<%BANNER%>

Emotional Working Memory

Permanent Link: http://ufdc.ufl.edu/UFE0022095/00001

Material Information

Title: Emotional Working Memory An Individual Differences Approach to Understanding Attention Control
Physical Description: 1 online resource (89 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: arousal, attention, buffer, control, differences, emotion, episodic, individual, memory, operation, processing, span, storage, valence, working
Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Emotion affects how we respond to events, how we think about them, and how we remember them. Researchers from a variety of domains have investigated these aspects of 'emotional cognition.' What seems to have received less attention in the literature, however, are the effects of emotion on short-term, working memory. Many commonly held ideas (i.e., 'I need to cool off before responding' or 'I wasn't thinking; I was caught up in the heat of the moment.') suggest there is an effect of emotion on working memory. Our study reviewed key research on the storage (STM) and process (executive functions) components of working memory as well as research on emotion and attention. After the brief review, a two-part experiment investigated the characteristics of an affective working memory. The study aimed at investigating individual differences in working memory capacity and their relationship to attention control. There is research suggesting that working memory measures actually reflect two aspects of attention control: the ability to maintain attention in the face of distraction, and the ability to maintain task set. Our study was intended to replicate these findings and explore how well these measures can predict performance on attention-intensive tasks that include emotional materials. Another aim of our study was to measure the impact that emotion has on working memory capacity. Can a measure of working memory that incorporates emotion predict attention control performance better than traditional working memory measures? Results showed that overall, individual differences on the emotional measure of working memory capacity were a better predictor of performance (for both errors and latencies) on an Emotional Stroop task than the neutral measure whereas the neutral measure of working memory was a better predictor (for latencies) in the classic Stroop task (color-naming). This pattern of results would seem to support the construct of emotional working memory. That is, one's ability to maintain emotional words in working memory was specifically a better predictor of performance on other tasks involving emotion, while the ability to maintain neutral words was a better predictor of attention control in the classic Stroop task.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Fischler, Ira S.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022095:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022095/00001

Material Information

Title: Emotional Working Memory An Individual Differences Approach to Understanding Attention Control
Physical Description: 1 online resource (89 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: arousal, attention, buffer, control, differences, emotion, episodic, individual, memory, operation, processing, span, storage, valence, working
Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Emotion affects how we respond to events, how we think about them, and how we remember them. Researchers from a variety of domains have investigated these aspects of 'emotional cognition.' What seems to have received less attention in the literature, however, are the effects of emotion on short-term, working memory. Many commonly held ideas (i.e., 'I need to cool off before responding' or 'I wasn't thinking; I was caught up in the heat of the moment.') suggest there is an effect of emotion on working memory. Our study reviewed key research on the storage (STM) and process (executive functions) components of working memory as well as research on emotion and attention. After the brief review, a two-part experiment investigated the characteristics of an affective working memory. The study aimed at investigating individual differences in working memory capacity and their relationship to attention control. There is research suggesting that working memory measures actually reflect two aspects of attention control: the ability to maintain attention in the face of distraction, and the ability to maintain task set. Our study was intended to replicate these findings and explore how well these measures can predict performance on attention-intensive tasks that include emotional materials. Another aim of our study was to measure the impact that emotion has on working memory capacity. Can a measure of working memory that incorporates emotion predict attention control performance better than traditional working memory measures? Results showed that overall, individual differences on the emotional measure of working memory capacity were a better predictor of performance (for both errors and latencies) on an Emotional Stroop task than the neutral measure whereas the neutral measure of working memory was a better predictor (for latencies) in the classic Stroop task (color-naming). This pattern of results would seem to support the construct of emotional working memory. That is, one's ability to maintain emotional words in working memory was specifically a better predictor of performance on other tasks involving emotion, while the ability to maintain neutral words was a better predictor of attention control in the classic Stroop task.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Fischler, Ira S.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022095:00001


This item has the following downloads:


Full Text

PAGE 1

1 EMOTIONAL WORKING MEMORY: AN INDI VIDUAL DIFFERENCES APPROACH TO UNDERSTANDING ATTENTION CONTROL By CYNTHIA ELIZABETH KASCHUB A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Cynthia Elizabeth Kaschub

PAGE 3

3 To my loving father who is my ever-encouraging beacon

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank m y disse rtation committee: Ira Fischler, Lise Abrams, Keith Berg, and William Perlstein. Without thei r support I would not have had th e flexibility to pursue this topic. I would also like to e xpress my extreme gratitude to all of the undergraduate Research Assistants who gathered my disse rtation data so meticulously: Sheila Hollands, Juliana Peters, Emily McAllister, Trudy Salmon, Nicole Smig ielski, Kathleen Viei ra, and Julianne Wong. I would also like to thank my friends and fam ily who have opened their ears and hearts to me on my doctoral path. It has been a long journey, one on which I am grateful for their encouraging words and actions. A special thanks to Jimmy for loving me as I am, the way only a best friend can.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES.........................................................................................................................8 ABSTRACT.....................................................................................................................................9 CHAP TER 1 INTRODUCTION..................................................................................................................11 The Current Study.............................................................................................................. .....16 Individual Differences in Working Memo ry Capa city and Attention Control.......................18 Emotion and Attention.......................................................................................................... ..21 Research Questions............................................................................................................. ....23 2 METHOD......................................................................................................................... ......26 Participants.............................................................................................................................26 Design.....................................................................................................................................26 Materials..........................................................................................................................26 Operation Span Task....................................................................................................... 27 Emotional Operation Span Task...................................................................................... 28 Emotional Stroop Task.................................................................................................... 28 Congruency Stroop Task................................................................................................. 28 Word Ratings Task.......................................................................................................... 29 Procedure................................................................................................................................30 3 RESULTS...............................................................................................................................31 Neutral and Emotional Operation S pan Task......................................................................... 31 Emotional St roop Task ...........................................................................................................32 Overall Error Rates............................................................................................................ .....33 Individual Differences in Error Rates..................................................................................... 34 Individual Differences in Neutral W orking Memory Capacity (OSPAN)............................. 34 Individual Differences in Emotiona l Working Mem ory Capacity (ESPAN)......................... 34 Overall Response Latencies.................................................................................................... 35 Individual Differences in Response Latencies....................................................................... 36 Individual Differences in Neutral W orking Memory Capacity (OSPAN)............................. 36 Individual Differences in Emotiona l Working Mem ory Capacity (ESPAN)......................... 36 Congruency Stroop.................................................................................................................37 Overall Error Interference..................................................................................................... ..38 Individual Differences in Neutral W orking Memory Capacity (OSPAN)............................. 39

PAGE 6

6 Individual Differences in Emotiona l Working Mem ory Capacity (ESPAN)......................... 39 Follow-up Regression Analyses on Error Interference.......................................................... 40 Overall Response Latencies.................................................................................................... 40 Response Time Interference................................................................................................... 41 Individual Differences in Neutral W orking Memory Capacity (OSPAN)............................. 41 Individual Differences in Emotiona l Working Mem ory Capacity (ESPAN)......................... 42 Follow-up Regression Analyses on Response Time Interference..........................................42 Overall Response Time Facilitation....................................................................................... 43 Individual Differences in Neutral W orking Memory Capacity (OSPAN)............................. 43 Individual Differences in Emotiona l Working Mem ory Capacity (ESPAN)......................... 44 Follow-up Regression Analyses on Response Time Facilitation........................................... 44 Word Rating Task............................................................................................................... ....45 4 DISCUSSION.........................................................................................................................63 APPENDIX A INFORMED CONSENT........................................................................................................70 B OPERATION SPAN STIMULI.............................................................................................72 C EMOTIONAL STROOP WORD STIMULI.......................................................................... 74 D CONGRUENCY STROOP ERROR IN TERFERENCE REGRESSION MODEL COMPARISONS ....................................................................................................................76 E CONGRUENCY STROOP RESPONSE TIME INTERFERENCE REGRESSION MODEL COMPARISONS ..................................................................................................... 78 F CONGRUENCY STROOP RESPONSE TIME FACILITA TION RE GRESSION MODEL COMPARISONS..................................................................................................... 80 LIST OF REFERENCES...............................................................................................................82 BIOGRAPHICAL SKETCH.........................................................................................................89

PAGE 7

7 LIST OF TABLES Table page 3-1 Frequency distribution of OSPAN and ESPAN gr oups.................................................... 61 3-2 Word ratings............................................................................................................... ........62

PAGE 8

8 LIST OF FIGURES Figure page 1-1 Experimental hypotheses:.................................................................................................. 25 3-1 Overall Emotional Stroop error rates ................................................................................. 47 3-2 Emotional Stroop: Individual differe nces in Error Rates for OSPAN groups ................... 48 3-3 Emotional Stroop: Individual differe nces in error rates for ES PAN groups..................... 49 3-4 Overall Emotional Stroop response latencies.................................................................... 50 3-5 Emotional Stroop: Individual differences in response latencies for OSPAN groups ........51 3-6 Emotional Stroop: Individual differen ces in response tim es for ESPAN groups.............. 52 3-7 Congruency Stroop: Overall mean error rates................................................................... 53 3-8 Congruency Stroop: Error interference for OSPAN groups.............................................. 54 3-9 Congruency Stroop: Error interference for ESPAN groups.............................................. 55 3-10 Congruency Stroop: Overal l response tim es by trial type................................................ 56 3-11 Congruency Stroop: Individual differen ces in response tim e interference by OSPAN groups.................................................................................................................................57 3-12 Congruency Stroop: Individual differen ces in response tim e interference by ESPAN groups.................................................................................................................................58 3-13 Congruency Stroop: Individual differences in response tim e facilitation by OSPAN groups.................................................................................................................................59 3-14 Congruency Stroop: Individual differences in response tim e facilitation by ESPAN groups.................................................................................................................................60

PAGE 9

9 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EMOTIONAL WORKING MEMORY: AN INDI VIDUAL DIFFERENCES APPROACH TO UNDERSTANDING ATTENTION CONTROL By Cynthia Elizabeth Kaschub May 2008 Chair: Ira Fischler Major: Psychology Emotion affects how we respond to events how we think about them, and how we remember them. Researchers from a variety of domains have investig ated these aspects of emotional cognition. What seems to have receive d less attention in the literature, however, are the effects of emotion on short-term, working me mory. Many commonly held ideas (i.e., I need to cool off before responding or I wasnt thinking; I wa s caught up in the heat of the moment.) suggest there is an effect of emo tion on working memory. This dissertation will review key research on the storage (STM) and process (executive functions) components of working memory as well as research on emotion a nd attention. After the brief review, a two-part experiment will be described that investigated the characteristics of an affective working memory. The study aimed at investigating individual differences in working memory capacity and their relationship to attention control. Ther e is research suggesting that working memory measures actually reflect two aspects of attention control: the ability to ma intain attention in the face of distraction, and the ability to maintain task set. The current investigation was intended to replicate these findings and explore how well these measures can predict performance on

PAGE 10

10 attention-intensive tasks that include emotional ma terials. Another aim of this investigation was to measure the impact that emotion has on worki ng memory capacity. Can a measure of working memory that incorporates emotion predict atten tion control performance better than traditional working memory measures? Results showed that overall, individual differences on the emotional measure of working memory capacity were a better pred ictor of performance (for both errors and latencies) on an Emotional Stroop task than the neutral measure whereas the neutral meas ure of working memory was a better predictor (for latencies) in the classic Stroop task (color-naming). This pattern of results would seem to support the construct of em otional working memory that is, one's ability to maintain emotional words in working memo ry was specifically a better predictor of performance on other tasks involving emotion, while the ability to maintain neutral words was a better predictor of attention cont rol in the classi c Stroop task.

PAGE 11

11 CHAPTER 1 INTRODUCTION The idea that em otion and cognition may be inte rtwined is not new. This view may have originated with Aristotles opposition to his mentor Plato, who deemed cognition, emotion, and motivation to be completely separatethe tripartite soul. Aristotle, in contrast, proposed an infinity of parts by which the soul can be divide d, all of which could be influenced by emotion. Although Platos view was widely accepted well into the Renaissance, in 1641 Descartes proposed the concept of dualism between the separate entities of mind and body. While Descartes suggested emotion and cognition were i nherently intertwined entities, he provided no mechanism or empirical support for this claim. As Descartes dualism approach flourished, the function of, and relation between these c onstructs became of great interest. Contemporary views of how emotion and c ognition are related have often taken an evolutionary perspective, beginning with Darwin ( The Expression of the Emotions in Man and Animals 1872/1999), who suggested that humans have adapted and changed over time to more successfully accommodate the ever-changing envir onment. This evolutionary view has led researchers to describe emotions as functi onal (i.e., taxonomy in Keltner & Haidt, 2001, pp. 196). The primary function of emotion is to mobilize the organism to deal quickly with important interpersonal encounters, prepared to do so by what types of activity have been adaptive in the past (Ekman, 1998). With respect to cognition, humans evolved an increasing ability to process complex materials and tasks. Consequently, this has been the focus of a substan tial amount of research on cognitive processes now known as executive functions. The term executive functions typically include planning, inhibition, task manageme nt, motivation, and working memory.

PAGE 12

12 If executive functions and emo tions serve in concert to guide behavior, then this should be evident in neuroanatomy as we ll, with specific regions or stru ctures engaged in both cognitive and affective processes. The prefrontal cortex (PFC) has been implicated in brain imaging research on executive functioning as well as re search on differential processing of affective relative to neutral stimuli. PFC activation has been shown in a variety of executive functions: working memory and content monitoring (Petrides et al., 1993), task switching while working memory is engaged (DEsposito et al., 1995), rule-driven worki ng memory (Smith, Patalano, & Jonides, 1998), computations in working memory (Smith & Jonide s, 1997), and in spatial selective attention and working memory tasks (Awh & Jonides, 2001). Cons istent with this idea, some researchers have even gone so far as to say that working memory capacity is actually a m easure of PFC integrity (Engle & Oransky, 1999; Kane & Engle, 1998). PFC activation is also found when affective stimuli are processed in a variety of situations: risk perception (Davidson & Irwin, 1999), the Emotional St roop task with words (slower times naming the color of emotional, th an neutral, words) (H errington et al., 2005), affective picture viewing (Northo ff et al., 2000), similarity judgme nts with task-i rrelevant faces (Bishop, Duncan, Brett, & Lawrence, 2004), and simple feature processing of affective pictures (Simpson et al., 2000). More important, however, is research showing interactive effects of affective processing and working memory in PFC activation. These include demonstrations of induced-mood effects on working memory performance (Gray, Braver, & Raichle, 2002), selective attention to subjective emotion responses (Lan e et al., 1997), and e ffects of affective stimuli on modified delayed-match to sample working memory perf ormance (Perlstein, Elbert, & Stenger, 2002).

PAGE 13

13 Such interactions are strong evidence for a relationship between cognition and emotion, since statistical independence suggests functional independence, even if the same brain region shows activation. This paper will build upon the finding of a similar underlying cognitive and neural architecture for emotion and working memory pro cesses with the goal of specifying the nature of this association, and describing possible research approaches th at may help in that effort. The research introduced above stresses the similarity of underlying neural structures for the processing of affective stimuli, execu tive functions broadly, and working memory specifically. While much of this paper elaborates this theme, there is a need to first put this research into a broader memory perspective. There has been a substantial amount of research investigating the effects of emotion on both l ong-term and short-term memory (Kleinsmith & Kaplan, 1963, 1964; Levinger & Clark, 1961; Queve do et al., 2003). Although there are some intriguing exceptions, these studies have generall y found little or no effect of emotion on shortterm memory (e.g., no difference in immediate reca ll of emotional vs. neutral stimuli), while there is often a substantial eff ect of emotion on long term memo ry (e.g., better delayed-recall of emotional vs. neutral stimuli). At a physiologica l level, long-term memory effects have been associated with amygdalar involvement for l ong-term memory conso lidation (Cahill, 1997; Cahill & McGaugh, 1996; Packard & Teather, 1998). Animal research suggests that the modulation of memory by emotion via the amygdala is present in long-term memory but not in working or short term memory, and is seemingly independent of these types of memory (Bianchi et al., 1999). The rate of cons olidation into long-term memory, independent of other encoding or rehearsal processes, also appears to be more efficient for emotional stimuli (Anderson & Phelps, 2001).

PAGE 14

14 While the amygdala may directly modulate memo ry for emotional stimuli, there are also numerous psychological factors th at may contribute to superior long-term memory for emotional stimuli (Buchanon & Adolphs, 2004). Affective s timuli may capture and sustain attention because they are perceptually more arousing (inter esting or important) than are neutral stimuli. Similarly, the depth of encoding an d elaborative processes may be greater for emotional relative to neutral stimuli. Additionally, emotional stimu li may be reviewed and rehearsed more often than neutral stimuli and receive th e benefit of increased retrieval pr actice, as every review of an item increases likelihood of later retrieval of an it em. Emotions may also serve as retrieval cues for information from long-term memory and facilitate retrieval. The research outlined so far supports what we may intuitively believe: Emotion affects how we respond to events, how we think about th em, and how we remember them. What seems to have received less attention in the literature, however, are the effects of emotion on short-term, working memory. Many commonly held ideas (i.e., I need to cool o ff before I respond. or I wasnt thinking. I was caught up in the heat of the moment .) suggest there is an effect of emotion on working memory. The goal of the cu rrent investigation is to determine how, and when, emotion may modulate storage (STM) and process (executive func tions) components of working memory. The use of individual difference variables has been a powerful tool in emotion and personality research. Some of th ese variables reflect potential diffe rences in processing, such as emotional valence or arousal focus (Feldman Barre tt & Gross, 2001), ability to regulate emotion (Gross, 1998), personality (Canli, Zhao, Desmo nd, Gross, & Gabrielli, 2001), gender (KeslerWest et al., 2001), and psychopa thology (Patrick, Cuthbert, & Lang, 1994; Winton, Clark, & Edelmann, 1995), to name a few. Similarly, cognitive researchers have used individual

PAGE 15

15 differences in working memory capacity to better understand the mechanisms involved in working memory (Engle, 2001). There has been little work, however, on i ndividual differences in emotional working memory. Privileged perceptual processing of emoti onal events provides a means of not only indexing occurrences of value but facilitating their availability to other cognitive domains (Dolan, 2002, pp. 1192). Research on the relati onship between emotion and attention suggest that emotion can affect attenti on allocation at the ear liest pre-attentive le vel (Keil et al., 2002; Schupp, Junghofer, Weike, & Hamm, 2003; Smit h, Caccioppo, Larsen, & Chartrand, 2003), as well as in controlled attenti on situations (Andersen, 2005; Ke il & Ihssen, 2005; Ohman, Flykt, & Esteves, 2001). Since emotion tends to bias attention, a process which has b een implicated as a crucial aspect of working memory, one would expect th at emotional stimuli would influence working memory processes as well. The potential usef ulness of the concept of "affective working memory" has been mentioned by some researcher s (Davidson, 1999; Mikels, Reuter-Lorenz, & Fredrickson, 2003; 2004) and addressed, to varying degrees, by othe rs (Gray, Braver, & Reichel, 2002; Kensinger & Corkin, 2003; Mikels, Lark in, Reuter-Lorenz, & Cartensen, 2005). Surprisingly, research efforts on this t opic have been relatively scarce. Recent research investigating the effects of emotion on working memory has generally taken three approaches: (a) manipulating mood and looking at subseque nt working memory performance (Elliman, Green, Rogers, & Finch, 1997; Gray, 2001; Spie s, Hesse, & Hummitzsch, 1996); (b) manipulating emotionality of materials used in non-traditional working memory tasks and looking for differences between neutral and emotional task performa nce (Mikels, ReuterLorenz, & Fredrickson, 2003; 2004); or c) manipulat ing the emotionality of traditional working

PAGE 16

16 memory task materials and looking at differences in performance between neutral and emotional (Kensinger & Corkin, 2003; Perlstein, Elbert, & Stenger, 2002). This investigation will be taking the latter approach. The Current Study Working m emory is considered to be a key ex ecutive function, impacting planning, inhibition, task management, motivation, and other cognitive pr ocesses. The specific definition of working memory varies by theory, but generally working memory is the on-line coordination and use of information for a given cognitive task. One widely accepted model of working memory was first proposed by Baddeley and Hitch (1974). The st ructure of the Baddeley and Hitch model pays tribute to earlier models of human memory proposed by Atkinson and Shiffrin (1968) and Broadbent (1958), which posited distinct stores th at buffered and transformed information en route to long-term memory. The Baddeley and Hitch model consists of a central executive that interacts with two temporary mode-specific inform ation stores, referred to as slave systems: the phonological loop is responsible for verbally coded information and the visuo-spatial sketchpad is responsible for information that has a sp atial representation to be coded. Both slave systems have an active rehearsal mechanism and passive store. The active component of the phonological loop is the articulatory loop, and the passive is calle d the phonological store. The active component of the visuo-spat ial sketchpad is the inner scribe and the passive mechanism is the visual cache. Empirical support for the di ssociation between verb al and visuo-spatial information has come from many areas of resear ch (e.g., Brooks, 1967; Jonides et al., 1996). For example, Brooks (1968) found there was greater interference on verbal tasks when processing verbal information than spatial information, and more interference on spatial tasks when processing spatial information. Additionally, J onides et al., (1996) found hemispheric activity

PAGE 17

17 differences when processing verbal and spatial information, with tasks utilizing verbal tasks associated with left and spatial tasks associated with the right hemisphere. Combined, the components within this model are thought to be necessary for the on-line use of information needed to retrieve, monitor, manipulate, and attain everyday cognitive goals (e.g., reasoning, learning, and comp rehension to name a few). Th is view of working memory places emphasis on the temporary storage and maintenance of information from long-term memory as well as encoding and integration of new information for the benefit of accomplishing current goals. The central executive, according to this model, does not store information, but serves to coordinate the slave systems by c ontrolling attention, encoding, retrieval, and manipulation of information held in both active a nd passive stores within each slave system. More recently, Baddeley (2000) added another component, the episodic buffer The buffer is a limited-capacity component which merg es information from the two slave systems to make a cohesive episodic representation for co mprehension. Since in troduced, the episodic buffer has not received much theoretical or em pirical attentio n; however, it may provide an interesting theoretical framework for investigating how emotion may affect on-line processing. The episodic buffer may be particularly vulnerabl e to emotional influences from both attention and storage mechanisms. Attenti on mechanisms may be impacted by emotion through the central executive. Consistent with the existing literature showing that attention is biased by emotional stimuli, the central executive may serve as an atte ntional filter for what type of information is attended to and put into temporary storage w ithin the episodic buffer. Additionally, emotion could serve as an organizing or prioritizing cue through which in formation in long-term memory could be integrated with temporarily held information, consistent with the integrative nature of storage in the episodic buffer, as Baddeley initially proposed.

PAGE 18

18 Individual Differences in Working Mem ory C apacity and Attention Control Kane, Engle, and Tuholski (1999, 2004) propose a two-factor theory of executive control to explain individual differences in working memo ry. This theory is derived from a series of studies suggesting that working memory capacity (WMC) may be "domain-general," with WMC span tasks predicting a wide range of abilities: note-taking (Kiewra & Benton, 1988), following directions (Engle, Carullo, & Collins, 1991), br idge playing (Clarkson-Smith & Hartley, 1990), computer-language learning (Shute, 1991), and novel reasoning (Kyllonen & Christal, 1991). The two-factor theory of executi ve attention retains domain-specifi c stores with their traditional duties of coding, storage, and rehearsal as well as the domain-gene ral executive attention component which interacts w ith the temporary stores. The two-factor theory of execu tive control model they (Kane, Engle, & Tuholski, 1999; 2004) propose does not differ in st ructure from that of Baddeleys original model; however, it does differ in function. It suggests that individua l differences in working memory are embedded in the functioning of the central executive, wh ich is responsible for goal maintenance and conflict resolution. Kane et al ., posit that working memory capacity is the capacity for controlled, sustained attention in the face of interference or distraction (p. 104, 1999). Accordingly, WMC measures (span tasks) are sensitive to individual differences in central executive functioning, and account for their ability to predict a wide range of higher-cognitive abilities. Those with higher WMC are better ab le to use attention to avoid distraction and perform well on span tasks. Those with low WMC ar e not as skilled at using attention to avoid distraction, and perform poorly on span tasks. This view clearly moves the emphasis away from storage and type of stimuli used in a given span task, and instead focuses on how differences in attention control, because of its crucial impor tance in maintaining current goals and response

PAGE 19

19 conflict resolution, could account for findings of span tasks predicting a variety of higher cognitive functions. Empirical evidence supporting the idea th at individual differences in WMC are attributable to the functioning of the central executive come from several widely-used attention paradigms: the anti-saccade task (Kane, Bleckle y, Conway, & Engle, 2001), Stroop task (Kane & Engle, 2001), negative priming (Engle, Conw ay, Tuholski, & Shistler, 1995) and dichotic listening task (Conway, Cowan, & Bunting, 2001). The findings from these studies highlight two separate functions of WMC: the ability to sustain a task set in memory and the ability to resolve conflicts between a pre-po tent response and the one necessary for the current task. Some of these studies will be reviewed below. Kane and Engle (2003) investigated whether there is a relationship between WMC and Stroop task performance. Briefly, a list of words written in different colors is presented and participants are asked to indicate the color of the word, for both congruent (RED written in red) and incongruent trials (RED written in blue). The typical finding is of faster responses on congruent trials than on incongruent trials. Ka ne and Engle hypothesized that successful Stroop performance requires both memory and attention pr ocesses: memory for the task goal, which was to say the word color, and atte ntion to inhibit the pre-potent re sponse to read the word on every trial, even the congruent ones. Consistent with their two-factor model, Stroop interference would be the result of (a) failure to maintain task set in memory, and (b) difficulty in conflict resolution on incongruent trials. Therefore, Kane and Engl e hypothesized that perfor mance on two different measures that reflect these two processes would be predicted by WMC. To test this idea they varied the list-wise proportion of congruent and incongruent trials within blocks (0% versus 75% congruent). Kane and Engle suggested that in purely congruent or incongruent blocks

PAGE 20

20 participants are able to adopt strategies to optimize perf ormance. For example, in purely incongruent blocks the goal is kept active because of the homogeneity of trials within the block: Every incongruent stimulus ther efore reinforces the goal, to i gnore the word, and so the task environment acts instead of the central executiv e. They predicted th at the 75% congruent blocks would encourage higher goal maintenance processes, relative to the 0% block. Kane and Engle took a categorical approach to classifying the working memory capacity of their participants by incl uding those scoring low or high on a widely-used WMC measure called Operation span (OSPAN). This a pproach is sometimes called extreme-group methodology. OSPAN is a measure of WMC in which participants are asked to read aloud a mathematical operation and verify whether the answer provided is correct, then to read the word immediately following the answer. One typical operation-word string is, Is (9/3) + 2 = 5? Drill. Participants continued reading items aloud, without pauses, until they are presented with a ????? prompt, indicating that they were to reca ll all words that follow ed the operations in a trial. The number of operat ion-word strings varied randomly between 2-5 items by trial. OSPAN score was determined by the number of sets with correct items recalled in the correct order, so scores range from 0-42 (across three sets of stimuli). The low and high WMC groups were defined as the bottom and top quartile of OSPAN participants. Typically this includes those scoring below 9 and above 19. Kane and Engle (2003) found that there were systematic differences between low and high working memory capacity groups. Specifica lly, in the 0% congruent condition, they found that the low-span group exhibited greater res ponse time interference than the high-span group that is, the low-span group took longer on the incongruent trials relative to the neutral colornaming trials, suggesting they were slower at resolving the conflict between the pre-potent

PAGE 21

21 reading response the task response of color-n aming. In the 75% congruent condition, the opposite pattern occurred, with greater respon se time interference for the high-span group relative to the low-span group, suggesting the high-span group wa s taking longer to resolve the conflict between the pre-potent re sponse and the task response. The response time data on their own would tell a confusing story, but when Kane a nd Engle analyzed the e rrors rates across each trial type, they found greater e rrors in the low-span group th an the high-span group for both conditions. Combined, this study suggests that individual differences in working memory capacity can predict performance across two types of tasks, one that requires sustained attention in the face of distraction (0% congruency condi tion), and another that requires sustained activation of the task set to perform with accuracy (75% condition). Emotion and Attention Researchers studying the relation of emoti on and cognition have been particularly interested in how attention is allocated in the face of dist raction. It has been found that presentation of task-irrelevant em otional stimuli tends to result in an i nvoluntarily allocation of attentional resources to those stimuli (i.e., atte ntional capture). Physio logical measures taken during such tasks have shown that the processing of emotional pictures differs from that of neutral pictures (Keil et al., 2002; Schupp, Junghofer, Weik e, & Hamm, 2003; Smith, Caccioppo, Larsen, & Chartrand, 2003). Pictures of threatening stimuli also result in faster detection in visual search task s: an angry face in a crowd (Han sen and Hansen, 1988), a snake in the grass (Ohman, Flykt, & Esteves, 2001), a sc hematic emotional face among schematic neutral faces (Ohman, Lundqvist, & Esteves, 2001), or a photograph of an emotional face among photographs of neutral faces (Jut h, Lundqvist, Karlsson, & Ohman, 2005). Further support for attentional capture by em otional stimuli comes from the so-called Emotional Stroop task, where color-naming of emotionally evocative words is slowed relative to

PAGE 22

22 neutral words (Dalgleish, 1995; Pratto & John, 1 991; Williams, et al, 1996). Recently, there has been debate about the similarity and differen ces in the processes involved in the emotional Stroop task (Algom, Chajut, & Lev, 2004; Chajut Lev, & Algom, 2005; Dalgleish, 2005). The details of this debate are beyond th e scope of this dissertation, but th e gist of the debate is rooted in whether the cognitive processe s that result in slowed color naming found in the traditional Stroop task (1935) are the same cognitive processe s involved in the emotional Stroop task. In the traditional Stroop task there are both congruent ( red written in red) and incongruent trials ( red written in blue) with faster responses on the former when compared to the latter. In contrast, in the emotional Stroop task, there are neither cong ruent nor incongruent tr ials, as defined above; rather, there are either emotional or neutral words printed in vary ing colors. An elegant way to investigate the relative attention processes invo lved in color naming of neutral and emotional words was reported by McKenna and Sharma (2004). McKenna and Sharma (2004) noted that ther e has been a wide range of interference effects or slowed responding to emotional wo rds, reported (-1 ms to 400 ms), across the emotional Stroop studies (as reviewed by W illiams, Mathews, and MacLeod, 1996). McKenna and Sharma noted that the wide range of interf erence may be due in part to group differences (psychopathologies) and/or to specific methodol ogies used: presentation (card, computer, Tscope), design format (blocked, missed, individual, simultaneous), stimuli (negative, positive, neutral, category neutral), stimulus repetition (a small set of words repeated often, or a larger set repeated less often), intertrial interval (ITI), and so on (p. 383). They posit that the widespread consensus of the emotional Stroop effects bei ng due to the automatic grabbing of attention (Pratto & John, 1991) by emotional words and the momentary suspension of current content processing is at best not the full st ory. As an alternative, they s uggest that there are two different

PAGE 23

23 temporal effects at work in emo tional Stroop; one is relatively fa st, occurring within a trial, and the other slow, occurring between trials. McKenna and Sharma (2004) directly tested the relative contributi ons of both fast and slow effects of emotion on word naming latency w ithin a modified emotional Stroop task (their Experiment 4). There were seven word positions, with all but the first holding neutral words; Position 1 was either a positive (e.g., ROMANT IC), negative (e.g., REJECTED), or neutral (transportation-related) word. Color-naming latency was measured for each position. Latency to name the emotional word at Position 1 was not slower than latency for positive or neutral words, suggesting that the fast, within-trial effects of word emotionality we re at best small. In contrast, latency to name neutral words in Position 2, was substantially slower when following an emotional word, than when following a neutral word. The size of the difference was surprisingly large (75 ms). McKenna and Sharma suggest that the slow effects tend to be the predominant effect seen in non-clinical samples, whereas the fast effects may be reliably found only in clinical samples (e.g., PTSD and Schizophrenia). Research Questions The main goal of this research was to e xplore the impact of emotion on working memory processes. To that end, a measure of working memory that incorporated emotion was developed. The emotional working memory task was identical to the Operation Span task used by Kane and Engle (2003), except the Emotional Operation Span task (ESPAN below) required participants to remember a sequence of emotionally unpleasant word s, instead of the emotionally neutral words used by Kane and Engle and others. Th ese words can be found in Appendix C. Using the extreme-group methodol ogy noted above, individual differences in emotional working memory capacity were related to pe rformance on two kinds of tasks involving potential interference and distract ion. The first involves the two factors of attention that Kane

PAGE 24

24 and Engle argued were central to working memory : maintenance of task set, and maintaining attention in the face of distraction (the Congruence Stroop task). The second involves maintenance of attention in the face of emotional words (the Emotional Stroop task), which previous research sugges ts distract attention. Within this framework, a number of specific hypotheses were proposed. These are outlined in Figure 1-1. First, I sought to re plicate two Stroop-related findi ngs: (a) Stroop interference is sensitive to context effects (i.e ., variations in the list-wise co ngruency proportion; Kane & Engle, 2003) and (b) the time course of Em otional Stroop effects may be slower than originally thought (McKenna & Sharma, 2004) (Hypotheses 1a and 1b). Second, I explored whether the ESPAN task w ould capture an aspect of working memory capacity that is different than traditionally neutral measures of working memory capacity (neutral working memory measure, OSPAN), and so predic ted that performance on the two tasks should be at least to some degree dissociable (Hypothesi s 2). Third, the current study sought to replicate the relationship between working memory capaci ty and attention control reported by Kane & Engle (2003) (Hypothesis 3a). Consistent with the hypothesis th at the ESPAN measure captures something unique about working memory, it should predict attention cont rol in the face of emotional distraction; specifica lly, individual differences in ESPAN should predict performance in the Modified Emotional Stroop task (Hypothesi s 3b). Moreover, ESPAN should be a stronger predictor for performance in the Emotional St roop task than the OSPAN measure (Hypothesis 4a), and the OSPAN will be a stronger predicto r of performance in the Congruency Stroop than the ESPAN task (Hypothesis 4b).

PAGE 25

25 Figure 1-1. Experimental hypot heses: (1a) Congruency Stroop will be replicated. (1b) Emotional Stroop will be replication. (2 ) OSPAN and ESPAN will be associated, but ESPAN will capture variance associated w ith emotion processing. (3a) Individual differences in OSPAN will predict Congruency Stroop performance. (3b) Individual differences in ESPAN will predict Emotiona l Stroop performance. (4a) Individual differences in OSPAN will predict Emotional Stroop performance. (4b) Individual differences in ESPAN will predict Congruency Stroop performance better than OSPAN. [H3a] [H2] [H4a] [H4b] CONGRUENCY STROOP EMOTIONALSTROOP OSPAN ESPAN [H1a] [H1b] [H3b]

PAGE 26

26 CHAPTER 2 METHOD Participants Upon approval from the University of Florid a Institutional Review Board (protocol 2007U-0821), one-hundred and forty-se ven undergraduates were recru ited from the Introductory Psychology Participant pool for par ticipation. Participants that did not complete all portions of the study were excluded from analyses resulting in a one-hundred and twenty-five participants. Design The present experim ent included two types of Stroop tasks: One was the Congruency Stroop and the other was the Emotional Stroop. Within the Congruency Stroop task there were two blocks of trials: One is the 0% Congruency Block that has two types of trials (incongruent and neutral), and the other is the 75% Congrue ncy Block that has three types of trials (incongruent, congruent, and neutral) Within the Emotional Stroop task there were three blocks of trials defined by the emotionali ty of the critical word in Pos ition 1 (Pleasant, Unpleasant, and Neutral Transportation). Each word in the se quence was given a position number of 1-6, with the critical word being position 1 and neutral fi ller words appearing randomly at positions 2-6 (See Appendix C). Participants also comple ted a neutral (OSPAN) and emotional working memory capacity measure (ESPAN). All participants completed all of the above tasks; however, to look at individual differences the sample is split (top third and bottom third) based on their OSPAN and ESPAN scores. Materials A Dell computer was used for all tasks: The m odified emotional Stroop task (McKenna & Sharma, 2004), Congruency Stroop tasks (Kane & E ngle, 2003), the neutral Operation span task

PAGE 27

27 used by previous researchers (Unsworth, He itz, Schrock, & Engle, 2005), the emotional Operation Span task developed for this study, and for the word ratings task. Operation Span Task Participants read aloud a m athematical operati on and a potential answ er ((9/3) + 2 =6) and indicated whether the answer provided was correct (true) or incorrect (fal se). They then read a neutral word (dock) and retain it until asked to recall all words since last recall. Participants continued reading items aloud, without pauses, until they were presen ted with a ?????? prompt. A recall screen followed, at which time they attempted to recall all words since the previous recall. They were asked to recall the words in the order in which they received them. For example, if they knew that they had seen fi ve words but could only remember three of them to leave the blanks empty for the words that they couldnt remember. Participants were encouraged to keep their overall mathematical performance above 85%, and received feedback as to their overall mathematical accuracy and mean response time. The number of operation strings (and words) participants received (i.e., me mory set) varied randomly from 2-5 items. This continued until all trials were completed (total of 48 trials, including practice trials). See Appendix B for the full set of stimuli. The main dependent measures of interest in study include: (1) Overall mathematical performance accuracy and (2) OSPAN score, define d as the number of sets with all presented items, in the correct order. There may be some other theoretically interesting data, such as the number of intrusions or omissions, but these will not be considered here (see Unsworth & Engle, 2007).

PAGE 28

28 Emotional Operation Span Task This task is identical to the Operations Span task described above, except instead of using neutral words as m emoranda, unpleasant words were used. All words were obtained from the Affective Norms for English Words (Bradley & Lang, 1991). Emotional Stroop Task Participan ts were asked to view words one at a time and indicate the color each word was printed in. The word was presen ted in blue, brown, red, or gree n. Responses were made with the index and middle finger of both hands. One of four colors was indicated on stickers (RED, GREEN, BROWN, and BLUE) over the keys of d, f, j, and k respectively. Participants were encouraged to indicate the color as quickly and accurately as possible, and not to correct any mistakes they made while completing the task. Pa rticipants received res ponse training before the experimental trials began. The training included making responses with the four keys for a total of 20 trials. For every sequence there were six po sitions (1-6). In position 1 a critical word is presented that is emotionally unpleasant, pleasant, or neutral (t ransportation-related) word. In positions 2-6 are neutral (e.g., non-transportation-re lated) words. Please see Appendix C for a complete list of task stimuli. The type of crit ical word presented in Position 1 was blocked. Within each block, five sequences were pres ented randomly six times. After each block, participants were given a brief br eak (30-60 seconds). In total ther e were 210 experimental trials. The dependent variables in this task were (a) correct color-naming latency, and (b) color-naming errors. Congruency Stroop Task This is the classic Stroop task (1935), wher e participants are asked to nam e the color of the presented word orally, as quickly and accurately as possible. Specifically, participants named the word color, and then trained research assistants indicated the participant response by

PAGE 29

29 keyboard press. Research assistants did not l ook at the computer scr een while participants named the word color, rather they focused on making their responses on one of three possible keys on the keyboard that were each labeled with one of three colors: RED, BLUE, or GREEN. The only way that this task differs from the origin al Stroop paradigm in that it consists of two blocks that differ in their list-wise proportion of congr uent trials. One has no congruent trials (where the word name and co lor are the same, 0% congruency) and the other has 75% congruency. Participants always received the 0%, followed by the 75% congruency block. There are three possible types of critical tria ls presented: congruent, incongruent, and neutral. The congruent trials include the words RED, BLUE, GREEN, in those colors, with each presented 12 times, for a total of 36 trials. The incongruent trials include the same words only presented in all of the possible conflicting colors, with each c onflicting color being presented 6 times. The neutral trials include the letter strings of JKM, XTQZ, and FPSTW, with each of them being presented 4 times in each of the three colors. In the 0% congruent block, there are 144 trials: the 36 critical incongruent trials, 36 critical neutral trials, and 72 filler trials that are incongruent. The 75% congruent bl ock has 288 trials, which includes al l three of the critical trial types (3 x 36 trials) and 180 filler trials that are congruent. The dependent variables in this task are: (a) correct color-naming latencies, and (b) color-naming errors. Word Ratings Task Participan ts viewed words, one at a time, and made valence and ar ousal ratings on a fivepoint Likert scale. For vale nce, a 1 indicates the particip ant perceives the word as very unpleasant, whereas a indicate s the participant vi ews the word as highly pleasant. For arousal, a indicates the pa rticipant perceives the word as non-arousing, whereas a indicates the participan t views the word as highly arousing.

PAGE 30

30 Procedure There were two experimental sessions, one on each of two consecutive days. Session I was one hour in duration, while Session II was an hour and a half in duration. The order in which participants completed the measures wi thin a given session was counterbalanced across subjects. Regardless of the order of measures, Session I consisted of participants giving their informed consent and then completing one of the Stroop Tasks (Congruenc y Stroop, or Modified Emotional Stroop), followed by one of the meas ures of working memory capacity (OSPAN, or ESPAN), and Session II consisted of participants completing the other Stroop Tasks, followed by the other measures of working memory capacity. Session II ended with the Word Rating Task. On each day, only one of the two main tasks invo lved emotional words; that is, the ESPAN task was always paired with the C ongruency Stroop tasks, and the OSPAN task was always paired with the Modified Emotional Stroop task.

PAGE 31

31 CHAPTER 3 RESULTS Neutral and Emotional Operation Span Task Participant perform ance was scored using a strict scoring method, whereby a given recall event was deemed correct if and only if they recalled all words in the order in which they received them. The range of possible scores fo r both the neutral and emotional Operation Span task was 0-42. In the current sample the range of Neutral Working Memory Capacity scores was 2-39 and the mean was 17.54 (standard deviation = 8.04). The range of scores for the Emotional Working Memory Capacity scores was 0-37 and the mean was 15.27 (standard deviation = 6.60). The difference in OSPAN and ESPAN scores was significant, (paired ttest, t(124) = 3.50, p < .01), suggesting that overall, having to maintain emotional words in working memory is more difficult than maintaining neutral words. Math ematical operation accuracy was equal for both ESPAN and OSPAN, approximately 95% (paired t-test, t (124) = -.57, p = .565), indicating that the processes being impacted by the inclusion of emotional stimuli is specific to the temporary maintenance of the words. The current study is aimed at investigating individual diffe rences in working memory capacity and attention control, the full sample was divided into groups using the extreme group methodology (consistent with Kane & Engle, 2003) Specifically, the top third and bottom third of performers were used in the individual diffe rences analyses presented below. Within the Neutral Operation Span task the low span group included participants who scored at or below 13 (n = 47), whereas the high span group included part icipants who scored at or above 19 (n = 52). For the Neutral Operation Span the mean span score for the low span group was 10 and the mean high span group score was 25. Within the Emotional Operation Span task the low span group included participants who scored at or below 12 (n = 49), whereas the high span group included

PAGE 32

32 participants who scores at or above 17 (n = 45). The mean sp an score for the low emotional working memory group was 9, and the mean for the high span group was 22. There was a correlation of .45 between the O SPAN and ESPAN task, t hus the individual differences analyses below reflect some degree of overlap in subjects, please see Table 3-1. There are no existing reliabilities for the ESPAN task because it was a novel measure created for assessing whether there is an emotional working memory. Prior research using the OSPAN task has shown reliabilities (Cronb achs alpha) to be approximate ly .65-.75 (Conway et al., 2002; Engle, Cantor, & Carrullo, 1992; Engle, Tuholsk i et al., 1999; LaPointe & Engle, 1990). Additionally, the test-retest measur e have demonstrated stability fo r a wide range of intervals: a few minutes (r = .77-.79)(Turley-Ames & Whitfiel d, 2002), 3 weeks (stability coefficient, r = .82) (Klein & Fiss, 1999), and 3 months (r = .76 ) (Klein & Fiss, 1999). The correlation between OSPAN and ESPAN addresses Hypothesis 2 in Figure 1-1. An informal comparison of the correlation between ESPAN and OSPAN obtained in the current study to previous test-retest reliabilities of the OSPAN suggest there is a fair amount of unique variance that can be attributed to the cognitive processes that the ESPAN is measuring. Emotional Stroop Task In this task, participan ts completed three bloc ks of the color-naming task with a series of six-word sequences, each with a critical word in the first position, and neutral words in positions 2-6. The blocks were defined by the emotionality of the position 1 critical word. Critical (position 1) word emotionality could be Pleasan t, Unpleasant, or Neutra l (transportation-related words). The mean response time for the colo r-naming task was found for each position within each block. All participants complete d each block of trials in a randomly determined order. It is important to note that, upon completion, none of th e participants spontane ously reported noticing any emotional words in the tasks. When probed as to whether they noticed the emotional words

PAGE 33

33 in the sequence, only a handful reported being aw are that there were Unpleasant or Pleasant words in the task. The results of the current study will be presen ted as follows: The color-naming error rates will be presented first, followed by the correct color-naming response latencies. The overall (n = 125) results will be presented first. The overall analyses presented below address Hypothesis 1b in Figure 1-1, which asked whether the current study replicated the time c ourse of the Emotional Stroop effects reported by McKenna and Sharma ( 2004). The individual di fference analyses that follow the overall analyses will be presented fo r both the neutral working memory groups (low OSPAN n = 47, high OSPAN n = 52) followed by th e emotional working memory groups (low ESPAN n = 49, high ESPAN n = 45). The individua l differences in neutral working memory analyses that follow address H ypothesis 4a in Figure 1-1, whereas the individual differences in emotional working memory analyses address Hypothesis 3b in Figure 1-1. Overall Error Rates As indicated above, McKenna and Sharm a found no differences in er ror rates across the three emotionality blocks, with rates of 3.63% for the Neutral (Transportation-related) blocks, and 3.50% for the Pleasant and Unpleasant blocks. In the current study the mean error rates showed a different pattern, with greatest erro rs in the Unpleasant block (5.2%), followed by Pleasant (3.5%) and Neutral (transportation-rela ted) (3.4%) blocks, indi cated by the significant main effect of emotional block (F(2, 123) = 8.21, p < .001), please see Figure 3-1. Pair-wise comparisons of errors between blocks reveal th at more errors were ma de in the Unpleasant blocks than in either Pleasant blocks (p < .001) or the Neutral (Transpor tation-related) blocks (both ps <.001), while the latter two did not differ (p = .687). There were no differences across position (p = .573), nor was there an y differential effect of block across positions (interaction of block and position, p = .568). Importantly, then, more errors were made during sequences which

PAGE 34

34 contained unpleasant words, but this effect was not limited to the critical Unpleasant words themselves. Individual Differences in Error Rates The error rates above reflect the whole sam ple. The sections below are the results of using the extreme-group methodology of comparing those that are high in working memory capacity, to those that are low in working memory capacit y. The results for the individual differences in working memory capacity will begin with the ne utral or traditional working memory capacity groupings, using the Operation Span task, followed by the results when the sample is divided on performance on the emotional work ing memory capacity measure, the Emotional Operation Span task. Individual Differences in Neutral Working Memory Capacity (OSP AN) As can be seen in Figure 3-2, the two OSPAN groups did differ in their overall er ror rate performance, seen in the significant between -group effect (F (1, 97) = 139.23, p < .001). Although the low OSPAN group produced more errors overall, the distribu tion of errors across emotional block and position was essentially id entical for the two Span groups. This is confirmed by the absence of interactions betw een Span group and the other two factors (Span x Emotional Block (F < 1, p = .907, the three-way interaction between Emotional Block x Position x Span Group (F(10, 97) = 1.22, p = .23). Individual Differences in Emotiona l Working Memory C apacity (ESPAN) In contrast to the division of participants by OSPAN, the small difference in overall accuracy between high and low ESPAN groups (c 1.5%) was not significant,( F(1, 92) = 133.41, p = .095), see Figure 3-3. Similarly, the effect of Unpleasant words in the sequence on error rates was overall not greater for the low ESPA N group (Span x Emotional Block interaction, F(1, 92) =1.22, p = .297), but importantly, was more sustained across position for the low-span group

PAGE 35

35 (three-way interaction between Emotional Bloc k x Position x ESPAN group (F(10, 92) = 1.91, p =.039). Follow-up simple effects of Emotion Bl ock at positions 1 and 2 showed no differences for either the low or high ESPAN groups. In co ntrast both groups showed an effect of Emotion Block at position 3 (low ESPAN, F (2, 96) = 7.31, p < .01; hi gh ESPAN, F(2, 88) = 3.16, p < .05 ). Pairwise comparisons showed for bot h the low and high ESPAN group that neutral (transport-related) block resulted in less errors than pleasant or the unpleasant words (ps < .01). The remaining positions of 4-6 failed to show any systematic ESPAN differences in errors. Overall Response Latencies McKenna and Sharm a found significant differe nces in correct-color naming response latencies between emotional blocks that varied by word position, with the greatest differences being at position 2 when participants were co lor-naming a neutral wor d. In the current study, the response times overall were somewhat slow er for the Pleasant block (Mean RT = 845) and the Unpleasant block (Mean RT = 844), relative to the transport block (Mean RT = 836), however these differences were not significant (main effect of block, F(2, 246) = 1.50, p = .224), please see Figure 3-4. Response times were slower in the earlier positio ns (main effect of position, F(5, 615) = 3.83, p = .002); pair-wise co mparisons of position showed that position 1 was significantly slower than all other pos itions (p-values rangi ng from .001-.050). The interaction between block and position was not significant (F (10, 123) = 1.74, p = .066), but there was a tendency for the emotional blocks to elicit slower response times across positions 24. Follow-up analyses looking at the effect of Emotional Block at positions 1, 2, and 3 were completed. These analyses showed no effect of emotion at position 1 (p = .835) and 3 (p = .371) but was on the cusp of significance for position 2 (F (2, 246) = 2.92, p = .055). This is similar to what McKenna and Sharma reported with the emotionality effect gr eatest at the position following the critical wor d, and absent for the critical word itself.

PAGE 36

36 Individual Differences in Response Latencies As was done with the error rates, effects of individual differences in W MC on correct colornaming response latencies will be presen ted below using two different measures of working memory capacity to divide the sample into high and low working memory capacity groups: One measure will be an emotionally ne utral measure (Operation Span), and the other measure will be include emotional stimuli (Emo tional Operation Span). These response time analyses will be discussed in turn below. Individual Differences in Neutral Working Memory Capacity (OSP AN) The effects of OSPAN capacity on correct co lor-naming in the modified Emotional Stroop task were analyzed in a mixed-model analysis of variance (ANOVA) w ith block and position as within-subject variables and sp an group (low and high OSPAN groups) as the between-subjects variable. As can be seen in Figure 3-5, the two OSPAN groups did not di ffer in overall response latencies (main effect of OSPAN group F(1, 97) = .58, p = .446), nor in the effects of emotional block, position, or their combination (Block x OSPAN group, Position x OSPAN group, Block x Position x OSPAN, all Fs < 1.00). Combined, these analyses suggest that individual differences in neutral working memory cannot predict response latencies in an Emotional Stroop task. Individual Differences in Emotiona l Working Memory C apacity (ESPAN) The effects of ESPAN capacity on correct colo r-naming latency in the Modified Emotional Stroop task were analyzed in a mixed-model analysis of variance (ANOVA) with block and position as within-subject variables and span group (low and high ESPAN groups) as the between-subjects variable. As can be seen in Fi gure 3-6, the two ESPAN groups did not differ in overall response latencies (main effect of Sp an group, F(1, 92) = 1.14, p = .287). Recall, that these group differences in response latencies are similar to those in the OSPAN analyses, with the low span groups having longer latencies over all than the high span groups. There were no

PAGE 37

37 ESPAN group differences in pattern of response latencies across either block or position (n.s. block x span group and position x span group). Ho wever, in contrast to the neutral working memory capacity analysis, there was a significa nt three-way interaction between block, position, and span group (F(10, 92) = 1.95, p = .035). Follow -up analyses of the high span group response latencies showed that th e greatest effect of Emotional Bloc k is at position 1 (F (2,88) = 3.15, p < .05), please see Figure 3-6. All other follow-up analyses on Emotional block at each position failed to reach significance. Thus for both error rates and response latencies, WMC as assessed by the ESPAN task, but not the OSPAN, is predictiv e of the effects of the presence of emotional words on color-naming performance. Congruency Stroop Kane and Engle (2003) found that the order in which participants received the two list-wise proportion blocks of 0% and 75% in fluenced whether individual di fferences in a ttention control were seen in the errors or the response latencies. If the partic ipants received the 0% congruent condition before the 75% condition, the individual differences in working m emory capacity were found in the response time interferen ce and not in the error rates. In contrast, if participants received the 75% congruency condition before th e 0% condition, the differences were found in error interference and not in the response latencies. In the current study participants always received the 0%, followed by the 75% congruency block, so we might expect that any effects of Span group would be seen on response latency rather than accuracy. Nonetheless, both depe ndent measures were analyzed. To determine whether individual differences in working memo ry capacity affected performance, a mixedmodel analysis of variance w ith span group as a between-sub ject factor, and trial type (incongruent and neutral) and congruency (0% or 75% congruent) as the within subject factors. The color-naming error rates will be presented first, followed by the response latencies. As with

PAGE 38

38 the Modified Emotional Stroop task above, the overall (n = 125) results for each dependent measure will be followed by the individual differe nce results for that measure, for both the neutral working memory groups (low OSPAN n = 47, high OSPAN n = 52), and then the emotional working memory groups (low ESPAN n = 49, high ESPAN n = 45). The analysis of individual differences in neutra l working memory that follow addr ess Hypothesis 3a in Figure 11, whereas that of the individual differences in emotional working memory address Hypothesis 4b in Figure 1-1. Overall Error Interference The overall effect of congruenc y on error rate interference is presented in Figure 3-7. As the Figure s uggests, the classic Stroop effect wa s observed, with incongruent trials resulting in greater errors relative to neut ral trials across span groups (incongruent error rates = .10 and neutral error rate = .02), suppor ted by a significant effect for trial type (F(1,124) = 106.16, p < .001). Additionally, error interference was much greater (on the incongruent trials relative to neutral trials) within th e 75% congruency condition (incongruent Mean Error Rate = .16) than in the 0% congruency condition (inc ongruent Mean Error Rate = .06) seen in the significant main effect of block type (F(1, 124) = 180.50, p < .001). Thus the incongruent trials are more difficult when they are presented in the context of a large proportion of congruous trials, supported by a significant interaction between trial type and congruency bl ock (F(1, 124) = 102.02, p < .001). In contrast, the neutral trials seem to be unaff ected by the list-wise congruency proportion (Mean Error Rate = .024 for both the 0% and 75% conditions). Interestingly, the congruent trials had slightly fewer errors (Mean E rror Rate = .02), suggesting that the match between the semantic content of the congruent trials wa s helping or facilitating performa nce slightly. This facilitation is mentioned in the results noted by Kane and E ngle (2003), which they expl ained as due to there being no semantic interference or response competi tion on these trials, or that the participants

PAGE 39

39 were simply reading the words and not color-naming, thus showing greater accuracy on the congruent, relative to the neutral trials. Individual Differences in Neutral Working Memory Capacity (OSP AN) In the results that follow, the interference effect will be presented directly in the figures as difference scores between the incongruent and ne utral trials, although tria l type (incongruent versus neutral) was included as a factor in the ANOVAs. The error interference scores for the low and high OSPAN groups are shown in Figure 3-8. The mixe d-model ANOVA (noted above) showed greater errors for the low span group across congruency condition (Mean interference for low-span group, 13%; mean interference for high span group, 7%), supported by a significant interaction between trial type and span group (F (1, 97) = 11.22, p < .001). There was no interaction between list-wise congrue ncy proportion and OSPAN groups (p = .279), nor was there a significant three-way interaction between trial type list-wise congruency proportion and OSPAN group (p = .143). There was, however, a significant in teraction between trial type and OSPAN group (F(1, 97) = 11.22, p < .001), w ith the low OSPAN group showing larger errors for the incongruent, relative to the neut ral trials, for both the 0% to 75% congruity conditions than did the high-span group (see Figure 3-8). Kane and Engle observed WMC effects on response latency and not error rates when the 75% congruent condition followed the 0% condition. The present effect on erro r rates replicates their findings, with low-span individuals less able to maintain attentiona l task set in the face of a large proportion of congruent trials. Individual Differences in Emotiona l Working Memory C apacity (ESPAN) The error interference scores for the low and high ESPAN groups are shown in Figure 3-9. The mixed-model ANOVA showed that the lo w and high span groups had similar error interference rates across Congrue ncy blocks (Mean interference for low-span group, 12%; mean interference for high span group, 9%). In contrast to the neutral working memory results, there

PAGE 40

40 was no significant interaction of sp an group with any variable (trial type x span group (F(1,92) = 2.14, p = .147), congruency proportion x span gr oup (F(1,92) = .001, p = .982), trial type x proportion x span group (F(1,92) = .39, p = .531). Follow-up Regression Analyses on Error Interference The following analyses address Hypothe sis 2. in Figure 1-1. Stepwise regression, using the full sam ple of participants (n=125), wa s used to compare the predictive power of two different models. Model A, looked at how well OSPAN scores alone and then OPSAN and ESPAN combined, predicted erro r interference in th e 0% and 75% Congruency Blocks. Model B looked at how well ESPAN alone, versus ESPAN and OSPAN combined could capture variance in predicting er ror interference in the 0% and 75% Congruency Blocks. Please see Appendix D for detailed regression tables. The regression results show that OSPAN scores account for about 4.2% of the variability in predicting error interferen ce for the 0% Congruency Block, whereas the full model including both OSPAN and ESPAN only accounts for 4.4%. Th is suggests that ESPAN scores are not contributing much toward predicting these scores. We see a similar pattern for the error rates in the 75% Congruency Block where the model with OSPAN alone accounts for 3.7%, whereas the model with both OSPAN and ESPAN only accounts for 4%. Model B was not a significant predictor of error rates for ei ther congruency block. Overall Response Latencies Mean correct response latencies for all participants were entered into an analysis of variance, with trial type (inc ongruent and neutral) and congrue ncy (0% or 75% congruent) as within-sub ject variables. Extreme latencies were corrected before any analyses, whereby trials with latencies faster than 250 ms were omitted, and latencies greater than 1.7 times the median were set to that value. Overall, across congruency blocks, there is a classic Stroop effect

PAGE 41

41 present, with color-naming task taking longer on the incongruent trials (Mean RT = 1311 ms) relative to the neutral trials (M ean RT = 1153 ms), supported by a si gnificant effect for trial type (F(1,124) = 876.47, p < .001), please see Figure 310. Additionally, there was a significant interaction between trial type and congru ency block (F(1, 124) = 128.83, p < .001) with incongruent trial response latencies longer in the 75% congruency bl ock (Mean RT = 153 ms) relative to the 0% congruency block (Mean RT = 138 ms). Interestingly, the neutral trial response times remain relatively constant across list-wise congruency blocks (CS0% Mean RT = 1147 and CS75% Mean RT = 1160). In the 75% c ongruency block, when congruent trials were presented, their response latencies are relativel y quick (Mean RT = 1083 ms), suggesting that when the lexical content of th e trials is consistent with the task goal of color-naming, performance is facilitated. This idea of facilitation will be discussed within the context of individual differences in working memory capacity and attention control below. Response Time Interference Response tim e interference scores were obt ained by subtracting the mean color-naming latency for neutral items from color-naming latency for incongruent items. The overall mean interference scores will be presented below, followed by individual differences analyses of the neutral traditional working memory capacity measure, and th en by the emotional working memory capacity measure. Response time interf erence increased with increased congruency proportion (CS0% RT interference = 114 ms, and CS 75% RT interference = 200 ms), as seen in the significant main effect of congrue ncy proportion (F(1, 124) = 122.07, p < .001). Individual Differences in Neutral Working Memory Capacity (OSP AN) The interaction between span group and congruency proportion failed to reach significance (F (1, 97) = 3.17, p = .078), and the interaction between trial type and span group was not significant (p = .993). The three-wa y interaction between trial type, list-wise

PAGE 42

42 congruency proportion, and OSPAN group, however, was significant (F (1, 97) = 5.37, p =.023): interference scores were slightly lower for the low-span than high-span group in the 0% congruency condition, but this difference was reve rsed in the CS75% condition (see Figure 3-11) suggesting that working memory capacity groups had different patterns of performance for each trial type between the two congruency proportion blocks. Individual Differences in Emotiona l Working Memory C apacity (ESPAN) Mean response latencies were entered into a mixed-model analysis of variance, with span group as a between-subject factor and trial type (incongruent and neutral) and congruency (0% or 75% congruent) as the within subject factor s. Response times were equivalent across Congruency blocks for the ESPAN groups, as seen in the lack of a significant interactions between span group and congrue ncy proportion (F (1, 92) = .23, p =.628). In contrast, the interaction between span group a nd trial type (F(1,92) = 4.74, p = .032) was significant, with the low span group response latencies longer than th e high span group for both the incongruent and neutral trials, see response time interference in Figure 3-12. The three-way interaction (F(1, 92) = 2.93, p =.090) was not significant. Combined, these effects suggest that the emotional working memory capacity measure may be a weak predic tor of response time performance within the Congruency Stroop task. Follow-up Regression Analyses on Response Time Interference The following analyses address Hypothesis 2 in Figure 1-1. Step wise regression, using the f ull sample of participants (n=125), was used to compare the predictive power of two different models. Model A, looked at how well OSPAN scores alon e versus OPSAN and ESPAN scores combined could predict interferen ce effects in response latencies in the 0% and 75% Congruency Blocks. Model B looked at how well ESPAN alone, versus ESPAN and

PAGE 43

43 OSPAN scores combined could capture variance in predicting response ti me interference in the 0% and 75% Congruency Blocks. Please see Appe ndix E for detailed regr ession tables. The regression results show that neither Mode l A nor Model B is a significant predictor of response time interference in the 75% Congruency Block. In contrast, for the 0% Congruency Block ESPAN is a better pred ictor than OSPAN, where in Model B, ESPAN accounts for 8.9% of the variability, with the addition of OSPAN to the model adding only .5% of variance accountability (resulting in 9.4%). When Model A is looked at, we see that OSPAN alone only accounts for 3.8%, before adding ESPAN to account for 9.4%. This suggests that EPSAN is in at least one instance measuring some aspect of a ttention control in a way that is more sensitive that the OSPAN measure to pred ict response time interf erence in the 0% Congruency this is likely the ability to maintain attention in the face of distraction. Overall Response Time Facilitation Facilitation is a difference sc ore obtained from subtracti ng the mean correct response latency of neutral trials from the mean res ponse latency of the congr uent trials. Individual differences in working memory are analyzed to see whether the amount of facilitation within the 75% congruent condition varied with ESPAN or OSPAN scores. The mean facilitation response times were analyzed in a mixed-model analysis of variance with span gr oup (low and high) as the between-subject factor and trial type (congruent and neutral) as the within-subject factor. As mentioned in earlier analyses, th e congruent trials had faster latencies than the neutral trials across working memory span groups (congruent mean RT = 1082 ms, neutral mean RT=1162 ms) and is supported by a significant main eff ect of trial type (F (1, 124) = 245.06, p < .000). Individual Differences in Neutral Working Memory Capacity (OSP AN) The main effect of span group was not significant (F(1,97) = 1.18, p = .280), please see Figure 3-13. There was a significant interaction between trial type and span group (F(1,97) =

PAGE 44

44 7.69, p = .007). The low span group had faster response times for the congruent (Mean RT = 1100 ms) than the neutral trials (Mean RT = 1189 ms), whereas the high span group showed the opposite pattern, with greater response times for the congruent (Mean RT = 1090 ms), relative to the neutral trials (Mean RT = 1149 ms) sugge sting that the high OSPAN group was doing the task, and not simply reading the words. This is an indirect measure of the status of the task-set or goals of each group, where the low group is not keeping th eir color-naming goal active, otherwise they would not have shown such signi ficant benefit in their response times for the congruent versus the neutral trials. Individual Differences in Emotiona l Working Memory C apacity (ESPAN) Similar to the analysis on the individual di fferences in neutral working memory capacity, the main effect of span group was not signi ficant (F(1,92) = .04, p = .841, and there was a significant interaction between trial type and ESPAN span group (F(1,92) = 7.23, p = .009), please see Figure 3-14. The low span group had fa ster response times for the congruent (Mean RT = 1072 ms) than the neutral trials (Mean RT = 1092 ms), whereas the high span group showed the opposite pattern, with greater res ponse times for the congruent (Mean RT = 1167 ms), relative to the neutral trials. (Mean RT = 1156 ms). As in the OSPAN groups, the ESPAN groups showed different facilitation patterns, s uggesting the low ESPAN gr oup is failing to keep their task-set or goal active, rather that they were simply reading the words. Conversely, we see that the High ESPAN group is keeping their goal ac tive, as seen by the smaller facilitation seen in their response latencies. Follow-up Regression Analyses on Response Time Facilitation The following analyses address Hypothesis 2. in Figure 1-1. Step wise regression, using the full sam ple of participants (n = 125), was used to compare the predictive power of two different models. Model A, looked at how well OSPAN scores alon e versus OPSAN and

PAGE 45

45 ESPAN scores combined could capture variance in response time facilitation in the 0% and 75% Congruency Blocks. Model B looked at how well ESPAN alone, versus ESPAN and OSPAN scores combined, would predict er ror interference predicted response time facilitation in the 0% and 75% Congruency Blocks. Please see Append ix F for detailed regression tables. The regression analyses show that ESPAN is a better predictor of response time facilitation because it accounted for 6.4% (compa red to 3.9% in Model A). Additionally when OSPAN is added to the model we see that the to tal variance that is accounted for in predicting response time facilitation is incr eased to 7.3%, suggesting that O SPAN doesnt contribute much of the total variance when the model contains both OSPAN and ESPAN. Word Rating Task Participants rated the words th at they encountered in the Neutral and Emotional W orking Memory Capacity Measures and the Emotional Stroop task on two dimensions, valence and arousal. The ratings were made on 5-point Like rt scales. For valence, a indicates the participant perceives the word as very unpleasant, whereas a indicates the participant views the word as highly pleasant. For arousal, a indicates the participant perceives the word as non-arousing, whereas a indicates the particip ant views the word as highly arousing. This rating task was included in the current investiga tion as a verification that the words chosen to represent the emotional categories within the task s used above, were perceived how we intended them to be perceived. The mean valence and ar ousal ratings for each cat egory of words can be seen in table (Table 3-1). From the table, we see that the words chosen for each word-type were rated and/or perceived as inte nded (Appendix C for exact word lists). For example, across all participants, the Unpleasant Emo tional Stroop words were rated to be the most unpleasant (Mean Valence Rating = 1.38, standard deviation = .38 ), relative to the Pleasant and Neutral (transportation-related) words, this is supported by a significant ma in effect of Word Type (F(3,

PAGE 46

46 372) = 1633.03, p < .01) and a signif icant linear effect of emoti on. Similarly with the arousal dimension, the Unpleasant words were rated as most arousing (Mean Arousal Rating = 4.64, standard deviation = .93), followed by the Pleasa nt words (Mean Arousal Rating = 3.21, standard deviation = 1.13), and Neutral words (Mean Ar ousal Rating = 2.35, standard deviation = .75), this is supported by a significant main effect of Word Type (F(3, 372) = 139.90, p < .01) and a significant quadratic effect.

PAGE 47

47 Figure 3-1. Overall Emo tional Stroop error rates

PAGE 48

48 A B Figure 3-2. Emotional Stroop: I ndividual differences in Error Rates for OSPAN groups A) Low OSPAN group, B) high OSPAN group

PAGE 49

49 A B Figure 3-3.Emotional Stroop: Indi vidual differences in error rates for ESPAN groups A) Low ESPAN group, B) High ESPAN group

PAGE 50

50 Figure 3-4. Overall Emotiona l Stroop response latencies

PAGE 51

51 A B Figure 3-5. Emotional Stroop: Individual differences in res ponse latencies for OSPAN groups A) Low OSPAN group, B) High OSPAN group

PAGE 52

52 A B Figure 3-6. Emotional Stroop: Individual differences in response times for ESPAN groups A) Low ESPAN group, B)High ESPAN group

PAGE 53

53 Figure 3-7. Congruency Stroop: Overall mean error ra tes. On the Left are the error rates for the 0% Congruent condition, and on the Right ar e the error rates fo r the 75% Congruent condition.

PAGE 54

54 Figure 3-8. Congruency Stroop: Error interference for OSPAN groups

PAGE 55

55 Figure 3-9. Congruency Stroop: E rror interference for ESPAN groups

PAGE 56

56 Figure 3-10. Congruency Stroop: Overall response times by trial type

PAGE 57

57 Figure 3-11. Congruency Stroop: Individual differences in response time interference by OSPAN groups

PAGE 58

58 Figure 3-12. Congruency Stroop: Individual differences in response time interference by ESPAN groups

PAGE 59

59 Figure 3-13. Congruency Stroop: Individual differences in res ponse time facilitation by OSPAN groups

PAGE 60

60 Figure 3-14. Congruency Stroop: Individual differences in res ponse time facilitation by ESPAN groups

PAGE 61

61 Table 3-1. Frequency distribut ion of OSPAN and ESPAN groups ESPAN group Low Middle High Total in OSPAN Group OSPAN Group Low 27 12 9 48 Middle 9 4 13 26 High 14 15 23 52 The frequency distribution shows the group as signments of ESPAN scores based on their OSPAN group. For example, the number of part icipants who had a high OSPAN score, yet a low ESPAN score was 14.

PAGE 62

62 Table 3-2. Word ratings Task Valence Arousal Neutral Operation Span 4.59(1.29) 3.10(.68) Emotional Operation Span 2.36(.67) 4.59(.54) Emotional Stroop Pleasant Block Position 1 4.63(.40) 3.21(1.13) Positions 2-6 3.19(.33) 2.30(.66) Transport Block Position 1 3.25(.45) 2.39(.77) Positions 2-6 3.10(.31) 2.34(.68) Unpleasant Block Position 1 1.38(.38) 3.85(.92) Positions 2-6 3.21(.33) 2.34(.67) *This is a table of mean valence and arousal ratings made for words presented within the neutral working memory (Operation Span), emotional working memory (Emotional Operation Span) and the Emotional Stroop tasks. The values in parentheses are the standard devi ations associated with each mean rating.

PAGE 63

63 CHAPTER 4 DISCUSSION The overarching goal of this investigation wa s to explore the concept of an em otional working memory. Working memory is a psycho logical construct repr esenting a temporary memory system that maintains information temporar ily, via code-specific stor es, and serves as an interface between perception and long term memo ry structures. Since the introduction of this construct, there has been little work done to investigate how emotional stimuli might impact the working memory system. The current research approached this question by investigating the impact that emotionally-charged stimuli have on the storage and processing capabilities of working memory. The storage question was approached from a psychometric standpoint, and the processing question was approached via performance in several attention tasks. To evaluate the role that emotion has on stor age capabilities of wo rking memory, a novel measure of working memory capacity was develope d that incorporates emotional stimuli, and an individual differences approach to performance on storage of emotional stimuli was used to predict performance in attention-rich tasks. The relative predictive power of the working memory capacity with emotional stimuli (the ESPAN task) was compared to that of a traditional working memory capacity measure (the OSPAN task ) that has only neutral stimuli for two types of attention tasks. The attention tasks were chos en to tap two specific atte ntion abilities that are thought to be measured by working memory capacity measures such as the Operation Span task, and its emotional modification, used in this study. The two abilitie s thought to be measured in the Operation span task are (a) th e ability to maintain attention in the face of distraction, and (b) the ability to maintain task set or goals across time. Both attention tasks used in this investigation were variations of the traditional color-naming inte rference task of Stroop (1935). One was a modified Emotional Stroop task, wher ein accuracy and latency of color-naming for

PAGE 64

64 words were compared across blocks of trials th at differed in whether an emotional or neutral word appeared in every sixth position (McKenna & Sharma, 2004). The other was the Congruency Stroop task, wherein a ccuracy and latency of color-naming for words was compared across blocks of trials where the list-wise propor tion on congruent (color words that matched the to-be-named color) trials was either low (0 %) or high (75%) (Kane & Engle, 2003). The Emotional Stroop task was chosen to tap the abi lity to maintain attention in the face of an emotional word, whose semantic content is like ly to distract attention. The Congruency Stroop was chosen to test both abilities measured by working memory capacity measures where the material to be maintained in working memory was not of an emotional nature. Among a sample of more than 100 students, working memory capacity (WMC) scores ranged widely, with some participants doing very well, and some, very poorly. The range of Neutral Working Memory Capacity scores was 2-39, and the range of scores for the Emotional Working Memory Capacity scores was 0-37. Wh ile the ranges were similar, overall capacity was significantly lower for ESPAN (Mean = 15.3) than ESPAN (Mean = 17.5). This effect is notable, since as reviewed in the introduction, emotion typically helps long-term memory, and has little effect on traditional s hort-term memory tasks. This is the first demonstration, to my knowledge, of a direct cost of emotion in working memory. Consistent with the notion that maintaining emotional material in working memory is a specific skill on which individuals may vary, th e correlation of OSPAN and ESPAN scores was +.45. Although significant, this corr elation is substantially lower th an the test-retest reliability typically reported for OSPAN itself over similar lags in time of test (+.66 +.85), and shows that only about 25% of the variance in ESPAN scores is accounted for by OSPAN scores. Additionally, the scores across both measures were significantly different from each other, yet

PAGE 65

65 the accuracy scores remained constant across both. This is remarkable, given that, except for the use of emotional words versus neutral words, the tasks were identical in every aspect. This is perhaps our strongest evidence in support of the reality of an emotional working memory. These findings support Hypothesis 2, presented in Figure 1-1. The ability of the ESPAN scores to predic t performance in two very different tasks involving attentional focus and control, and i ndeed to be more closely associated with performance in the task that also involved emotional materials (the Emotional Stroop) than in the task that did not (the Congruency Stroop), furt her demonstrates the va lue of the emotional working-memory construct. (Importantly, the va riability of ESPAN and OSPAN scores in our sample was comparable, so any difference in thei r association with target variables is unlikely due to any range restriction issues.) The findi ngs from each of these target tasks will be summarized in turn. McKenna and Sharma (2004) had found that the emotional Stroop effect slower colornaming latencies to emotional words than neut ral words was not seen in latency to the emotional word itself, but to words downstream from it, when the response-stimulus interval (RSI) was minimal (< 20 m). They argued that this delayed effect of emotionality reflected not a rapid, automatic interference from the emo tional words processing, but a slower, more temporally spread-out attentional triggered by th at automatic processing of the words meaning. Moreover, they found this differe nce only for blocks where the critical word was unpleasant, and not pleasant. The current study aimed to re plicate their findings, please see Hypothesis 1b., Figure 1-1. In broad terms, the current study did replicate these findings, as the greatest differences in color-naming latencies for type of emotional bl ock were found on color-naming of neutral words

PAGE 66

66 that appeared just after the emotional word was responded to (Position 2, see Figure 3-4); but in contrast to the results of McKenna & Sharma, it was an overall emotionality effect (found for both Pleasant and Unpleasant blocks ) that resulted in faster, not longer latencies. Moreover, the effect appeared more sustained across subse quent positions, with a trend, a least, toward continued interference at positions 3 and 4. Also in contrast to McKenna & Sharma, the current investigation found an emo tionality effect for errors, with si gnificantly more errors for blocks with Unpleasant critical words then for Pleasant or Neutral (Tra nsportation-related). The effect on errors was even more diffuse and sustained acr oss position than was that for response latency (see Figure 3-2). The extension of the McKenna and Sharma (2004) Emotional Stroop study was to determine whether there are individual differenc es in working memory capacity on color-naming performance that would be better predicted when the WMC measure its elf involved emotional materials. In the current investigation, then, two measures of working memory capacity (neutral and emotional) were used to predict performa nce within the Emotional Stroop task (Hypothesis 4a. and 3b respectively, Figure 1-1). Results showed that overall, individual differences in the emotional measure of working memory capacity (E SPAN) were better predictor of performance (for both errors and latencies) in the Emotional Stroop task than the neutral measure (OSPAN). Although the detailed patterns of how WMC affected the profile of emotionality effects across position in this task are complex, ESPAN as a fact or was a significant modulator of this pattern for both error scores and latency, while OSP AN was not. (As noted above, the difference in outcome for ESPAN and OSPAN me asures cannot be due to gr eater variance in the ESPAN scores, since they were in fact comparable. Also as will be reviewed below, the OSPAN factor

PAGE 67

67 was more predictive than ESPAN scores, for the neutral Congruency Stroop task, on several measures.) An additional goal of the current study wa s to evaluate whether a novel measure of working memory capacity that incorporates emo tional stimuli could predict performance in a Stroop task with nonemotional stim uli, where the list-wise propor tion of congruent trials was varied between blocks (Hypothesis 4b., Figure 1-1). This type of Stroop task was chosen to test the ability of participants to maintain the colo r-naming task goal across a series of trials where accuracy could be obtained without doing the co lor-naming, but rather by reading the words presented. The results of this analysis were mixed. On the one hand, OSPAN as a factor was significantly associated with the pattern of interference effects across the 0% and 75% congruence conditions for response latencies (H ypothesis 3a., Figure 1-1), while ESPAN was not. The current study showed that the emo tional working memory measure was not a good predictor of performance, neither error rates nor re sponse latencies, on this t ype of attention task. On the other hand, regression analyses using th e full range of scores and sample of 125 individuals suggested that ESPAN scores may be better predictors of some of interference (0%) and facilitation effects in this task than is OSPAN. Overall, however, the present pattern of working memory capacity analyses suggest a dissociation between emotional and neutral working memory c onsistent with the overarching hypothesis of this dissertation: individual differe nces in emotional work ing memory capacity (Emotional Operation Span Task) were better predictors of performance in the Emotional Stroop task; conversely, individual differences in perf ormance on a neutral work ing memory capacity (Operation Span Task) can pred ict error rates and response in terference latencies in the Congruency Stroop task, whereas differences in the emotional working memory measure were

PAGE 68

68 less able to do so. The one index of performa nce that both OSPAN and ESPAN groups show individual differences on is response f acilitation in the Congruency Stroop task. This pattern of results would seem to suppor t the construct of emo tional working memory laid out in the introduction; one's ability to ma intain emotional words in working memory (the ESPAN task) was specifically a be tter predictor of performance in other tasks involving emotion, while the ability to maintain neutral words was a better predictor of attention control in the classic Stroop task. The current investigation looked at the imp act of emotional stimuli on working memory capacity and attention control. The approach used in the current study was to vary the emotionality of stimuli within the working me mory capacity measure as well as within the Stroop paradigm. The measurement and impact of emotional stimuli was assessed using verbal stimuli. It may be that emotional words requ ire more cognitive processing for the emotionality dimension to be culled, thus making differences by emotionality a bit more difficult to detect. Future research should evaluate the role of em otion on working memory and attention allocation using other stimuli, such as pictures or videos where the emotionality of the stimuli can be derived faster, without the need to verbally process the stimuli. Future research should also employ a variety of working memory measures. In the current task the Operation Span and its emotional derivative, the emoti onal Operation Span task, were the only measures of working memory capacity empl oyed. It seems likely th at with a variety of working memory capacity of measures, and thei r emotional derivatives, a better picture of emotions impact on measurement could be better understood. Additionally, by employing multiple measures of working memory capacity, more sophisticated statistical routines could be employed where more of the variance in predictive power could be captured because presumably

PAGE 69

69 there would be some different c ognitive processes being used in one measure, and not in another measure. Another interesting approach to studying emotional working memory would be to see how it is related to general intelligence. In other words: How much error variance can a measure such as ESPAN capture within a latent variable analys is aiming to measure general intelligence? The general goal of latent variable analysis is to account for error variance across two sources: error associated with multiple measurements of the same construct, as well as error variance attributed to the individuals tested on each of the measures. The theory behind this analysis is that after the error variance from the measurement tools and individuals performance on these measures is found and extracted, only error variance associated with the tasks used to measure the construct of interest is left. It is this remaining erro r which provides insight as to the nature of the construct. Kane et al., (2004) conducted a latent variable analysis on a variety of span measures in an effort to determine whether WMC is a doma in specific or domain general construct, and to explore the relationship of WMC to other constr ucts (e.g., general intell igence and short term memory). An extension of this type of study c ould be done to determine whether a measure of emotional working memory can contribute to meas uring the construct of general intelligence. Finally, there has been a good d eal of interest recently in the concept of emotional intelligence. It would be interesting to see if measures like ESPAN, targeted to emotional working memory, might be especially useful in predicting variation in executive function tasks requiring such emotional cognition. Some people, presumably, with greater control over emotional working memory, are bett er able to avoid having to say, I dont know what I was thinking; I was caught up in the heat of the moment.

PAGE 70

70 APPENDIX A INFORMED CONSENT Project Title: Working Memory: An individual differences approach to understanding attention control. Principal Investigator: Cynthia E. Kaschub Department of Psychology PSY 052/ (352) 392-0601 Ext. 367 ckaschub@ufl.edu This study is conducted under the supervision of Dr. Ira Fischler, as part of my doctoral training in cognitive psychology. In this study, we are interested in working memory and your ability to control your attention within tasks. We will ask you to participate in two experimental sessions, the first of which will last about one hour, and the second sessi on about one and a half hours. In both sessions you will be asked to complete a color-naming task, followed by a working memory task. The color-naming task requires you to indicate the color th at each of a series of words is printed in. In the working memory task, you verify if a simple mathema tical equation (for example, (6 / 2) + 1 = 4?) is correct, then read and try to remember a word that is presented after your response. You will retain the words that appear after each answer-verification until you receive the recall prompt, at which time you will be asked to type the words that you received, in the same order, since the last recall. Precise instructions for the experiment will be shown on the computer screen before you start the experiment. In the first experimental session, the color-naming task will include words representing a variety of concepts, objects and events, some of which may be considered emotionally evocative. In the second experimental session the color-naming words will be em otionally neutral. In the second experimental session the working memory task will include emotionally evocative words. After you complete this task, you will be shown a word and will be asked to rate th e word on two nine-point scales. The scale will be provided to you on the computer screen following the word presentation. Precise instructions for the task will be shown on the computer screen before you start the experiment. You will receive five credits towards the research pa rticipation requirements in your class and have a chance to learn about the use of working memory meas ures to study attention-related processes. The data from this experiment will help us to better understand individual differences in attention control. Time required for this experiment is about two and a half hours total. If you have any question about the purposes or procedures of the experiment that need to be clarified before you give your consent to participate, I would be happy to answer them now. Remember that at any point during the experiment you are free to ask further questions or to withdraw your consent and discontinue participation without penalty. Your identity will be kept confidential to the extent provided by law. Your information will be assigned a code number. Your name will not be used in any report. After your data is collected, we destroy any linking of your name and the data files. There are no direct benefits from participating in this experiment. There are no more than minimal risks nor are there any anticipated direct benefits to you during this study. If you have questions about the experiment please contact: Cynthia Kaschub at 392-0601 x 367 or my supervisor Ira Fischler at 392-0601 x 228 Dept. of Psychology, University of Florida, Gainesville, Fl. 32611. Questions or concerns about the research par ticipants rights can be directed to the UFIRB office, PO Box 112250, University of Florida, Gainesville, FL 32611-2250.

PAGE 71

71 Project title: Working Memory: An individual differences appr oach to understanding attention control. PARTICIPANTS CONSENT: I have read the procedure descri bed above. I voluntarily agree to participate in the proce dure and I have received a c opy of this description. Participant: ___________________________________________ Date: _______ Principal Investigat or: ___________________________________ Date: _______

PAGE 72

72 APPENDIX B OPERATION SPAN STIMULI OPER ATIONS OSPAN WORDS ESPAN WORDS (9 / 3) 2 = 2 humble rape (8 / 4) 1 = 1 stomach roach (6 / 2) + 1 = 4 book anger (6 x 3) 2 = 11 jelly tumor (4 x 2) + 1 = 9 unit bomb (10 / 2) + 4 = 9 ketchup torture (10 / 2) 3 = 2 village revolt (10 / 10) 1 = 2 truck killer (7 / 1) + 2 = 7 custom hostage (3 / 1) 2 = 3 gender anxious (2 x 1) 1 = 1 ballot hatred (10 / 1) + 3 = 13 integer leprosy (9 x 2) + 1 = 18 rattle tornado (9 / 1) 7 = 4 hawk betray (8 x 4) 2 = 32 writer python (9 x 3) 3 = 24 phase panic (4 / 1) + 1 = 4 statue ulcer (10 / 1) 1 = 9 news assault (8 x 4) + 2 = 34 horse slap (6 x 3) + 2 = 17 salad surgery (6 / 3) + 2 = 5 army bees (6 x 2) 3 = 10 column snake (8 / 2) + 4 = 2 invest scream (8 / 2) 1 = 3 frog evil (9 / 1) 5 = 4 chair quarrel (6 / 2) 2 = 2 quiet burn (7 x 2) 1 = 14 paint fear (6 x 2) 2 = 10 tower thief (2 x 2) + 1 = 4 cannon violent (7 x 1) + 6 = 13 engine dreary (3 / 1) + 3 = 6 context poison (10 / 1) + 1 = 10 farm divorce (4 x 4) + 1 = 17 mantel horror (3 x 3) 1 = 8 detail brutal (3 x 1) + 2 = 2 circle chaos (4 / 2) + 1 = 6 lamp demons (5 / 5) + 1 = 2 minute cancer (2 x 3) + 1 = 4 salute failure (9 / 3) 2 = 1 humble pain (10 / 2) 4 = 3 locker stress (5 / 1) + 4 = 9 council crash (10 x 2) + 3 = 23 watch trauma (7 / 1) + 6 = 12 milk danger

PAGE 73

73 (3 x 2) + 1 = 6 decade unhappy (6 x 4) + 1 = 25 highwa y abuse (9 / 3) 1 = 2 century danger (8 / 1) 6 = 4 glass rage (9 x 1) + 9 = 1 number victim

PAGE 74

74 APPENDIX C EMOTIONAL STROOP WORD STIMULI

PAGE 75

75 1 2 3 4 5 6 Unpleasant Words REJECTEDSYMPHONYQUANTITY REDUCING ELECTRONSUITABLE FAILURE CLOTHES KITCHEN PROJECT SOMEONE SOMEONE SUICIDE BUFFALO COLLECT CAMPING FLOWING OBSCURE ABUSE HURT BOOTS ITEM BRICK LAWN CREAM BOOK ESSAY COIN LEMON TOOL SAD ARC BAG ERA ICE ODD Pleasant Words ROMANTICCOVERING MOUNTAINCOMPOSERSENTENCEFRONTIER SUCCESS BALANCE DEVELOP OBVIOUS MANAGER PRODUCT MIRACLE ANGULAR CHANNEL CONFIRM FARMING GLIMPSE LUCKY LOVE ALIKE FLAG BRASS HAWK CLOCK MEEK DAIRY CELL INPUT BOWL JOY AIM ROW CUP FLY NET Transport (Neutral) Words AIRPLANE BASEMENT EQUATION THURSDAY FREQUENTFOURTEEN STATION COMPLEX MEASURE QUICKLY ATTEMPT SITTING RAILWAY CABINET COMBINE EMBASSY GRADUAL PACKING FERRY BOAT AWAKE ROCK CHEEK KNOT CROWN PART LABEL NEWS LOBBY CHIN BUS COW BAY FED MUD SUM

PAGE 76

76 APPENDIX D CONGRUENCY STROOP ERROR IN TERFERENCE REGRESSION MODEL COMPARIS ONS

PAGE 77

77 Overall Model Predictor Variable Model A Step R2 R Adj R2 F p-value OSPAN ESPAN 0% Congruent Error Interference 1 0.042 0.035 5.449 0.021 -0.206 2 0.044 0.002 0.029 2.827 0.063 -0.184 -0.049 Overall Model Predictor Variable Model B Step R2 R Adj R2 F p-value ESPAN OSPAN 0% Congruent Error Interference 1 0.017 0.01 2.19 0.142 -0.132 2 0.044 0.027 0.029 2.827 0.063 -0.049 -0.184 Overall Model Predictor Variable Model A Step R2 R Adj R2 F p-value OSPAN ESPAN 75% Congruent Error Interference 1 0.037 0.029 4.767 0.031 -0.193 2 0.04 0.003 0.024 2.541 0.083 -0.167 -0.058 Overall Model Predictor Variable Model B Step R2 R Adj R2 F p-value ESPAN OSPAN 75% Congruent Error Interference 1 0.018 0.018 2.248 0.136 -0.134 2 0.04 0.022 0.04 2.541 0.083 -0.058 -0.167 Model A compares OSPAN as a sole predictor to a model that includes OSPAN and ESPAN as predictors of the dependent variable pre sent. Model B compares ESPAN as a sole predictor to a model that includes ESPAN and OSPAN as predictors of the dependent variable. For Step 1, dfs = 1, 124; for Step 2, dfs = 2, 124.

PAGE 78

78 APPENDIX E CONGRUENCY STROOP RESPONSE TI ME INTERFERENCE REGRESSION MODEL COMPARIS ONS

PAGE 79

79 Overall Model Predictor Variable Model A Step R2 R Adj R2 F p-value OSPAN ESPAN 0% Congruent Response Time Interference 1 0.038 0.031 4.901 0.029 -0.196 2 0.094 0.056 0.079 6.302 0.002 -0.075 -0.264 Overall Model Predictor Variable Model B Step R2 R Adj R2 F p-value ESPAN OSPAN 0% Congruent Response Time Interference 1 0.089 0.082 12.035 0.001 -0.299 2 0.094 0.005 0.079 6.302 0.002 -0.264 -0.075 Overall Model Predictor Variable Model A Step R2 R Adj R2 F p-value OSPAN ESPAN 75% Congruent Response Time Interference 1 0.003 -0.005 0.413 0.52 0.058 2 0.015 0.012 -0.001 0.954 0.388 0.114 -0.123 Overall Model Predictor Variable Model B Step R2 R Adj R2 F p-value ESPAN OSPAN 75% Congruent Response Time Interference 1 0.005 -0.003 0.626 0.43 -0.071 2 0.015 0.01 -0.001 0.954 0.388 -0.123 0.114 Model A compares OSPAN as a sole predictor to a model that includes OSPAN and ESPAN as predictors of the dependent variable pre sent. Model B compares ESPAN as a sole predictor to a model that includes ESPAN and OSPAN as predictors of the dependent variable. For Step 1, dfs = 1, 124; for Step 2, dfs = 2, 124.

PAGE 80

80 APPENDIX F CONGRUENCY STROOP RESPONSE TI ME FACILITATION RE GRESSION MODEL COMPARISONS

PAGE 81

81 Overall Model Predictor Variable Model A Step R2 R Adj R2 F p-value OSPAN ESPAN 75% Congruent Response Time Facilitation 1 0.039 0.031 4.99 0.027 0.198 2 0.073 0.034 0.058 4.79 0.01 0.103 0.207 Overall Model Predictor Variable Model B Step R2 R Adj R2 F p-value ESPAN OSPAN 75% Congruent Response Time Facilitation 1 0.064 0.057 8.46 0.004 0.254 2 0.073 0.009 0.058 4.79 0.01 0.207 0.103 Model A compares OSPAN as a sole predicto r to a model that includes OSPAN and ESPAN as predictors of the dependent variable present. Model B compares ESPAN as a sole predictor to a mode l that includes ESPAN and OSPAN as predictors of the dependent variable. For Step 1, dfs = 1, 124; for Step 2, dfs = 2, 124.

PAGE 82

82 LIST OF REFERENCES Algom D., Chajut, S., & Lev, S. (2004). A rati onal look at the Emotional Stroop phenomenon: A general slowdown, not a Stroop effect. Journal of Experimental Psychology: General, 133, 323-338. Anderson, A. K. (2005). Affective influences on the attentional dynamics supporting awareness. Journal of Experimental Psychology: General, 134 258-281. Atkinson, R. C. & Shiffrin R. M. Human memory: A proposed system and its control processes. In The psychology of lear ning and motivation: II Spence, Kenneth W .; Spence, Janet T.; Oxford, England: Academic Press, 1968. Awh, E., & Jonides, J. (2001). Overlapping m echanism of attention and spatial working memory. Trends in Cognitive Sciences, 5, 119-126. Baddeley, A. D. (2000). The episodic buffer: A new component of working memory. Trends in Cognitive Sciences, 4, 417-423. Baddeley, A. D. (2002). Is working memory still working? European Psychologist, 7, 85-97. Baddeley, A. D. (2003). Working memory or working attention? In Baddeley, A. D & Weiskrantz, L. (Eds.), Attention: Selection, awarene ss, and control: A tribute to Donald Broadbent. New York, NY: Clarendon Press pp. 152-170. Baddeley, A. D. & Logie, R. H. Working memory: The multiple-component model. In A. Miyake & P. Shah. (Eds.), Models of working memory: Mechanisms of active maintenance and executive control. New York, NY: Cambri dge University Press, 1999. Bianchin, M., Mello-e-Souza, T., Medina, J. H ., & Izquierdo, I. (1999). The amygdala is involved in the modulation of long-term me mory but not in working or short-term memory. Neurobiology of Learning and Memory, 71 127-1310 Bishop, S., Duncan, J., Brett, M ., & Lawrence, A. D. (2004). Pref rontal cortical function and anxiety: controlling attention to fearful faces. Nature Neuroscience, 7 184-188 Broadbent, D. E. (1958). Perception and Communication New York, NY: Pergamon Press. Brooks, L. R. (1967). The suppression of visualization by reading. Quarterly Journal of Experimental Psychology, 19, 289-299. Buchanon, T. W., Adolphs, R. (2004). The neuroana tomy of emotional memory in humans. In Reisburg, D., & Hertel, P. (Eds.), Memory and Emotion New York: Oxford University Press, pp 42-75.

PAGE 83

83 Canli, T., Zhao, Z., Desmond, J. E., Gross, J. J. & Gabrielli, J. D. E. (2001). An fMRI study of personality influences in brain r eactivity to emotional stimuli. Behavioral Neuroscience, 115, 33-42. Chajut, S., Lev, S., & Algom, D (2005). Vicissitudes of a misnomer: Reply to Dalgleish Journal of Experimental Psychology: General, 134, 592-595. Clarkson-Smith, L. & Hartley, A. A. (1990) The game of bridge as an exercise in working m emory and reasoning. Journals of Gerontology, 45 233-238. Conway, A. R.,Cowan, N. & Bunting, M. F. (2001). The cocktail party phenomenon revisited: The im portance of working memory capacity. Psychonomic Bulletin & Review, 8, 331-335. Conway, A. R.,Cowan, N., Bun ting, M. F., Therriault, D., & Minkoff, S. (2002). A latent variable analysis of working memory capacity, short term memory capacity, processing speed, and general fluid intelligence. Intelligence, 30, 163-183. Cowan, N (1988). Evolving conceptions of memory st orage, selective atten tion, and their mutual constraints within the hu man information-processing. Psychological Bulletin, 10, 163-191. Cowan, N. An e mbedded-processes model of worki ng memory. In (Eds.), A. Miyake & P. Shah. Models of working memory: Mechanis ms of active maintenance and executive control New York, NY: Cambridge University Press, 1999. Dalgleish, T. (1995). Performance on the emoti onal Stroop task in groups of anxious, expert, and control subjects: A co mparison of computer and card presentation formats. Cognition and Emotion, 9, 341-362. Dalgleish, T. (2005). Putting some feeling into itThe conceptual and empirical relationships between the classic and emotional Stroop ta sks: Comment on Algom, Chajut, and Lev (2004), Journal of Experimental Psychology: General,134, 585-591. Daneman, M. & Carpenter, A. (1980). Individual differences in working memory and reading. Journal of Verbal Learning and Verbal Behavior, 19, 450-466. Darwin, C. (1999). The expression of the emotions in man and animals. (Introduction by Paul Ekman). London: Harper Collins. (original work published in 1872). Davidson, R. J. & Irwin, W. (1999). The functio nal neuroanatomy of emotion and affective style. Trends in Cognitive Sciences, 3 11-21. DEsposito, M., Aguirre, G. K., Zarahn, E., Ballar d, D., Shin, R. K., & Lease, J. (1998). Functional MRI of spatial a nd nonspatial working memory. Cognitive Brain Research, 7, 11-13. Dolan, J. (2002). Emotion, Cognition and Behavior. Science, 298 1191-1194.

PAGE 84

84 Dolcos, F., & McCarthy, G. M. (2006). Brai n systems mediating cognitive interference by emotional distraction. Journal of Neuroscience, 26, 2072-2079. Duncan, F. (1990). Goal weighting and the choice of behavior in a complex world. Ergonomics, 33, 1265-1279. Ekman, P. Basic emotions. In T. Dalgleish and M. Powers (Eds.), Handbook of Cognition and Emotion. Sussex, U. K.: John Wiley & Sons, Ltd., 1999. Elliman, N. A., Green, M. W., Rogers, P. J., & Finch, G. MJ. (1997). Processing-efficiency theory and the working memory system: Im pairments associated with sub-clinical anxiety. Personality and Individual Differences, 23 31-35. Engle, R. W., Carullo, J. J., Collins, K. W. (1991). Individual differences in working m emory for comprehension and following directions. Journal of Educational Research, 84, 253-262. Engle, R. W., Cantor, J., & Carullo, J. J. (1992) Individual differences in working memory and comprehension: A test of four hypotheses. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 972-992. Engle, R. W., Conway, A. R. A. Tuholski, S. W. & Shisler, R. J. (1995). A resource account of inhibition. Psychological Science, 6, 122-125. Feldman Barrett, L. A. & Gross, J. J. Emo tional intelligence: A process model of emotion representation and regulation. In T. J. Mayne & G. A. Bonanno (Eds.), Emotions: Current Issues and Future Directions New York: The Guilford Press, pp. 286310, 2001. Gray, J. R. (2001). Emotional modulation of cognitive control: Appro ach-withdrawal states double dissociate spatial from verbal two-back task performance. Journal of experimental Psychol ogy: General, 130, 436-452. Gray, J. R., Braver, T. S., Raichle, M. E. ( 2002). Integration of emotion and cognition in the lateral prefrontal cortex. Proceedings of the National Academy of Sciences, 99 4115-4120. Gross, J. J. (1998). The emerging field of em otion regulation: An integrative review. Review of General Psychology, 2, 271-299. Herrington, J. D. Mohanty, A., Koven, N. S., Fish er, J. E., Stewart, J. L., Banich, M. T., Webb, A. Miller, A. G., Heller, W. (2005). Em otion-modulated performance and activity in left dorsolateral prefrontal. Emotion, 5, 200-207. Jha, A. P., Fabiani, S. A., & Aguirre, G. K. (2004). The role of the prefront al cortex in resolving distractor interference. Cognitive, Affective & Be havioral Neuroscience, 4, 517527.

PAGE 85

85 Jonides, J., Reuter-Lorenz, P. A., Smith, E. E ., Awh, E., Barnes, L. L. Drain, M., Glass, J., Lauber, E. J., Patalano, A., & Schumacher, E., (1996). Verbal and spatial working memory in humans. In D. L. Medin (Ed.), The Psychology of Learning and Motivation, Vol. 35. Jonides, J., Reuter-Lorenz, P. A., Smith, E. E., Awh, E. Barnes, L. L., Drain, M., Glass, J., Lauber, E.J., Patalano, A. L., Schumacher, E. H., Verbal and spatial working m emory in humans. The psychology of learning and motivation: Advances in research and theory, Vol. 35. Medin, Douglas L. (Ed); pp. 43-88. San Diego, CA, US: Academ ic Press, 1996. Juth, P., Lundqvist, D., Karlsson, A., Ohman, A. (2005). Looking for foes and friends: Perceptual and emotional factors when finding a face in the crowd. Emotion, 5 379-395. Kane, M. J. Bleckley, M. K., & Conway, A. R. A. (2001). A controlled-attention view of working memory capacity. Journal of Experimental Psychology: General,130,169-183. Kane, M. J. & Engle, R. W. (2003). Working memo ry capacity and the control of attention: The contributions of goal neglect, response competition, and task set to Stroop interference. Journal of Experimental Psychology: General, 132, 47-70. Kane, M. J., Engle, R. W. & Tuholski, S. W. Individual differences in working memory capacity and what they tell us about controlled attention, general fluid intelligence, and f unctions of the prefrontal cortex. In Miyake, Akira; Shah, Priti (Eds): Models of working memory: Mechanisms of acti ve maintenance and executive control. New York, NY, US: Cambridge University Press, 1999. pp. 102-134. Kane, M. J., Hambrick, D.Z., Tuholski, S. W., Wilhelm, O., Payne, T. W. & Engle, R. W. (2004). The generality of working memory capacity: A latent-variable approach to verbal and visuouspatial memory span and reasoning. Journal of Experimental Psychology: General, 133, 189-217. Keil, A., Bradley, M. M., Hauk, O., Rickstrohl, B., Elbert, T., & Lang, P. J. (2002). Large-scale neural correlates of aff ective picture processing. Psychophysiology, 39 641-649. Keil, A., & Ihssen, N. (2004). Identification f acilitation for emotionally arousing verbs during the attentional blink. Emotion, 4 23-35. Kensinger, E. & Corkin, S. (2003). Effect of Negative Emotional Content on Working Memory and Long-Term Memory. Emotion, 3, 378-393. Kesler-West, M. L., Andersen, A. H., Smith, C. D., Avison, M. J., Davis, C. E., Kryscio, R. J., Blonder, L. X., (2001). Neur al substrates of facial emotion processing in fMRI. Cognitive Brain Research, 11 213-226

PAGE 86

86 Kiewra, K. A. & Benton, S. L. (1988) The relationship between info rm ation-processing ability and notetaking. Contemporary Educational Psychology, 13, 33-44. Keltner, D., & Haidt, J. (2001). Social functions of emotions. In T. J. Mayne & G. A. Bonanno (Eds.), Emotions: Current Issues and Future Directions New York: The Guilford Press, pp. 192-213. Klein, K., & Fiss, W. H. (1999). The reliability and stability of the Turner and Engle working memory task. Behavior Research Methods, Instruments and Computers, 31, 429432. Kleinsmith, L. J. & Kaplan, S. (1963). Paired associate learning as a function of arousal and interpolated interval. Journal of Experiment al Psychology 65. 190-193. Kleinsmith, L. J. & Kaplan, S. (1964). Interac tion of arousal and recall interval in nonsense syllable paired associate learning. Journal of Experimental Psychology,67, 124126. Kyllonen, P. C. & Cristal, R. E. (1991). Reasoning is (little more than) working memory capacity?! Intelligence, 14, 389-433. Lane, R. (1997). The neuroanatomical correl ates of happiness, sadness, and disgust. American Journal of Psychiatry, 154, 926-933. Lane, R., Reiman, E., Bradley, M., Lang, P. J., Ahern, G.L., Davidson, R. J., & Schwartz G.E. (1997). Neuropsychologia, 35, 1437-1444. LaPointe, L. B., & Engle, R.W. (1990). Simp le and complex word spans as a measure of working memory capacity. Journal of Experiment al Psychology: Learning, Memory, and Cognition, 16, 1118-1133. Levinger, G., Clark, J.(1961). Emotional factors in th e forgetting of word associations. Journal of Abnormal & Social Psychology, 62 99-105. Mikels, J. A., Larkin, G. R., Reuter-Lorenz, P. A., & Cartensen, L. L. (2005). Divergent trajectories in the aging m ind: Changes in working memory for Affective versus visual information with age. Psychology and Aging, 20, 542-553. Mikels, J. A., Reuter-Lorenz, P. A., & Fredricks on, B. L. (2003). Hold on to that feeling: An empirical analysis of aff ective working memory. Post er presented at the annual meeting of the Psychonomic Society. Mikels, J. A., Reuter-Lorenz, P. A., & Fredrick son, B. L. (2004). Dissociable orbitofrontal and limbic correlates of affectiv e working memory. Poster presented at the annual meeting for the Society for Neuroscience. McKenna, F. P.(1986). Effects of unattended emotional stimuli on color-naming performance. Current Psychological Research & Reviews, 5, 3-9.

PAGE 87

87 McKenna, F. P., & Sharma, D. (1995). Intrusive cognitions: An investigation of the emotional Stroop task. Journal of Experimental Psyc hology: Learning, Memory & Cognition, 21, 1595-1607. McKenna, F. P., & Sharma, D. (2004). Reversing th e emotional Stroop effect reveals that it is not what it seems: The role of fast and slow components. Journal of Experimental Psychology, 30, 382-392. Northoff, G., Richter, A., Gessner, M., Schlagenhauf, F., Fell, J., Baumgart, F., Kaulisch, T., Ktter, R., Stephan, K., Leschinger, A., Hagner, T., Bargel, B., Witzel, T., Hinrichs, H., Bogerts, B., Scheich, H., Heinze, H. (2000). Functional dissociation between m edial and lateral prefrontal cortical spatiotemporal activation in negative and positive emotions: A combined fMRI/MEG study. Cerebral Cortex 10, 93-107. Ohm an, A., Flykt, A., & Esteves, F. (2001). Emotion drives attenti on: Detecting the snake in the grass. Journal of Experimental Psychology: General. 130 466-478. Ohman, A., Lundqvist, D., & Esteves, F. (2001). The face in the crowd effect revisited: A threat advantage with schematic faces. Journal of Personality and Social Psychology, 80, 381-396. Packard, M. G. & Teather, L. A. (1998). Amygda la modulation of multiple memory systems, hippocampus and caudate putamen. Neurobiology of Learning and Memory, 69 163-203. Patrick, C. J., Cuthbert, B. N. & Lang, P. J. ( 1994). Emotion in the criminal psychopath: Fear image processing. Journal of Abnormal Psychology, 103, 523-534. Petrides, M. & Milner, B. (1982). Deficits on subject-ordered tasks after frontaland temporallobe lesions in man, Neuropsychologia, 20 249-262. Perlstein, W. M., Elbert, T., & Stenger, A. (2002) Dissociation in human prefrontal cortex of affective influences on working memory-related activity. Proceedings of the National Academy of Sciences, 99, 1736-1741. Perlstein, W. M., Carter, C. S., Barch, D. M., & Baird, J. (1998). The Stroop task and attention deficits in schizophrenia: A critical evaluation of car d and single-trial Stroop m ethodologies. Neuropsychology, 12, 414-425. Pratto, F. & John, O. P. (1991). Automatic vigilance: the attention-gr abbing power of negative social information. Journal of Personality and Social Psychology, 61, 380-91. Quevedo, J., Sant, M., Madruga, M., Lovato, I., de-Paris, F., Kapczinkski, F., Izquierdo, I., & Cahill, L. (2003). Differential effects of emotional arousal in shortand longterm memory in healthy adults. Neurobiology of Learning and Memory, 79, 132135.

PAGE 88

88 Schupp, H. T., Junghofer, M., Weike, A. I., & Hamm, A. O. (2003). Emotional facilitation of sensory processing in the visual cortex. Psychological Science, 14 7-13. Shute, V. J. (1991). Who is likely to acquire programming skills? Journal of Educational Computing Research, 7, 1-24. Simpson, J.R., ngr, D., Akbudak, E., Conturo, T. E., Ollinger, J.M., Snyder, A. Z., Gusnard, D. A., Raichle, M. E. .(2000). The emotional modulation of cognitive processing: An fMRI study. Journal of Cognitive Neuroscience, 12 ,157-170. Sm ith, K. N., Cacioppo, J. T., Larsen, J. T. & Ch artrand, T. (2003). May I have your attention, please: Electrocortical responses to positive and negative stimuli. Neuropsychologia, 41, 171-183. Smith, E. E., Patalano, A. L., & Jonides, J. ( 1998). Alternative strate gies for categorization. Cognition, 65, 167196. Smith, E. E., & Jonides, J. (1997). Wo rking memory: A view from neuroimaging. Cognitive Psychology, 33, 5-42. Spies K., Hesse, F. W. & Hummitzsch, C. (1996) Mood and capacity in Baddeleys model of human memory, Zeitchrift fur Psychologie, 204, 367-381. Stroop, J. R. (1935). Studies of interf erence in serial verbal reactions. Journal of Experimental Psychology, 18, 643-662. Turley-Ames, K. J., & Whitfield,M. M. (2002). Strategy training and working memory task performance. Manuscript submitted for publication. Unsworth, N. & Engle, R., W. ( 2007). The nature of individual differences in working memory capacity: Active maintenance in primary memory and controlled search from secondary memory. Psychological Review, 114, 104-132. Unsworth, N., Heitz, R., P., Schrock, J., C., & E ngle, R., W. (2005). An automated version of the operation span task. Behavior Research Methods, 37 498-505. Williams, J. M.G., Mathews, A., & MacLeod, C. (1996). The emotional Stroop task and psychopathology. Psychological Bulletin, 120 3-24. Winton, W. C., Clark, D. M. & Edelmann, R. J. (1995). Social anxiety, fear of negative evaluation, and the detection of negative emotion in others. Behaviour Research and Therapy, 33, 193-196.

PAGE 89

89 BIOGRAPHICAL SKETCH Cynthia Elizabeth Kaschub was raised in C onnecticut. She did her undergraduate work at Boston College, where she first tasted em pirical research methods. After Boston College, she began her graduate studies at Villanova Univer sity, where she earned her Master of Science degree in 2004. After completing her Ph.D. at th e University of Florida she will serve as a Researcher at the North Atlant ic Treaty Organization Naval U ndersea Research Centre in La Spezia, Italy for the summer of 2008. Upon re turning from Italy, Dr. Kaschub will begin her Cognitive Engineering post-doc Department at th e Applied Physics Laboratory at Johns Hopkins University.