeared only
once in the experiment. Subject recognition for the non-verbal targets
thus became automatic over trials, resulting in much faster and less
variable reaction times than for verbal targets. These data were thus
not comparable to the verbal target conditions and were excluded from
further analysis.
Episodic Priming Analyses. A2X2X2X3X2 factorial design
again included the between-subject factors of Imagery Ability,
Instructions, and Startle Probe Group. The within-subjects variables
consisted of Affective Content, and Prime Type (Study Word, New Word).
As stated earlier (page 19), XXXXX-Study pairs were excluded from
analyses of priming condition. Important differences between the XXXXX
and study word primes (verbal vs. non-verbal stimuli; repetition both
within and between test lists; different levels of processing
requirements for decoding by subjects) served as the rationale for their
elimination as a control condition.
Recognition accuracy. The effects of the experimental factors on
accuracy in the recognition RT task were analyzed in order to address
the question of speed-accuracy trade-off as a possible factor in
recognition performance. The accuracy measure was calculated as percent
errors; analyses of these data paralleled those for recognition RT.

RESULTS
Encoding Measures
Read Time
Table 1 presents mean read times for the early, middle, and late
sentence pairs of paragraphs with affectively pleasant, neutral, or
unpleasant content. Consistent with findings in the text processing
literature, subjects read the later sentences in the paragraphs more
rapidly than earlier sentences, F (2,96) = 59.71, p < 011.
Furthermore, paragraphs with pleasant content were generally read more
quickly than unpleasant texts, yielding a linear relationship between
read time and pleasant versus unpleasant affective content, Valence
F (2,92) = 2.17, p = .12; Valence (linear) F (1,46) = 9.46, p < .01.
Finally, on the average, good imagers tended to read all sentences of
the paragraphs faster (mean per sentence = 4.7 secs) than poor imagers
(mean per sentence = 5.6 secs), F (1,46) = 3.55, p = .06.
Affective content and sentence interacted significantly,
F (4,184) = 3.97, p < .01 in the omnibus test. As can be seen in
Table 1, read times for early sentences followed a quadratic pattern
with affective content, Valence (quadratic) X Sentence F (2,92) = 5.86,
1 Where appropriate, reported p values reflect application of the
Greenhouse-Geisser correction.
31

32
Table 1
Mean read times (in
seconds) for early, middle
, and late
sentence pairs
of texts with affectively pleasant
, neutral,
and unpleasant content
Pleasant
Neutral
Unpleasant
Sentence Pair
Mean
Early
5.49
5.35
5.59
5.48
Middle
5.08
5.27
5.30
5.22
Late
4.63
4.72
4.79
4.71
Mean
5.07
5.12
5.23
2 < .01. That is, for these first two sentences only, mean reading
times for both emotional contents (pleasant and unpleasant) were longer
than for neutral paragraphs. The overall linear valence effect--
shortest read times for pleasant paragraphs--developed in the later
sentences (Valence (linear) Fs (1,46) = 1.57, 9.64, and 7.07; ps < .22,
.01, and .02; for early, middle and late sentence pairs, respectively).
In subsequent tests, the Valence X Sentence interaction was also
found to vary with subgroup: Thus, good imagers showed the early
quadratic valence effect more strongly than poor imagers, Valence
(quadratic) X Imagery Ability X Sentence F (2,92) = 3.47, p < .01. For
poor imagers the linear valence effect (with the fastest read times for
pleasant content) was the dominant response across all trials (see
Table 7, Appendix D).
No reliable main effects or interactions involving instructions
were observed for read time.

33
Heart Rate
Baseline. A preliminary 3 X 2 X 2 X 2 ANOVA (Valence X Imagery
Ability X Instructions X Probe Group) of heart rate baseline values
yielded no significant main effects or interactions of heart rate change
with the independent variables. Thus, baseline heart rate was not
systematically related to any of the factors in this study.
Encoding. Heart rate decreased from baseline during paragraph
encoding, F (2,52) = 52.17, p < .01. As can be seen in Table 2, this
deceleration was less during encoding of pleasant than unpleasant
content, Valence F (2,51) = 2.57, p= .11 in the omnibus test; Valence
(linear) F (1,25) = 15.52, p < .01. Examination of the Table also
suggests that the above effect was stronger late in the paragraph. The
Valence X Sentence interaction did not, however, reach statistical
significance in the omnibus test (F (4,104) = 2.14, p = .12).
Table 2
Mean heart rate deceleration for early, middle, and late sentence pairs
of texts with affectively unpleasant, neutral, and pleasant content
Pleasant
Neutral
Unpleasant
ice Pair
Mean
Early
-1.02
-1.88
-1.45
-1 .45
Middle
-2.74
-3.11
-3.49
-3.11
Late
-3.02
-3.34
-3.69
-3.35
-2.26
-2.78
-2.88
Mean

34
Nevertheless, based on prior hypotheses, separate tests were undertaken
for each sentence pair. These tests revealed that while the effect of
valence on heart rate change was not significant early in the paragraph,
Valence (linear) F (1,25) = 2.58, p = .12, pleasant contents resulted in
reliably less heart rate deceleration than unpleasant contents for
middle and late sentence pairs, Valence (linear) Fs (1,25) = 20.65 and
8.24, respectively; both ps < .01.
Table 3 suggests that subjects receiving verbal instructions
showed greater heart rate deceleration compared to subjects receiving
imagery instructions; however, the statistical test only approached
significance, F (1,25) = 3.89, p = .06. On the other hand, instructions
significantly modulated the affect/heart rate relationship, yielding a
Valence (linear) X Instructions interaction, F (1,25) = 7.01, p < .02.
The greater relative deceleration for unpleasant than pleasant texts,
found overall, was more pronounced in subjects receiving verbal
instructions.
Table 3
Mean heart rate change (beats per minute) for texts with
pleasant, neutral and unpleasant content, for subjects receiving
verbal and imagery instructions
Instructions
Verbal
Imagery
Pleasant
Neutral
Unpleasant
Mean
-2.92
CO
CO
ro
1
-3.91
-3.57
-1.61
-1.67
-1 .85
-1 .71
-2.26
-2.78
-2.88
Mean

35
The stronger valence effect for verbal instructed subjects was
greater for those who received startle probes, resulting in a Valence
(linear) X Instructions X Probe Group interaction, F (1,25) = 5.63,
p < .03 in the omnibus test (see Table 19, Appendix D). Furthermore,
startle probes potentiated a stronger linear valence effect in poor
imagers, Valence (linear) X Imagery Ability X Startle Group interaction,
F (1,25) = 5.87, e < -03 in the omnibus test (see Table 18, Appendix D).
To summarize the effects of these moderator variables on heart rate and
affect, the subjects who showed the strongest linear valence pattern
(least deceleration to pleasant content and greatest to unpleasant) were
poor imagers, under verbal instructions, who also received startle
probes.
Startle Reflex
Baseline EMG level. Analysis of baseline values for the startle
response yielded no significant main effects or interactions of the
independent variables in this study.
Startle magnitude. The omnibus analysis indicated that startle
responses did not vary significantly with text valence overall, F < 1 .
The Startle Position X Valence (linear) interaction, however, approached
significance, F (1,15) = 4.27, p = -057. As can be seen in Table 4,
mean blink magnitudes increased linearly from pleasant to unpleasant
paragraphs for late probes, valence (linear) F (1,15) = 4.81, p < -05.
Neither a significant linear nor quadratic valence effect was found for
early probes, F < 1.

36
No main effects or interactions involving instructions or imagery
ability occurred for this measure.
Table 4
Mean startle magnitude (arbitrary A-D units) for probes measured
early and late in affectively pleasant, neutral and unpleasant texts
Pleasant Neutral Unpleasant
Startle Position Mean
Early
114.6
133.9
111.2
119.9
Late
134.0
149.8
165.1
149.6
Mean
124.3
141 .9
138.2
Onset latency. Compared to the magnitude data, startle latency
showed a closerbut still not statistically significant--overall
relationship with affective content (Pleasant = 55.2 msec,
Neutral = 52.9 msec, Unpleasant = 52.7 msec), F (2,28) = 1.63, p = .21
in the omnibus test; Valence (linear) F = 2.11, p = .17. Emotional
content affected blink latency differently, however, for subjects
receiving verbal versus imagery processing instructions, and
particularly, depending on whether the startle probe occurred early or
late in the paragraph (see Figure 2), Valence X Instructions X Startle
Position interaction, F (2,28) = 3.72, p = .053 in the omnibus test;
Valence (linear) X Instructions X Startle Position, F (1,14) = 5.99,
p < .03.
As can be seen in the left panel of Figure 2, instructions
interacted with emotional content to affect startle latency early in the
encoding process, Valence (linear) X Instructions F (1,14) = 5.41,

o
0)
CO
E
65
60
55
50
# Imagery
O Verbal
Pleasant Neutral Unpleasant
Pleasant Neutral Unpleasant
EARLY LATE
Figure 2. Startle latency for verbal versus imagery instructions, for probes presented
early and late in affectively pleasant, neutral and unpleasant texts.
LO
^3

38
2 < .04. While subjects receiving imagery instructions showed the
expected linear valence relationship, those receiving verbal
instructions showed a deviant pattern. For late probes, both
instruction groups startled faster to unpleasant content, Valence
(linear) F (1,14) = 5.61, p < .04.
Imagery ability did not modulate startle latency for pleasant and
unpleasant content. The only effect of imagery ability was a generally
faster early probe startle for good imagers, compared to poor imagersa
difference that disappeared with the late probes, Startle Position
F (1,14) = 11.88, p < .02 (see Table 22, Appendix D).
Retrieval Measures
Free Recall
Consistent with study predictions, paragraphs with emotionally
evocative content, both pleasant and unpleasant, were recalled in
greater proportion compared to affectively neutral texts (pleasant =
.41, neutral = .29, unpleasant = .42), Valence F (2,100) = 11.20,
p < .01 in the omnibus analysis; Valence (quadratic) F (1,50) = 18.20,
p < .01. Instructions, imagery ability, and presence or absence of
startle probes did not affect free recall.
Episodic Recognition
As described earlier, analyses of recognition reaction time data
were accomplished only for word targets and primes. The effects of the
experimental factors (Valence, Instructions, Imagery Ability, Probe
Group) were tested as they interacted with target type (study words
versus new words) and priming condition (study-study, new-study).

39
Target recognition latency. Study word targets were words that
appeared in the preceding text and required a "yes" response. New word
targets were non-text-related words which required a "no" response.
Recognition reaction time for study and new word targets was not
significantly different, F < 1. However, emotional content, imagery
ability, and instructions interacted with target type, F (2,92) = 4.36,
p < .03 in the omnibus test, Valence (quadratic) X Instructions X
Imagery Ability X Target Type F (1,46) = 4.62, p < .04. Thus, separate
follow-up analyses were done for new word and study word targets. As
expected, no significant effects of the experimental factors were found
for new word targets, all Fs < 1.9, all ps > .17.
Recognition memory for study word targets followed a pattern over
affective contents similar to free recall (pleasant = 857 msec, neutral
= 881 msec, unpleasant = 853 msec). In this case, however, the
statistical test was not significant, Valence F (2,92) = 1.90, p = .17;
Valence (quadratic) F (1,46) = 2.37, p = .13. Nevertheless, latencies
did interact significantly with all of the main experimental factors,
Valence X Instructions X Imagery Ability F (2,92) = 4.24, p < .03,
Valence (quadratic) X Instructions X Imagery Ability F (1,46) = 5.30,
p < .03. Examination of Table 4 reveals that target recognition was
clearly faster for emotional than neutral content only for good imagers
receiving imagery instructions, valence (quadratic), F (1,46) = 5.94,
p < .02). Subsequent separate tests of affective content differences
for other groups were not significant, F (1,46) = 2.04, p = .16 for poor
imagers receiving verbal instructions,; all other Fs < 1.

40
Target recognition errors. Subjects made a larger percentage of
errors for study word compared to new word targets (study = 8.40
percent; new = 3.93 percent), F (1,46) = 59.39, p < .01. Target type
also interacted with valence and instructions to affect error rate,
Valence X Instructions X Target Type F (2,92) = 4.02, p < .04 in the
omnibus test; Valence (quadratic) X Instructions X Target Type,
F = 5.22, p < .03. Thus, separate follow-up analyses were carried out
for new word and study word targets.
Table 4
Mean recognition reaction time (msec) for study word targets
for affectively pleasant, neutral, and unpleasant texts
for good and poor imagers receiving imagery and verbal instructions
Imagery Instructions
Imagery Ability
Pleasant
Neutral
Unpleasant
Mean
Good
805
890
819
838
Poor
854
851
861
856
Mean
830
871
840
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
856
823
829
836
Poor
912
961
902
925
Mean
884
892
865

41
Subjects made more errors in rejecting new word targets with
imagery compared to verbal instructions (imagery = 4.71 percent, verbal
= 3.16 percent), F (1,46) = 4.37, p < .05. Other tests were not
significant: Errors for new word targets did not vary with emotional
content of the paragraphs, presence or absence of startle probes, or
imagery ability of the subjects.
Overall, study word errors did not vary significantly with valence
(F < 1 ). Examination of Table 5 does suggest that subjects receiving
verbal instructions made more errors in recognition of study word
targets from affectively neutral paragraphs compared to emotional (both
pleasant and unpleasant) texts. Statistical support for this
interpretation was, however, marginal, Valence X Instructions
interaction, F (2,92) = 3.27, p = .06; Valence (quadratic) X
Instructions F (1,46) = 3.43, p = .07.
Errors in recognition of study word targets did not vary reliably
with imagery ability, or presence or absence of startle probes.
There is no evidence in these data that the observed recognition latency
differences in affective contents can be attributed to a speed-accuracy
trade-off.
Priming condition and target recognition latency. A subset of the
reaction times to study words were re-examined, considering the effect
of the preceding prime. Specifically, the analysis compared recognition
latency for study word targets primed by another study word (SS) with
recognition latency to study words primed by a new word (NS). The same
experimental factors (affective valence, instructions, imagery ability,
presence or absence of startle probes) were assessed, as previously.

42
Table 5
Mean percent errors for recognition of study word targets from
pleasant, neutral and unpleasant texts, for subjects receiving
verbal and imagery instructions
Pleasant
Neutral
Unpleasant
Instructions
Mean
Imagery
8.65
6.79
8.00
7.81
Verbal
7.05
11.19
8.70
8.98
Mean
7.85
9.00
8.45
In the overall analysis, emotional content interacted with priming
condition to affect reaction time (see Figure 3), Valence X Priming
Condition F (2,92) = 8.52, p < .01; Valence (linear) X Priming
Condition, F (1,46) = 6.56, p < .02; Valence (quadratic) X Priming
Condition, F (1,46) = 9.52, p < .01. These effects were followed up by
separate tests for each priming condition.
Study-Study. Recognition reaction time was fastest for pleasant
and slowest for unpleasant content when the prime-target pair consisted
of study words (SS) (pleasant = 832 msec, neutral = 849 msec, unpleasant
= 859 msec), Valence F (2,92) = 1.14, p = .32 in the omnibus analysis;
Valence (linear) F (1,46) = 4.71, p < .04. Recognition memory for SS
prime-target pairs did not vary with imagery ability, instructions, or
startle probe condition.
New-Studv. Reaction time for recognition of study word targets
primed by new words (NS) was affected by emotional content in a
different manner--reaction time was relatively slower for neutral
compared to emotional (either pleasant or unpleasant) content
(pleasant = 845 msec, neutral = 907, unpleasant = 825 msec),

43
AFFECTIVE CATEGORY
Figure 3. Mean reaction times for Study-Study and New-Study
prime-target pairs following affectively pleasant, neutral and
unpleasant paragraphs.

44
Valence F (2,92) = 6.08, g < .01 in the omnibus test; Valence
(quadratic) F (1,46) = 6.97, g < .01. Other experimental factors did
not reliably affect reaction time for NS pairs.
Recognition errors were least when a study word target was primed
by another study word; performance declined somewhat with new word
primes, (SS = 6.23 percent, NS = 8.38), Priming Condition F (1,46) =
5.66, g < .03. While error rate was unaffected by imagery ability, or
the presence or absence of startle probes, error rates for SS and NS
prime-target pairs were differentially affected by emotional content and
instructions, Valence X Instructions X Priming Condition interaction
F (2,92) = 3.12, g = .06; Valence (quadratic) X Instructions X Priming
Condition F (1,46) = 4.28, g < .05. This interaction was followed up by
separate analyses of each priming condition.
Study-Study. Recognition of study words primed by other study
words was not affected by emotional content of the paragraphs, Valence F
< 1 in the omnibus test. However, instructions interacted with imagery
ability to affect subject performance, F (1,46) = 6.05, g < .02. Good
imagers receiving imagery instructions made significantly more errors
than poor imagers (good = 8.69 percent, poor = 3.98 percent), F (1,26) =
7.95, g < .01. Under the verbal instructions condition, however,
imagery ability did not affect error rates (good imagers = 5.44 percent,
poor imagers = 6.82 percent), F < 1; (see Table 48, Appendix D) .
Comparison of the error data results with the reaction time
analyses does not support a speed-accuracy trade-off explanation for the
valence effects on recognition reaction time described earlier.

45
New-Studv. For study words primed by new words, no main effects
or interactions of the experimental factors on error rate were observed.
Episodic Priming
Facilitation of recognition RT for SS relative to NS prime-target
pairs was not reliable when averaged across experimental conditions,
(SS = 847 msec; NS = 859 msec), F(1,46) = 1.65, p > .20. As noted
previously (p. 44), however, priming did vary with emotional content of
the paragraphs, Valence X Priming Condition F (2,92) = 8.52, p < .01
(see also Figure 3). Thus, individual tests were conducted for each
affective category. Consistent with findings in the episodic priming
literature, reliable priming occurred for neutral texts, (56 ms), F
(1,46) = 10.79, p = .01. However, for affectively pleasant texts, there
was no significant priming (F < 1); and for affectively unpleasant
texts, RT latency was actually 34 ms longer for study-study compared to
new-study prime-target pairs, F (1,46) = 5.17, p < .03.

DISCUSSION
The research provided broad support for the hypotheses that both
the encoding and retrieval of narrative text vary significantly with the
emotionality of story content. Thus at the learning stage, subjects
tended to read pleasant texts faster, and unpleasant texts slower, than
they did the neutral narratives. Furthermore, this effect of emotional
valence was more pronounced later in encoding, even though reading times
were then faster overall. Slower reading times for unpleasant
paragraphs were accompanied by strong decelerative heart rate responses,
and in response to startle probes, faster and larger blink reflexes.
Again, these effects were reliably stated only late in the subjects'
reading of the paragraph, when the story schema was already well
developed.
Memory measures also showed the effects of emotional content.
Thus, pleasant and unpleasant texts were better remembered than neutral
paragraphs at free recall. A similar pattern over affective contents
was found for recognition reaction time, which was most evident for
imagery instructed good imagers. Examination of primed recognition
suggests that non-text primes distracted subjects, producing a
performance deficit for neutral content, but did not interfere with the
recognition of words from affective paragraphs.
46

47
Encoding Measures
According to text processing theory (Kintsch & van Dijk, 1978),
the basic structure of a text schema is constructed during the initial
stages of encoding. In later sentences, the processing task shifts to
elaboration of the story framework. The greater cognitive processing
demands of schema development have been associated with longer reading
times for earlier sentences of logical narratives, with later sentences
being read more quickly. Consistent with study predictions and the text
processing literature, earlier sentences of the study paragraphs were
read more slowly than later sentences.
Emotional valence of the paragraphs was also expected to modulate
reading time in this study. It was hypothesized that affective text
contains information on visceral and motor activation, resulting in more
elaborate and complex representations, compared to memory for neutral
content. However, since this additional information does not increase
the logical complexity of the texta factor known to increase encoding
time in text processing--it was not clear how emotional content might
modulate encoding of textual materials.
The results of this study suggested that emotional content
affected reading time for early sentences differently than for later
sentences. The tendency to read affectively unpleasant paragraphs more
slowly than pleasant texts was significant overall, although this effect
only developed later in the encoding process. In addition, this overall
effect was strongest for subjects with self-reported poor imagery
ability, who showed the linear effect of valence consistently throughout
the narrative.

48
On the other hand, early in the encoding process, good imagers
read sentences from emotional (both pleasant and unpleasant) paragraphs
more slowly than those with neutral content. This effect was replaced
with the overall linear effect of valence in later sentences. Because
the early quadratic pattern of valence effects (unpleasant and pleasant
slower than neutral) occurred during the initial sentences of the
paragraphs for subjects reporting good but not poor imagery ability, it
may be argued that these results reflect individual differences in
schema development. It was suggested earlier that representations of
affective experience include additional information consisting of the
visceral/motor activation inherent in emotion (Lang, 1979). Hence,
construction of the schema for emotionally evocative text may result in
longer processing time (as reflected by read time) for affective
compared to neutral text, but only for subjects with a predisposition to
include the "action set" of emotion in the encoding process.
Heart rate decreased reliably during text encoding, consistent
with theories that deceleration is associated with information intake
(Lacey & Lacey, 1974). In addition, heart rate interacted with valence,
resulting in greater heart rate deceleration for affectively unpleasant
and less deceleration for pleasant contents. Consistent with read time
results, this effect emerged later in the encoding process.
Study predictions and prior research predicted that imaginal
versus verbal encoding strategies would result in different patterns of
heart rate change over sentences. Heart rate decrease was predicted for
verbal instructions, reflecting an information intake orientation. This
result is consistent with studies of heart rate response to presentation

49
of affective picture slides (Greenwald et al., 1989). However,
research also exists in which subjects listened to affective and neutral
narratives over headphones (Cook et al., 1988) and showed heart rate
acceleration in response to unpleasant content, despite the fact that
they were engaged in a task requiring information intake. The visual
aspect of the experimental task is common to both studies showing heart
rate deceleration to unpleasant stimuli, suggesting that the sensory
aspect of the experimental task may modulate the direction of heart rate
change in response to affective content.
Heart rate was expected to increase late in the encoding process
as a result of the internal processing demands in the imagery
instructions condition, consistent with prior research on imaginal
processing of emotional content (Lang et al., 1980; Vrana & Lang, 1990).
While study results confirmed the predicted heart rate decrease in the
verbal instructions condition, heart rate did not increase for the
imagery instructions condition. Nevertheless, the lack of heart rate
deceleration in the imagery instructions condition might be construed as
evidence for acceleratory effects cancelling out the deceleration
component.
While the read time and heart rate results showed consistent
effects of emotional content on subject responses, these measures are
nevertheless limited as direct measures of emotion because they are
inherently susceptible to a variety of other factors, as noted above.
Read time and heart rate are measures of perceptual processing and
stimulus intake, and thus vary with instructions, characteristics of the
experimental materials, and individual predisposition of the subjects.

50
The startle reflex, on the other hand, has been shown in recent studies
to be a robust measure of emotional response that is invariant across a
variety of experimental tasks and sensory modalities.
Lang has conceptualized startle as a protective reflex that is
primed or facilitated in negative affective contexts (Lang et al.,
1990). That is, the response to a noxious stimulus (the startle probe)
would be expected to be larger when the organism is in an affectively
negative state, and smaller during an affectively positive state. Thus,
the startle response served as a manipulation check in the present
study, to assess the development of affect in the subject during the
encoding process.
Consistent with the hypothesis that schema development occurs over
time, emotional content only affected startles presented late in the
encoding process. Startle magnitude showed the expected pattern of
greater potentiation for unpleasant relative to pleasant content.
However, encoding strategy did not affect this aspect of the startle
response.
Consistent with the startle magnitude results, startle latency was
faster for probes presented late in affectively unpleasant compared to
pleasant paragraphs. These effects were evidenced more for subjects
receiving imagery instructions, who showed this linear effect for early
probe startles as well. Subject receiving verbal instructions, on the
other hand, only developed this effect late in the encoding process.
These results suggest that imagery instructions may result in earlier
contact with the emotional content of text than instructions that focus
on the verbal characteristics of the narrative.

51
In summary, development of the affective information structure, or
text schema, was hypothesized to occur over time during encoding. Late
in the encoding process, a consistent pattern of results emerged in
which affectively unpleasant content resulted in longer read times,
greater heart rate decrease, and larger and faster startle responses,
compared to paragraphs rated as pleasant. For read time and heart rate,
the negative versus positive valence effect was most evident for
subjects reporting poor imagery ability and those receiving verbal
instructions, while imagery processing strategies tended to minimize
these effects. Instructions and individual differences variables did
not interact with startle magnitude effects, but for startle latency,
good imagers showed the linear effect of emotional content earlier in
the processing task.
Retrieval Measures
While measures of encoding of affective texts appeared to vary
with the negative versus positive valence of the materials, retrieval
measures varied more according to whether or not paragraphs were
emotional in content. Both recognition and free recall measures showed
greater facilitation for affective compared to neutral paragraphs.
Recognition of study word targets varied with presence or absence
of emotional content as well as imagery ability and instructions. Good
imagers receiving imagery instructions showed faster and more accurate
recognition for words from emotionally evocative compared to neutral
paragraphs. These results are consistent with the idea that imaginal
encoding strategies are more likely to activate arousing emotional

52
response information. Consistent with the bio-informational hypotheses,
and with other recent studies, more arousing materials were remembered
faster and more reliably (Bradley et al., in press; Heuer & Reisberg,
1990) .
In addition, recognition RT for new word targets showed no
relationship to either affective content or encoding strategy,
suggesting that the representation of text in memory must be activated
(by processing of a study word target) in order for the experimental
factors to affect performance. It may then be reasoned that the
recognition RT measure does in fact reflect the characteristics of the
information network represented in memory rather than nonspecific,
generalized effects of emotional arousal.
Analyses of the priming data yielded significant facilitation of
reaction time for study-study compared to new-study prime-target pairs
for neutral texts, but priming did not occur for affective texts.
However, Figure 2 clearly shows that the lack of priming for affective
materials resulted from a change in RT for New-Study pairs. That is,
recognition RT was faster for New-Study pairs from affective than from
neutral paragraphs. It appears from these data that a new word prime
has a greater inhibitory effect on recognition of a study word target
from an affectively neutral text, relative to when the study word target
is part of an affective information structure. Lang has suggested that
the associative structure of affective information, particularly when
processed in an imagery mode, contains a greater amount and different
types of information (motor and visceral activation codes) than the
materials typically employed in priming experiments. It appears that

53
one characteristic of such affective information structures is that they
facilitate both probability and speed of memory retrieval.
However, one important measure of episodic priming deviated from
this general pattern of valence effects. When study words primed by
other study words were considered independently, reaction times were
slower for affectively unpleasant content relative to pleasant content.
This result does not conform to the study predictions for memory
performance. This linear effect of valence (slower for unpleasant and
faster for pleasant content) study-study reaction times is more
consistent with the pattern of valence effects observed for the encoding
measures during the later sentences. For reading time and heart rate,
this linear valence effect was associated with poor imagery ability and
verbal instructions, and was less pronounced for good imagers and
imagery instructions. These results suggest that a perceptual
orientation to the text processing task may moderate the valence effects
on a variety of dependent measures. Given that the perceptual aspect of
the episodic recognition task is inherently enhanced when the exact
information is presented at retrieval that was present at encoding, the
study-study reaction time data may reflect processes similar to those
acting on read time and heart rate late in the encoding process.
Accuracy data from this experiment showed a similar pattern of
results to the recognition RT data. Recognition was both faster and
more accurate when study words had previously appeared in an emotionally
evocative text, and when subjects employed imagery during encoding.

54
Overall Conclusions
In summary, the results of this study show that while encoding
measures varied along the valence dimension, retrieval from memory
appeared to be largely mediated by the presence or absence of emotional
content in the texts. These results were consistent for all measures of
encoding, including read time, heart rate, and blink magnitude and
latency. In addition, recognition and recall measures of memory
retrieval were influenced in a similar fashion by the arousal dimension
of the study texts. An important exception to this general trend was
study word targets primed by study word primes. This research suggests
that variation in emotional content and manipulation of information
processing strategies (e.g., imaginal versus verbal encoding
instructions) work separately and in interaction to affect both the
encoding and retrieval for textual materials. Further study of the
effects of these variables on verbal learning and memory promises to
extend our understanding of these cognitive processes in important ways.

APPENDIX A
QUESTIONNAIRES, INSTRUCTIONS AND DEBRIEFING STATEMENT
OMI Vividness of Imagery Scale
The aim of this test is to determine the vividness of your
imagery. The items of the test will bring certain images to your mind.
You are to rate the vividness of each image by reference to the
accompanying rating scale which is shown at the bottom of the page. For
example, if your image is "vague and dim" you give it a rating of 5.
Record your answer in the brackets provided after each item. Just write
the appropriate number after each item. Before you turn to the items on
the next page, familiarize yourself with the different categories on the
rating scale. Throughout the test, refer to the rating scale when
judging the vividness of each image.
* Rating 1 Perfectly clear and as vivid as the actual experience *
* Rating 2 Very clear and comparable in vividness to the actual *
* experience *
* Rating 3 Moderately clear and vivid *
* Rating 4 Not clear or vivid, but recognizable *
* Rating 5 Vague and dim *
* Rating 6 So vague and dim as to be hardly discernable *
* Rating 7 No image present at all, you're only "knowing that" you *
* are thinking of the object *
55

56
An example of an item on the test would be one which asked you to
consider an image which comes to your mind's eye of a red apple. If
your visual image was moderately clear and vivid you would check the
rating scale and mark "3" in the brackets as follows:
Item Rating
5. A red apple (3)
Now turn to the next page when you have understood these
instructions and begin the test.
Think of some relative or friend whom you frequently see,
considering carefully the picture that rises before your mind's eye.
Classify the imagers suggested by each of the following questions as
indicated by the degree of clearness and vividness specified on the
Rating Scale.
1 The exact contour of face, head, shoulders, and body ( )
2. Characteristic poses of head, attitude of body, etc. ( )
3. The precise carriage, length of step, etc., in walking ( )
4. The different colors in some familiar costume ( )
Think of seeing the following, considering carefully the picture
which comes before your mind's eye, and classify the image suggested by
the following question as indicated by the degrees of clearness and
vividness specified on the Rating Scale.
5.The sun as it is sinking below the horizon ( )
Think of each of the following sounds, considering carefully the
image which comes to your mind's eye, and classify the images suggested
by each of the following questions as indicated by the degree of
clearness and vividness specified on the Rating Scale.

57
6. The whistle of a locomotive ( )
7. The honk of an automobile ( )
8. The mewing of a cat ( )
9. The sound of escaping steam ( )
10. The clapping of hands in applause ( )
Think of "feeling" or touching each of the following, considering
carefully the image which comes to your mind's touch, and classify the
images suggested by each of the following questions as indicated by the
degrees of clearness and vividness specified on the Rating Scale.
Rating
11. Sand ( )
12. Linen ( )
13. Fur ( )
14. The prick of a pin ( )
15. The warmth of a tepid bath ( )
Think of performing each of the following acts, considering
carefully the image which comes to your mind's arms, legs, lips, etc.,
and classify the images suggested as indicated by the degree of
clearness and vividness specified on the Rating Scale.
16. Running upstairs ( )
17. Springing across a gutter ( )
18. Drawing a circle on paper ( )
19. Reaching up to a high shelf ( )
20. Kicking something out of your way ( )

58
Think of tasting each of the following, considering carefully the
image which comes to your mind's mouth, and classify the images
suggested by each of the following as indicated by the degree of
clearness and vividness specified on the Rating Scale.
21. Salt
( )
22.
Granulated (white) sugar
( )
23.
Oranges
( )
24.
Jelly
( )
25.
Your favorite soup
( )
Think of smelling each of the following, considering carefully the
image which comes to your mind's nose, and classify the images suggested
by each of the following questions as indicated by the degree of
clearness and vividness specified on the Rating Scale.
26. An ill-ventilated room ( )
27. Cooking cabbage ( )
28. Roast beef ( )
29. Fresh paint ( )
30. New leather ( )
Think of each of the following sensations, considering carefully
the image which comes before your mind, and classify the images
suggested as indicated by the degree of clearness and vividness
specified on the Rating Scale.
( )
31. Fatigue

59
32.
Hunger
(
)
33.
A sore throat
(
)
34.
Drowsiness
(
)
35.
Feeling full after a large meal
(
)

60
Instructions to Subjects
In this study, you will be reading paragraphs that will appear on
this monitor screen, and then respond yes or no to a list of words,
depending on whether or not they were in those paragraphs.
The exact procedure will be as follows: You will hold these two
buttons, one in each hand (get subject to hold buttons), and use them to
indicate yes or no to the test words as they appear. You should press
the right button if the word is in the paragraph and the left button if
the word is not in the paragraph. You should press a button as fast as
possible while still making accurate decisions. You will also use the
buttons to present the paragraphs line by line.
After you press both buttons to start, the first line of a
paragraph will appear on the monitor screen. Read through the sentence
at a comfortable pace, and then press both buttons again to move to the
next sentence.
IMAGERY INSTRUCTIONS:
As you read the sentences, please make a special effort to form as
vivid an image of the events described as you can. Place yourself in
the image as a participant in the storyas if you were actually
experiencing it all personally. The more you can experience the actions
and feelings, the more vivid your image will be. Read through the lines
at your own pace, constructing your image as you read. The more vivid
your image of the events described in the paragraph, the more it is a
personal experience, the better you'll be able to remember it for the
subsequent test.

61
VERBAL INSTRUCTIONS:
As you read the sentences, you should concentrate on the specific
words used in each sentence. Read through the lines of the paragraph at
your own pace. Try to remember each sentence in the paragraph and the
specific vocabulary employed. As you read, try to retain the precise
wording of each sentence. The better your memory of the exact language
used, the better your performance will be on the word test that comes
after.
MAIN INSTRUCTIONS:
After you press the buttons to remove the last sentence, a word
pair will be presented on the screen. The first word will appear very
briefly, so you must stay alert and pay close attention. Read this
first word, but do not press a button. Wait for the second word to
appear. When the second word appears, your task is to press one of the
buttons to indicate whether it was in the paragraph you just studied.
Press the button in your right hand for yes, and the button in your left
hand for no. After you respond, another pair of words will appear. If
you are wrong, an error message will appear on the screen, and then the
next test pair will appear as normal. The test lists move very fast, so
it's important for you to stay alert and to respond as quickly and
accurately as possible. After the test list there will be a short pause,
and then a message will appear on the screen instructing you to press
both buttons to start the next trial. At the end of the session a
message will come on the screen to let you know you have finished, and
I'll come in shortly after that. In order to make your reaction times
more precise, please hold your thumbs lightly on the button throughout

62
the trials--so that you don't have to reach for them when you want to
make a response.
We will go through two practice trials now to be sure you
understand how the experiment goes.
Do you have any questions?
Present practice paragraphs and continue specialized instructions:
IMAGERY INSTRUCTIONS:
Tell me about the image you formed.
Were you aware of knocking on the door? Movement? Feeling of knocking?
Could you see the faces? The candles on the cake? Were you aware of
feeling surprised? Happy? REINFORCE AND INSTRUCT AS PER IMAGERY 18
VERBAL INSTRUCTIONS:
Repeat the first sentence in the paragraph as precisely as you can.
What age was mentioned in the paragraph?
What had you agreed to do that evening?
Did you smile with happiness or delight?
REINFORCE AND INSTRUCT FOR EXACT WORDS AND IGNORE ANY IMAGERY BEHAVIOR
OR RESPONSES.
Remind subject about baseline periods, sitting quietly not to disturb
physiological recordings, and tell them you'll be right next door.

63
Debriefing Statement
This study is concerned with determining whether people form
different mental representations of emotional vs. non-emotional texts.
You may have noticed that some of the paragraphs that you read were
pleasant, some were unpleasant, and some were relatively neutral. We
were measuring your reaction time to target words taken from those
paragraphs. Some target words were preceded (or primed) by words taken
from the same sentence, some target words were primed by words taken
from another sentence in the paragraph, and some targets were primed by
words that did not appear in the paragraph at all. Usually reaction
time is faster when a word is primed by a word that is closely related
(when the prime and target were from the same sentence in this case),
and measuring the changes in reaction time based on the logical distance
between the prime and the target can provide information about the
structure of the mental representation you have constructed from the
paragraph.
We think that mental representations of emotional paragraphs may
be more complex and extensive than those of neutral paragraphs, because
emotional stimuli more often include representations of motor activity
(jumping away from the palmetto bug) and of visceral activation (your
heart is pounding from the scare). This more extensive information
network would result in longer reaction times for affective paragraphs,
which is what we have found in an earlier study.
We also think that people who form vivid images may form
particularly extensive information structures, and so we would expect
those persons to have longer reaction times also.

64
This research is part of a larger program of experiments which
concern the study of modes of information processing and emotion in
persons who vary in imagery ability. We are working to develop a model
of how affective memories are represented, stored and retrieved in the
brain.
It is important that subjects participating in these studies not
have preconceived ideas about the experimenter's hypotheses and
expectations, in order to obtain valid experimental data. Therefore,
please do not discuss this experiment with other psychology students
until after the term is over.
I have read this form and had any questions concerning the study
answered to my satisfaction.
SUBJECT SIGNATURE

APPENDIX B
TEXT MATERIALS AND TEST LIST WORDS
Instructions and Practice Paragraph
Experimenter Instructions
This is a warm-up exercise to teach you special symbols used in
this study. This one (press right button to display *******) signals
you to make a "yes" response. This one (press left button to display
=======) signals you to make a "no" response. Try it a few times:
******** ======== ******** ========. These symbols will
appear in the test lists which will follow each paragraph. What does
======= signal you to do? What does ******* signal you to do?
Practice Paragraph
Today you turned twenty, and you wonder if any of your friends
will notice. You go out running after class and go over to study at a
friend's house later. Knocking at her door, you wonder if she has any
idea that it's your birthday. As you enter, people seem to jump out
from everywhere shouting "SURPRISE!" You gasp with delight as you see a
big cake with candles and balloons all over. You smile with affection
as they laugh gleefully that you suspected nothing.
STUDY WORDS: NOTICE, BIRTHDAY, IDEA, PEOPLE, DELIGHT, CAKE, NOTHING
NEW WORDS: POSTER, ERASER, BOLT, PERFUME, FLINCH, HYDROGEN, BLOOMS
65

66
Text Materials
1. Evening Walk
Unpleasant. One night you and your roommate go for a walk through
a bad neighborhood. You notice several dark figures slip quietly
between two buildings just ahead. You stop and listen closely; beads of
sweat break out on your forehead. Your breath catches as you hear soft
footsteps nearby, and you both panic. Echoes of your footsteps break
the night's stillness as you run for your car. At the car you're
fumbling with your keys when two guys grab your roommate.
Neutral. One night you and your roommate take a relaxed walk
through the neighborhood. You've been studying all evening and it feels
good to get up and walk around. There is a light sweat on your forehead
and the cool breeze feels refreshing. Quiet and preoccupied with your
studies, you have no particular desire to talk. A lone car drives by
momentarily creating echoes in the night's stillness. You walk back
home, thinking maybe you'll read one more chapter before bed.
STUDY WORDS: ROOMMATE, WALK, SWEAT, FOREHEAD, STILLNESS, ECHOES,
NEIGHBORHOOD
NEW WORDS: RECORD, WINNERS, PARROT, IMAGINATION, BARREL, GEESE,
PLASTER

67
2. Bicycle Ride
Unpleasant. You are riding your bicycle to school early in the
morning, in a rush. You are late for an exam, but you can't seem to get
free of all the cars. You're stuck on a narrow street in a line of
traffic moving slowly up the hill. Exasperated, you bolt for the other
sidewalk between cars, at top speed. You look around to a loud horn,
but the wind blows hair across your face. Suddenly you are struck hard
from the side, slamming you onto the pavement.
Neutral. You are riding your bicycle to school, relaxed and in no
particular rush. Pedaling smoothly along, you mentally review your
schedule for that morning. You go along steadily in a line of bikes
moving with the traffic up a hill. Picking up speed as you reach a wide
sidewalk on campus, you feel more relaxed. You coast downhill, lifting
your head to let the wind blow through your hair. Arriving at the
library you climb off your bike and lock it up in the rack.
STUDY WORDS: BICYCLE, SCHOOL, RUSH, TRAFFIC, HILL, SPEED, SIDEWALK
NEW WORDS:
NEPHEW, THIRST, MINK, CLAMP, HAM, SPARROW, COAL

68
3. Hot Summer Night
Unpleasant. Lying in the dark naked on a night too hot for
covers, you can't get any sleep. You have the ceiling fan on high
trying to get comfortable in the humid heat. Suddenly you feel a huge
palmetto bug drop from the ceiling onto your chest. You can feel the
insect crawl rapidly over your chest and up onto your neck. You leap
onto the carpet in panic and disgust, quickly switching on the lamp.
Scanning the bedclothes, you hesitantly lift the sheets looking for the
insect.
Neutral. You are in bed, drifting off to sleep on a summer night
too warm for covers. The ceiling fan cools off the summer heat, and the
ice cream helped too. You lie there relaxing, watching occasional car
lights play on the ceiling. The weekend starts tomorrow, and you think
about things you'd like to do. The phone ringsyou switch on the lamp
and walk across the carpet to get it. It's a good friend calling long
distance, and you sit down to have a talk.
STUDY WORDS: SLEEP, COVERS, FAN, HEAT, LAMP, CARPET, CEILING
NEW WORDS:
EAGLE, STEAK, VICTORY, WALNUT, LAKE, BOWL, TREE

69
4. Dental Appointment
Unpleasant. Your anxiety is raised by a call confirming your
appointment with the dentist. Walking in, you are instantly aware of
the whine of the drill from a back room. As you climb into the chair
you glimpse the pointed instruments on the tray. You wince as he drags
the sharp steel probe along your sensitive gumline. Hands wet with
sweat, you feel the long needle pierce the back of your mouth. As he
approaches you with the whining drill you grip the chair hard.
Neutral. You get a phone call that your appointment has been
cancelled with the dentist. Now you have some extra time and you decide
to have a leisurely breakfast. You settle into a comfortable chair,
reaching for a bagel on the tray. You turn on the morning news to get
the latest developments on the Iran probe. Putting the last bite into
your mouth you wipe your hands on your napkin. You look through the
paper and read the funnies before leaving for class.
STUDY WORDS: APPOINTMENT, DENTIST, CHAIR, TRAY, HANDS, MOUTH, PROBE
NEW WORDS: SOCKS, PUPPY, KNOB, RAILWAY, BOXES, DINNER, PLANET

70
5. Airplane Trip
Unpleasant. You startle as the plane lurches violently in
midflight on your way home. Confused, the passengers stir restlessly as
the cabin lights go out. Suddenly there's an odd noisethe engines'
sputtering silences all talk. Grabbing the chair arms hard, you feel
the hull shudder violently. The hostess is thrown sideways, spilling
the drinks she was carrying. Screams pierce the dark as the plane tilts
sideways and starts to fall.
Pleasant. You are flushed with excitement as the plane soars
high, taking you home. Soon you will be with your boyfriend again after
a long separation. The noise of small talk around you is of no
interest, you are so happy. Smiling with a rush of pleasure, you
imagine him in your arms again. When the hostess comes by with drinks,
you decide on champagne to celebrate. As you approach your destination
you smile happily, full of anticipation.
STUDY WORDS: PLANE, HOME, NOISE, TALK, DRINKS, HOSTESS, ARMS
NEW WORDS: REMAINDER, COUNTER, PEPPER, APPLES, GRAPEFRUIT, FRONTIER,
CAPTIVE

71
6. Football Game
Unpleasant. The football game is a few minutes past halftime
inside the packed stadium. You have a bad headache and the hot sun
makes you feel worse by the minute. The crowd roars and tension is
high--they are only yards away from a touchdown. You're sick and
uncomfortable on the hard seat as play continues on the field. Your
stomach churns painfully as the stands pulsate with the roaring cheers.
Your head throbs and you're sick to your stomach as the game finally
ends.
Pleasant. The football game is in the last few minutes inside the
packed stadium. You sweat freely in the hot sun as your team moves into
playing position. Tension is high--you are one point behind, and thirty
yards from a touchdown. You jump from your seat cheering wildly as the
play moves down the field. The stands pulsate with the crowd's cheers
as the halfback gets the ball. Touchdown! You thrust your arms
overhead as the crowd roars with excitement.
STUDY WORDS: MINUTES, STADIUM, TENSION, YARDS, SEAT, FIELD, PLAY
NEW WORDS: SUGAR, MONSTER, ECONOMY, MAYOR, PRISON, HAT, SPINACH

72
7. First Date
Unpleasant. Out on a first date your shyness around this new
person is painfully obvious. Sitting stiffly in your seat, there is no
conversation before the movie starts. You are aware of your sweaty
palms and feel a slow blush covering your face. The silence drags on
endlessly as you nervously search for something to say. You finally
manage to make a joke and fall silent after the brief response. You sit
there silent and miserable, wishing the evening were over.
Pleasant. Out on a date, you're having a wonderful time with a
very special person. Conversation seems effortless as you laugh and
talk before the movie starts. You feel a blush of excitement and
attraction as you make plans for later on. You agree to go dancing
later in the evening as silence falls over the theater. You lean over
close with a last little jokehe squeezes your hand in response. You
sit there, warm and excited, wishing the evening would go on forever.
STUDY WORDS: DATE, PERSON, CONVERSATION, MOVIE, JOKE, RESPONSE, BLUSH
NEW WORDS:
BUSHEL, STRAND, SCARF, CEREAL, SWEATER, SALUTE, REMINDER

73
8. Beach Picnic
Unpleasant. You're having a great time on a picnic by the ocean
with some good friends. Some are out swimming, and you look twice when
you think you spot sharks. Your feet pound through the surf as you run
calling, the fins now obvious. Unaware, your friends don't hear you,
and you are gripped with horror. Your heart pounding in your chest, you
run across the sand shouting and waving. You scream in desperation as
someone is jerked under, and the water turns red.
Pleasant. You're having a great time on a picnic by the ocean
with some good friends. Some are out swimming, and you wave to them as
you walk along the surf. Walking along, you stumble as your feet trip
over a pair of lost swim fins. You look up to see a frisbee sailing in
your direction, and you go after it. Your heart pounds in your chest as
you run hard across the sand to catch it. You laugh and shout to your
friends as you grab it, full of high spirits.
STUDY WORDS: PICNIC, OCEAN, FEET, FINS, HEART, SAND, FRIENDS
NEW WORDS: DEGREE, WAGON, PINT, CHARM, TENT, CALF, SCISSORS

74
9. End of School Term
Neutral. It's the end of another school year and you are going
home for the summer. You have been packing, and just finished finals
for your courses this term. You start carrying plants and clothes and
suitcases out to your parents' car. Everything is finally loaded up and
you drive away, waving to your roommate. Leafing through a magazine in
the back seat you think about a summer job. Thinking of the long trip
ahead, you settle down in the comfortable car.
Unpleasant. It's the end of the school year and you have to go
home for the summer. You feel down, having just told your parents you
failed two finals this term. You wearily drag plants and clothes and
suitcases out to their car in silence. Your father says that maybe you
should quit school and you feel more depressed. Sitting in the car, you
think about the boring job you'll have to start soon. You sigh and
settle down in the back seat, dreading the long trip ahead of you.
STUDY WORDS: YEAR, SUMMER, FINALS, TERM, PLANTS, CLOTHES, SUITCASES
NEW WORDS: CHIMNEY, LANTERN, GLUE, MELON, RECEIPT, CANAL, SILVER

75
10. Elevator
Neutral. You are part of a crowd of shoppers, running errands in
a new shopping mall. Walking around, you discover the mall has a glass
elevator and a nice fountain. Your errands are run, and you notice you
feel slightly on edge from hunger. On your way out, you stop and treat
yourself to a delicious cold milkshake. It's a relief to get out into
the cool darkness after the busy confusion there. As you drive away,
you start to relax and think about watching a movie at home.
Unpleasant. You are part of a crowd of shoppers, running errands
in a large shopping mall. The elevator in the mall is full, but you're
in a hurry and squeeze in anyway. People are pressed in around you; you
feel uncomfortably hot and on edge. Suddenly the elevator jerks to a
stop; you startle as a loud alarm goes off. There is no relief as panic
spreads in the darkness in the tight enclosure. You clench your teeth
as a baby screams; you feel yourself sweating heavily.
STUDY WORDS: SHOPPERS, ERRANDS, MALL, ELEVATOR, EDGE, RELIEF, DARKNESS
NEW WORDS: SUNSET, RANCH, BALLET, VINES, LIQUID, CARAMEL, EGGS

76
11. Father Waits Up
Neutral. Coining home from racquetball and a movie with a friend,
it's gotten late. Although it's midnight your father is still up, so
you introduce your friend. As you close the front door you offer a
vague explanation for your late return. You groan as you sit on the
couch suddenly aware of sore muscles from the game. Your father has a
sleepy look on his face as he smiles and says goodnight. Pretty tired
yourselves, your friend soon leaves and you head off for bed.
Unpleasant. Coming back from a late movie with a girlfriend, it's
gotten very late. Her father has been waiting up, and you realize that
it's hours after midnight. He slams the door shut after you come in,
and loudly demands an explanation. Your muscles tense as he rudely
orders you in and sits you both on the couch. You feel nervous and a
little scared from the furious look on his face. You feel yourself
sweating and your face getting red as he goes on yelling.
STUDY WORDS: FATHER, MIDNIGHT, EXPLANATION, DOOR, MUSCLES, COUCH, FACE
NEW WORDS: SALMON, PURSE, PASTURE, TOAST, MARE, PRUNE, BRICK

77
12. Parents' Pool Party
Neutral. The evening of your parents' pool party the guests are
not very exciting. Finally tired of hearing the old familiar stories,
you head for your room. You take off your watch, your shoes and
clothes, and put on your bathing suit. Things have quieted down outside
by the pool, and you dive in for a quick swim. You swim vigorously
through the cool refreshing water using your full strength. You come
out of the pool feeling much better and head off for the refreshments.
Unpleasant. It's late and some of the guests at the pool party
are getting very drunk. You ask them to quiet down but they laugh,
grabbing your arms and both legs. You panicyou're wearing expensive
shoes and a watch that isn't waterproof. As they drag you towards the
pool you yell that your watch will be ruined. Angrily, you fight with
all your strength as they swing you out over the water. You plunge into
the cold water and come up choking, completely furious.
STUDY WORDS: EVENING, GUESTS, WATCH, SHOES, POOL, STRENGTH, WATER
NEW WORDS: SPACE, DWARF, CREAM, MIRROR, ATMOSPHERE, ONION, SONG

78
13. Canoe Ride
Neutral. You lean forward as the canoe glides smoothly along
among the water lilies. Slowing the boat, you scan the surface of the
water close to the shore. You spot a duck nesting on her eggs over to
the side, hidden among the reeds. You hurriedly check the camera--the
timing is critical in this situation. You watch closely, adjusting the
focus as the full nest comes into view. Delighted with your luck, you
get several excellent shots of this rare find.
Unpleasant. You slip out of the canoe into the brackish water
among the water lilies. Unnoticed, a huge alligator glides off the
shore, sliding under the surface. You jump as you spot something big
moving in the reeds over to the side of you. Panic shoots through you
as you suddenly know you're in a dangerous situation. Your heart stops
as you see the alligator now gliding towards you in full view.
Desperate to escape, you run hard, crashing through the reeds and
lilies.
STUDY WORDS: CANOE, LILIES, SURFACE, SHORE, SIDE, REEDS, SITUATION
NEW WORDS: DESK, KITCHEN, ORANGES, PORCH, STEAM, TEMPLE, CRUST

79
14. Tennis Game
Unpleasant. A cute guy you met invited you to play tennis but
you're not in good form. As the game begins you make clumsy mistakes
and feel more and more embarrassed. You can't seem to get the feel of
the game as the ball lands in the net again. You breathe hard, trying
to catch your wind between volleys, your face flushed. As you lose the
match, you shout hollow congratulations and manage a smile. You redden
as you see people watching you; you just want to get off the court.
Pleasant. You are playing tennis with a good friend, feeling
energetic and in top form. Running hard across the court, every move
feels rightyou're playing well. You delight in the solid feel of each
shot as you slam the ball across the net. You breathe hard, catching
your wind between volleys, your face flushed. Your partner shouts
congratulations with a big smile as you win another match. You walk off
the court feeling great, ready for a shower and a cold drink.
STUDY WORDS: TENNIS, FORM, FEEL, NET, SMILE, MATCH, VOLLEYS
NEW WORDS: JUDGE, PILLOW, ROBIN, POUND, JUNGLE, BRACELET, CUPS

80
15. Favorite Rock Star
Unpleasant. You are over visiting with friends, while the radio
plays in the background. Suddenly the DJ says that your favorite
musician was just shot and killed. You sit frozen and numb; some of the
others sit quietly crying in the room. You can't believe this is true.
Your spirit sinks as you realize how empty this makes the future. The
news is so upsetting that you don't know what to do next.
Pleasant. You are visiting with friends, while the radio plays in
the background. Suddenly the DJ announces that you won a trip to meet
your favorite musician. You sit thrilled and amazed as the others
scream and dance around the room. You can't believe it! Your spirit
soars as you realize the future holds this great meeting. The news is
so exciting that you don't know what to do next.
STUDY WORDS: RADIO, BACKGROUND, OTHERS, ROOM, SPIRIT, FUTURE, MUSICIAN
NEW WORDS: DIRT, TRIFLE, ABILITY, COTTAGE, POTATO, CANDY, BOTTLE

81
16. Driving in Traffic
Unpleasant. You are driving your car across town in heavy
traffic, on a hot summer day. The windows are down, and you are
sweating heavily in the hot summer sun. A wasp blows in striking you in
the face, and you accidentally hit the horn. You jerk the wheel,
swerving dangerously out of your lane, and just miss a car. You wave
around your head and neck, anxiously aware of the buzzing behind you.
You feel a sharp sting on your neck and strike at it, trying to get it
off.
Pleasant. You are out for a drive heading across town in your
beautiful new sports car. The windows and top are down; you bask in the
sun and some admiring glances. You notice an old friend walking along
the sidewalk and you blow your horn. His head turns towards you and you
notice the buzzing interest among his group. Turning the wheel, you
move neatly into the curb lane, and drive up beside him. The day is now
perfecthe climbs in and you head to the country for a ride.
STUDY WORDS: TOWN, CAR, HEAD, BUZZING, WHEEL, LANE, HORN
NEW WORDS: BOOT, OVEN, STAGE, MACKEREL, BLEACH, SOLDIER, BUBBLE

82
17. Hotel Party
Unpleasant. You are at a large exclusive party which packs
several floors of an old hotel. Out a window you suddenly see fire
exploding into the sky up on the hotel roof. Someone runs in yelling
that the hotel is on fire from the fireworks up there. Incredibly fast,
the halls fill with smoke and panic spreads through the crowd. In
disbelief, you stare at windows shattering in the repeated explosions.
Your lungs and eyes burn as you struggle towards a balcony, gasping for
air.
Pleasant. At the hotel, you two leave the party and wander up to
the rooftop garden. It's a beautiful clear night; you stand on the roof
gazing up at the sky. Under the full moon, soft kisses along your
throat fill you with warmth. The noisy crowds down in the halls filled
with smoke seem very far away now. Back inside, you lose yourselves in
soft caresses and explosions of desire. His long, deep kisses send
thrills of pleasure all through your body.
STUDY WORDS: PARTY, HOTEL, ROOF, SKY, HALLS, SMOKE, EXPLOSIONS
NEW WORDS: CLIP, PASTRY, BASEBALL, WEED, TODDLER, NOVEL, GRASS

83
18. Ski Slopes
Unpleasant. It is the dead of winter, a windy day with dark skies
and freezing temperature. You glumly climb into the chair on the ski
lift, behind your best friend. Far below you great clouds of snow are
being blown up off the deserted slopes. Jerking up the hill, a strong
gust of icy wind hits the ski lift hard. You watch helplessly as your
friend's chair swings violently on the cable. You scream, staring in
disbelief, as your friend is thrown from the lift.
Pleasant. It is winter and a beautiful day with clear blue skies
and mild temperature. Full of energy, you gaily hop onto the ski lift
with your best friend. You watch expert skiers far below raise up
clouds of snow on the long slopes. As you glide up on the lift a sudden
gust of wind knocks your friend's hat off. The chair swings on the
cable as your friend reaches out to grab it, laughing. Soon the hat is
only a small bright spot on the sparkling white hill below.
STUDY WORDS: WINTER, TEMPERATURE, CLOUDS, SNOW, GUST, LIFT, SLOPES
NEW WORDS: HAMBURGER, CAKE, NATION, BEAR, CUBICLE, MERCHANT, COMBS

84
19. Mexican Vacation
Neutral. You are on vacation on the Mexican coast, arranging some
deep sea fishing. Standing on the deck, you are aware of the constant
rocking motion of the boat. The guide explains that big tuna inhabit
the deep shadows around the shoals. You reflect that speaking Spanish
all the time will probably get to be a pain. You mentally check your
supplies, adding the pills that prevent sea sickness. Hopefully the
weather will be good, but you'll pack a raincoat just in case.
Unpleasant. You are on vacation on the Mexican coast, out doing
some deep sea fishing. Standing on the deck, the constant motion of the
boat makes you lightheaded. Nausea creeps over you as you search for
tuna in the shadows under the waves. Suddenly another tourist makes a
bad cast and a hook sinks into your cheek. You scream with pain and
several Mexicans start speaking rapid Spanish at you. Blood drips down
onto your shirt and you feel like you're going to be sick.
STUDY WORDS: VACATION, COAST, DECK, MOTION, TUNA, SHADOWS, PAIN
NEW WORDS: BLADE, VENTURE, CHIMNEY, PORTIONS, HAWK, GERMS, CAVE

85
20. Shower
Pleasant. It is early morning as you reach for a towel to dry off
after your shower. The fluffy towel feels pleasantly soft as you dry
yourself with long strokes. You rub your hair lightly and dry your face
and neck, feeling more awake now. Stepping out of the shower stall you
catch the fragrant odors in the bathroom. Finally, you're done; your
skin feels smooth, your body feels fresh and clean. Refreshed and ready
for the day, you put on your robe and go back to your room.
Unpleasant. You dry off from your shower with a thin cotton towel
in the old locker room. The towel turns into a small soggy lump after a
few half-hearted strokes. You are vaguely aware of something crawling
on your hair or maybe your neck. The narrow stall is dark, humid, and
smells of mold and stale body odors. You jump and your skin crawls as
you feel a bug scuttle down your naked body. Your heart pounds as you
nervously scan the walls and floor in the dark stall.
STUDY WORDS: SHOWER, TOWEL, STALL, ODORS, SKIN, BODY, STROKES
NEW WORDS:
LOG, HINGES, SATIN, DRILL, VEST, CHALK, KIDNEY

86
21. Moving into Apartment/ Rat and Babv
Neutral. The apartment is on a busy, tree-lined avenue in a
pleasant part of the city. The guys are gradually unpacking the truck
parked in front along the curb. They carry in the crib, the beds, and
chairs, while you go out for food. It is near dinner time and everyone
is ready to relax for a minute and eat. You give the kids some toys and
a cookie, and see that the baby is asleep. As you come back with dinner
you see the truck is empty and everyone's inside.
Unpleasant. The apartment is on a crowded, dirty avenue in the
slums of the city. It is late, but no adults are home in the dark,
empty apartment. A large rat climbs up into the baby's crib, sniffing
around for food. The baby draws back and cries anxiously as the ugly
animal sniffs around. Roaches crawl over a broken cookie on the floor
amid scattered toys. The rat approaches the frightened, crying child in
the dark room.
STUDY WORDS: APARTMENT, AVENUE, CRIB, FOOD, COOKIE, TOYS, BABY
NEW WORDS:
MONKEY, TORNADO, THEME, BUTTER, BASEMENT, FUNGUS, GARMENT

87
22. Radio Station/Blood Donation
Neutral. You are touring the radio station on campus, fascinated
by all the equipment. They show you everything from the phonograph
needle to the cot for naps. You are chosen as a volunteer to play a
record and blood rushes to your cheeks. You manage to get it started
and playing with only a little difficulty. As you pull the lever to
lift the needle off, the announcer gives you a smile. You stand back
and relax, watching the next person play another record.
Unpleasant. You are sitting in the blood donor station, noticing
all the medical equipment. The sheer size of the long needles makes you
sick as you sit waiting on the cot. You stare at the dark blood filling
the syringe of the volunteer next to you. It is your turn now, and you
smile with difficulty as she tightens the tourniquet. You feel your
hands sweat as she prepares the syringe; it's taking her forever. A
sharp pain shoots through your arm as the needle slowly pierces your
skin.
STUDY WORDS: STATION, EQUIPMENT, NEEDLE, COT, BLOOD, VOLUNTEER,
DIFFICULTY
NEW WORDS: CRACK, SQUIRREL, CAMEL, INVASION, MEADOW, SALAD, DISPUTE

88
23. Graduation
Pleasant. Coming down the stairs the morning of your graduation
you're happy and excited. As you sit down to breakfast your parents are
smiling at each other a lot. Looking down, your eyes fix on a set of
car keys sitting in your plate. You jump up from your chair, grabbing
them, and run to the front door. Sitting by the curb is an incredible
presenta beautiful, shining new car. You just can't believe it! You
are so excited you hardly know what to do.
Unpleasant. You move slowly down the stairs the morning of your
graduation feeling sick. Determined to attend the ceremony that
afternoon, you sit down to breakfast. Noise from the television set
hurts your head as you stare at your plate. You feel miserable as you
realize you're too sick to even eat your toast. Looking out you see a
friend stepping up on the curb with a present for you. Your head throbs
and you feel terribleyou are going to miss your graduation.
STUDY WORDS: STAIRS, MORNING, SET, PLATE, CURB, PRESENT, BREAKFAST
NEW WORDS: BONE, CABBAGE, SKATER, SHIRT, STATUE, TILE, CABINET

89
24. Partv/Nighttime Intruder
Pleasant. Arriving home at night, you hear sounds of a party from
across your back yard. You just finished your last final; you are up
and really ready to celebrate. Your hand moves fast to pick up the
telephone by your bed when it rings. You hear laughter and music behind
the voiceit's the cute neighbor you like. Your heart beats with joy
as you rush out, the screen door slamming behind you. You walk into the
apartment full of music and dancing, and your mood soars.
Unpleasant. Alone and lying awake at night you hear sounds of
something out in the yard. You become instantly alert and scarcely
breathe as you lie rigid, listening. You hand is lightly shaking as you
reach for the telephone sitting by the bed. Your voice dies in your
throat as you think you see a shadow moving outside. Your heart pounds
in your chest as you hear someone cutting the screen door. The police
are finally on the line, and you struggle to give your address.
STUDY WORDS: SOUNDS, YARD, TELEPHONE, BED, HEART, SCREEN, VOICE
NEW WORDS: DOLLAR, CHEMIST, HUNGER, MINT, GOOSE, PLOW, FRUIT

90
25. Sailboat
Unpleasant. You duck your head as waves crash over the bow,
drenching the whole crew. The wind tears powerfully at the sails,
pulling you over in a sideways motion. Holding the rail, you brace
yourself as the boat lurches unsteadily forward. The day is cold, the
sea dark and churning with dirty caps of gray foam. You take a sharp
breath; the ice cold spray cuts hard against your face. You clutch the
mast with all your strengthyou are going under.
Pleasant. You laugh as the waves race past the bow, spraying most
of the crew. The wind pulls powerfully at the sails, increasing the
boat's forward motion. Holding to the mast, you steady yourself as your
boat surges into the lead. The day is perfect, the sea shining bright
blue with flecks of foam. You take a deep breath; the spray feels
invigorating, fresh against your face. Around you everyone is shouting
and cheeringyou are winning the race!
STUDY WORDS: WAVES, BOW, DAY, FOAM, BREATH, SPRAY, CREW
NEW WORDS:
TOPIC, CORNER, FEVER, MAP, HELMET, RUBY, OWL

91
26. Walk Across Campus
Pleasant. Walking across campus in a hurry, you thread your way
among crowds of students. With a shock you spot the attractive guy you
met at the library last night. Your attention is drawn to him as he
walks the across the lawn towards you. He comes closer and looks up,
and as your eyes meet your heart skips a beat. Your pace slows quickly,
and you finally come to almost a complete stop. You thrill with delight
as he flashes you a smile and stops to say hello.
Unpleasant. Walking across campus in a hurry, you thread your way
amid crowds of students. With a shock, you spot your boyfriend and one
of your friends holding hands. In disbelief, your attention fixes on
them talking closely on the lawn. Your heart pounds as you walk up to
them, and you feel upset and agitated. Your pace slows as they look
over, startled, and their talk comes to a stop. Taking a deep breath,
you fight back your feelings and force yourself to speak.
STUDY WORDS: CAMPUS, STUDENTS, ATTENTION, LAWN, PACE, STOP, SHOCK
NEW WORDS: CELERY, CRANE, MUD, SPOON, PICKLE, LEMONADE, MAGAZINES

92
27. Coming Home from Library
Neutral. Coming back from a good study session at the library,
it's early evening. Your neck and shoulders feel tight since you've
been reading for two hours. You stretch with relief as you put down
your booksyou're ready to relax. Turning on the radio, you sit down
in your favorite chair and open your mail. Later on you test yourself,
reviewing parts of the information you covered. Most of it comes back
to you, but you decide to review your notes once again.
Unpleasant. Coming back from a long cramming session at the
library, it's late at night. Your shoulders and neck feel stiff from
hunching over your work for hours. You put down your books, aware of a
dull tension headache behind both eyes. As you sit down to open your
mail, the fatigue and worry are overwhelming. You try to rehearse parts
of the information you covered, but you can't think. You just can't
seem to stop worrying about the big test and your bad grades.
STUDY WORDS: SESSION, LIBRARY, SHOULDERS, HOURS, PARTS, INFORMATION,
MAIL
NEW WORDS:
CONQUEST, STONE, DREAM, COMPUTER, GREASE, LUMBER, THOUGHT

93
28. Store Clerk
Neutral. You are opening the register as you start your shift in
a nice clothing store. Only a few customers are in your area of the
store, so you can take your time. You turn off the alarm, find the key
to the drawer lock, and get out a pen. You are all set up now, and you
wander over to look through the new blouses. There a beautiful white
silk one with a front pocket that you especially like. A customer is
browsing through the skirts nearby and you go over to help her.
Unpleasant. You are closing the register after the late shift at
a convenience store. The store is in a secluded area, and you don't
like being there alone at 2 a.m. At last it's time to close; you turn
the lock on the door and set the alarm. Suddenly you look up to see a
man's face watching you through the window. You freeze with fear,
staring, as he pulls a gun out of his jacket pocket. You jump as he
shouts something at you, motioning with a jerk towards the door.
STUDY WORDS: REGISTER, SHIFT, STORE, AREA, LOCK, ALARM, POCKET
NEW WORDS:
TACKLE, SINK, DIME, ORGAN, RIBBON, BRIDE, BUMPS

94
29. First Day of Class/Final Exam
Neutral. You walk into the classroom on the first day of class,
full of confidence. Sitting down, you put your books and papers under
your chair and sit back. A booklet is handed out, and a few homework
questions for the first assignment. You take out pencil and paper as
the organization of the course is outlined. Your brain is full of
information as you note down the date of the first test. Walking out
after class you reflect that this looks like an interesting course.
Unpleasant. You walk into the classroom to take your final with
no confidence at all. Behind in the course, you crammed all night, but
there was too much to learn. Opening the test booklet you pale with
apprehension as the questions look hard. You find you don't even have a
pencil and, embarrassed, have to borrow one. Your brain is filled with
confusion as you read the test; you just can't think. Avoiding your
professor's eyes, you turn in the half empty exam and leave.
STUDY WORDS: CLASSROOM, CONFIDENCE, BOOKLET, QUESTIONS, BRAIN, TEST,
PENCIL
NEW WORDS: VENDOR, RELUCTANCE, PEANUTS, MEDICINE, LUNCH, MOTEL,
SPINACH

95
30. Motorist Survev/Traffic Ticket
Neutral. You have an assignment to survey motorist opinions of
the 55 mph speed limit. With two classmates, you flag down cars on a
street corner under the lights. You stop a policeman for fun and laugh
with surprise when he says he'll answer. He talks about how speeding is
a big mistake, even when the roads are empty. You listen with interest
as he quotes a lot of statistics about accidents. As he drives away,
you think what a great addition to your survey this will be.
Unpleasant. Driving along, a motorist swerves past you going well
over the speed limit. Looking up, you see a police car close behind you
with flashing red lights. You feel a shock of surprise and nervousness
as the policeman pulls you over. Unaware of his mistake, he flatly
accuses you of speeding and reckless driving. As you try to explain, he
angrily cuts you off and tells you to speak in court. You fight to
control your angry feelings as he silently writes you a ticket.
STUDY WORDS: MOTORIST, LIMIT, SURPRISE, POLICEMEN, SPEEDING, MISTAKE,
LIGHTS
NEW WORDS: MOUNTAIN, STRAWBERRY, OXYGEN, KITCHEN, MIXTURE, GRADE,
FLOCKS

96
31. Driving on Highway/Accident
Neutral. Your comfortable automobile drives on steadily after the
brief summer storm. The rain has stopped and you reach over to turn off
the windshield wipers. Your hands lightly steady the wheel as the road
follows a wide smooth curve. Keeping an eye on the curve, you turn the
radio on to your favorite station. Driving without effort, you notice
the wild flowers in the ditch by the road. You yawn and stretch as you
check the clock againjust one more hour to go.
Unpleasant. You tense up as the small automobile swerves
dangerously in the heavy storm. The windshield wipers are useless as
rain blows in sheets across the road. Your hands grip the wheel as you
peer through the dark and see a curve ahead. You turn hard to make the
curve--you can feel the car skidding sideways. In a desperate effort to
stop the car, you run off the road into the ditch. Your head slams hard
against the window as you feel the car pitch sideways.
STUDY WORDS: AUTOMOBILE, STORM, RAIN, WIPERS, EFFORT, DITCH, ROAD
NEW WORDS:
PIPE, FROST, PEARS, CHUNK, CLOTH, SUIT, SNAIL

97
32. Staff Meeting/Partv
Pleasant. The staff drifts in and the party gets started, now
that the meeting is over. Your mouth waters as you open the pizza and
get out the ice cold cokes. With no hesitation, you put away your
project notes and grab a cold drink. You feel very hungry as you
breathe in the delicious smell of the hot pizza. The clock hits seven,
and silence falls over the room as you all start to eat. The ice cold
coke perfectly complements the pizza's spicy sauce and cheese.
Unpleasant. You feel tense as the staff gathers for a meeting in
the conference room. Your stomach churns and your mouth is dry as you
nervously wait to speak. With some hesitation you look over your notes
again, your hands shaking. You cringe with embarrassment as you stumble
badly over the introduction. They all sit in silence staring at you;
the clock ticks loudly in the silence. Your heart pounds and your face
is burning as you start over once more.
STUDY WORDS: STAFF, MEETING, HESITATION, NOTES, CLOCK, SILENCE, MOUTH
NEW WORDS:
MUSEUM, YARN, WIDOW, RULER, COLISEUM, STOCKINGS, VIOLENCE

98
33. Rollercoaster
Neutral. You slam the door shut for your little cousins at the
rollercoaster ride. You are having an enjoyable day at an amusement
park with your relatives. Looking up from below, you hear faint laughs
and shrieks as they top the crest. Your stomach growls with hunger as
you follow their course down the incline. While they are still zipping
around curves on the ride, you go get a snack. With hot dog and coke in
hand, you greet them as they walk down the ramp.
Unpleasant. You and your cousins are climbing into your seats on
the rollercoaster ride. You are feeling very anxious and nervous, but
try to hide it from the others. You sit pale among the laughs and
shrieks, the car jerking up to the crest. Your stomach drops as the car
plunges, roaring fast down the steep incline. You whip hard around two
sharp curves, feeling more and more dizzy and sick. Your hands are
white as you grip the bar hard, trying to take a deep breath.
STUDY WORDS: COUSINS, RIDE, LAUGHS, SHRIEKS, STOMACH, INCLINE, CREST
NEW WORDS: BUTTONS, ENVELOPE, FLAME, CARTON, MAID, PAINT, OLIVE

99
34. Hike through Woods
Pleasant. It's late spring, and you and your group have hiked
miles into the woods. Suddenly, a beautiful baby fawn steps uncertainly
onto the path just ahead. It gazes at you in silence with deep brown
eyes, munching on some leaves. You glance sideways to meet the look of
pure delight on the others' faces. Caught with this wonderful sight,
you stand completely still on the path. You marvel at this small,
spotted brown fawn standing trustingly before you.
Unpleasant. It's late spring, and you and your group have hiked
miles into the woods. You jump violently off the path realizing you've
just stepped on a rattlesnake. It coils back among some leaves fixing
you in silence with expressionless eyes. You glance sideways at the
looks of fear and helplessness on the others' faces. Frozen at the
sight of the snake's coiled body, you are too afraid to move. Striking
with lightning speed, you feel its fangs sink deep into your calf.
STUDY WORDS: PATH, EYES, SPRING, WOODS, LEAVES, MILES, GROUP
NEW WORDS: CORN, FRAUD, FLOOD, BELLS, PARSLEY, PEBBLE, CORK

100
35. Sauna
Pleasant. You are walking out, having just finished your favorite
class at your gym. The muscles in your abdomen and legs feel tight and
you head for the sauna. Stepping into the hot steam of the sauna, you
can feel yourself relaxing. Only a few other people are there, and you
stretch out comfortably on a bench. The heat penetrating your tired
limbs is a wonderfully soothing feeling. You step out into the cool air
again, feeling refreshed and energized.
Unpleasant. You have just finished a strenuous class at a rundown
old gym in town. Your legs are tired and shaking; you have a sick
feeling in your abdomen. Stepping into the hot steam of the sauna you
sink onto the bench, exhausted. You lean back against the wall, head
pounding, realizing you've overdone it. The overwhelming heat gives you
a suffocated feeling; you need some fresh air. Walking unsteadily
towards the door, you lean against it but it won't open.
STUDY WORDS: CLASS, GYM, ABDOMEN, LEGS, STEAM, BENCH, FEELING
NEW WORDS: BOOTS, THEME, DELIGHT, BARN, SLIPPER, ISLAND, ANGLE

101
36. Tour of New York City
Pleasant. You are on a trip to New York City, touring the sights
for the first time. Your tour route ends on the deck at the Lincoln
Center overlooking the city. You are breathless in admiration at the
spectacular skyscrapers all around. Another special building is a few
streets away and you decide to go there too. Soon you start thinking
about the evening's entertainment and buy a newspaper. You're ready for
another night of exploring great restaurants and nightclubs.
Unpleasant. You are on a trip to New York City, touring the
sights for the first time. Your tour route ends on the deck at the
Lincoln Center overlooking the city. It's very windy up there, and you
feel uncomfortable as the building sways. You feel dizzy, sensing the
building move as you look far down at the streets. You pale at the
sight of the city so far below, and start to lose your balance. Your
breathing is shallow and rapid, your face feels cold and sweaty.
STUDY WORDS: SIGHTS, TIME, ROUTE, CITY, BUILDING, STREETS, TRIP
NEW WORDS: PRIZE, GLASSES, SPONGE, CARROT, VEST, PRAIRIE, LAWYER

APPENDIX C
IMAGERY ABILITY SCORES FOR INSTRUCTION GROUPS
Imagery Instructions
Verbal
Instructions
Good Imagers
Poor Imagers
Good Imagers
Poor Imagers
40
81
37
84
45
86
43
85
48
87
49
85
58
93
50
92
58
94
51
95
59
95
56
96
62
97
57
98
63
101
60
104
67
105
61
113
69
105
65
121
76
108
76
123
78
111
76
125
79
114
76
79
117
77
147
79
102

APPENDIX D
TABLES OF MEANS FOR DEPENDENT MEASURES
Table 6
Mean reading time (sec) for early, middle, and late sentence pairs
of affectively pleasant, neutral and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.49
5.35
5.59
5.48
Middle
5.08
5.27
5.30
5.22
Late
4.63
4.72
4.79
4.71
Mean
5.07
5.12
5.23
5.14
Table 7
Mean reading
time (sec)
for early, middle, and late sentence
of affectively pleasant, neutral
and unpleasant
texts
for qood and
poor imagery ability subjects
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.10
4.77
5.16
5.01
Middle
4.63
4.65
4.84
4.70
Late
4.35
4.44
4.39
4.39
Mean
4.69
4.62
4.80
4.70
103

104
Table 7continued
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.87
5.93
6.01
5.94
Middle
5.53
5.90
5.77
5.73
Late
4.90
5.00
5.19
5.03
Mean
5.44
5.61
5.66
5.57
Table
8
Mean reading
time (sec) for
early, i
middle, and late
sentence
of affectively pleasant, neutral and unpleasant texts
for subiects receiving verbal
and imagery instructions
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.39
5.36
5.58
5.44
Middle
5.12
5.21
5.39
5.24
Late
4.69
4.76
4.92
4.79
Mean
5.07
5.11
5.30
5.16
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.59
5.34
5.60
5.51
Middle
5.04
5.34
5.22
5.20
Late
4.56
4.68
4.66
4.63
Mean
5.06
5.12
5.16
5.11

105
Table 9
Mean reading time (sec) for early, middle, and late sentence
pairs
of affectively pleasant, neutral and unpleasant texts
with and without startle probes
No Startle Subjects
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.41
5.34
5.36
5.37
Middle
5.09
5.06
5.16
5.10
Late
4.55
4.64
4.76
4.65
Mean
5.01
5.01 5.10
Startle Subjects
5.04
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.57
5.37
5.82
5.59
Middle
5.08
5.48
5.44
5.33
Late
4.71
4.80
4.82
4.78
Mean
5.12
5.22
5.36
5.23

106
Table 10
Mean reading time (sec) for early, middle, and late sentence pairs
of affectively pleasant, neutral and unpleasant texts, for good and poor
imagery ability subjects receiving verbal and imagery instructions
IMAGERY INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.11
4.95
5.22
5.09
Middle
4.83
4.90
5.09
4.94
Late
4.56
4.64
4.72
4.64
Mean
4.84
4.83
5.01
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.66
5.78
5.95
5.79
Middle
5.42
5.53
5.68
5.54
Late
4.82
4.88
5.12
4.94
Mean
5.30
5.40
5.58

107
Table 10continued
VERBAL INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.09
4.60
5.11
4.93
Middle
4.43
4.40
4.58
4.47
Late
4.14
4.23
4.06
4.14
Mean
4.55
4.41
4.59
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
6.09
6.08
6.08
6.08
Middle
5.65
6.27
5.85
5.93
Late
4.99
5.13
5.26
5.12
Mean
5.58
5.83
5.73
5.58

108
Table 11
Mean reading time (sec) for early, middle, and late sentence pairs
of affectively pleasant, neutral and unpleasant texts, with and without
startle
probes, for qood and poor
imaqerv ability
subiects
NO STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
4.97
4.50
4.87
4.78
Middle
4.58
4.36
4.67
4.54
Late
4.20
4.21
4.31
4.24
Mean
4.58
4.36
4.62
4.52
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.85
6.17
5.84
5.95
Middle
5.59
5.77
5.66
5.67
Late
4.90
5.06
5.20
5.05
Mean
5.44
5.67
5.57
5.56

109
Table 11 continued
STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.24
5.04
5.46
5.25
Middle
4.67
4.93
5.00
4.87
Late
4.50
4.66
4.46
4.54
Mean
4.80
4.88
4.97
4.89
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.90
5.69
6.19
5.92
Middle
5.48
6.03
5.88
5.80
Late
4.91
4.95
5.18
5.01
Mean
5.43
5.55
5.75
5.58

110
Table 12
Mean reading time (sec) for early, middle, and late sentence pairs
of affectively pleasant, neutral and unpleasant texts, with and without
startle probes, for subjects receiving verbal and imagery instructions
NO STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.74
5.49
5.73
5.65
Middle
5.50
5.24
5.56
5.43
Late
4.87
4.82
5.11
4.93
Mean
5.37 5.18 5.47
Verbal Instructions
5.34
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.07
5.18
4.98
5.08
Middle
4.67
4.89
4.76
4.78
Late
4.23
4.46
4.40
4.36
Mean
4.66
4.84
4.72
4.74

Table 12continued
111
STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.03
5.24
5.43
5.23
Middle
4.75
5.19
5.21
5.05
Late
4.52
4.71
4.73
4.65
Mean
4.77
5.04
5.12
4.98
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
6.11
5.49
6.21
5.94
Middle
5.41
5.78
5.67
5.62
Late
4.90
4.90
4.92
4.90
Mean
5.47
5.39
5.60
5.49

112
Table 13
Mean heart rate change (beats per minute) for early, middle, and late
sentence pairs of affectively pleasant, neutral and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1 .02
-1 .88
-1 .45
-1.45
Middle
-2.74
-3.11
-3.49
-3.11
Late
-3.02
-3.34
-3.69
-3.35
Mean
-2.26
-2.78
-2.88
Table 14
Mean heart rate
chancre (beats
per minute)
for earlv.
middle, and late
sentence pairs of affectively
pleasant, neutral and unpleasant texts
for
subjects with
qood and poor imaqerv
ability
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1 .61
-2.80
-1 .80
-2.07
Middle
-3.08
-3.73
-3.71
-3.51
Late
-3.37
-4.08
-3.89
-3.78
Mean
-2.69
-3.54
-3.13
-3.12
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.43
-0.95
-1.10
-0.83
Middle
-2.40
-2.48
-3.27
-2.72
Late
-2.67
-2.60
-3.48
-2.92
Mean
-1 .83
-2.01
-2.62
-2.16

113
Table 15
Mean heart rate change (beats per minute) for early, middle, and late
sentence pairs of affectively pleasant, neutral and unpleasant texts
for subjects receiving verbal and imagery instructions
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.41
CO
V£>
O
-0.61
-0.57
Middle
-1.98
-2.15
-2.44
-2.19
Late
-2.43
-2.17
-2.49
-2.36
Mean
-1.61 -1.67 -1.85
Verbal Instructions
-1 .71
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1 .62
-3.08
-2.30
-2.33
Middle
-3.51
-4.06
-4.54
-4.04
Late
-3.62
-4.50
-4.88
-4.33
Mean
-2.92
-3.88
-3.91
-3.57

114
Table 16
Mean heart rate
change (beats
per minute)
for early.
middle, and late
sentence pairs
of affectively
pleasant, neutral and unpleasant texts
with and
without startle probes
No
Startle Subjects
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1.11
-2.06
-1 .25
-1.47
Middle
-2.70
-3.09
-3.24
-3.01
Late
-2.92
-3.12
-3.56
-3.20
Mean
-2.24
-2.76
-2.68
-2.56
Startle Subjects
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.92
-1 .70
-1.66
-1 .43
Middle
-2.79
-3.12
-3.75
-3.22
Late
-3.13
-3.56
-3.81
-3.50
Mean
-2.28
-2.79
-3.07
-2.72

115
Table 17
Mean heart rate change (beats per minute) for early, middle, and late
sentence pairs of affectively pleasant, neutral and unpleasant texts
for good and poor imagers receiving verbal and imagery instructions
IMAGERY INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.26
-1.18
-0.19
-0.54
Middle
-1.27
-2.47
-1 .88
-1 .87
Late
-1.59
-2.47
-1.65
-1.90
Mean
-1 .04
-2.04
-1 .24
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.57
-0.17
-1 .03
-0.59
Middle
-2.69
-1 .84
-3.00
-2.51
Late
-3.26
-1 .88
-3.34
-2.83
Mean
-2.18
-1 .30
-2.45

116
Table 17continued
VERBAL INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-2.96
-4.43
-3.42
-3.60
Middle
-4.90
-4.99
-5.53
-5.14
Late
-5.16
-5.68
-6.14
-5.66
Mean
-4.34
-5.03
-5.03
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.28
-1.73
-1.17
-1 .06
Middle
-2.11
-3.12
-3.55
-2.93
Late
-2.09
-3.32
-3.63
-3.01
Mean
-1.49
-2.72
-2.78

117
Table 18
Mean heart rate change (beats per minute) for early, middle, and late
sentence pairs of affectively pleasant, neutral and unpleasant
texts, with and without startle probes, for good and poor imagers
NO STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1.20
-3.06
-1 .24
-1 .83
Middle
-2.38
-3.37
-3.12
-2.96
Late
-2.48
-3.83
-3.49
-3.27
Mean
-2.02
-3.42
-2.62
-2.69
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1 .03
-1.06
-1.25
-1.11
Middle
-3.01
-2.82
-3.35
-3.06
Late
-3.35
-2.41
-3.64
-3.13
Mean
-2.46
-2.10
-2.75
-2.46

118
Table 18--continued
STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-2.01
-2.55
-2.37
-2.31
Middle
-3.79
-4.10
-4.30
-4.06
Late
-4.26
-4.33
-4.30
-4.30
Mean
-3.35
-3.66
-3.65
-3.55
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
+0.17
-0.85
-0.94
-0.54
Middle
-1 .79
-2.14
-3.20
-2.38
Late
-2.00
-2.78
-3.32
-2.70
Mean
-1 .21
-1 .92
-2.49
-1.87

119
Table 19
Mean heart rate change (beats per minute) for early, middle, and late
sentence pairs of affectively pleasant, neutral and unpleasant texts.
with and without startle probes, for verbal and imagery instructions
NO STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
+ 0.66
-0.54
+0.26
+0.12
Middle
-1 .35
-2.22
-1.68
-1 .75
Late
-1 .50
-1 .66
-1 .92
-1.69
Mean
-0.73 -1.47 -1.12
Verbal Instructions
-1.10
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-2.88
-3.58
-2.75
-3.07
Middle
-4.04
-3.96
-4.79
-4.26
Late
-4.34
-4.58
-5.20
-4.71
Mean
-3.76
-4.04
-4.25
-4.01

120
Table 19continued
STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1.49
-0.81
-1 .47
-1 .26
Middle
-2.61
-2.08
-3.20
-2.63
Late
-3.36
-2.69
-3.06
-3.04
Mean
-2.48 -1.86 -2.58
Verbal Instructions
-2.31
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.35
-2.58
-1 .84
-1 .59
Middle
-2.97
-4.15
-4.29
-3.81
Late
-2.90
-4.42
-4.56
-3.96
Mean
-2.08
-3.72
-3.57
-3.12

121
Table 20
Mean heart rate change (beats per minute) for affectively pleasant,
neutral and unpleasant texts, with and without startle probes.
for good and poor imagers receiving verbal and imagery instructions
NO STARTLE SUBJECTS
Imagery Instructions
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
-0.28
1
u>
o
CD
-0.99
-1 .45
Poor
-1.18
+ 0.13
-1.24
-0.76
Mean
-0.73
-1 .47
-1.12
-1.10
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
-3.76
-3.75
-4.24
-3.92
Poor
-3.74
-4.33
-4.25
-4.11
Mean
-3.76
-4.04
-4.25
-4.01

122
Table 20--continued
STARTLE SUBJECTS
Imagery Instructions
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
-1 .79
o
o
-1.48
-1.43
Poor
-3.17
-2.73
-3.67
-3.19
Mean
-2.48
-1 .86
-2.58
-2.31
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
-4.91
-6.38
-5.83
-5.68
Poor
+ 0.76
-1.12
-1.31
-0.56
Mean
-2.08
-3.72
-3.57
-3.12

123
Table 21
Mean startle magnitude (arbitrary A-D units) for probes presented
early and late in affectively pleasant, neutral, and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
114.6
133.9
111.2
119.9
Late
134.0
149.8
165.1
149.6
Mean
124.3
141.9
138.2
134.8
Table 22
Mean startle
magnitude (arbitrary A-D
units) for probes presented
early and late
in
affectively pleasant.
neutral, and unpleasant texts
for
good and
poor imagery
ability subjects
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
82.0
62.8
102.8
82.5
Late
101.1
125.2
133.3
119.9
Mean
91 .6
94.0
118.1
101.2
Pleasant
Startle Position
Early 147.2
Late 166.9
157.1
Poor Imagers
Neutral
Unpleasant
Mean
204.9
119.6
157.2
174.4
196.9
179.4
189.6
158.3
168.3
Mean

124
Table 23
Mean startle magnitude (arbitrary A-D units) for probes presented
early and late in affectively pleasant, neutral, and unpleasant texts
for good and poor imagery ability subjects
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
158.7
190.6
142.0
163.8
Late
167.2
195.7
200.7
187.9
Mean
163.0
193.2
171.4
175.9
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
70.5
77.2
80.4
76.0
Late
100.7
103.9
129.6
111.4
Mean
85.6
90.6
105.0
93.7
Table 24
Mean startle
latency (msec)
for probes
: presented early and
in affectively pleasant, neutral.
and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
55.3
51 .2
55.3
53.9
Late
55.1
54.6
50.2
53.3
Mean
55.2
52.9
52.7
53.6

125
Table 25
Mean startle
latency (msec) for probes
presented
early and late
in affectively pleasant, neutral.
and unpleasant texts
for qood and
poor imaqerv ability subiects
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
59.2
53.5
56.5
56.4
Late
55.0
51.6
52.0
52.9
Mean
57.1
52.6
54.2
54.6
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
51 .4
48.9
54.0
51 .4
Late
55.2
57.6
48.5
53.8
Mean
53.3
53.2
51.3
52.6
Table 26
Mean startle
latency (msec) for probes
presented early and late
in affectively pleasant, neutral.
and unpleasant texts
for subiects receivinq verbal and
imaqery instructions
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
54.6
53.1
49.3
52.3
Late
50.6
49.6
47.0
49.1
Mean
52.6
51 .4
48.2
50.7

126
Table 26--continued
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Startle Position
Early
56.0
49.3
61 .2
55.5
Late
59.6
59.6
53.4
57.6
Mean
57.8
54.4
57.3
56.5
Table 27
Mean recognition
reaction time
(msec) for
study word and
new word
targets for affectively oleasant. neutral.
and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Target Type
Study
857
881
853
864
New
868
887
887
877
Mean
862
884
865
870

127
Table 28
Mean recognition reaction time (msec) for study word and new word
targets for affectively pleasant, neutral, and unpleasant texts
for good and poor imagery ability subjects
Study Word Targets
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
831
856
824
837
Poor
883
906
882
890
Mean
857
881
853
864
New Word Targets
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
856
876
865
866
Poor
879
898
889
889
Mean
868
887
877
877

128
Table 29
Mean recognition reaction time (msec) for study word and new
word targets for affectively pleasant, neutral, and unpleasant texts
for subjects receiving imagery and verbal instructions
Study Word Targets
Pleasant Neutral Unpleasant Mean
Instructions
Imagery
830
871
840
847
Verbal
884
892
865
880
Mean
857
882
852
864
New Word Targets
Pleasant
Neutral
Unpleasant
Mean
Instructions
Imagery
840
859
845
848
Verbal
895
914
909
906
Mean
867
887
877
877

129
Table 30
Mean recognition reaction time (msec) for study word and new word
targets for affectively pleasant, neutral, and unpleasant texts
with and without startle probes
New Word Targets
Pleasant
Neutral
Unpleasant
Mean
Startle Group
No Startle
857
883
868
869
Startle
878
891
886
885
Mean
868
887
877
877
Pleasant
Study Word Targets
Neutral Unpleasant
Mean
Startle Group
No Startle
863
906
869
879
Startle
851
857
836
848
Mean
857
882
852
864

130
Table 31
Mean recognition reaction time (msec) for study word and new word
targets for affectively pleasant, neutral, and unpleasant texts
for good and poor imagers receiving imagery and verbal instructions
NEW WORD TARGETS
Imagery Instructions
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
848
858
839
848
Poor
833
860
851
848
Mean
840
859
845
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
864
894
891
883
Poor
925
935
927
929
895 914 909
Mean

131
Table 31continued
STUDY WORD TARGETS
Imagery Instructions
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
805
890
819
838
Poor
854
851
861
856
Mean
830
871
840
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
856
823
829
836
Poor
912
961
902
925
Mean
884
892
865

132
Table 32
Mean recognition reaction time (msec) for study word targets
primed by a study word, a new word, or XXXXX
for affectively pleasant, neutral, or unpleasant texts
Affective Category
Pleasant Neutral Unpleasant Mean
Prime Type
Study Word
832
849
859
847
New Word
845
907
825
859
XXXXX
893
888
874
885
N-S Priming
+ 13
+58
-34
+ 9
X-S Priming
+61
+ 39
+ 15
+ 38

133
Table 33
primed bv a
study word.
a new word.
or XXXXX for affectively
leasant, neutral, or unpleasant texts
; for good and
poor imaqers
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
808
813
830
817
New Word
808
894
799
834
xxxxx
876
862
842
860
N-S Priming
0
+81
-31
+ 17
X-S Priming
+ 68
+ 49
+ 12
+ 43
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
857
884
889
877
New Word
882
920
850
884
XXXXX
911
914
906
910
N-S Priming
+25
+ 36
-39
+ 7
X-S Priming
+ 54
+ 30
+ 17
+ 34

134
Table 34
Mean recognition
reaction
time (msec) for studv word
. targets primed bv
a studv word, a
new word
, or XXXXX for
affectively
pleasant, neutral.
or unpleasant texts for imagery
and verbal instructions
Imagery Instructions
Pleasant Neutral
Unpleasant
Mean
Prime Type
Study Word
810
863
850
841
New Word
830
897
791
839
xxxxx
849
852
879
860
N-S Priming
+20
+ 34
-59
- 2
X-S Priming
+ 39
-11
+29
+ 20
Verbal Instructions
Pleasant Neutral
Unpleasant
Mean
Prime Type
Study Word
855
834
869
853
New Word
860
917
859
878
XXXXX
938
924
869
911
N-S Priming
+ 5
+83
-10
+26
X-S Priming
+83
+90
0
+ 57

135
Table 35
Mean recognition reaction time (msec) for study words primed by
a study word, a new word, or XXXXX for affectively pleasant, neutral.
or unpleasant texts, with and without startle probes
Type
No
Pleasant
Startle
Neutral
Subjects
Unpleasant
Mean
Study Word
844
876
870
863
New Word
838
933
842
871
XXXXX
906
908
896
903
N-S Priming
- 6
+ 57
-28
+ 17
X-S Priming
+62
+ 32
+ 26
+ 40
Startle Subjects
Pleasant Neutral Unpleasant Mean
Prime Type
Study Word
821
821
849
830
New Word
852
881
807
847
XXXXX
881
868
852
867
N-S Priming
+ 31
+ 60
-42
+ 17
X-S Priming
+ 47
+61
+ 3
+ 37

136
Table 36
Mean recognition RT (msec) for study words primed by a study word, a new
word, or XXXXX for affectively pleasant, neutral, or unpleasant texts.
with and without startle probes, for good and poor imagers
NO STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
828
863
859
850
New Word
818
941
838
866
XXXXX
942
909
899
917
N-S Priming
-10
+ 78
-21
+ 16
X-S Priming
+ 114
+ 46
+ 40
+ 67
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
861
890
880
877
New Word
858
924
846
876
XXXXX
869
907
893
890
N-S Priming
- 3
+ 34
-34
- 1
X-S Priming
+ 8
+ 17
+ 13
+ 13

137
Table 36continued
STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
787
763
800
784
New Word
797
847
760
801
XXXXX
810
815
785
803
N-S Priming
+ 10
+84
-40
+ 18
X-S Priming
+ 23
+52
-15
+20
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
854
878
897
876
New Word
907
916
855
892
XXXXX
952
921
920
931
N-S Priming
+ 53
+ 38
-42
+ 17
X-S Priming
+98
+ 43
+ 23
+ 55

138
Table 37
Mean recognition RT (msec) for study words primed by a study word, a new
word, or XXXXX for affectively pleasant, neutral, or unpleasant texts,
with and without startle probes, for imagery and verbal instructions
NO STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
827
911
858
865
New Word
821
936
808
855
XXXXX
877
872
908
886
N-S Priming
- 6
+25
-50
-10
X-S Priming
+ 50
-39
+ 50
+21
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
862
841
882
862
New Word
856
930
876
887
XXXXX
934
944
883
920
N-S Priming
- 6
+89
- 6
+ 25
X-S Priming
+ 72
+ 103
+ 1
+ 59

139
Table 37continued
STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
793
815
842
816
New Word
840
859
774
824
XXXXX
820
831
850
834
N-S Priming
+ 47
+ 44
-68
+ 7
X-S Priming
+ 27
+ 16
+ 8
+ 17
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
848
827
856
844
New Word
864
904
841
870
XXXXX
942
905
855
901
N-S Priming
+ 16
+ 77
-15
+ 26
X-S Priming
+ 94
+ 78
- 1
+ 57

140
Table 38
Mean recognition RT (msec) for study words primed by a study word, a new
word, or XXXXX for affectively pleasant, neutral, or unpleasant texts.
for good and poor imagers receiving verbal and imagery instructions
IMAGERY INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
778
863
827
823
New Word
791
939
782
837
XXXXX
845
868
847
853
Mean
805
890
819
838
N-S Priming
+ 13
+76
-45
X-S Priming
+ 67
+ 5
+ 20
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
842
863
872
859
New Word
869
855
800
841
XXXXX
852
836
912
867
Mean
854
851
861
856
N-S Priming
+ 27
- 8
-72
X-S Priming
+ 10
-27
+ 40

141
Table 38continued
VERBAL INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
837
763
832
811
New Word
824
849
817
830
XXXXX
907
856
837
867
Mean
856
823
829
836
N-S Priming
-13
+86
-15
X-S Priming
+ 70
+93
+ 5
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
872
905
905
894
New Word
896
984
900
927
XXXXX
969
992
901
954
Mean
912
960
902
925
N-S Priming
+24
+ 79
- 5
X-S Priming
+ 97
+87
- 4

142
Table 39
Mean percent errors for new word and study word targets,
for affectively pleasant, neutral, and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Target Type
New
3.68
3.83
4.29
3.93
Study
7.85
8.99
8.35
8.40
Mean
5.77
6.41
6.32
6.17
Table 40
Mean percent errors
for new word
and study word
targets
for affectively pleasant, neutral
, and unpleasant texts
for
qood and
poor imaaerv
ability subjects
New Word Targets
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
4.19
4.12
4.94
4.42
Poor
3.18
3.54
3.64
3.45
Mean
3.68
3.83
4.29
3.93
Study Word Targets
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
8.18
8.82
8.83
8.61
Poor
7.52
9.15
7.87
8.18
Mean
7.85
9.00
8.35
8.40

143
Table 41
Mean percent errors for new word and study word targets,
for affectively pleasant, neutral, and unpleasant texts.
for subjects receiving imagery and verbal instructions
New Word Targets
Pleasant Neutral Unpleasant Mean
Instructions
Imagery
4.61
5.08
4.43
4.71
Verbal
2.75
2.58
4.15
3.16
Mean
3.68
3.83
4.29
3.93
Study Word
Targets
Pleasant
Neutral
Unpleasant
Mean
Instructions
Imagery
8.65
6.79
8.00
7.81
Verbal
7.05
11.19
8.70
8.98
Mean
7.85
9.00
8.45
8.40

144
Table 42
Mean percent errors for new word and study word targets,
for affectively pleasant, neutral, and unpleasant texts.
with and without presentation of startle probes
New Word Targets
Pleasant
Neutral
Unpleasant
Mean
Startle Group
No Startle
3.47
3.56
3.47
3.50
Startle
3.90
4.10
5.11
4.37
Mean
3.68
3.83
4.29
3.93
Pleasant
Study Word
Neutral
Targets
Unpleasant
Mean
Startle Group
No Startle
7.25
7.71
7.88
7.61
Startle
8.45
10.27
8.82
9.18
Mean
7.85
9.00
8.35
8.40

145
Table 43
Mean percent errors for new word and study word targets.
for affectively pleasant, neutral, and unpleasant texts,
for good and poor imagers, receiving imagery and verbal instructions
NEW WORD TARGETS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
5.38
5.96
4.43
5.25
Poor
3.85
4.20
4.43
4.16
Mean
4.61 5.08 4.43
Verbal Instructions
4.71
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
3.00
2.29
5.46
3.58
Poor
2.51
2.87
2.84
2.74
Mean
2.75
2.58
4.15
3.16

146
Table 43continued
STUDY WORD TARGETS
Imagery Instructions
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
9.88
7.00
9.18
8.69
Poor
7.42
6.58
6.82
6.94
Mean
8.65
6.79
8.00
7.81
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
6.48
10.64
8.49
8.54
Poor
7.62
11.73
8.92
9.42
Mean
7.05
11.19
8.70
8.98

147
Table 44
Mean percent errors for studv
words primed bv a studv
word, a new word.
or XXXXX for affectively pleasant.
neutral, or unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Prime Type
Study Word
6.81
5.05
6.83
6.23
New Word
7.92
9.00
8.22
8.38
XXXXX
10.32
12.91
8.49
10.58
Mean
8.35
8.99
7.85
8.40
Table
45
Mean percent errors for studv
words primed bv a studv
word, a new word.
or XXXXX for
affectively pleasant.
neutral, or unpleasant texts,
for qood and poor imaqers
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Prime Type
Study Word
7.44
5.28
8.47
7.06
New Word
8.04
7.41
8.17
7.87
XXXXX
9.06
13.78
9.86
10.90
Mean
8.18
8.82
8.83
8.61
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Prime Type
Study Word
6.22
4.82
5.15
5.40
New Word
8.41
10.59
7.67
8.89
XXXXX
7.92
12.04
10.79
10.25
Mean
7.52
9.15
7.87
8.18

148
Table 46
Mean percent errors for study words primed by a study word, a new word,
or XXXXX for affectively pleasant, neutral, or unpleasant texts for
subjects receiving imagery and verbal instructions
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
7.15
6.34
5.52
6.33
New Word
9.96
6.85
9.31
8.70
XXXXX
8.84
7.18
9.18
8.40
Mean
8.65
6.79
8.00
7.81
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
6.51
3.76
8.11
6.13
New Word
6.49
11.15
6.53
8.06
XXXXX
8.14
18.64
11.47
12.75
Mean
7.05
11.19
8.70
8.98

149
Table 47
Mean percent errors for study words primed by a study word, a new word,
or XXXXX for affectively pleasant, neutral, or unpleasant texts,
with and without startle probes
No Startle Subjects
Prime Type
Pleasant
Neutral
Unpleasant
Mean
Study Word
5.36
5.46
5.42
5.41
New Word
9.21
9.05
8.53
8.93
XXXXX
7.17
8.62
9.71
8.50
Mean
7.25
7.71
7.88
7.61
Startle Subjects
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
8.30
4.65
8.20
7.05
New Word
7.24
8.95
7.31
7.83
XXXXX
9.82
17.20
10.94
12.65
Mean
8.45
10.27
8.45
9.18

150
Table 48
Mean percent errors for study words primed by a study word, a new word,
or XXXXX for affectively pleasant, neutral, or unpleasant texts,
for good and poor imagers receiving verbal and imagery instructions
IMAGERY INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
9.51
8.18
8.37
8.69
New Word
9.29
7.20
11.01
9.17
XXXXX
10.84
5.62
8.15
8.20
Mean
9.88
7.00
9.18
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
4.78
4.50
2.66
3.98
New Word
10.63
6.50
7.60
8.24
XXXXX
6.84
8.74
10.22
8.60
Mean
7.42
6.58
6.82

151
Table 48continued
VERBAL INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
5.36
2.38
8.56
5.44
New Word
6.80
7.62
5.33
6.58
xxxxx
7.28
21 .93
11.57
13.60
Mean
6.48
10.64
8.49
Poor Imagers
Pleasant Neutral Unpleasant Mean
Prime Type
Study Word
7.66
5.15
7.65
6.82
New Word
6.18
14.69
7.74
9.54
XXXXX
9.01
15.35
11.36
11 .90
Mean
7.62
11.73
8.92

152
Table 49
Mean percent errors for study words primed by a study word, a new word,
or XXXXX for affectively pleasant, neutral, or unpleasant texts.
with and without startle probes, for good and poor imagers
NO STARTLE SUBJECTS
Good Imagers
Pleasant Neutral Unpleasant Mean
Prime Type
Study Word
6.34
7.66
7.67
7.22
New Word
6.95
9.19
9.52
8.56
XXXXX
7.39
5.88
8.16
7.14
Mean
6.89
7.58
8.45
7.64
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
4.38
3.26
3.17
3.60
New Word
11.35
8.90
7.53
9.30
XXXXX
6.95
11.35
11.25
9.85
Mean
7.60
7.84
7.32
7.58

153
Table 49continued
STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
8.53
2.90
9.26
6.90
New Word
9.14
5.62
6.82
7.19
xxxxx
10.74
21.67
11.56
14.66
Mean
9.47
10.07
9.21
9.58
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
8.07
6.39
7.14
7.20
New Word
5.35
12.28
7.80
8.48
xxxxx
8.90
12.74
10.32
10.65
Mean
7.44
10.47
8.42
8.78

154
Table 50
Mean percent errors for study words primed by a study word, a new word,
or XXXXX for affectively pleasant, neutral, or unpleasant texts,
with and without startle probes, for imagery and verbal instructions
NO STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
3.36
6.17
4.32
4.62
New Word
10.89
4.67
9.67
8.41
XXXXX
5.51
5.82
8.29
6.54
Mean
6.59
5.55
7.43
6.52
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
7.36
4.75
6.52
6.21
New Word
7.52
13.42
7.38
9.44
XXXXX
8.83
11.41
11.12
10.45
Mean
7.91
9.86
8.34
8.70

155
Table 50continued
STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Prime Type
Study Word
10.93
6.51
6.71
8.05
New Word
9.03
9.02
8.94
9.00
xxxxx
12.17
8.53
10.07
10.26
Mean
10.71
8.02
8.58
9.10
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Prime Type
Study Word
5.67
2.78
9.69
6.05
New Word
5.45
8.88
5.68
6.67
XXXXX
7.46
25.87
11.81
15.05
Mean
9.06
12.51
6.19
9.26
Table 51
Mean proportions
of affectively pleasant
. neutral. and unpleasant
recalled at
the end of
the experiment
Affective Category
Pleasant
Neutral
Unpleasant
Mean
All subjects
.41
.42
.29
.37

APPENDIX E
AFFECTIVE RATINGS FOR PARAGRAPHS
Affective Ratings
Ratings provided subjects in the present study for the valence,
arousal, and dominance scales were subjected to an analysis of variance
in order to determine whether any of the individual difference or
instructional variables systematically influenced ratings. The factors
included in the 2X2X3 ANOVA were QMI (good, poor) X Instructions
(verbal, imagery) X Valence (pleasant, neutral, unpleasant).
Valence Ratings
Table 52 summarizes mean ratings for texts classified as pleasant,
neutral, and unpleasant on the SAM valence, arousal, and dominance
scales. A significant effect was obtained for valence ratings used to
classify affective texts in the omnibus test, F (2,94) = 3695.67,
p < .0001.
TABLE 52
Mean valence, arousal, and control ratings for affective texts
Rating Scale
Pleasant
Neutral
Unpleasant
Valence
8.29
4.94
1 .86
Arousal
6.25
4.58
7.46
Dominance
7.01
4.95
2.42
156

157
Subjects with poor imagery ability rated unpleasant paragraphs as
less unpleasant and pleasant paragraphs as less pleasant than subjects
with good imagery ability, yielding an Imagery Ability X Valence
interaction: F (1,47) = 8.51, p < .01 (See Table 53).
TABLE 53
Mean valence ratings for pleasant, neutral, and unpleasant texts
for subjects reporting good and poor imagery ability
Pleasant
Neutral
Unpleasant
Valence Ratings
Good
8.4
5.0
1 .7
Poor
8.2
4.9
2.0
Arousal Ratings
Arousal ratings also differed reliably among affective categories
F(2,94) = 98.49, p < .0001. Affectively unpleasant texts were
significantly more arousing than pleasant texts, F (1,47) = 88.40,
p < .0001, and pleasant and unpleasant texts received higher arousal
ratings than neutral texts, F (1,47) = 100.99, p < .0001.
Imagery ability also affected arousal ratings. Table 54
summarizes mean arousal ratings for affectively negative, neutral, and
positive texts for subjects reporting good and poor imagery ability.
Good imagers rated texts as more arousing than poor imagers, F (1,47) =
5.35, p < .03.

158
TABLE 54
Mean arousal ratings for pleasant, neutral, and unpleasant texts
for subjects reporting good and poor imagery ability
Pleasant
Neutral
Unpleasant
Arousal Ratings
Good
6.7
4.6
7.8
Poor
5.8
4.5
7.2
Dominance Ratings
Mean dominance ratings are summarized in Table 55 for affectively
pleasant, neutral, and unpleasant texts. Dominance ratings were highest
for pleasant texts and decreased for texts rated as more unpleasant,
yielding a significant overall effect of valence, F (2,94) = 347.39,
p < .0001. While imagery ability did not affect dominance ratings for
affectively pleasant or unpleasant texts, good imagers gave higher
dominance ratings than poor imagers for neutral paragraphs in the verbal
instructions condition (Good = 5.5; Poor = 4.3). However, in the
imagery instructions condition, poor imagers gave higher dominance
ratings than good imagers for neutral texts (Good = 4.9; Poor = 5.2),
resulting in a significant Valence X Instructions X Imagery Ability
interaction: F (2,94) = 3.63; p < .04.
Stability of text ratings between a priori and subject classification:
A contingency table was calculated in order to determine how many
texts changed affective classification, and which affective categories
showed greater change, comparing the previous validation study to the
present experiment (see Table 56).

159
TABLE 55
Mean dominance ratings for pleasant, neutral, and unpleasant texts
for subjects with good and poor imagery ability
receiving verbal and imagery instructions
Pleasant Neutral Unpleasant
Instructions
Good
Poor
Good
Poor
Good
Poor
Imagery
7.4
6.7
4.9
5.2
2.5
2.6
Verbal
7.1
6.6
5.5
4.3
2.2
2.4
TABLE 56
Contingency table comparing affective classification of texts
based on a priori (pilot-based) and experimental subject rating data
A PRIORI RATINGS
SUBJECT RATINGS
Pleasant
Neutral
Unpleasant
Total
Pleasant
474
267
36
777
Neutral
33
213
144
390
Unpleasant
9
18
858
885
Total
1038
498
516
2052
Numbers of texts: (57 subjects X 36 texts = 2052 total ratings)
The number of texts that remained in the a priori affective
categories according to ratings by the present sample appears on the
diagonal of this table. The number of texts that changed categories for
the present sample appears in the rows. For example, of the texts
classified as affectively unpleasant in the validation study, 858

160
ratings remained in this category for the present sample, while in 144
instances the texts were rated as neutral, and a priori unpleasant texts
were rated as pleasant in 36 instances by the present sample.
Examination of this table reveals that while the affective classifi
cations of text content were relatively more stable across samples for
texts rated as pleasant or unpleasant, ratings for neutral texts were
less stable and tended to move to the affectively pleasant category.

APPENDIX F
SELF-ASSESSMENT MANIKIN (SAM)

Figure 4. Self-Assessment Manikin (SAM) used to obtain affective ratings of texts.

163

APPENDIX G
SUMMARY OF RESULTS FOR PILOT PRIMING STUDIES
The two pilot studies completed prior to the present experiment
suggested that affective content altered the episodic priming effects
demonstrated by Ratcliff and McKoon. The design and results of the two
studies are briefly summarized below:
Study 1
The initial study was adapted from a priming paradigm utilized by
Ratcliff and McKoon (1981), in which subjects read sets of two three-
line paragraphs followed by a test list of prime-target pairs. Subjects
responded to the target word in prime-target pairs, and recognition RT
was compared for study words primed by a word from the same paragraph,
compared to RT for study words primed by a word from the other
paragraph.
Ratcliff and McKoon's paradigm was replicated as closely as
possible. A copy of the texts and test lists was obtained from those
authors and their methods of presentation for these materials were
followed as closely as possible, with the exception that in some cases
emotionally evocative paragraphs were substituted for their texts.
Study 2
Another paradigm developed by Ratcliff and McKoon (1978) was
adapted for the second study. In this paradigm, episodic priming was
164

165
used to demonstrate logical distance between propositions in text.
Subjects read 6-line paragraphs and then responded to a test list of
prime-target pairs. Recognition RT for study words primed by another
study word from the same proposition was compared to RT for study words
primed by another study word taken from a different proposition in the
paragraph. Ratcliff and McKoon's design and methods were again closely
replicated in our adaptation of their study. However, rather than
combine their textual materials with affective texts developed in our
laboratory, paragraphs were employed that had been developed in a
previous validation study (Spence, Bradley, & Lang, 1985). These
materials were the same as those employed in the present experiment.
Consistent with the results of Study 1, recognition RT was slower
for study-study prime-target pairs from affectively negative texts
compared to affectively positive and neutral texts. No differential
priming was found between near and far propositions. In a free recall
task similar to that carried out in the dissertation experiment,
subjects were more likely to recall affectively negative texts than
positive or neutral texts.
The results of the present dissertation experiment are consistent
with these earlier studies in that recognition RT was significantly
slower for study-study prime-target pairs from affectively negative
compared to positive and neutral texts.

REFERENCES
Anderson, J. R. (1976). Language, memory, and thought. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Anderson, J. R. (1983). The architecture of cognition. Cambridge, MA:
Harvard University Press.
Anderson, J. R., & Bower, G. H. (1973). Human associative memory.
Washington, D.C.: Hemisphere Press.
Anthony, B. J. (1985). In the blink of an eye: Implications of reflex
modification for information processing. In P. K. Ackels, J. R.
Jennings, & M. G. H. Coles (Eds.), Advances in psychophysiology
(Vol. 1, pp. 167-218). Greenwich, CT: JAI Press.
Balaban, M. T., Losito, B., Simons, R. F., & Graham, F. K. (1986).
Offline latency and amplitude scoring of the human reflex eyeblink
with Fortran IV (computer program abstract). Psychophysiology.
23, 612.
Berg, W. K., & Davis, M. (1984). Diazepam blocks fear-enhanced startle
elicted electrically from the brainstem. Physiology and Behavior.
32, 333-336.
Bower, G. H., Gilligan, S. G., & Monteiro, K. P. (1981). Selectivity
of learning caused by affective states. Journal of Experimental
Psychology: General. 110. 451-473.
Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1990). Startle reflex
modification: Emotion or attention? Psychophysiology, 27, 513
522.
Bradley, M. M., Greenwald, M. K., Petry, M. C., & Lang, P. j. (1991).
Remembering pictures: Pleasure and arousal in memory.
Unpublished manuscript, University of Florida.
Bradley, M. M., Spence, E. L., & Lang, P. J. (1992). Retrieval effects
for affective texts. Unpublished manuscript, University of
Florida.
Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition. 5, 73-
99.
166

167
Carroll, M., & Kirsner, K. (1982). Context and repetition effects in
lexical decision and recognition memory. Journal of Verbal
Learning and Verbal Behavior. 21. 55-69.
Christianson, S.-, & Loftus, E. F. (1987). Memory for traumatic
events. Applied cognitive psychology. 1, 225-239.
Cirillo, R. K., & Foss, D. J. (1980). Text structure and reading time
for sentences. Journal of Verbal Learning and Verbal Behavior.
19, 96-109.
Cook, E. W., Ill, Hawk, L. W., Davis, T. D., & Stevenson, V. E. (1991).
Affective individual differences and startle reflex modulation.
Journal of Abnormal Psychology, 100. 5-13.
Cook, E. W., Ill, Melamed, B. G., Cuthbert, B. N., McNeil, D. W., Lang,
P. J. (1988). Emotional imagery and the differential diagnosis
of anxiety. Journal of Consulting and Clinical Psychology. 56.
734-740.
Cook, E. W., Ill, Davis, T. L., Hawk, L. W., Jr., Spence, E. L., &
Gautier, C. H. (in press). Fearfulness and startle potentiation
during aversive visual stimuli. Psychophysiology.
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A
framework for memory research. Journal of Verbal Learning and
Verbal Behavior. 11. 671-684.
Davis, M. (1984). The mammalian startle response. In R. C. Eaton
(Ed.) Neural mechanisms of startle behavior (pp. 287-351). New
York: Plenum Press.
Galambos, J. A., & Rips, L. J. (1982). Memory for routines. Journal
of Verbal Learning and Verbal Memory, 21. 260-281 .
Graham, F. K. (1979). Distinguishing among orienting, defense, and
startle reflexes. In H. D. Kimmel, E. H. van Olst, & J. F
Orlebeke (Eds.). The orienting reflex in humans (pp. 137-167).
Hillsdale, NJ: Erlbaum.
Graham, F. K., & Clifton, R. K. (1966). Heart-rate change as a
component of the orienting response. Psychological Bulletin. 65,
305-320.
Hamm, A. 0., Globisch, J., Cuthbert, B. N., & Vaitl, D. (1991). Startle
reflex modulation in simple phobics and normals. Psycho
physiology. 27, S37.
Hebb, D. 0. (1968). Concerning imagery. Psychological Review. 75.
466-477.

168
Heuer, F., & Reisberg, D. (1990). Vivid memories of emotional events:
The accuracy of remembered minutiae. Memory and Cognition. 18.
496-506.
Jacobson, E. (1930). Psychology and the integrative action of the
nervous system. Acta Symbolica. 1, 31-35.
Jacoby, L. L. (1982). Knowing and remembering: Some parallels in the
behavior of Korsakoff patients and normals. In L. S. Cermak
(Ed.), Human memory and amnesia (pp. 97-122). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Jonides, J., & Mack, R. (1984). On the cost and benefit of cost and
benefit. Psychological Bulletin. 96, 29-44.
Kieras, D. (1978). Beyond pictures and words: Alternate information
processing models for imagery effects in verbal memory.
Psychological Bulletin. 85, 532-554.
Kintsch, W. (1974). The representation of meaning in memory.
Hillsdale, NJ: Lawrence Erlbaum Associates.
Kintsch, W., & van Dijk, T. A. (1978). Toward a model of text
comprehension and production. Psychological Review. 85, 363-394.
Klorman, R., Weisberg, R. P., & Wiesenfeld, A. R. (1977). Individual
differences in fear and autonomic reactions to affective
stimulation. Psychophysiology. 14. 45-51.
Klorman, R., Wiesenfeld, A. R., & Austin, M. L. (1975). Autonomic
responses to affective visual stimuli. Psychophysiology. 12. 553-
560.
Lacey, J. I., & Lacey, B. C. (1974). Studies of heart rate and other
bodily processes in sensorimotor behavior. In P. A. Obrist, A. H.
Black, J. Brener, & L. V. DiCara (Eds.), Cardiovascular psycho
physiology (pp. 538-564). Chicago: Aldine.
Lang, P. J. (1964). Experimental studies of desensitization in
psychotherapy. In J. Wolpe (Ed.), The conditioning therapies.
New York: Holt, Rinehart, & Winston.
Lang, P. J. (1977). Psychophysiological assessment of anxiety and
fear. In J. D. Cone & R. P. Hawkins (Eds.), Behavioral
assessment: New directions in clinical psychology. New York:
Bruner-Mazel.
Lang, P. J. (1979). A bio-informational theory of emotional imagery.
Psychophysiology. 16. 495-512.

169
Lang, P. J. (1984). Cognition in emotion: Concept and action. In C.
Izard, J. Kagan, & R. Zajonc (Eds.), Emotion, cognition, and
behavior (pp. 192-228). New York: Cambridge University Press.
Lang, P. J. (1985). The cognitive psychophysiology of emotion: fear
and anxiety. In A. H. Turna & J. D. Maser (Eds.), Anxiety and the
anxiety disorders (pp. 131-170). Hillsdale, NJ: Lawrence Erlbaum
Associates.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1990). Emotion,
attention, and the startle reflex. Psychological Review, 97, 377-
395.
Lang, P. J., Bradley, M. M., Cuthbert, B. N., & Patrick, C. J. (In
press). Emotion and psychopathology: A startle probe analysis.
In L. Chapman & D. Fowles (Eds.), Progress in experimental
personality and psychopathology research: Models and methods of
psychopathology. New York: Springer.
Lang, P. J., Kozak, M. J., Miller, G. A., Levin, D. N., & McLean, A.,
Jr. (1980). Emotional imagery: Conceptual structure and pattern
of somato-visceral response. Psychophysiology, 17. 179-192.
Lang, P. J., Levin, D. N., Miller, G. A., Kozak, M. J. (1983). Fear
behavior, fear imagery, and the psychophysiology of emotion: The
problem of affective response integration. Journal of Abnormal
Psychology, 92, 276-306.
Lang, P. J., Ohman, A., & Simons, R. F. (1978). The psychophysiology
of anticipation. In J. Requin (Ed.), Attention and performance
(Vol. 7). New York: Academic Press.
Lusebrink, V. B., & McGuigan, F. J. (1989). Psychophysiological
components of imagery. Pavlovian Journal of Biological Science,
24, 58-62.
McGuigan, F. J., & Schoonover, R. A. (1973). Psychophysiology of
thinking: Studies of covert processes. New York: Academic Press.
McKoon, G. & Ratcliff, R. (1980). Priming in item recognition: The
organization of propositions in memory for text. Journal of
Verbal Learning and Verbal Behavior, 19, 369-386.
Mehrabian, A. & Russell, J. A. (1974). An approach to environmental
psychology. Cambridge, MA: MIT Press.
Miller, G. A., Levin, D. N., Kozak, M. J., Cook, E. W. Ill, McLean, A.,
& Lang, P.J. (1987). Individual differences in imagery and the
psychophysiology of emotion. Cognition and Emotion. 1, 367-390.
Obrist, P. A. (1981). Cardiovascular psychophysiology: A perspective.
New York: Plenum Press.

170
Osgood, C., Suci, G., & Tannenbaum, P. (1957). The measurement of
meaning. Urbana, IL: University of Illinois Press.
Paivio, A. (1971). Imagery and verbal processes. New York: Holt,
Rinehart, & Winston.
Pylyshyn, Z. (1973). What the mind's eye tells the mind's brain: A
critique of mental imagery. Psychological Bulletin. 80, 1-22.
Ratcliff, R., & McKoon, G. (1978). Priming in item recognition:
Evidence for the propositional structure of sentences. Journal of
Verbal Learning and Verbal Behavior. 17. 403-417.
Ratcliff, R., & McKoon, G. (1981). Automatic and strategic priming in
recognition. Journal of Verbal Learning and Verbal Behavior, 20.
204-215.
Schank, R. C. (1970). A conceptual dependency representation for a
computer-oriented semantics. (Doctoral dissertation, University
of Texas, 1970). Dissertation Abstracts International. 30,
2994A-2995A.
Sheehan, P. W. (1967). A shortened form of Betts' questionnaire upon
mental imagery. Journal of Clinical Psychology. 223, 380-389.
Simons, R. F., & Zelson, M. F. (1985). Engaging visual stimuli and
reflex blink modification. Psychophysiology, 22, 44-49.
Sokolov, Y. N. (1963). Perception and the conditioned reflex (S. W.
Waydenfeld, Trans.). New York: Macmillan. (Original work
published 1958).
Spence, E. L., Bradley, M., & Lang, P. J. (1985). Imagery ability and
psychophysiological response to affective texts. Psycho
physiology. 22, 613. Abstract.
Strosahl, K. D. & Ascough, J. C. (1981) Clinical uses of mental
imagery: Experimental foundations, theoretical misconceptions,
and research issues. Psychological Bulletin, 89, 422-438.
Vrana, S. R., & Lang, P. J. (1990). Fear imagery and the startle-probe
reflex. Journal of Abnormal Psychology. 99, 189-197.
Vrana, S. R., Spence, E. L., & Lang, P. J. (1988). The startle probe
response: A new measure of emotion? Journal of Abnormal
Psychology. 97. 487-491.
Yerkes, R. M., & Dodson, J. D. (1908). The relation of strength of
stimulus to rapidity of habit-formation. Journal of Comparative
and Neurology and Psychology. 18. 459-482.

BIOGRAPHICAL SKETCH
I completed my undergraduate education at the Ohio State
University in 1968, where my major field of study was English
Literature, my minor was French Language and Literature. I was admitted
to the masters program in general/experimental psychology at the
California State University, Northridge, in 1978. My masters research
was a psychophysiological study of the relationship between subjects'
electrodermal lability and their ability to control their heart rate
with biofeedback training.
I was admitted to the graduate program in the Department of
Clinical and Health Psychology at the University of Florida in 1982.
During my graduate training I served as research assistant in Dr. Lang's
psychophysiology laboratory. In this setting I gained experience in the
design, execution, and analysis of a series of experiments which were
part of a larger program of research involving the study of various
aspects of affective cognition. The present dissertation represents the
culmination of this pilot work which validated the affective properties
of the experimental materials employed in the present research, and
provided early replications of several of the phenomena reported in this
work.
I completed my clinical internship at the University of Alabama at
Birmingham Psychology Training Consortium where I developed a
171

172
specialization in child clinical psychology. In future research, I plan
to combine psychophysiological, cognitive, and behavioral measures to
explore the interaction of affective and attentional factors in
attentional deficits and learning disorders in pediatric populations.

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, a^~a dissertation for the degree of
Doctor of Philosophy.
Pet^fc ¡y. ¡Lang, Cha
Graduate Resear
Psychology
ssor of Clinical and Health
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
M.
lussell M. Bauer
Associate Professor of Clinical and Health Psychology
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
W. Keith Ber
Professor of Psychology
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Dawn Bowers
Associate Professor of Clinical and Health Psychology

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Lra Fischler
Professor of Psychology
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
This dissertation was submitted to the Graduate Faculty of the
College of Health Related Professions and to the Graduate School and was
accepted as partial fulfillment of the requirements for the degree of
Doctor of Philosophy.
May, 1992
P-
Dean, College of Health Related Professions
Dean, Graduate School

UNIVERSITY OF FLORIDA
3 1262 08554 8138



153
Table 49continued
STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
8.53
2.90
9.26
6.90
New Word
9.14
5.62
6.82
7.19
xxxxx
10.74
21.67
11.56
14.66
Mean
9.47
10.07
9.21
9.58
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
8.07
6.39
7.14
7.20
New Word
5.35
12.28
7.80
8.48
xxxxx
8.90
12.74
10.32
10.65
Mean
7.44
10.47
8.42
8.78


83
18. Ski Slopes
Unpleasant. It is the dead of winter, a windy day with dark skies
and freezing temperature. You glumly climb into the chair on the ski
lift, behind your best friend. Far below you great clouds of snow are
being blown up off the deserted slopes. Jerking up the hill, a strong
gust of icy wind hits the ski lift hard. You watch helplessly as your
friend's chair swings violently on the cable. You scream, staring in
disbelief, as your friend is thrown from the lift.
Pleasant. It is winter and a beautiful day with clear blue skies
and mild temperature. Full of energy, you gaily hop onto the ski lift
with your best friend. You watch expert skiers far below raise up
clouds of snow on the long slopes. As you glide up on the lift a sudden
gust of wind knocks your friend's hat off. The chair swings on the
cable as your friend reaches out to grab it, laughing. Soon the hat is
only a small bright spot on the sparkling white hill below.
STUDY WORDS: WINTER, TEMPERATURE, CLOUDS, SNOW, GUST, LIFT, SLOPES
NEW WORDS: HAMBURGER, CAKE, NATION, BEAR, CUBICLE, MERCHANT, COMBS


58
Think of tasting each of the following, considering carefully the
image which comes to your mind's mouth, and classify the images
suggested by each of the following as indicated by the degree of
clearness and vividness specified on the Rating Scale.
21. Salt
( )
22.
Granulated (white) sugar
( )
23.
Oranges
( )
24.
Jelly
( )
25.
Your favorite soup
( )
Think of smelling each of the following, considering carefully the
image which comes to your mind's nose, and classify the images suggested
by each of the following questions as indicated by the degree of
clearness and vividness specified on the Rating Scale.
26. An ill-ventilated room ( )
27. Cooking cabbage ( )
28. Roast beef ( )
29. Fresh paint ( )
30. New leather ( )
Think of each of the following sensations, considering carefully
the image which comes before your mind, and classify the images
suggested as indicated by the degree of clearness and vividness
specified on the Rating Scale.
( )
31. Fatigue


132
Table 32
Mean recognition reaction time (msec) for study word targets
primed by a study word, a new word, or XXXXX
for affectively pleasant, neutral, or unpleasant texts
Affective Category
Pleasant Neutral Unpleasant Mean
Prime Type
Study Word
832
849
859
847
New Word
845
907
825
859
XXXXX
893
888
874
885
N-S Priming
+ 13
+58
-34
+ 9
X-S Priming
+61
+ 39
+ 15
+ 38


122
Table 20--continued
STARTLE SUBJECTS
Imagery Instructions
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
-1 .79
o
o
-1.48
-1.43
Poor
-3.17
-2.73
-3.67
-3.19
Mean
-2.48
-1 .86
-2.58
-2.31
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
-4.91
-6.38
-5.83
-5.68
Poor
+ 0.76
-1.12
-1.31
-0.56
Mean
-2.08
-3.72
-3.57
-3.12


47
Encoding Measures
According to text processing theory (Kintsch & van Dijk, 1978),
the basic structure of a text schema is constructed during the initial
stages of encoding. In later sentences, the processing task shifts to
elaboration of the story framework. The greater cognitive processing
demands of schema development have been associated with longer reading
times for earlier sentences of logical narratives, with later sentences
being read more quickly. Consistent with study predictions and the text
processing literature, earlier sentences of the study paragraphs were
read more slowly than later sentences.
Emotional valence of the paragraphs was also expected to modulate
reading time in this study. It was hypothesized that affective text
contains information on visceral and motor activation, resulting in more
elaborate and complex representations, compared to memory for neutral
content. However, since this additional information does not increase
the logical complexity of the texta factor known to increase encoding
time in text processing--it was not clear how emotional content might
modulate encoding of textual materials.
The results of this study suggested that emotional content
affected reading time for early sentences differently than for later
sentences. The tendency to read affectively unpleasant paragraphs more
slowly than pleasant texts was significant overall, although this effect
only developed later in the encoding process. In addition, this overall
effect was strongest for subjects with self-reported poor imagery
ability, who showed the linear effect of valence consistently throughout
the narrative.


169
Lang, P. J. (1984). Cognition in emotion: Concept and action. In C.
Izard, J. Kagan, & R. Zajonc (Eds.), Emotion, cognition, and
behavior (pp. 192-228). New York: Cambridge University Press.
Lang, P. J. (1985). The cognitive psychophysiology of emotion: fear
and anxiety. In A. H. Turna & J. D. Maser (Eds.), Anxiety and the
anxiety disorders (pp. 131-170). Hillsdale, NJ: Lawrence Erlbaum
Associates.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1990). Emotion,
attention, and the startle reflex. Psychological Review, 97, 377-
395.
Lang, P. J., Bradley, M. M., Cuthbert, B. N., & Patrick, C. J. (In
press). Emotion and psychopathology: A startle probe analysis.
In L. Chapman & D. Fowles (Eds.), Progress in experimental
personality and psychopathology research: Models and methods of
psychopathology. New York: Springer.
Lang, P. J., Kozak, M. J., Miller, G. A., Levin, D. N., & McLean, A.,
Jr. (1980). Emotional imagery: Conceptual structure and pattern
of somato-visceral response. Psychophysiology, 17. 179-192.
Lang, P. J., Levin, D. N., Miller, G. A., Kozak, M. J. (1983). Fear
behavior, fear imagery, and the psychophysiology of emotion: The
problem of affective response integration. Journal of Abnormal
Psychology, 92, 276-306.
Lang, P. J., Ohman, A., & Simons, R. F. (1978). The psychophysiology
of anticipation. In J. Requin (Ed.), Attention and performance
(Vol. 7). New York: Academic Press.
Lusebrink, V. B., & McGuigan, F. J. (1989). Psychophysiological
components of imagery. Pavlovian Journal of Biological Science,
24, 58-62.
McGuigan, F. J., & Schoonover, R. A. (1973). Psychophysiology of
thinking: Studies of covert processes. New York: Academic Press.
McKoon, G. & Ratcliff, R. (1980). Priming in item recognition: The
organization of propositions in memory for text. Journal of
Verbal Learning and Verbal Behavior, 19, 369-386.
Mehrabian, A. & Russell, J. A. (1974). An approach to environmental
psychology. Cambridge, MA: MIT Press.
Miller, G. A., Levin, D. N., Kozak, M. J., Cook, E. W. Ill, McLean, A.,
& Lang, P.J. (1987). Individual differences in imagery and the
psychophysiology of emotion. Cognition and Emotion. 1, 367-390.
Obrist, P. A. (1981). Cardiovascular psychophysiology: A perspective.
New York: Plenum Press.


108
Table 11
Mean reading time (sec) for early, middle, and late sentence pairs
of affectively pleasant, neutral and unpleasant texts, with and without
startle
probes, for qood and poor
imaqerv ability
subiects
NO STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
4.97
4.50
4.87
4.78
Middle
4.58
4.36
4.67
4.54
Late
4.20
4.21
4.31
4.24
Mean
4.58
4.36
4.62
4.52
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.85
6.17
5.84
5.95
Middle
5.59
5.77
5.66
5.67
Late
4.90
5.06
5.20
5.05
Mean
5.44
5.67
5.57
5.56


85
20. Shower
Pleasant. It is early morning as you reach for a towel to dry off
after your shower. The fluffy towel feels pleasantly soft as you dry
yourself with long strokes. You rub your hair lightly and dry your face
and neck, feeling more awake now. Stepping out of the shower stall you
catch the fragrant odors in the bathroom. Finally, you're done; your
skin feels smooth, your body feels fresh and clean. Refreshed and ready
for the day, you put on your robe and go back to your room.
Unpleasant. You dry off from your shower with a thin cotton towel
in the old locker room. The towel turns into a small soggy lump after a
few half-hearted strokes. You are vaguely aware of something crawling
on your hair or maybe your neck. The narrow stall is dark, humid, and
smells of mold and stale body odors. You jump and your skin crawls as
you feel a bug scuttle down your naked body. Your heart pounds as you
nervously scan the walls and floor in the dark stall.
STUDY WORDS: SHOWER, TOWEL, STALL, ODORS, SKIN, BODY, STROKES
NEW WORDS:
LOG, HINGES, SATIN, DRILL, VEST, CHALK, KIDNEY


127
Table 28
Mean recognition reaction time (msec) for study word and new word
targets for affectively pleasant, neutral, and unpleasant texts
for good and poor imagery ability subjects
Study Word Targets
Pleasant Neutral Unpleasant Mean
Imagery Ability
Good
831
856
824
837
Poor
883
906
882
890
Mean
857
881
853
864
New Word Targets
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
856
876
865
866
Poor
879
898
889
889
Mean
868
887
877
877


I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, a^~a dissertation for the degree of
Doctor of Philosophy.
Pet^fc ¡y. ¡Lang, Cha
Graduate Resear
Psychology
ssor of Clinical and Health
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
M.
lussell M. Bauer
Associate Professor of Clinical and Health Psychology
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
W. Keith Ber
Professor of Psychology
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Dawn Bowers
Associate Professor of Clinical and Health Psychology


57
6. The whistle of a locomotive ( )
7. The honk of an automobile ( )
8. The mewing of a cat ( )
9. The sound of escaping steam ( )
10. The clapping of hands in applause ( )
Think of "feeling" or touching each of the following, considering
carefully the image which comes to your mind's touch, and classify the
images suggested by each of the following questions as indicated by the
degrees of clearness and vividness specified on the Rating Scale.
Rating
11. Sand ( )
12. Linen ( )
13. Fur ( )
14. The prick of a pin ( )
15. The warmth of a tepid bath ( )
Think of performing each of the following acts, considering
carefully the image which comes to your mind's arms, legs, lips, etc.,
and classify the images suggested as indicated by the degree of
clearness and vividness specified on the Rating Scale.
16. Running upstairs ( )
17. Springing across a gutter ( )
18. Drawing a circle on paper ( )
19. Reaching up to a high shelf ( )
20. Kicking something out of your way ( )


8
Weisberg, & Wiesenfeld, 1977). This may reflect greater attention to
aversive stimuli. On the other hand, heart rate increases have been
observed during imagery of aversive events, relative to neutral
mentation (Lang, Kozak, Miller, Levin, & McLean, 1980), perhaps indexing
a more intensive internal processing. Perception and imagery are not,
of course, wholly independent. In order to be processed, any stimulus,
text or picture, must initially be an object of orientation and
attention. However, if the material subsequently contacts relevant
representations in memory, the cognitive work of image construction may
begin. Presuming that heart rate is a reliable covariate of these
processes, we might expect (1) that the response to a stimulus would
change over time, and (2) that subjects who quickly develop images,
because of the particular pertinence of the stimulus or their general
disposition to image, are more likely to show the heart rate pattern of
imagery than that of attention. An example of the latter phenomenon has
been observed in phobic subjects (Hamm, Globisch, Cuthbert, & Vaitl,
1991; Klorman, Wiesenfeld, & Austin, 1975), who readily access a vivid
fear image, and show a strong accelerative response (rather than the
usual deceleration in aversion) to slides of their own primary fear
objects.
In the present study heart rate response should vary with both
encoding strategy and activation of an affective "action set." The
above considerations suggest the following working hypotheses:
1. It is predicted that heart rate will decrease during text
encoding, consistent with the experimental task of information intake.


59
32.
Hunger
(
)
33.
A sore throat
(
)
34.
Drowsiness
(
)
35.
Feeling full after a large meal
(
)


109
Table 11 continued
STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.24
5.04
5.46
5.25
Middle
4.67
4.93
5.00
4.87
Late
4.50
4.66
4.46
4.54
Mean
4.80
4.88
4.97
4.89
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
5.90
5.69
6.19
5.92
Middle
5.48
6.03
5.88
5.80
Late
4.91
4.95
5.18
5.01
Mean
5.43
5.55
5.75
5.58


79
14. Tennis Game
Unpleasant. A cute guy you met invited you to play tennis but
you're not in good form. As the game begins you make clumsy mistakes
and feel more and more embarrassed. You can't seem to get the feel of
the game as the ball lands in the net again. You breathe hard, trying
to catch your wind between volleys, your face flushed. As you lose the
match, you shout hollow congratulations and manage a smile. You redden
as you see people watching you; you just want to get off the court.
Pleasant. You are playing tennis with a good friend, feeling
energetic and in top form. Running hard across the court, every move
feels rightyou're playing well. You delight in the solid feel of each
shot as you slam the ball across the net. You breathe hard, catching
your wind between volleys, your face flushed. Your partner shouts
congratulations with a big smile as you win another match. You walk off
the court feeling great, ready for a shower and a cold drink.
STUDY WORDS: TENNIS, FORM, FEEL, NET, SMILE, MATCH, VOLLEYS
NEW WORDS: JUDGE, PILLOW, ROBIN, POUND, JUNGLE, BRACELET, CUPS


149
Table 47
Mean percent errors for study words primed by a study word, a new word,
or XXXXX for affectively pleasant, neutral, or unpleasant texts,
with and without startle probes
No Startle Subjects
Prime Type
Pleasant
Neutral
Unpleasant
Mean
Study Word
5.36
5.46
5.42
5.41
New Word
9.21
9.05
8.53
8.93
XXXXX
7.17
8.62
9.71
8.50
Mean
7.25
7.71
7.88
7.61
Startle Subjects
Pleasant
Neutral
Unpleasant
Mean
Type
Study Word
8.30
4.65
8.20
7.05
New Word
7.24
8.95
7.31
7.83
XXXXX
9.82
17.20
10.94
12.65
Mean
8.45
10.27
8.45
9.18


24
where the prime was presented. After the "+" was removed, the prime
appeared immediately and remained on the screen for 300 ms. After the
prime was removed from the screen, the target was presented on the line
immediately below where the prime had appeared, and remained on the
screen until the subject responded. When the subject gave an incorrect
response, the word "ERROR" appeared on the screen for 2 seconds. After
the subject response there was another 200 ms pause, after which the "+"
appeared again for 150 ms, followed by the prime and target stimuli, as
described for the previous test pair. This sequence continued until all
12 prime-target pairs were presented. A variable inter-trial interval
occurred between text presentations ranging from 15 to 25 seconds in
length. After the last text had disappeared from the screen, the
following message appeared on the screen: "The session has been
completed. Thank you for your participation."
After the subject had completed this part of the study, the
physiological sensors were removed, and she was taken to another room
where she was asked to recall and write down a brief phrase
characterizing the content of as many of the previously presented texts
as possible. This task began approximately 10 minutes after the end of
the experiment. Subjects generally required about 5 or 10 minutes to
complete this task, although they were given as much time as needed.
After the free recall task was completed, the subject rated the
texts for affective valence, arousal, and dominance. Ratings were made
using a paper-and-pencil version of the Self-Assessment Manikin (SAM;
Hodes, Cook, & Lang, 1985; Lang, 1980; see also Appendix F). Typed
versions of the texts were interleaved in the SAM ratings booklet; the


9
2. Greater heart rate deceleration is expected for subjects
receiving verbal encoding instructions and subjects who report poor
imagery ability. As these subjects are primarily "attending" and are
less likely to be engaging in imaginal processing, greater heart rate
deceleration is predicted for unpleasant text content.
3. Imagery instructions and self-reported vivid imagery are
predicted to result in heart rate acceleration during the encoding
process, reflecting the internal cognitive effort of image generation.
The pattern of heart rate change associated with emotional imagery
(greater heart rate increase for more aversive content) is predicted for
these subjects.
4. Consistent with text processing theory, heart rate change
associated with the development of affective schemata is more probable
late in the encoding process.
The startle reflex is a generalized flexor response elicited by
sudden, intense stimuli. The response complex entails a large number of
protective elements, such as eyeblink, forward and downward head
movements, raising of the shoulders, contraction of the abdomen, and
flexion of the limbs (Davis, 1984). Latency and magnitude of the
eyeblink component of this reflex are modulated by both the attentional
and the emotional state of the organism (Anthony, 1985; Graham, 1979;
Lang, Bradley, & Cuthbert, 1990). Inhibition of the reflex response to
an acoustic startle probe occurs during visual attention to pleasant,
interesting stimuli (Simons & Zelson, 1985). On the other hand, startle
probe responses are exaggerated in aversive emotional states (Berg &
Davis, 1984; Bradley, Cuthbert, & Lang, 1990; Cook, Hawk, Davis, &


10
Stevenson, 1991; Cook, Davis, Hawk, Spence, & Gautier, in press; Vrana &
Lang, 1990; Vrana, Spence, & Lang, 1988; see Lang, Bradley, Cuthbert, &
Patrick, in press, for a review). In the present study, startle probes
are presented unpredictably while subjects read the study paragraphs.
The blink component of the startle response is recorded. It is expected
that variations in its latency and magnitude will track the development
of the affective schema during the encoding process. Specifically, it
is predicted that:
1. Potentiation of the startle reflex (greater magnitude and
faster latency) is expected as a result of activation of an aversive
emotional schema; relative inhibition is expected for pleasant schemata.
2. These effects will be enhanced late in the encoding process,
after the affective schema is well developed.
3. Imagery instructions and good imagery ability will result in
greater affective modulation of the startle response, relative to verbal
instructions and self-reported poor imagery ability.
Retrieval Hypotheses
Bio-information theory states that emotional images have an extra
set of concepts in the networknamely those involving response
information. Network theory states that greater elaboration of
representations in memory enhances retrieval (Anderson, 1976). This
mechanism may underlie the corollary finding that moderately arousing
affective events are better and more accurately remembered than less
arousing events (e.g., Bradley, Greenwald, Petry, & Lang, 1991; Heuer &
Reisberg, 1990). These findings are also consistent with the Yerkes-


4
Emotion and Memory
The view taken here (see Lang, 1985; Lang, Bradley & Cuthbert,
1990) is that emotions are action dispositions, which are either
appetitive--involving approach and consummatory-related behaviors, or
they are aversiveinvolving defensive, avoidant, and protective
behavioral strategies. Both the appetitive and aversive emotional
states vary in intensity or arousal, which is associated with increased
efferent activity and a greater overt action potential.
The responses in emotion include expressive language, overt
behavioral acts, and a supporting somatic and visceral physiology (Lang,
1964; Lang, 1978). Therefore, the network of an emotional memory image
is held to include, in addition to response concepts relevant to the
remembered perceptual context (e.g., eye movements, postural
adjustments), response representations of remembered emotional
reactions. Furthermore, while overt manifestations may be inhibited by
the context of recall, the memory image should generate an affective
physiological pattern, aversive or appetitive, moderate or intense,
depending on the emotional context being recalled (Lang et al., 1983).
The Research Problem
The present experiment examines the encoding and subsequent memory
retrieval of affective text. Subjects initially completed self report
measures of imagery ability and were then told they would be asked to
learn a series of narrative paragraphs. Half of the subjects received
imagery instructions. They were told to imagine the content of the
paragraphs as a vivid personal experience, clearly imaging the events


84
19. Mexican Vacation
Neutral. You are on vacation on the Mexican coast, arranging some
deep sea fishing. Standing on the deck, you are aware of the constant
rocking motion of the boat. The guide explains that big tuna inhabit
the deep shadows around the shoals. You reflect that speaking Spanish
all the time will probably get to be a pain. You mentally check your
supplies, adding the pills that prevent sea sickness. Hopefully the
weather will be good, but you'll pack a raincoat just in case.
Unpleasant. You are on vacation on the Mexican coast, out doing
some deep sea fishing. Standing on the deck, the constant motion of the
boat makes you lightheaded. Nausea creeps over you as you search for
tuna in the shadows under the waves. Suddenly another tourist makes a
bad cast and a hook sinks into your cheek. You scream with pain and
several Mexicans start speaking rapid Spanish at you. Blood drips down
onto your shirt and you feel like you're going to be sick.
STUDY WORDS: VACATION, COAST, DECK, MOTION, TUNA, SHADOWS, PAIN
NEW WORDS: BLADE, VENTURE, CHIMNEY, PORTIONS, HAWK, GERMS, CAVE


155
Table 50continued
STARTLE SUBJECTS
Imagery Instructions
Pleasant
Neutral
Unpleasant
Mean
Prime Type
Study Word
10.93
6.51
6.71
8.05
New Word
9.03
9.02
8.94
9.00
xxxxx
12.17
8.53
10.07
10.26
Mean
10.71
8.02
8.58
9.10
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Prime Type
Study Word
5.67
2.78
9.69
6.05
New Word
5.45
8.88
5.68
6.67
XXXXX
7.46
25.87
11.81
15.05
Mean
9.06
12.51
6.19
9.26
Table 51
Mean proportions
of affectively pleasant
. neutral. and unpleasant
recalled at
the end of
the experiment
Affective Category
Pleasant
Neutral
Unpleasant
Mean
All subjects
.41
.42
.29
.37


142
Table 39
Mean percent errors for new word and study word targets,
for affectively pleasant, neutral, and unpleasant texts
Affective Category
Pleasant
Neutral
Unpleasant
Mean
Target Type
New
3.68
3.83
4.29
3.93
Study
7.85
8.99
8.35
8.40
Mean
5.77
6.41
6.32
6.17
Table 40
Mean percent errors
for new word
and study word
targets
for affectively pleasant, neutral
, and unpleasant texts
for
qood and
poor imaaerv
ability subjects
New Word Targets
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
4.19
4.12
4.94
4.42
Poor
3.18
3.54
3.64
3.45
Mean
3.68
3.83
4.29
3.93
Study Word Targets
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
8.18
8.82
8.83
8.61
Poor
7.52
9.15
7.87
8.18
Mean
7.85
9.00
8.35
8.40


116
Table 17continued
VERBAL INSTRUCTIONS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-2.96
-4.43
-3.42
-3.60
Middle
-4.90
-4.99
-5.53
-5.14
Late
-5.16
-5.68
-6.14
-5.66
Mean
-4.34
-5.03
-5.03
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.28
-1.73
-1.17
-1 .06
Middle
-2.11
-3.12
-3.55
-2.93
Late
-2.09
-3.32
-3.63
-3.01
Mean
-1.49
-2.72
-2.78


23
Instructions were read to each subject for verbal or imagery
processing, and the subject was given practice in reading texts and
responding to test lists. For startle subjects, headphones were placed
on the subject's head and instructions were given to ignore the brief
bursts of white noise that would occasionally occur while she was
reading the study texts. The consent form, imagery questionnaire (QMI)
and the exact instructions read to each subject are included in
Appendix A.
After the experimenter had insured that the subject understood the
instructions and had verified that the physiological measures were being
recorded accurately, the subject was left alone in the experimental
chamber to complete the experiment independently. The duration of this
part of the experiment was approximately 45 minutes.
Prior to the onset of each trial, a 3-second pre-trial baseline
period occurred. At the end of this period, the following message
appeared on the subject's monitor screen: "Please press button to
begin." This message signalled the subject to bring the first sentence
of the study text onto the monitor screen by pressing one or both
buttons. Each successive button press presented an additional sentence
of the text, until the entire paragraph was displayed on the screen.
For a subset of subjects, a brief burst of white noise was presented
over headphones at either 2.0, 2.25, or 2.50 seconds after onset of one
of the first two sentences (12 texts) or last two sentences (12 texts)
of the texts. When the subject had read the entire text, a final button
press cleared the screen. After a 200 ms pause, the test list began.
Initially a "+" appeared on the screen for 150 ms, centered on the line


3
past events, this mental imagery is reliably accompanied by somatic and
visceral changes which are similar in pattern to those occurring in the
referent perceptual contexts. An analysis of imagery based on the
activation of efferent information was developed in a neurophysiological
framework by Hebb (1968), and elaborated in cognitive terms by Lang's
(1979) bio-informational theory of emotional imagery.
Bio-Information Theory
The bio-informational model of memory imagery assumes the
essentials of an associative network model. It differs from other views
in that response information is included as a concept category. That is
to say, an image network would include sensory concepts (e.g., color,
object representations, spatial locations) and perhaps also meaning
concepts (e.g., logical relationships, categorical or other semantic
information), as in other network theories. In this view, however,
images would also include response concepts (e.g., overt behavior
patterns and covert physiological events). This latter information
would be in the same associative net, and its activation would have
similar consequences for overall network firing. Furthermore, there is
a natural neural connection between response concepts in memory (which
might be activated, for example, by text content) and the actual
efferent events to which they refer. Thus, while imagery is in
principle a state in which overt behavior is inhibited, "efferent
leakage" --a sub-overt activation of visceral organs and somatic muscle
intrinsic to the represented behavioris a normal part of the image
process. In point of fact, Lang (1979) argues that in the intact
organism image processing is defined by its efferent consequences.


16
"Matched" means that the texts were similar in topic and sentence
construction, but varied in affective content. For example, a text
describing a beach picnic was matched with a similar text describing a
shark attack at the beach. In all cases, each pair of matched texts was
followed by the same test list. To illustrate with the previous
example, the study word "fins" appeared in the test list following the
matched texts pertaining to a visit to the beach. However, in the
affectively unpleasant text, "fins" referred to shark fins spotted prior
to a shark attack; in the pleasant text, the word referred to swim fins
that were lying on the beach.
The use of matched texts and test lists was employed to control
for inadvertent effects on reaction time (RT) that might be considered
material-specific. That is, it could be argued that the words in test
lists derived from affective texts might speed or slow recognition RT
because of a number of factors, including concreteness versus
abstractness of the nouns, differential frequency characteristics, or
affective valence of the test words themselves. The text matching
strategy controlled for this confound. See Appendix B for a listing of
study text pairs and their associated test list words.
These texts were also rated by subjects in the present study at
the end of the experimental session, using the booklet form of the SAM
ratings scale--see Appendix F for an illustration. Line the automated
SAM system, the booklet form included the affective dimensions of
valence, arousal, and dominance. Each of these dimensions was set to a
9-point scale. A rating of 1 represented the lowest possible value for
the valence (i.e., unpleasant), arousal, or dominance scale; and a


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
PSYCHOPHYSIOLOGY AND MEMORY:
THE ENCODING AND RETRIEVAL OF AFFECTIVE TEXT
By
Ellen Louise Spence
May, 1992
Chairman: Peter J. Lang, Ph.D.
Major Department: Clinical and Health Psychology
The conceptual framework for this research was derived from Lang's
bio-informational theory of emotion which states that affective memory
representations contain information about physiological mobilization and
action. Imaginal processing of emotional text content was posited to
further increase the representation of physiological activation in
memory. The resulting affective information structure contains more
information than representations of neutral content, and supports faster
and more probable retrieval from memory.
Subjects were trained in a verbal or imaginal encoding strategy.
They then read brief narratives that depicted pleasant, neutral or
unpleasant experiences. Verbal instructions emphasized attention to the
exact words used in the text, while imagery instructions focussed on the
events described as a vivid personal experience. Read time was employed
as a measure of the time required to construct the text schema, while
v


63
Debriefing Statement
This study is concerned with determining whether people form
different mental representations of emotional vs. non-emotional texts.
You may have noticed that some of the paragraphs that you read were
pleasant, some were unpleasant, and some were relatively neutral. We
were measuring your reaction time to target words taken from those
paragraphs. Some target words were preceded (or primed) by words taken
from the same sentence, some target words were primed by words taken
from another sentence in the paragraph, and some targets were primed by
words that did not appear in the paragraph at all. Usually reaction
time is faster when a word is primed by a word that is closely related
(when the prime and target were from the same sentence in this case),
and measuring the changes in reaction time based on the logical distance
between the prime and the target can provide information about the
structure of the mental representation you have constructed from the
paragraph.
We think that mental representations of emotional paragraphs may
be more complex and extensive than those of neutral paragraphs, because
emotional stimuli more often include representations of motor activity
(jumping away from the palmetto bug) and of visceral activation (your
heart is pounding from the scare). This more extensive information
network would result in longer reaction times for affective paragraphs,
which is what we have found in an earlier study.
We also think that people who form vivid images may form
particularly extensive information structures, and so we would expect
those persons to have longer reaction times also.


118
Table 18--continued
STARTLE SUBJECTS
Good Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-2.01
-2.55
-2.37
-2.31
Middle
-3.79
-4.10
-4.30
-4.06
Late
-4.26
-4.33
-4.30
-4.30
Mean
-3.35
-3.66
-3.65
-3.55
Poor Imagers
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
+0.17
-0.85
-0.94
-0.54
Middle
-1 .79
-2.14
-3.20
-2.38
Late
-2.00
-2.78
-3.32
-2.70
Mean
-1 .21
-1 .92
-2.49
-1.87


52
response information. Consistent with the bio-informational hypotheses,
and with other recent studies, more arousing materials were remembered
faster and more reliably (Bradley et al., in press; Heuer & Reisberg,
1990) .
In addition, recognition RT for new word targets showed no
relationship to either affective content or encoding strategy,
suggesting that the representation of text in memory must be activated
(by processing of a study word target) in order for the experimental
factors to affect performance. It may then be reasoned that the
recognition RT measure does in fact reflect the characteristics of the
information network represented in memory rather than nonspecific,
generalized effects of emotional arousal.
Analyses of the priming data yielded significant facilitation of
reaction time for study-study compared to new-study prime-target pairs
for neutral texts, but priming did not occur for affective texts.
However, Figure 2 clearly shows that the lack of priming for affective
materials resulted from a change in RT for New-Study pairs. That is,
recognition RT was faster for New-Study pairs from affective than from
neutral paragraphs. It appears from these data that a new word prime
has a greater inhibitory effect on recognition of a study word target
from an affectively neutral text, relative to when the study word target
is part of an affective information structure. Lang has suggested that
the associative structure of affective information, particularly when
processed in an imagery mode, contains a greater amount and different
types of information (motor and visceral activation codes) than the
materials typically employed in priming experiments. It appears that


41
Subjects made more errors in rejecting new word targets with
imagery compared to verbal instructions (imagery = 4.71 percent, verbal
= 3.16 percent), F (1,46) = 4.37, p < .05. Other tests were not
significant: Errors for new word targets did not vary with emotional
content of the paragraphs, presence or absence of startle probes, or
imagery ability of the subjects.
Overall, study word errors did not vary significantly with valence
(F < 1 ). Examination of Table 5 does suggest that subjects receiving
verbal instructions made more errors in recognition of study word
targets from affectively neutral paragraphs compared to emotional (both
pleasant and unpleasant) texts. Statistical support for this
interpretation was, however, marginal, Valence X Instructions
interaction, F (2,92) = 3.27, p = .06; Valence (quadratic) X
Instructions F (1,46) = 3.43, p = .07.
Errors in recognition of study word targets did not vary reliably
with imagery ability, or presence or absence of startle probes.
There is no evidence in these data that the observed recognition latency
differences in affective contents can be attributed to a speed-accuracy
trade-off.
Priming condition and target recognition latency. A subset of the
reaction times to study words were re-examined, considering the effect
of the preceding prime. Specifically, the analysis compared recognition
latency for study word targets primed by another study word (SS) with
recognition latency to study words primed by a new word (NS). The same
experimental factors (affective valence, instructions, imagery ability,
presence or absence of startle probes) were assessed, as previously.


40
Target recognition errors. Subjects made a larger percentage of
errors for study word compared to new word targets (study = 8.40
percent; new = 3.93 percent), F (1,46) = 59.39, p < .01. Target type
also interacted with valence and instructions to affect error rate,
Valence X Instructions X Target Type F (2,92) = 4.02, p < .04 in the
omnibus test; Valence (quadratic) X Instructions X Target Type,
F = 5.22, p < .03. Thus, separate follow-up analyses were carried out
for new word and study word targets.
Table 4
Mean recognition reaction time (msec) for study word targets
for affectively pleasant, neutral, and unpleasant texts
for good and poor imagers receiving imagery and verbal instructions
Imagery Instructions
Imagery Ability
Pleasant
Neutral
Unpleasant
Mean
Good
805
890
819
838
Poor
854
851
861
856
Mean
830
871
840
Verbal Instructions
Pleasant
Neutral
Unpleasant
Mean
Imagery Ability
Good
856
823
829
836
Poor
912
961
902
925
Mean
884
892
865


64
This research is part of a larger program of experiments which
concern the study of modes of information processing and emotion in
persons who vary in imagery ability. We are working to develop a model
of how affective memories are represented, stored and retrieved in the
brain.
It is important that subjects participating in these studies not
have preconceived ideas about the experimenter's hypotheses and
expectations, in order to obtain valid experimental data. Therefore,
please do not discuss this experiment with other psychology students
until after the term is over.
I have read this form and had any questions concerning the study
answered to my satisfaction.
SUBJECT SIGNATURE


114
Table 16
Mean heart rate
change (beats
per minute)
for early.
middle, and late
sentence pairs
of affectively
pleasant, neutral and unpleasant texts
with and
without startle probes
No
Startle Subjects
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-1.11
-2.06
-1 .25
-1.47
Middle
-2.70
-3.09
-3.24
-3.01
Late
-2.92
-3.12
-3.56
-3.20
Mean
-2.24
-2.76
-2.68
-2.56
Startle Subjects
Pleasant
Neutral
Unpleasant
Mean
Sentence Pair
Early
-0.92
-1 .70
-1.66
-1 .43
Middle
-2.79
-3.12
-3.75
-3.22
Late
-3.13
-3.56
-3.81
-3.50
Mean
-2.28
-2.79
-3.07
-2.72


APPENDIX E
AFFECTIVE RATINGS FOR PARAGRAPHS
Affective Ratings
Ratings provided subjects in the present study for the valence,
arousal, and dominance scales were subjected to an analysis of variance
in order to determine whether any of the individual difference or
instructional variables systematically influenced ratings. The factors
included in the 2X2X3 ANOVA were QMI (good, poor) X Instructions
(verbal, imagery) X Valence (pleasant, neutral, unpleasant).
Valence Ratings
Table 52 summarizes mean ratings for texts classified as pleasant,
neutral, and unpleasant on the SAM valence, arousal, and dominance
scales. A significant effect was obtained for valence ratings used to
classify affective texts in the omnibus test, F (2,94) = 3695.67,
p < .0001.
TABLE 52
Mean valence, arousal, and control ratings for affective texts
Rating Scale
Pleasant
Neutral
Unpleasant
Valence
8.29
4.94
1 .86
Arousal
6.25
4.58
7.46
Dominance
7.01
4.95
2.42
156