<%BANNER%>

Interference Effects of Anxiety and Affective Processing on Working Memory: Behavioral and Electrophysiological Accounts


PAGE 1

INTERFERENCE EFFECTS OF ANXIETY AND AFFECTIVE PROCESSING ON WORKING MEMORY: BEHAVIOR AL AND ELECTROPHYSIOLOGICAL ACCOUNTS By DAVID ANDREW STIGGE-KAUFMAN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by DAVID ANDREW STIGGE-KAUFMAN

PAGE 3

ACKNOWLEDGMENTS I thank my advisor, William M. Perlstein, for his excellent guidance and support during this project. I also thank the other members of the Clinical-Cognitive Neuroscience Lab for their assistance. Lastly, I thank my wife and family for their love and support. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT.......................................................................................................................ix CHAPTER 1 INTRODUCTION........................................................................................................1 Phobias and Fear...........................................................................................................1 Attentional and Perceptual Components of Emotional Processing..............................2 Emotional Connections with Working Memory...........................................................5 Event-Related Potentials...............................................................................................8 Neural Correlates of Cognitive and Affective Picture Processing.............................10 Summary and Rationale for the Current Study...........................................................12 Predictions..................................................................................................................14 WM Task Performance........................................................................................14 ERP Reflections of Task-Irrelevant Interference Stimulus Processing...............15 2 METHOD...................................................................................................................16 Participants.................................................................................................................16 Materials and Procedure.............................................................................................18 EEG Acquisition and Reduction.................................................................................21 EEG Data Acquisition.........................................................................................21 EEG Data Reduction...........................................................................................22 Statistical Analyses.....................................................................................................23 WM Performance Data........................................................................................24 Picture Response Data.........................................................................................24 ERP Data.............................................................................................................24 3 RESULTS...................................................................................................................26 Task Performance.......................................................................................................26 Effects of WM Load............................................................................................26 iv

PAGE 5

Effects of Interference on WM Task Accuracy...................................................27 Effects of Interference on WM Task Reaction Time..........................................28 Picture Response Data................................................................................................31 Interfering Picture Detection Data.......................................................................31 Viewing Times During Picture Rating................................................................32 Manipulation Check: Picture Rating Data...........................................................33 Valence ratings.............................................................................................33 Arousal ratings.............................................................................................34 Event-Related Potential (ERP) Data...........................................................................36 Effects of WM Load............................................................................................36 Effects of Picture Interference.............................................................................40 4 DISCUSSION.............................................................................................................48 Interference Effects on WM Task Performance.........................................................48 Picture Response Data................................................................................................50 ERP Reflections of Interference-Stimulus Processing...............................................51 WM Load Effects on Interference-Related Activity...........................................51 Valence-Related Effects Interference Activity....................................................52 Emerging Patterns.......................................................................................................53 Alternative Explanations and Possible Limitations....................................................54 Future Directions........................................................................................................57 Summary.....................................................................................................................58 LIST OF REFERENCES...................................................................................................59 BIOGRAPHICAL SKETCH.............................................................................................65 v

PAGE 6

LIST OF TABLES Table page 2-1. Demographic and emotional functioning data for all participants.............................18 2-2. Intercorelations between measures of anxiety and depression..................................18 3-1. ANOVA statistics from the WM task accuracy data.................................................28 3-2. Valence effects in the WM task accuracy data..........................................................28 3-3. ANOVA statistics from the probe reaction time data................................................29 3-4. Valence effects in the probe reaction time data.........................................................29 3-5. ANOVA statistics from the ERP components...........................................................42 3-6. Valence effects in the ERP components....................................................................43 vi

PAGE 7

LIST OF FIGURES Figure page 1-1. Extraction of the ERP waveform from ongoing EEG................................................9 2-1. Overview of each trial, showing the time course of each cue, picture interference, and probe.............................................................................................20 2-2. Sensor layout of the 64-channel geodesic sensor net...............................................22 3-1. WM task performance for each WM load during no-interference trials..................27 3-2. Error rates by interference category, and fear group................................................29 3-3. Probe reaction time by interference category, and fear group.................................31 3-4. Mean picture detection reaction times by valence and fear group...........................32 3-5. Mean RTs for picture viewing during the rating procedure at the end of the experiment................................................................................................................33 3-6. Subjective ratings for picture valence......................................................................34 3-7. Subjective ratings for picture arousal.......................................................................35 3-8. Grand-averaged ERPs for all scalp sites during interference picture processing in the low (blue) and high (red) WM load conditions..................................................37 3-9. Spherical-spline interpolated scalp voltage maps representing the differences in neural processing between the high and low WM loads for the early and late LPPs and slow wave.................................................................................................38 3-10. Grand-averaged ERPs for site #34 during interference picture processing in the low (blue) and high (red) WM load conditions........................................................38 3-11. Mean ERP amplitudes for low and high WM load conditions for the early and late LPP and slow wave...........................................................................................39 3-12. Grand-averaged ERPs for all scalp sites during interference picture processing in the neutral (blue), pleasant (black), unpleasant (green), and threat (red) conditions. ................................................................................................................41 vii

PAGE 8

3-13. Spherical-spline interpolated scalp voltage maps representing the differences in neural processing between affective and neutral pictures for the early and late LPPs and slow waves of high load trials..................................................................42 3-14. Grand-averaged ERPs for site #34 during interference picture processing in the neutral (blue), pleasant (black), unpleasant (green), and evolutionary threat (red) conditions at high WM loads, collapsed across groups............................................43 3-15. Mean ERP amplitudes for the early and LPP and slow wave for controls..............44 3-16. Mean ERP amplitudes for the early and LPP and slow wave for high-fear participants...............................................................................................................45 3-17. Grand-averaged ERPs for controls and high-fear participants during threat and unpleasant interference picture processing in the at high WM load........................45 viii

PAGE 9

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science INTERFERENCE EFFECTS OF ANXIETY AND AFFECTIVE PROCESSING ON WORKING MEMORY: BEHAVIORAL AND ELECTROPHYSIOLOGICAL ACCOUNTS By David Andrew Stigge-Kaufman May, 2005 Chair: William M. Perlstein Major Department: Clinical and Health Psychology Emotionally salient information is often given preferential processing in attention and perception. While this offers many benefits, it can become problematic in disorders of anxiety when threat-related processing is excessive and dysfunctional. Affective processing has been shown to modulate higher-level cognitive functions like working memory, although the nature of these effects is poorly understood for both anxious and healthy individuals alike. This study examined whether affective interference impairs representations in working memory (WM) and whether this impairment is greater in response to interfering emotional stimuli that have intrinsic clinical or evolutionary relevance. High-fear participants and low-fear controls performed a novel delayed matching-to-sample (DMS) task involving high and low WM loads and the presentation of interference pictures of varying valence (pleasant, neutral, unpleasant, and clinically/evolutionarily-relevant threat) during the delay. Error rates, reaction time (RT), ix

PAGE 10

and high-density event-related potentials (ERP) were acquired while participants performed the task. Performance data from WM task revealed that WM performance is disrupted by affective interference and that arousal and anxiety play key roles in that interference. ERP data suggest that increased WM load and the processing of task-irrelevant affective interference compromise neural resources of attention. Deficits in the neural processes involved in sustained attention appear to be associated with enhanced threat-based interference of WM in high-fear participants, which may have important implications for broader behavioral symptoms associated with anxiety. In the end, these findings offer new insights into the complex interplay between competing cognitive and affective demands on attention, and suggest ways that anxiety disrupts higher-level cognitive processes involved in WM. x

PAGE 11

CHAPTER 1 INTRODUCTION Phobias and Fear Recent epidemiological studies estimate that phobias are the single most common mental disorder in the United States, affecting approximately 5 to 10 percent of the population (Sadock & Sadock, 2003). Distress associated with phobias often results in a restricted lifestyle and can lead to many other psychiatric complications, including other anxiety disorders, major depressive disorder, and substance-related disorders (American Psychiatric Association [APA], 2000). Although phobias are extremely common, many persons with phobias do not seek help to overcome their phobias or are misdiagnosed when given medical or psychiatric attention (Sadock & Sadock, 2003). Specific phobics have been shown to exhibit cognitive biases for threat-related pictures and words (Kindt & Brosschot, 1997), which likely play a role in the maintenance of fear. Additionally, individuals with subclinical levels of anxiety exhibit notable deficits in the cognitive processing of threatening information (Fox, Russo, Bowles, & Dutton, 2001; Ohman & Soares, 1998). Animals that pose evolutionary threat (e.g., snakes, spiders) are common objects of intense fear, reported by as many as 38% of females and 12% of males sampled (Ohman & Mineka, 2003). Given their high prevalence, the functional consequences of phobias and high levels of fear are worthy of considerable attention. Previous investigations have examined the perceptual and attentional processing of emotional information, while other lines of research have begun to focus on connections between emotional processing and 1

PAGE 12

2 higher levels of cognition, including working memory. The combined study of affective and cognitive processes is complex, but previous findings now lay the foundation for a better understanding of the neural processes involved in affective influences on cognition, which makes it possible to more carefully study the functional impact of fear and other emotional factors on higher cognitive processes like working memory. Attentional and Perceptual Components of Emotional Processing Examinations of emotions from an evolutionary perspective are quick to highlight the importance of allocating preferential perceptual processing to certain environmental stimuli (Dolan, 2002). For millions of years, species survival has depended on the ability to efficiently process the wide variety of sensory cues that exist in the natural world (Ledoux, 1996). Emotional stimuli that occur naturally in the environment (such as snakes and spiders) are detected faster than non-emotional stimuli, suggesting that evolutionarily relevant threatening stimuli are effective in capturing visual attention quickly and consistently (Ohman, Flykt, & Esteves, 2001). Enhanced attention has even been shown to spread to non-emotional stimuli that merely appear within spatial proximity to emotional cues (Williams, Watts, MacLeod, & Matthews, 1997). Participants of spatial orienting tasks respond faster to non-emotional targets appearing on the same side or same location as an emotional cue (e.g., positive or negative faces), while a slower response results for targets appearing on the opposite side or different location (Armony & Dolan, 2002), suggesting preferential processing for emotionally salient information. Despite the obvious benefits that accompany enhanced perceptual processing of emotional stimuli, excessive levels of attention may give rise to the emotional dysfunction present in certain psychological disorders. It has been suggested that at least

PAGE 13

3 some anxiety disorders are caused by automatic biases of attention that cause excessive engagement to threat-related stimuli (Ledoux, 1996; Williams et al., 1997). These claims have been fueled in part by results from tasks that measure the reaction time of affectively modulated attentional processing. In studies involving the dot-probe paradigm, for example, anxious participants detected a target dot more quickly after it appeared in the location previously occupied by a threat-related word (MacLeod, Matthews & Tata, 1986; Fox, 1993). Other studies employing the emotional Stroop paradigm have found that color naming was slower on anxiety-related words compared to neutral words (MacLeod, 1991, Vrana, Roodman, & Beckham, 1995), suggesting that an interference effect can be facilitated by emotional processing. However, not all researchers attribute anxiety to hypervigilence of the attentional system. Fox and colleagues (2001) demonstrated that threatening words and faces may not serve to attract attention, but instead disrupt the disengagement of attention by using a modified version of the exogenous cueing paradigm (Posner, Inhoff, Friedrich, & Cohen, 1987). They concluded that threatening cues increase the amount of time that attention is allocated to a stimulus by making it difficult to disengage attention from that stimulus. While initial disengagement may constitute an early stage of the attentional bias for threatening stimuli, research in spider phobics has suggested that high levels of anxiety can also cause rapid disengagement at later processing stages (Hermans, Vansteenwegen, & Eelen, 1999). Although the debate over the mechanism underlying these findings has not been resolved, it seems likely that emotionally-motivated attention is mediated by a complex interplay between multiple, overlapping neural systems that mediate different kinds of psychological processing (Compton & Banich, 2003).

PAGE 14

4 Capturing the attentional system does not appear to be the only way that emotional stimuli influence perception. In situations where attention has been systematically limited, emotional stimuli continue to show preferential processing. Visual backward masking paradigms tap the automatic regulation of attention by presenting an threatening target stimulus briefly (30 milliseconds) that is then masked by an immediately following second stimulus. Even when the emotional target (e.g., pictures of snakes or spiders) is not consciously perceived, physiological processing (as measured by skin conductance responses) reveals differential autonomic processing in individuals who are fearful of those stimuli (Ohman & Soares, 1994). Furthermore, other studies have found similar effects in non-fearful subjects who were aversively conditioned to fear the masked stimuli (Ohman & Soares, 1993). From these findings, it appears that certain threat-provoking stimuli are selectively processed by autonomic and central nervous systems even in the absence of conscious awareness. A brain structure that plays a key role in automatic processing of threat is the amygdala, a subcortical structure of the limbic system located deep in the anterior medial temporal lobe (Armony & LeDoux, 1997). Neuroimaging studies consistently find the amygdala to be active during the evaluation of fearful stimuli, such as fearful facial expressions (Morris et al., 1996) and affective pictures (Lane et al., 1997). Patients with amygdala lesions show significant deficits in recognition of fearful facial expressions (Adolphs et al., 2005). Anatomical analyses reveal that the amygdala receives inputs from the cortex as well as direct sensory inputs from the thalamus, suggesting that the amygdala can process emotional information independently of conscious processing streams (Ledoux, 1996). Rapid processing of emotional information is highly adaptive

PAGE 15

5 when the brain needs to evaluate threatening information; however, this automaticity comes at the cost of more in-depth cortical processing of stimuli (Armony & LeDoux, 1997). With its anatomical connections and its role in evaluating fearful stimuli, the amygdala appears to serve the evolutionary role of the brains threat detector (Ohman & Mineka, 2001). All in all, there is strong evidence that humans give preferential attention to emotional stimuli. This bias in attention allows for rapid adaptation to environmental cues that promotes species survival. However, humans possess higher levels of cognitive control that allow for more flexible adaptation of behavior to the contextual demands of our environments (Miller & Cohen, 2001). Are these higher levels of cognitive functioning also affected by threat-provoking cues that are present in the environment, or is affective modulation of cognition limited only to the basic levels of attention and perception? Emotional Connections with Working Memory The ability to internally store and manipulate information provides additional survival benefits beyond those afforded by perception and attention alone. These benefits have been provided by a fundamental set of cognitive processes known as working memory. Working memory (WM) involves the active maintenance and manipulation of selective symbolic information while inhibiting other information, utilizing two components: short-term storage and executive processes (Baddeley, 1986). Through its dependence on processing symbolic representations, WM serves to free an organism from a dependence on the presence of environmental cues and is critical for behavioral flexibility, internal monitoring, and guidance of contextually appropriate action (Miller & Cohen, 2001). In experimental tasks that measure WM, participants are required to

PAGE 16

6 actively maintain representations or manipulate information over the course of a delay. Anatomical analyses in primates have concluded that WM is mediated by a network of brain structures that includes the prefrontal cortex (e.g., Goldman-Rakic, 1987). Functional neuroimaging research in humans has shown that multiple frontal regions are active during short-term WM storage (including Brocas area and motor areas), while the dorsolateral prefrontal cortex (dlPFC) is involved in active maintenance of representations (Cohen, et al., 1997; Fuster, 1997; Smith & Jonides, 1999). Recent findings suggest that the dlPFC aids in the maintenance of information by directing attention toward internal representations of sensory and motor information that are stored in more posterior processing centers of the brain (Curtis & DEsposito, 2003). Understanding the neurochemical functioning of the PFC provides clues as to how WM may be integrated with affective processing. Research has shown that normal PFC activity is dependent on dopamine (DA), a catecholanime neurotransmitter that is modulated by the amygdala during times of stress (Arnsten, 1998). Both stress and pleasant affect can trigger the release of DA, which also increases the activity of the amygdala, in turn triggering more production of DA in the PFC (Ashby, Isen, & Turken, 1999). Unfortunately, when D1 receptors receive too much DA in the PFC, cognitive dysfunction such as poor attention, impaired response inhibition, and disruptions in WM often result (Arnsten, 1998, Goldman-Rakic, 1996). Paradoxically, mild levels of positive affect have been shown to enhance performance on WM tasks (Ashby et al., 1999), while anxiety can serve to impair both visuospatial (Lavric, Rippon, & Gray, 2003) and verbal WM (Ikeda, Iwanaga, and Seiwa, 1996). The end result of this appears to be that both too little and too much dopamine D1 stimulation are detrimental to

PAGE 17

7 prefrontal function (Goldman-Rakic, 1996, p. 13478), which may serve to take the dlPFC offline (Arnsten, 1998). If the dlPFC is taken offline, subcortical structures like the amygdala may have the opportunity to guide behavior with more automatic and reflexive affective processing and responding. Recent evidence has found that task-irrelevant pictures that are emotionally arousing cause greater interference effects in cognitive tasks (e.g., solving math problems) than emotional pictures that represent evolutionary threat or general negativity in low-fear individuals (Schimmack, 2005). This suggests that emotional arousal may be the key ingredient for interference of cognitive functioning; however, little is known about the underlying neural processes that are involved with affective interference on higher cognitive functions. It is also not clear if these effects would generalize to high-fear individuals, who show more susceptibility for threat-evoked decreases in dlPFC activity than controls (Carlsson et al., 2004). Further clues about the dissociations between affective and cognitive processing have been studied using fMRI in attentionand WM-related tasks. Recent research has found that the middle frontal gyrus (i.e., dorsolateral region) of the PFC is actively involved in the processing of cognitive targets but deactivated during the detection of novel, non-target emotional stimuli (Yamasaki, LaBar, & McCarthy, 2002). Perlstein, Elbert, and Stenger (2002) found that WM-related dlPFC activity was influenced by the emotional characteristics of task-relevant stimuli, but only when brought on-line by task demands that required active maintenance. Furthermore, they found that unpleasant affective content reduced WM-related brain activation in the dlPFC relative to neutral and pleasant content. Other researchers have sought to examine the effects of emotional

PAGE 18

8 states, finding that pleasant states enhance verbal WM and decrease spatial WM performance, while unpleasant states enhance spatial WM and decrease verbal WM performance (Gray, Braver, & Raichle, 2002). Overall, the lateral PFC was highly correlated with these behavioral changes, suggesting that this region of the brain is important for emotional-cognitive integration. This research has yielded many insights into the interactions between cognition and emotion, but a key disadvantage of fMRI research is that it offers less sensitivity to rapid changes in the brain compared to direct measures of neural activity, such as electrophysiological techniques like event-related potentials. Event-Related Potentials Before examining the contributions of event-related potential (ERP) methodology to the study of emotion, one must understand the basic assumptions behind ERP methods. One assumption is that the distribution of electrical activity across the scalp is reflective of the activities of underlying neural structures. A second assumption is that this neural activity corresponds with specific cognitive states and processes. As an extension of these assumptions, electrical potentials then represent information regarding cognitive states and processes (Kutas & Dale, 1995). The electrical activity of the brain can be measured non-invasively across the scalp using electrodes. The electroencephalogram (EEG) is the record of the volume-conducted electrical activity of the brain. EEGs can be used to observe ongoing electrical brain activity. Alternatively, electrical activity can be averaged in association with the presentation of specific events of interest. Initially, the event-related response associated with the presentation of a stimulus is embedded in the ongoing EEG activity. Extracting an ERP waveform associated with a specific stimulus is accomplished by averaging

PAGE 19

9 multiple samples of the EEG that are time-locked to repeated occurrences of the stimulus (see Figure 1-1). The benefit of averaging is that the ERPs should remain somewhat consistent from trial to trial, while the ongoing background EEG is random and will be averaged out. Figure 1-1. Extraction of the ERP waveform from ongoing EEG. (a) Stimuli (1N) are presented while the EEG is being recorded, but the specific response to each stimulus is too small to be seen in the much larger EEG. (b) To isolate the ERP from the ongoing EEG, the EEG segments following each stimulus are extracted and averaged together to create the averaged ERP waveform. Taken from Luck, Woodman, & Vogel (2000). ERPs are highly sensitive to changes in neural activity on the level of milliseconds (ms), making them the gold standard among noninvasive imaging methods in terms of temporal resolution (Fabiani, Gratton, & Coles, 2000). ERP waveforms usually consist of discrete voltage deflections that can be positive or negative, which are often followed by longer lasting potentials that are called slow-waves.

PAGE 20

10 Specific components of ERP waveforms are usually named in accordance with their polarity (positive or negative) and latency (in ms). A common example is P300, which refers to an ERP component with a positive peak that has latency of approximately 300 ms. Sometimes it is more appropriate for descriptors to use broader latency terminology, referring to deflections as early (100-200 ms) or late potentials (300-600 ms). Other descriptors may incorporate scalp location at which the component is maximal, such as late centro-parietal potentials. The neural activity associated with ERP activity is attributed primarily to post-synaptic potentials in pyramidal neurons of the cortex (Williamson & Kaufman, 1990). Neurons and the extracellular space that surrounds them are filled with charged ions like sodium (Na + ) and potassium (K + ). While neurons are at rest, Na + ions are more plentiful in the extracellular space than in the neurons, but this changes when action potentials cause the post-synaptic membrane to depolarize. When dendrites of a neuron depolarize, Na + ions flow into the cell, making the extracellular space more negative (known as the current sink). These Na + ions that enter the cell repel other positive ions (like K + ) away from the dendrites, creating a current that sends them toward the cell body. The accumulation of positively charged ions in the cell body (known now as the current source) repels positively charged ions away from the surrounding region in the extracellular space, sending those extracellular positive ions back toward the current sink. This process creates a dipolar extracellular current that accumulates with large populations of neurons, giving rise to detectable scalp potentials (Coles & Rugg, 1995). Neural Correlates of Cognitive and Affective Picture Processing Since phobia and specific fear often have strong visual components, much can be learned about their neural correlates by examining the neural processes underlying

PAGE 21

11 affective picture processing. Past investigations have demonstrated that a variety of ERP components have larger amplitudes to emotionally salient or arousing pictures (pleasant or unpleasant) compared to neutral pictures. This finding has been consistently associated with enhanced amplitudes of late positive potentials (including P300) and slow wave components (Cuthbert, Schupp, Bradley, Birbaumer, & Lang, 2000; Johnson, Miller, & Burleson, 1986). Some researchers have claimed that these effects in picture viewing paradigms have theoretical links to motivated attention, in which attentional resources are aroused and directed by motivationally relevant stimuli (Lang, Bradley, & Cuthbert, 1997). Research has shown that affective pictures elicit larger amplitudes for late positive potentials (LPPs) that typically begin 300-400 ms after picture onset (Cuthbert et al., 2000). These LPPs show greater amplitude differences when the pictures are more arousing and emotionally salient, even when the pictures are only presented briefly (Schupp et al., 2004). Visual processing of pictures associated with threat has evoked larger deflections in early LPPs (Baas, Kenemans, Bocker, & Verbaten, 2001). Early components of LPPs have been interpreted as a response of capturing attention, recognition, and stimulus evaluation (Donchin & Coles, 1988). As a result, ERPs can be used to investigate the amount of resources available for attentionally-dependent processing. Schupp and colleagues (1997) found that the attentional resources engaged while viewing affective pictures can serve to detract from the attentional resources that are available for other forms of processing, leading to a diminished probe startle ERP to a burst of noise. From this finding, it appears that the resources of attention that are available for a given task can be limited by other forms of attentional processing, even

PAGE 22

12 across different sensory domains. A provocative question that remains is how the attentional demands of higher-order cognitive processes, like WM, could potentially limit the attentional processing available for processing affective content. LPPs are not the only ERP components that are sensitive to affective stimuli. Enhanced ERP positivity from affective pictures may last up to 5 seconds after the picture onset in the form of slow waves (Cuthbert et al., 2000). Ruchkin and colleagues (1988) have proposed that positive slow waves vary with amount of memory storage required for processing. Positive slow potentials have been interpreted as reflecting increased sustained attention (Cuthbert et al., 2000) and possible roles in memory storage (Donchin & Coles, 1988). Given their apparent functional significance and capacity for lengthy durations, positive slow waves are likely to have meaningful effects subsequent objects of attention or motivation. Summary and Rationale for the Current Study Previous research has paved the way for a more comprehensive understanding of the complex interplay between cognitive and affective processing systems of the brain and the additional role that anxiety plays in these interactions. Emotional stimuli receive preferential processing on the perceptual and attentional level, and can even be processed automatically without conscious awareness. Emotional stimuli activate the amygdala, which modulates PFC concentrations of dopamine and may serve to disrupt prefrontally-mediated WM functioning in times of heightened emotional arousal. The dlPFC is involved in active maintenance of representations in WM, yet recent studies have also shown that its activity can be modulated by affective or emotional context, as well. Additionally, threat has been shown to modulate WM performance in both visuospatial and verbal domains.

PAGE 23

13 ERP methods can be used to detect rapid neural changes in the affective processing in order to map out a time-course of brain responses that are difficult to assess using other techniques. The method of ERP is well suited for measuring changes in neural processing that reflect changes in attentional processing, yet little research has used this technique to study cognitive-emotional interactions in this regard. Certain forms of affective processing (e.g., picture viewing) can serve to modulate the amount of attentional resources available for other forms of emotional behavior (e.g., startle), as evidenced by LPPs. However, much still needs to be learned about the role that anxiety and other forms of emotion play in modulating neural and cognitive behavior. The current study examined the potential role of task-irrelevant emotionally evocative interference stimulation on PFC-mediated WM processes. The primary objective was to determine the effects of task-irrelevant affective (pleasant and unpleasant) stimuli presented during the delay period of a task requiring active maintenance of stimulus representation in WM. Key questions regarding this issue include: 1) does affective interference capture attention and, thereby, lead to a degradation of the strength of task-relevant stimulus representations in WM, resulting in impaired performance; 2) does the emotional valence or arousal of these interfering stimuli differ in the extent to which they draw resources away from the active maintenance of task-relevant representations; 3) do evolutionarily-relevant threat stimuli (i.e., snakes, spiders) disproportionately capture attention and impair WM performance; and 4) does active maintenance of task-relevant representations in WM draw resources away from the ability to process task-irrelevant interfering stimuli? A secondary goal was to determine if individuals with high fear of these evolutionarily-relevant threat

PAGE 24

14 stimuli (e.g., specific phobics/sub-clinical phobics) evidence a disproportionate attentional capture by these stimuli and a consequent increase in WM impairment. To address these aims, highand controls performed high and low load conditions of a novel visual delayed matching-to-sample (DMS) task in which task-irrelevant pictorial interfering stimuli of different valence categories (pleasant, neutral, unpleasant, evolutionary threat) were presented during the delay or retention period. Task performance was measured by error rate and reaction time (RT); high-density event-related potentials (ERPs) indexed the extent to which task-irrelevant interfering stimuli were processed. Predictions WM Task Performance It is predicted that error rates and probe RTs during the WM task will be modulated by the emotional valence of the interfering pictures. Emotionally arousing interference pictures (pleasant, unpleasant, evolutionary threat) are expected to cause greater task performance decrements than neutral pictures. Furthermore, it is predicted that high-fear participants will show a disproportionate decrement in task performance to the fear specific (i.e., evolutionary threat) interference stimuli than controls, and that this dissociation will be most evident at the high WM load condition when active maintenance of task-relevant stimulus representations is greatest. Finally, it is not anticipated that high-fear subjects will show deficits in baseline WM task performance, as there is little evidence to suggest that under non-arousing conditions they have impaired dlPFC/WM function.

PAGE 25

15 ERP Reflections of Task-Irrelevant Interference Stimulus Processing If, as predicted, the active maintenance of task-relevant representations in WM draws resources away from the ability to process task-irrelevant interfering stimuli, it is expected that there will be load-dependent effects on the LPP components and slow waves evoked during the presentation of interference pictures. That is, the LPP components will be smaller in amplitude during the highthan low-load task conditions. Furthermore, it is predicted that the LPPs and slow wave amplitudes evoked by the interfering pictures will be greater in amplitude for emotionally-arousing than neutral interference stimuli, indicative of a greater capture-of-attention effect by emotionally-arousing interference stimuli. Finally, it is also predicted that high-fear participants will show greater LPP and slow wave amplitudes to the evolutionary threat interference compared to interference of other valence categories and compared to controls.

PAGE 26

CHAPTER 2 METHOD Participants Thirty individuals (16 female) between the ages of 19 and 45 participated in the study in exchange for course credit or financial compensation. All participants provided written informed consent in accordance with procedures of the University of Florida Health Science Center Institutional Review Board. All participants were given the Specific Phobia section of the Anxiety Disorders Interview Schedule for the DSM-IV (ADIS-IV; Brown et al., 1994), along with other relevant sections necessary to assess comorbid psychopathology. The ADIS-IV provides diagnostic information pertaining to all anxiety disorder categories and the full range of mood, substance-related, and somatoform disorders. Based on the results of the ADIS-IV, 14 of the participants were judged to have high levels of fear to either snakes or spiders, ranging from moderate to very severe. Six of the high fear participants also met diagnostic criteria for other anxiety disorders, including Panic Disorder without Agoraphobia, Generalized Anxiety Disorder, and Social Phobia. Gender makeup was not significantly different between the high-fear and control groups [ 2 (1, N = 30) = 1.27, p > .25], although 64% of the high-fear participants were female, compared to 44% of the controls. Age and educational background were also equated across fear group, as shown in Table 2-1. All of the participants completed the interview and WM task, yet one of the controls had to be excluded from the ERP analysis due to technical problems. This subject is included in the sample statistics for 16

PAGE 27

17 demographic information, emotional assessment, and WM task performance (N = 30), while the ERP data does not include this subject (N = 29). Although the ADIS-IV provided the diagnostic information necessary to form the two participant groups, several other assessments were used to quantify each participants level of emotional functioning and affective health. The Snake Anxiety Questionnaire and Spider Questionnaire (SNAQ and SPQ; Klorman, Hastings, Weerts, Melamed, & Lang, 1974) further assessed each participants fear of snakes or spiders. The State-Trait Anxiety Inventory (STAI; Spielberger, Gorsuch, & Lushene, 1970) was also used to provide broader information regarding the participants general levels of anxiety as manifest in temporary states of distress and more long-term personality traits. Finally the Beck Depression Inventory Second Edition, (BDI-II; Beck, Steer, & Brown, 1996) was used to assess for elevated levels of depressive symptoms. Participants were excluded if they reported depressive symptoms sufficient for a current diagnosis of major depressive disorder, previous neurological disease, traumatic brain injury (TBI), or current psychotropic medication use. A total of seven participants were excluded from participation due to excessive symptoms of depression or substance abuse. As shown in Table 2-1, participants in the high fear group exhibited significantly higher levels of specific fear, state anxiety, trait anxiety, and depressive symptoms compared to the controls. Levels of specific fear, state anxiety, trait anxiety, and depressive symptoms were highly correlated with each other across all subjects, as shown in Table 2-2. Mean BDI scores, while different between groups, were below clinical cut-off levels considered to reflect depression (e.g., 14 for mild depression; Beck et al., 1996).

PAGE 28

18 Table 2-1. Demographic and emotional functioning data for all participants Mean (SD) High Fear (n=14) Low Fear (n=16) t-statistic Age 25.3 (7.5) 25.0 (5.8) .117 ns Education 15.6 (1.7) 15.7 (2.1) -.165 ns SNAQ/SPQ 10.6 (2.6) 2.7 (2.9) 7.93 ** STAI-State 37.7 (10.1) 29.1 (11.4) 2.17 STAI-Trait 42.9 (11.3) 29.7 (6.7) 3.96 ** BDI 10.2 (5.1) 5.0 (5.4) 2.70 ns p > .85. p < .05. ** p < .001 Table 2-2. Intercorelations between measures of anxiety and depression STAI-State STAI-Trait BDI SNAQ/SPQ .391 .536 ** .538 ** STAI-State --.572 ** .539 ** STAI-Trait ----.719 *** p < .05. ** p < .01. p < .001. df = 28. Materials and Procedure Participants performed a visual delayed match to sample (DMS) task, using stimuli modified from Low, Rockstroh, Harsch, Berg, & Cohen (2000). Participants were seated in front of a computer screen (approximately one meter from screen to nose) and a keyboard. In this task, geometric shapes were presented as cue-probe pairs with pictorial interference occurring during the delay period (Figure 2-1). Participants gave forced-choice responses indicating if they saw the probe stimuli in the previously presented cue. Each participant completed a total of 400 trials. Each trial began with the computer presentation of a visual cue, which remained on the screen for 2500 ms. The cues consisted of two difficulty levels, including one diamond (low load) and three diamonds (high load). The high and low load trials were presented randomly with equal probability. The diamonds varied by size and orientation, and participants were instructed to remember each diamond in the cue by its individual combination of size and orientation features.

PAGE 29

19 A task-irrelevant interfering picture was presented in 80% of the trials. This interfering picture was presented 1000 ms following cue offset and remained on the screen for 750 ms, during which the participants were instructed to press a button to ensure that the picture was detected and encoded. The pictures were taken from the International Affective Picture System (IAPS; Lang, Bradley, & Cuthbert, 1998), and consist of four distinct valences, or emotional categories: pleasant, neutral, unpleasant, and evolutionary threat (snakes or spiders). Each of the four valence categories consisted of 25 pictures that were presented multiple times in a pseudo-randomized fashion over the course of the experiment. 1 Because of limited numbers of IAPS pictures that consist of snakes or spiders, this category of pictures were supplemented with several non-IAPS pictures. IAPS pictures have standardized ratings for pleasantness (valence) and arousal on a 1 to 9 scale (9 equals most pleasant and most arousing). The ratings for the pleasant, neutral, and unpleasant pictures used in this experiment can be seen in Table 2-3. With regard to pleasantness, pleasant pictures were rated significantly higher than neutral, t(48) = 18.8, p < .001, which were rated significantly higher than unpleasant, t(48) = 22.0, p < .001. With regard to arousal, pleasant and unpleasant pictures did not differ, t(24) = .22, p > .80, but these pictures were rated significantly higher than neutral, ts(24) > 10.2, ps < .001. In 20% of the trials, participants viewed a continuous fixation point in place of the interfering picture in order to create a no interference condition. 1 IAPS identification numbers for pleasant pictures are 1440, 1460, 1463, 1604, 1710, 2040, 2070, 2071, 4599, 4608, 4623, 4680, 5470, 5621, 5623, 7230, 7260, 7330, 8030, 8370, 8380, 8470, 8501, 8510; for neutral pictures: 2191, 2235, 2372, 2383, 2393, 2394, 2515, 2745.1, 2840, 2850, 2980, 5390, 5740, 7036, 7100, 7140, 7180, 7217, 7234, 7285, 7491, 7496, 7595, 7710, 8311; for unpleasant pictures: 2095, 2205, 2375.1, 2683, 2800, 6213, 6550, 6570, 6831, 6834, 6838, 8485, 9050, 9220, 9280, 9342, 9400, 9410, 9470, 9471, 9520, 9561, 9830, 9910, 9921; for evolutionary threat pictures: 1019, 1022, 1050, 1051, 1052, 1070, 1080, 1090, 1101, 1110, 1111, 1112, 1113, 1114, 1120, 1201, 1220.

PAGE 30

20 Table 2-3. Mean ratings of each group of IAPS pictures (from Lang et al., 1998) Mean Ratings (SD) Pleasant Neutral Unpleasant Valence 7.59 (.40) 5.25 (.47) 2.44 (.43) Arousal 5.51 (.92) 3.29 (.58) 5.56 (.88) A memory probe (single diamond in a box) was presented on the screen 2750 ms following the offset of the original cue. The probe remained on the screen for 1500 ms, during which the participants were asked to determine if the diamond shown was an exact match in both size and orientation to one of the previously shown diamonds for that trial. Participants pressed m to indicate that the probe diamond was a match to the cue and n to indicate that the probe diamond was not a match. 1000 ms following the offset of the probe, the next trial began with a new cue. Participants were instructed to look at a fixation point (+) during the time intervals between the cue, picture, and probe. Figure 2-1. Overview of each trial, showing the time course of each cue, picture interference, and probe. Since the probe diamond in this example matches one of the cue diamonds exactly in its size and orientation, the correct response for the trial shown here is m.

PAGE 31

21 After completing the experimental task, participants then rated each picture that they viewed during the course of the experiment in order to provide a manipulation check on effects of valence and arousal. These ratings were computerized, using the Self-Assessment Manikin (SAM) rating system developed by Lang (1980). Participants were instructed to view each picture for as long as they needed in order to be able to rate the pictures, and they self-terminated their viewing of each picture with a button press. Participants then rated each picture separately for valence and arousal, using the same 1 to 9 scale used in the standardized IAPS ratings (9 equals most pleasant and most arousing). EEG Acquisition and Reduction EEG Data Acquisition EEG was recorded from 64 scalp sites using a 64-channel geodesic sensor net (Figure 2-1) and amplified at 20K using an Electrical Geodesics Incorporated (EGI) amplifier system (nominal bandpass .10 100Hz). Electrode placements enabled recording vertical and horizontal eye movements reflected in electro-oculographic (EOG) activity: one placed above and below each eye and centered around the pupil to record vertical eye movements; the others placed at the outer canthus of each eye for recording horizontal eye movements. EEG was referenced to Cz and was digitized continuously at 250 Hz with a 16-bit analog-to-digital converter. A right posterior electrode served as common ground. The impedance of all electrodes was maintained below 50 k, consistent with procedures suggested by the manufacturer.

PAGE 32

22 Figure 2-2. Sensor layout of the 64-channel geodesic sensor net. Electrode #34 (outlined) was used for measurement of picture-related ERPs. EEG Data Reduction Due to the volume-conducting nature of the brain, no single scalp site can be considered an inactive reference site (Tucker, Liotti, Potts, Russell, & Posner, 1994); therefore, data were mathematically re-referenced against an average reference (Bertrand, Perrin, & Pernier, 1989). In this procedure, the activity of each electrode site is reflected as the difference between itself and the average of all the other recording sites. Editing of the EEG for movement, electromyographic muscle artifact, electro-ocular eye movement, and blink artifacts was performed by computer algorithm in Brain Electrical Source Analysis software (BESA version 5.0; Scherg, 1990). EEG during which voltage exceeded 150 V and with point-to-point transitions exceeded 125 V were excluded from averaging.

PAGE 33

23 Individual-subject event-related potentials (ERPs) were extracted and averaged together from the ongoing EEG recording in discrete temporal windows that coincided with the onset of each stimulus (refer back to Figure 1-1). ERP averages from each subject were divided into three categories: cue activity, interfering picture activity, and probe activity. Stimulus-locked epochs were extracted with a duration of 100ms prior to stimulus presentation (constituting a 100ms pre-stimulus baseline period) and 1000ms post-stimulus presentation. Collapsing across certain conditions was necessary because of insufficient numbers of responses to conduct specific analyses. All averaged ERP epochs were baseline corrected using a 100 ms window prior to stimulus onset and digitally filtered at 15 Hz low-pass and a .5 Hz high-pass. Although ERPs were acquired for stimulus-related activity for the cues, probes, and interfering pictures, only the ERPs for interfering pictures will be discussed in this study. Statistical Analyses WM task performance, picture responses, and ERP amplitudes were analyzed using repeated measures analyses of variance (ANOVAs). Because of the a priori hypotheses regarding group differences on these measures, planned contrasts were used to decompose interaction effects within the ANOVAs that were performed. When applicable, these contrasts utilized orthogonal comparisons of different levels of affective picture categories: 1) affective vs. neutral, 2) pleasant vs. negative (unpleasant and threat), and 3) unpleasant vs. threat. In addition to these planned contrasts, follow-up contrasts were employed and, where appropriate, adjusted for multiple comparisons using the modified Bonferroni method (Keppel, 1982). For ANOVAs where there were more than two levels of a within-subject factor, the Huynh-Feldt epsilon adjustment (Huynh & Feldt, 1976) was used; uncorrected degrees of freedom and corrected p-values are

PAGE 34

24 reported. In addition, two-tailed Pearson product-moment correlations were performed to determine associations between continuous variables of interest. WM Performance Data For each participant, error rates and probe RTs were calculated for each WM load and interfering picture valence category. For all RT measurements, median RTs were initially calculated for each participant in order to better accommodate outliers (Ratcliff, 1993), and then means of these individual RTs were compared in subsequent inferential analyses. Error rates and probe RTs for the no-interference condition were then examined for an effect of WM load, using a 2-Group x 2 Load repeated measures ANOVA. After detecting a significant effect of WM load on task performance, error rates and probe RTs were analyzed by subjecting them both to a 2-Group x 4-Valence ANOVA. Picture Response Data A 2-Group x 4-Valence repeated measures ANOVA was performed on the RTs of the interfering picture detection responses, as well as the mean viewing times during the picture rating procedure. Post-experimental picture ratings were subjected to 2-Group x 4-Valence repeated measures ANOVAs separately for ratings of pleasantness and arousal. ERP Data Analysis of ERP waveforms focused on activity reflecting processing of the interfering pictures. Statistical analyses of ERP waveforms assessed the mean voltages over specified temporal windows (epochs) of individual subject ERPs extracted from individual electrode sites for correct trials. Scoring windows and electrode positions for each condition of interest were determined by examination of grand-averaged ERP

PAGE 35

25 waveforms and the grand mean global power of the voltage obtained over all electrode sites. Picture-related ERP activity was quantified at electrode site #34 (Figure 2-2) and was examined over three different time periods believed to represent different aspects of visual processing. The first time period was 296 ms to 356 ms, which corresponds to the early portion of late positive potentials (LPP) associated with affective picture processing, while the second time period was 452 ms to 512 ms, corresponding to the late portion of LPP (Schupp et. al., 2003). The third time period was 660 ms to 760 ms, corresponding to the slow wave component of visual processing (Keil, et al., 2002). Thus, three ERP components associated with processing interfering pictures were examined: early LPP, late LPP, and slow wave. A 2-Group x 2 Load repeated measures ANOVA was performed on the amplitudes of each ERP component to determine the overall effects of WM load on ERPs, collapsing across different picture types to determine the effects of WM load on ERPs to the interfering pictures. Subsequently, amplitudes of each ERP component were analyzed in 2-Group x 4-Valence repeated measures ANOVAs to determine the effects of interfering picture valence on ERPs to the interfering pictures.

PAGE 36

CHAPTER 3 RESULTS Task Performance Initial analyses were performed to examine the possibility of a speed/accuracy trade-off by correlating error rates on the WM task and reaction times to the probe. These analyses revealed that probe reaction time (RT) and error-rate were not significantly correlated, r(27) = .145, p > .40, indicating that speed-accuracy trade-off does not account for the behavioral findings reported below. Effects of WM Load An important aspect of the present research was to demonstrate that the WM load manipulation was indeed associated with alterations in behavioral performance. To examine effects of WM load in the absence of interfering pictures, 2-Group x 2 Load repeated measures ANOVAs were performed on error rates and probe RTs for the no-interference trials. As expected, participants as a whole exhibited more errors, F(1,28) = 106.26, p < .001, 2 = .791, and longer probe RT, F(1,28) = 181.46, p < .001, 2 = .866, in high WM load trials compared to low load trials, as shown in Figure 3-1. Fear group did not exert a significant main effect on error rate, F(1,28) = .01, p > .90, 2 < .001, or probe RT, F(1,28) = 3.44, p > .07, 2 = .109. Furthermore, examination of Group x Load interactions revealed that high-fear participants did not significantly differ from controls in their load-related WM performance with regard to error rates, F(1,28) = .13, p > .70, 2 = .005, or probe RTs, F(1,28)=2.92, p > .10, 2 = .094, for no-interference trials. 26

PAGE 37

27 Effects of WM Load on Accurac00.050.10.150.20.250.30.350.4ControlsHigh-FearGroupProportion of Errors y Effects of WM Load on Probe RT0200400600800100012001400ControlsHigh-FearGroupTime (ms) Low Load High Load Figure 3-1. WM task performance for each WM load during no-interference trials. Error bars represent standard errors. These findings suggest that high WM load trials required more active maintenance of the stimulus representations (three diamonds) in WM than low load trials (one diamond), and that high levels of fear did not impair WM performance. Since the pictures presented during the delay were predicted to interfere with the active maintenance of WM representations, it is likely that the interference will have its maximal effect on the high WM load trials. As a result, subsequent analyses will focus on high WM load trials only. Effects of Interference on WM Task Accuracy A 2-Group x 4-Valence repeated measures ANOVA was performed on the error rates for high WM load trials to determine the effects of the interfering pictures. Planned orthogonal contrasts were used to test a priori hypotheses about the effects of interfering picture valence. Statistics for the ANOVA main effects and interactions can be found in Table 3-1, while results from the contrasts can be found in Table 3-2.

PAGE 38

28 Table 3-1. ANOVA statistics from the WM task accuracy data Effect F-Ratio 2 Group (G) a 1.25 .043 Valence (V) b 3.19 .102 G x V b 1.60 .054 a df = 1,28. b df = 3,84. p < .05. Table 3-2. Valence effects in the WM task accuracy data Valence (Across Group) Valence x Group (Across Load) Valence Contrast a F-Ratio 2 F-Ratio 2 Affective vs. Neutral 9.49 ** .253 1.56 .053 Negative vs. Pleasant .53 .018 3.11 t .100 Threat vs. Unpleasant .02 .001 .28 .010 a df = 1,28. t p < .10. ** p < .01. Means and standard errors of the error rates as a function of fear group and interference category can be seen in Figure 3-2. A main effect of picture valence was present for the error rates of the high WM load interference trials. Planned contrasts revealed that participants as a whole made more errors in each of the three affective picture trials compared to neutral. A trend-level Group x Valence interaction emerged, with high-fear participants making more errors than controls during negative (unpleasant and threat) picture trials relative to pleasant. Follow-up group-wise contrasts using modified Bonferroni-corrected comparisons at each valence level (critical p = .0188) revealed that controls exhibited more errors during pleasant picture trials than neutral (p < .0188). Error rates for high-fear participants did not significantly differ among the interfering picture valences or in comparison with those of the controls. Effects of Interference on WM Task Reaction Time A 2-Group x 4-Valence repeated measures ANOVA was performed on the RTs for high WM load trials to determine the effects of the interfering pictures. Planned orthogonal contrasts were used to test a priori hypotheses about the effects of interfering

PAGE 39

29 picture valence. Statistics for the ANOVA main effects and interactions can be found in Table 3-3, while results from the contrasts can be found in Table 3-4. Effects of Interference on Accuracy00.050.10.150.20.250.30.350.4NeutralPleas.Unpleas.ThreatInterference ConditionProportion of Errors Control High-Fear Figure 3-2. Error rates by interference category, and fear group. Error bars represent standard errors. Table 3-3. ANOVA statistics from the probe reaction time data Effect F-Ratio 2 Group (G) a 1.58 .053 Valence (V) b 3.76 .118 G x V b 1.78 .060 a df = 1,28. b df = 3,84. p < .05. Table 3-4. Valence effects in the probe reaction time data Valence (Across Group) Valence x Group (Across Load) Valence Contrast a F-Ratio 2 F-Ratio 2 Affective vs. Neutral .76 .026 5.60 .167 Negative vs. Pleasant 8.27 ** .228 .17 .006 Threat vs. Unpleasant 1.32 .045 .02 .001 a df = 1,28. p < .05. ** p < .01. A main effect of picture valence was present for the probe RTs of the high WM load interference trials, and planned contrasts revealed that participants took longer to

PAGE 40

30 respond to the probe during negative (unpleasant and threat) picture trials compared to pleasant. Means and standard errors of the reaction times to the probes as a function of fear group and interference category can be seen in Figure 3-3. Follow-up contrasts clarified that this negative picture effect was actually due to the threat pictures, as interfering threat picture trials resulting in significantly longer probe RTs than pleasant, (p < .0188), while unpleasant picture trials did not (p = .05). Follow-up group-wise contrasts using modified Bonferroni-corrected comparisons revealed that high-fear participants took longer to respond to the probe during threat picture trials than neutral (p < .0188), while controls failed to show this pattern (p > .200). Instead, controls exhibited faster probe RTs during pleasant picture trials compared to neutral (p < .0188). Faster probe RTs for pleasant picture trials in controls were not significantly correlated with increased error rates, r(27) = .013, p > .95, suggesting that they did not exhibit a speed/accuracy tradeoff for pleasant picture trials. Although high-fear participants took longer to respond in every category of interfering picture trial, their probe RTs did not significantly differ from controls. To summarize the WM task performance data, WM load had very strong effects on both accuracy and RT, with participants making more errors and taking longer to respond to the probe during the high WM load trials. Participants as a whole made more errors during trials involving affective picture interference compared to those involving neutral pictures. Controls made more errors and took less time to respond to the probe during pleasant picture trials compared to neutral. High-fear participants exhibited a trend for making more errors in the negatively-valenced (unpleasant and threat) picture trials compared to pleasant and exhibited significantly longer probe RTs to threat picture

PAGE 41

31 trials compared to neutral, showing that they demonstrated the expected pattern of WM impairment from the threat picture interference. Effects of Interference on Probe RT850900950100010501100115012001250NeutralPleas.Unpleas.ThreatInterference ConditionTime (ms) Control High-Fear Figure 3-3. Probe reaction time by interference category, and fear group. Error bars represent standard errors. Picture Response Data Interfering Picture Detection Data Timed responses were measured as participants indicated that they detected the interfering pictures during the WM task. While the primary goal of requiring participants to button-press to the presentation of interfering pictures was to increase the likelihood that they did, indeed, actually view the pictures, RTs to the presentation of interfering pictures could potentially shed light on aspects of picture processing. Thus, a 2-Group x 4-Valence repeated measures ANOVA was performed on the mean picture detection RTs. Means and standard errors of the reaction times to the pictures as a function of fear group and valence can be seen in Figure 3-4. A significant main effect of picture valence was seen in the interference picture detection, F(3,81) = 13.07, p < .001, 2 = .326.

PAGE 42

32 Participants were significantly faster in responding to interfering threat pictures compared to neutral, pleasant, and unpleasant pictures (ps < .0188). A main effect of fear group was also seen in these RTs, F(1,27) = 5.25, p < .05, 2 = .163, with high-fear participants taking significantly longer to detect the interfering pictures as a whole than controls. Mean Picture Detection RT460480500520540560580600620NeutralPleas.Unpleas.ThreatPicture CategoryTime (ms) Control High-Fear Figure 3-4. Mean picture detection reaction times by valence and fear group. Error bars represent standard errors. Viewing Times During Picture Rating Timed responses to pictures were measured as participants viewed the self-terminating pictures prior to making their subjective ratings at the end of the experiment. Means and standard errors of the picture viewing time as a function of fear group and valence can be seen in Figure 3-5. A significant main effect of valence emerged, F(3,84) = 7.48, p < .01, 2 = .211, as participants as a whole chose to view unpleasant pictures longer than pleasant (p < .0188) and threat (p < .0188) pictures. Although high-fear participants took more time than controls to view the pictures as a whole before rating them, this difference was not significant, F(1,28) = 1.72, p = .20, 2 = .058.

PAGE 43

33 Mean Picture Viewing Time16002100260031003600410046005100NeutralPleas.Unpleas.ThreatPicture CategoryTime (ms) Control High-Fear Figure 3-5. Mean RTs for picture viewing during the rating procedure at the end of the experiment. Error bars represent standard errors. Manipulation Check: Picture Rating Data An important premise of the current research with respect to modulation of WM by emotionally interfering stimuli, and examination of evolutionarily-relevant threat stimuli in lowand high-fear participants, is that the interfering stimuli were, indeed, perceived as both emotionally-arousing and as producing the intended emotional valence. These manipulation checks are examined below. Valence ratings Significant main effects were found for participants ratings of the pictures pleasantness, F(3,84) = 174.94, p < .001, 2 = .862, which can be seen in Figure 3-6. As expected, participants rated pleasant pictures as more pleasant compared to ratings of neutral, unpleasant, and threat pictures (ps < .0188). Also as expected, participants rated unpleasant pictures as more unpleasant than neutral and pleasant pictures (ps < .0188). However, this main effect of valence was qualified by a Group x Valence interaction, F(3,84) = 5.86, p < .01, 2 = .173, as high-fear participants rated threat pictures as more

PAGE 44

34 unpleasant than controls (p < .0188). Controls rated threat pictures as less unpleasant than the unpleasant pictures (p < .0188), while high-fear participants rated threat and unpleasant pictures as equally unpleasant (p > .95). Picture Valence Ratings123456789NeutralPleas.Unpleas.ThreatPicture CategoryPleasantness Rating Control High-Fear Pleasant Unpleasant Figure 3-6. Subjective ratings for picture valence. Error bars represent standard errors. p < .001 Arousal ratings Significant main effects were found for participants ratings of the pictures level of arousal, F(3,84) = 19.83, p < .001, 2 = .415, which can be seen in Figure 3-7. Participants rated the pleasant, unpleasant, and threat pictures as more arousing than the neutral (ps < .05). However, this main effect of arousal was qualified by a Group x Valence interaction, F(3,84) = 8.98, p < .001, 2 = .243, as controls rated pleasant pictures as more arousing than high-fear participants (p < .0188), while high-fear participants rated threat pictures as significantly more arousing than controls (p < .0188). Additionally, high-fear participants rated threat pictures as more arousing than neutral pictures (p < .0188), while controls did not (p > .35), but instead rated pleasant pictures

PAGE 45

35 as more arousing than neutral (p < .0188). There was no significant difference between high-fear participants arousal ratings of threat pictures and controls arousing pictures of pleasant pictures, t(28) = 1.138, p > .26, although Levenes test was significant, F(1,28) = 11.05, p < .01, indicating that these two sets of ratings did not show equal variances and this result should be interpreted with caution. Picture Arousal Ratings123456789NeutralPleas.Unpleas.ThreatPicture CategoryArousal Rating Control High-Fear w Hi g h ** L o Figure 3-7. Subjective ratings for picture arousal. Error bars represent standard errors. p < .05, ** p < .001 To summarize the picture response data, participants detected interfering threat pictures during the WM task faster than the other categories, and high-fear participants took longer than controls to signify their detection of interfering pictures. Picture SAM ratings, which provide a manipulation check on participants judgments of valence and arousal, demonstrated that the interfering pictures indeed were effective in producing the intended arousal and valence effects. Participants ratings of valence and arousal differentiated neutral, pleasant, and unpleasant pictures in ways that were generally expected. High-fear participants rated threat pictures as significantly less pleasant than

PAGE 46

36 the neutral pictures, giving the threat pictures more arousing ratings than controls. Controls surprisingly rated pleasant pictures as significantly more arousing than high-fear participants. These arousal ratings show interesting correspondence with the task performance data, as high-fear participants made exhibited errors and longer probe RTs during threat picture trials, while controls exhibited more errors and shorter RTs during pleasant picture trials. Finally, during SAM ratings, participants viewed unpleasant pictures the longest prior to making their ratings. Event-Related Potential (ERP) Data Effects of WM Load ERP responses to interfering pictures were acquired from all scalp electrodes, and converted into grand averaged waveforms of picture interference by load (Figure 3-8). Electrode #34 showed clearly defined LPP and slow wave components. Spherical-spline interpolated scalp voltage maps showed that a centro-parietal region, encompassing electrode #34, was sensitive to WM load-related effects in the early and late LPP epochs (Figure 3-9). Thus, for simplicity, and because a primary aim of the research was to examine affective interference effects on active maintenance in WM, electrode #34 was used in all of the WM-related ERP analyses that follow. Differential load-related activity can clearly be seen over the centro-parietal region of the scalp in the earlier stages of picture processing (LPP), yet this effect is absent in the slow wave (see Figure 3-9). Figure 3-10 illustrates the grand-averaged ERP waveforms for the low and high WM load trials. 2-Group x 2 Load repeated measures ANOVAs were performed to determine the overall effects of WM load on ERPs in the early LPP, late LPP, and slow wave epochs from the interfering pictures, collapsing across different

PAGE 47

37 picture types. As shown in Figure 3-11, voltage amplitudes measured from electrode #34 were found to be more positive during low WM load trials compared to the high load for the LPP time windows, yet no load effect was seen in the slow wave, F(1,27) = 1.484, p > .20, 2 = .052. The WM load effects on LPP were stronger for the early, F(1,27) = 28.367, p < .001, 2 = .512, than late, F(1,27) = 7.239, p < .05, 2 = .211, time window. Controls and high-fear participants were equally affected by these WM load effects, as no significant Group x Load interaction emerged, Fs(1,27) < .423, ps > .50, 2 < .015. both l 1 c or-vec.avr; 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 0 1000 -4.6 5.9 0 Time [ms] Front Low Load High Load Left Right Plot (V) both l 3 c or-vec.avr; Back Figure 3-8. Grand-averaged ERPs for all scalp sites during interference picture processing in the low (blue) and high (red) WM load conditions.

PAGE 48

38 Early LPP Late LPP Slow Wave Front: Side: Back: _ + + + Figure 3-9. Spherical-spline interpolated scalp voltage maps representing the differences in neural processing between the high and low WM loads for the early and late LPPs and slow wave. Voltage difference maps were calculated by subtracting high load activity from low load, and were calculated at 326ms (early LPP), 482ms (late LPP), and 710ms (slow wave) after interference picture onset. Positive voltage differences are indicated in red, showing strongest load-related differences in the early LPP and the late LPP. 0 250 500 750 1000 0 1 2 3 Time [ms] Early LPP Late LPP Slow Wave Low Load High Load Amplitude (V) Figure 3-10. Grand-averaged ERPs for site #34 during interference picture processing in the low (blue) and high (red) WM load conditions.

PAGE 49

39 Effects of WM Load on ERP Amplitudes-1012345Early LPPLate LPPSlow WaveERP ComponentAmplitude (uV) Low Load High Load Figure 3-11. Mean ERP amplitudes for low and high WM load conditions for the early and late LPP and slow wave. Error bars represent standard errors. Figure 3-11 clearly shows that early stages (LPP) of interference picture processing show greater effect of WM load than later stages (slow wave), which fits the scalp distribution maps of the load-related differences (Figure 3-9). In order to compare the effect of WM load on these different time-locked ERP components, a 2-Group x 2-Load x 3-Epoch repeated measures ANOVA was performed. Significant main effects of load, F(1,27) = 12.49, p = .001, 2 = .316, and epoch, F(1.64, 44.13) = 33.45, p < .001, 2 = .546, were found, but these were qualified by a significant Load x Epoch interaction, F(1.95, 52.54) = 11.05, p < .001, 2 = .290. Planned contrasts revealed that the effect of WM load on slow wave amplitudes was significantly smaller than it was for the two LPP epochs, F(1,27) = 13.42, p .001, 2 = .332. Of the two LPP time windows, the early LPP was modulated more by WM load than the late LPP, F(1,27) = 8.46, p < .01, 2 = .239. These load effects did not differ between controls and high-fear participants, as there was

PAGE 50

40 no main effect of group or group interactions with regard to load or epoch effects, Fs < .80, ps > .45, 2 < .030. Overall, the effects of WM load on the ERP components of interest support the prediction that fewer attentional resources are available for processing interfering picture stimuli during the delay period of the high WM load condition relative to the low load. That is, less neural activity is engaged in the capture of attention for the pictures while participants actively maintain greater stimulus representations during the high load conditions. Since the pictures and WM representations are both competing for similar reservoirs of neural processing, this finding suggests that more neural resources are allocated to the active maintenance of high load representations (three diamonds) compared to the low load (one diamond). As the high WM load engages more neural resources, the opportunity is optimal for interference effects caused by the pictures. As a result, only ERP components taken from high WM load trials will be examined in the analyses that follow. Effects of Picture Interference ERP responses to interfering pictures were acquired from all scalp electrodes, and converted into grand averaged waveforms of picture interference for each picture category (Figure 3-12). Spherical-spline interpolated scalp voltage maps showed that a central region (including electrode #34) was differentially sensitive to affective pictures relative to neutral in the early LPP epoch; however, this region of highest differential sensitivity shifted more centro-frontally during the late LPP and slow wave time windows (Figure 3-13). In determining the electrode site to use for evaluating the effects of picture interference, consideration was given to prior research involving affective pictures and

PAGE 51

41 the earlier findings of this study relating to WM-related effects. Since previous research has found centro-parietal regions to show maximal positive potentials in response to affective pictures and this region also showed sensitivity to load-related effects in this study, electrode #34 was used for the picture-related analyses that follow. Plot [V] Plot [V] Plot [V] both n eut c or-vec.avr; both p lea c or-vec.avr; both u npl c or-vec.avr; 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 0 1000 -4.6 7.3 0 Time [ms] Front Neutral Pleasant Unpleasant Thre at Left Right Plot (V) both c lin c or-vec.avr; Back Figure 3-12. Grand-averaged ERPs for all scalp sites during interference picture processing in the neutral (blue), pleasant (black), unpleasant (green), and threat (red) conditions. In line with the WM task performance data, 2-Group x 4-Valence repeated measures ANOVAs were performed on mean amplitudes taken from electrode #34 for the early LPP, late LPP, and slow wave time periods in order to determine the effects of

PAGE 52

42 the interfering pictures at the high WM load. Planned orthogonal contrasts were used to test a priori hypotheses regarding effects of interfering picture valence. Statistics for the ANOVA main effects and interactions can be found in Table 3-5, while results from the contrasts can be found in Table 3-6. Early LPP Late LPP Slow Wave Front: Side: Back: _ + + + Figure 3-13. Spherical-spline interpolated scalp voltage maps representing the differences in neural processing between affective and neutral pictures for the early and late LPPs and slow waves of high load trials. Voltage difference maps were calculated by subtracting neutral from affective picture activity, and were calculated at 326ms (early LPP), 482ms (late LPP), and 710ms (slow wave) after picture onset. Positive voltage differences are indicated in red, showing strongest affective-related differences in the early LPP and the late LPP. Table 3-5. ANOVA statistics from the ERP components Early LPP Late LPP Slow Wave Effect F-Ratio 2 F-Ratio 2 F-Ratio 2 Group (G) a .025 .001 .194 .007 .24 .009 Valence (V) b 6.70 *** .199 6.67 *** .198 12.64 *** .319 G x V b .173 .006 .619 .022 1.62 .056 a df = 1,27. b df = 3,81. *** p < .001.

PAGE 53

43 Table 3-6. Valence effects in the ERP components Valence (Across Group) Valence x Group (Across Load) Valence Contrasts a F-Ratio 2 F-Ratio 2 Early LPP Affective vs. Neutral 4.26 .136 .001 .001 Negative vs. Pleasant 16.47 *** .379 .641 .023 Threat vs. Unpleasant 2.20 .075 .009 .001 Late LPP Affective vs. Neutral 4.743 .149 .337 .012 Negative vs. Pleasant 15.37 *** .363 .591 .021 Threat vs. Unpleasant .056 .814 .886 .032 Slow Wave Affective vs. Neutral .021 .001 .08 .003 Negative vs. Pleasant 11.05 ** .290 .756 .027 Threat vs. Unpleasant 32.96 *** .550 5.38 .166 a df = 1,28. p < .05. ** p < .01. *** p < .001. 0 250 500 750 1000 -1 0 1 2 3 Time [ms] Early LPP Late LPP Neutral Pleasant Unpleasant Thre at Amplitude (V) Slow Wave Figure 3-14. Grand-averaged ERPs for site #34 during interference picture processing in the neutral (blue), pleasant (black), unpleasant (green), and evolutionary threat (red) conditions at high WM loads, collapsed across groups. Figure 3-14 illustrates the grand-averaged ERP waveforms as a function of interfering picture valence category during high WM load trials. As can be seen, valence category of the interfering pictures modulated the amplitudes of all ERP epochs. Planned contrasts revealed that participants as a whole exhibited greater positive deflections in both LPP time windows for each of the three affective picture types compared to neutral.

PAGE 54

44 Both LPPs and slow waves had elevated positive amplitudes for negative (unpleasant and threat) compared to pleasant interfering pictures. Additionally, slow waves were more positive for unpleasant compared to threat pictures. This slow wave effect was qualified by planned contrasts that revealed a significant Group x Valence interaction, with controls exhibiting significantly more negative slow waves than high-fear participants for threat relative to unpleasant pictures. Put another way, this indicates that the slow waves of high-fear participants were more positive than controls with respect to unpleasant slow waves. Figures 3-15 and 3-16 show the overall effects of all picture valence categories on ERP amplitudes for controls and high-fear participants, while Figure 3-17 shows grand-averaged ERP waveforms that reveal the Group x Valence interaction where high-fear participants show more positive slow wave for threat pictures (relative to unpleasant) than controls. Effects of Picture Valence on ERPs: Controls-3-2-1012345ERP EpochMean Amplitude (uV) Neutral Threat Pleas. Unpleas. Earl y LPPLate LPPSlow wave Figure 3-15. Mean ERP amplitudes for the early and LPP and slow wave for controls. Error bars represent standard errors.

PAGE 55

45 Effects of Picture Valence on ERPs:High-Fear-2-1012345ERP EpochMean Amplitude (uV) Neutral Threat Pleas. Unpleas. Earl y LPPLate LPPSlow wave Figure 3-16. Mean ERP amplitudes for the early and LPP and slow wave for high-fear participants. Error bars represent standard errors. 0 250 500 750 1000 -2 3 Time [ms] Early LPP Late LPP Slow Wave Figure 3-17. Grand-averaged ERPs for controls and high-fear participants during threat and unpleasant interference picture processing in the at high WM load. Control Threat High-Fear Threat Control Unpleas. High-Fear Unpleas. -1 0 1 2 Amplitude (V) The enhanced positivity exhibited in high-fear participants slow waves for the threat pictures is intriguing because it may account for group differences in sustained attention to the threat picture interference. The slow wave time window used for the

PAGE 56

46 present analysis was taken from 660-760 ms following picture onset, but as Figure 3-17 shows, the group difference in threat picture slow waves increases dramatically beyond this point. If the time window for slow wave analysis was extended to 1000 ms following picture onset, the group difference found in the slow waves would likely be much more powerful than the present analyses indicate. Unfortunately, activity recorded after 750 ms is confounded by the fact that a fixation point replaced the pictures at this time, making it difficult to interpret this activity as purely reflecting sustained attention and not some degree of response to the picture offset. One final analysis was conducted to determine if the increased positivity seen in the fear participants slow waves for the threat pictures was associated with behavioral symptomatology reported by the participants in the self-report measures. To accomplish this, mean slow wave amplitudes for threat pictures were correlated with mean scores on the SNAQ and SPQ, STAI-trait, STAI-state, and BDI. Correlations revealed that slow wave responses to the threat pictures were moderately correlated to phobic symptoms reported in the SNAQ/SPQ, r(27) = .363, p = .053, and the state anxiety reported in the STAI, r(27) = .408, p < .05. Correlations further revealed associations between threat slow waves and trait anxiety reported in the STAI, [r(27) = .417, p < .05 and depressive symptoms reported in the BDI, r(27) = .402, p < .05. These findings suggest that sustained attentional processing to interfering threat pictures is associated with behavioral symptoms of negative affect (including specific fear), which may account for group differences in WM impairment caused by threat picture interference. To summarize the ERP results, WM load manipulation had a strong effect on the neural activity devoted to processing the interfering pictures. This effect mirrors that

PAGE 57

47 seen in the WM task data. Low WM load trials evoked more positive early and late LPPs compared to the high load, suggesting that less attentional resources were available to process the pictures during high load trials. Differential load-related activity was centro-parietal in its scalp distribution; however, this effect diminished over time and was not significant in the slow wave components. With regard to the valence effects of the interfering pictures, LPPs had larger positive amplitudes for emotionally salient compared to neutral interfering pictures. This also reflects effects that were seen in the accuracy and RT data, suggesting that the greater the positivity of the interfering picture response, the greater the disruption in the maintenance of the WM representation. All ERP components showed greater centro-parietal positivity for negative pictures (unpleasant and threat) compared to neutral. Slow waves were most positive for unpleasant pictures in the centro-parietal region, yet controls and high-fear participants differed in their slow waves such that high-fear participants showed more positive activity for the interfering threat pictures than controls. The enhanced slow wave positivity for threat pictures was associated with self-reported behavioral symptoms of negative affect, including specific fear of the threatening content.

PAGE 58

CHAPTER 4 DISCUSSION This study investigated the effects of affective interference on visual working memory (WM). The main objectives were: 1) to determine if emotionally-arousing interference degrades the active maintenance of task-relevant representations in WM and, if so, 2) whether this degradation is disproportionately greater to interfering evolutionarily-relevant threat stimuli. Additionally, this study sought to determine if high-fear participants showed evidence of greater WM-related disruption to clinically-relevant interfering stimuli. Evidence from WM task performance and electrophysiological data converge to suggest that the representations in WM were differentially modulated by task-irrelevant affective interference. Changes in task performance attributable to both WM load and affective interference appear to be associated with rapidly changing neural processes. Taken together, these findings raise new questions about interactions between affect and higher-level cognitive processes like WM, and provide insight into the disruptive nature of fear and anxiety. Interference Effects on WM Task Performance WM load had very strong effects on both accuracy and RT, with participants making more errors and taking longer to respond to the probe during the highthan low-load conditions. This indicates that participants had more difficulty with the more challenging 3-diamond cues and engaged more WM resources to actively maintain representations of the higher load stimuli. Valence of the interfering picture had different group-related effects on accuracy and probe RT, as controls made more errors and took 48

PAGE 59

49 less time to respond to the probe during trials that involved pleasant compared to neutral picture interference; high-fear participants took more time to respond to the probe during threat picture trials and showed a trend-level increase compared to controls in errors during the threat and unpleasant relative to pleasant picture interference. The finding that controls made more errors and took less time to respond during pleasant picture trials is somewhat surprising at first; however, controls rated pleasant pictures as the most arousing category of pictures, significantly higher than neutral. Correlations between error rates and probe RTs for the pleasant picture trials were not significant, suggesting that they did not exhibit a speed-accuracy trade-off for the pleasant picture trials. Therefore, it appears instead that the controls WM task performance was influenced by arousal in such a way as to cause greater interference in WM representations while at the same time facilitating faster processing of the probe. Ashby (1999) proposed that moderate levels of positive affect enhance WM, but the pleasant pictures used in this study may have been too arousing for this benefit to be realized. This finding fits a pattern similar to that of the classic Yerkes-Dodson law (Yerkes & Dodson, 1908) that performance is related to arousal in an inverted U-shaped pattern such that too low or high arousal leads to impairment in cognitive performance. In this light, the controls arousal ratings may suggest that the pictures were too arousing and may have overwhelmed the PFC with excessive levels of dopamine (Arnsten, 1998; Goldman-Rakic, 1996), thereby disrupting WM function and leading to more impulsive responding. Task-related findings for high-fear participants fit some of the predictions that their WM performance would be most impaired during trials that involved threat pictures.

PAGE 60

50 Increased reaction time in probe responses may be indicative of a reduced cognitive capacity to actively maintain the representation of the cue stimuli and/or execute the recognition decision and response. Since high-fear participants rated threat pictures as the most arousing, this pattern also appears to fit the pattern of the Yerkes-Dodson law. However, it is interesting that this effect is opposite to that seen in controls where pleasant pictures actually facilitated faster response to the probe. This may suggest that the underlying neural processing involved in threat-based affective interference in high-fear participants may be functionally different from those involved in pleasant affective interference in controls. It seems plausible that this threat-based effect may be linked to a greater level of disengaging attention bias invoked by the threatening stimuli (Fox et al., 2001), which may not occur for pleasant pictures. This would account for the ERP findings of more positive slow waves to threat pictures in high-fear participants. Nevertheless, the dissociation between different valence-specific arousal effects cannot be fully explained with these data alone. Picture Response Data Participants detected interfering threat pictures during the WM task faster than the other categories, and high-fear participants took longer than controls to detect interfering pictures as a whole. This is an interesting finding, as it shows that threat pictures successfully captured the participants attention better than any other type of picture regardless of the their level of fear and that higher levels of anxiety has a cost on overall efficiency of attentional processing. This supports previous work that has found that threat cues evoke automatic attention biases (Ohman, Flykt, & Esteves, 2001), and may be explained by neural pathways that are present for both healthy and anxious individuals that route threatening information directly to the amygdala for automatic processing

PAGE 61

51 (Ledoux, 1996). It appears that the picture detection response (which involved a keypress) may have made an impact on the late LPP component, as the peak latencies for each valence group seem to coincide somewhat with the picture detection RTs. While making their ratings at the end of the experiment, participants chose to view unpleasant pictures the longest prior to making their ratings. One likely explanation for this is that the unpleasant pictures were more complex and required more sustained attention in order to rate them. Instead of simple pictures of snakes, smiling babies, or non-affective items such as an ironing board, unpleasant pictures included some complex scenes such as accidents, natural disasters, and saddening events. It is also possible that they depicted more novelty and provoked more curiosity in the participants. Participants ratings of valence and arousal differentiated neutral, pleasant, and unpleasant pictures in line with the expected patterns, revealing that the key manipulation of picture valence was successfully executed. High-fear participants rated threat pictures as significantly more arousing and less pleasant than controls, while controls rated pleasant pictures as significantly more arousing than other picture categories. ERP Reflections of Interference Stimulus Processing WM Load Effects on Interference-Related Activity As predicted, there was a clear load-related effect on ERPs to the task-irrelevant interfering stimuli. Both the early and late LPP was smaller during the highthan low-load condition. This finding suggests that fewer resources were available during the high-load maintenance interval to process interfering stimuli and extends previous work that observed a similar compromise of attentional resources with regard to picture processing and startle modulation (Schupp et al., 1997). However, the WM load modulation of picture-related LPPs decayed over time and did not affect the slow wave

PAGE 62

52 ERP component. Load-related WM modulation of interference picture processing was most pronounced for the earlier stages of attentional processing and, electrophysiologically, was less reflected during later stages of more sustained attention. Valence-Related Effects Interference Activity As predicted, the valence characteristics of the interfering stimuli influenced the amplitudes of the early and late LPP and slow wave. LPPs were larger for emotionally-arousing than neutral interference pictures, which supports previous research that affective pictures selectively engage attention based on motivational importance (Cuthbert et al., 2000; Schupp et al., 2003, Baas et al., 2001). Additionally, all three ERP components were larger in amplitude to negative (unpleasant and threat) than pleasant pictures. This finding is consistent with the hypothesis that the unpleasant and threat pictures recruited more attentional processing than did pleasant pictures. Slow waves were most positive for unpleasant pictures, suggesting that unpleasant pictures engaged more sustained attention than the other picture categories. Slow waves were very responsive to differences in picture valence category, but they appeared to be cut short by the offset of the picture; that is, there is a clear offset potential evoked by picture termination, and this offset potential disrupted the slow wave making it difficult to evaluate the duration of differential processing associated with emotional valence. A significant Group x Valence interaction revealed that controls and high-fear participants differed in their slow waves. High-fear participants showed more positive activity than controls for threat pictures relative to unpleasant. Unpleasant pictures still evoked the most positive slow waves in both groups, but high-fear participants exhibited dramatically more sustained processing of the threat pictures compared to controls. This finding may help to explain the increased WM impairment for high-fear participants

PAGE 63

53 caused by the threat picture interference. Surprisingly, no ERP effects seemed to correlate with the controls WM impairments caused by the pleasant picture interference. Interestingly, the scalp-voltage distribution maps shown in Figure 3-13 suggest that emotionally arousing versus neutral scalp-voltage difference maps became progressively more frontal over time. While this effect has not yet been quantitatively examined, there are several potentially important implications. First, the choice of electrode site for measuring valence-related effects may not have been optimal for sensitively detecting differences in the different components. Second, the progressively more frontal distribution over time may be consistent with the hypothesis that emotionally-arousing interference stimuli both disrupt active maintenance of task-relevant representations in WM and, furthermore, that representations of these task-irrelevant interfering stimuli become preferentially processed over representations of the task-relevant stimuli. That is, it is possible that representations of the task-irrelevant interference stimuli may actually become task relevant and displace those of the task-relevant stimuli in the dlPFC. Emerging Patterns The data suggest that affective interference disrupts the active maintenance of representations in WM and that this impairment is associated with the arousal characteristics of the interfering stimuli. High-fear participants exhibited greater allocation of sustained attention for threat pictures than controls, which corresponded with a longer RT to make recognition decisions during the WM task. This may suggest that high-fear participants have difficulty disengaging their attention from clinically-relevant threat stimuli, which may lead to maintenance of their fear. Controls also showed task-related patterns of WM impairment from interfering pleasant pictures

PAGE 64

54 (which they rated as highly arousing), but these impairments did not show the neural signature that was detected for an overall attention bias for threatening pictures or the threat-induced WM impairments in high-fear participants. Taken with the task performance data, the ERP results from this study suggest several important findings. First, highand low-fear individuals both show selective processing of affectively arousing stimuli, even in the context of a challenging cognitive task. Second, clinically-relevant threat interference has the capacity to disrupt neural processes of sustained attention that are needed for higher-level cognitive processes such as WM, and these changes in sustained attention are significantly associated with behavioral symptoms of negative affect, including specific fear. Third, when attentional resources are devoted to a challenging visual WM task, they are effectively subtracted from the attentional capacity that is available for the processing of incoming affective pictures, indicating that WM can diminish the degree of affective processing resources that are available. Taken together, these ERP effects provide novel insights into neural processes involved in attention, WM, and emotion that have not been previously examined. Future studies are needed to clarify the nature of these cognitive-affective interactions and expand these findings to other domains of cognitive and affective processing. Alternative Explanations and Possible Limitations The results of this study offer new hope into better understanding nature of affective interference of WM and the potential role that anxiety plays on higher-level cognition. However, alternative explanations and possible limitations need to be considered. The operational definitions associated with any form of emotion research are complex. Lang (1998) has postulated that emotions can be measured on two dimensions:

PAGE 65

55 valence and arousal. As a result, this study sought to define affective stimuli along these two dimensions. However, it is possible that the pictures varied on other criteria that were not controlled (e.g., brightness, complexity, human content) and may have affected the outcome. It is also possible that the sample selected for participation may have had life experiences that were not assessed that may have influenced their reactions to the pictures. Group differences may have also been influenced by demand characteristics that were placed on the high-fear participants (i.e., special attention given to them during recruitment and interview because of their higher levels of fear). Alternatively, it is also possible that the high-fear participants were not fearful enough for robust group effects to emerge, as other studies have found individuals with spider or snake phobias score much higher on the SNAQ/SPQ than the high-fear participants in this study. The mean SNAQ/SPQ scores for high-fear participants in this study was 10.6, while snake and spider phobics have obtained mean scores over 23 in other studies (e.g., Fredrikson, 1983). Although this study repeated pictures during the WM task, it did not have a sufficient number of trials to examine potential habituation effects; that is, repeated exposure to each of the pictures may reasonably be assumed to be associated with decreased arousal to subsequent exposures. However, given the limited number of trials per condition, there was not adequate signal-to-noise ratio in the ERP data to examine effects of repeated exposure. An additional concern is that the baseline accuracy condition for comparing groups was measured as performance no-interference trials, yet this condition was randomly mixed with the four categories of picture interference, making the no-interference condition a bit of an odd-ball that may not truly represent

PAGE 66

56 baseline WM performance. Lastly, this study had a small sample size that unfortunately could not be completely balanced on gender in each group due to the low numbers of available high-fear individuals. Several additional potential limitations of the ERP data must be considered. First, terms like late positive potential have somewhat limited interpretational value aside from discussions of general attentional resources. Part of the reason for this is that volume conduction of electrical signals in the brain makes it very difficult to map sources of brain activity with precise spatial resolution, although new technologies are emerging that make source localization more feasible and effective. A considerable amount of ERP variability exists from participant to participant, and some participants exhibit the same ERP components at slightly different latencies. Grand-averaging subjects together in certain cases might actually average out a key component of interest. Also, due to the number of trials that were available, no ERP analyses were performed on incorrect trials, meaning that accuracy could not be considered as a variable in the ERP analysis. Another concern emerged after analyzing the data taken from the centro-parietal electrode, as the sensitivity to affective picture characteristics shifted more frontally and may have been more pronounced if the slow wave had been analyzed from a more frontal electrode site. It appears that 750 ms may not have not been a long enough picture presentation to assess full slow wave effects. Finally, since analysis of the cue and probe waveforms went beyond the scope of this project, it is difficult to completely interpret the picture processing ERPs without a more complete understanding of the neural responses to the task-relevant stimuli associated with the WM task.

PAGE 67

57 Future Directions Despite its limitations, this study provides important new information about the interfering effects of emotional arousal on active maintenance of task-relevant representations within WM. For the sake of complexity, and to address the primary aims of the current research, task-relevant stimulus-evoked activity (i.e., cues and probes) was not examined. These data will be examined in the future to more fully elucidate the role of affective interference on WM-related processes. Additionally, as the arousal level within a valence category has been shown to modulate startle probe and LPP amplitudes during emotional picture viewing (i.e., larger LPP and startle with higher arousal ratings within category; Schupp et al., 1997), it would be useful to examine arousal-related effects more parametrically. By necessity, in the current study, arousal and emotional versus neutral stimuli are confounded. That is, pleasant and unpleasant stimuli were both rated as high in arousal, while neutral stimuli were rated as low in arousal. While localization of the neural sources giving rise to the observed interference effects was not a primary aim of the study, such information will certainly be useful. Findings by Perlstein et al. (2002) and Gray et al. (2002) have indeed shown that WM-related dlPFC activity is modulated by emotionally arousing contexts. Future efforts will be aimed at applying source localization routines to the current data using Brain Electric Source Analysis (BESA; Scherg, 1990) to determine if the effects observed in the current study are similarly localizable to the dlPFC. Also, the paradigm used for the present study is currently being adapted for the functional MRI context, in order to facilitate examination of brain activity that offers a higher degree of spatial resolution. Future extensions of the project could also include measuring different types of specific fear or anxiety (e.g., needle phobia, PTSD, etc.), to learn whether the findings in the current

PAGE 68

58 study generalize to other clinical populations. Furthermore, this paradigm could potentially be used to assess behavioral and electrophysiological effects of therapy that focuses on reduction of fear symptoms in phobic individuals. Summary This study sought to investigate the effects of affective interference on visual working memory (WM), testing whether affective interference impairs representations in working memory (WM) and the degree to which this interference is due to effects of arousal brought about by the clinical or evolutionary relevance of threat in participants with high levels of fear. Performance data from the WM task revealed that WM performance is modulated by affective interference and that arousal plays a key role in that interference, causing high-fear participants to respond more slowly as a result of threat interference and controls to make more errors and respond more quickly as a result of pleasant picture interference. Subtle differences in sustained attention appear to be the only source of group differences in the neural effects associated with threat-based interference of WM, but no neural effects appear to explain the effects of pleasant picture WM interference in controls. As a whole, participants demonstrated selective detection of threatening stimuli, which corresponded to enhanced positivity in late positive potentials measured at the centro-parietal region of the scalp. Importantly, ERP results showed that increased WM load and the processing of affective interference compromised the neural resources of attention. In the end, these findings help to better illustrate the complexity of cognitive-affective interactions, and indicate that anxiety can play a key role in disrupting higher-level cognitive processes involved in WM.

PAGE 69

LIST OF REFERENCES Adolphs, R., Gosselin, F., Buchanan, T.W., Tranel, D., Schyns, P., & Damasio, A.R. (2005). A mechanism for impaired fear recognition after amygdala damage. Nature, 433 (7021), 68-72. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4 th ed., text revision). Washington, DC: Author. Armony, J.L. & Dolan, R.J. (2002). Modulation of spatial attention by fear-conditioned stimuli: An event-related fMRI study. Neuropsychologia, 40, 817-826. Armony J.L. & Ledoux, J.E. (1997). How the brain processes emotional information. Annals of the New York Academy of Sciences, 821, 259-270. Arnsten, A.F.T. (1998). Catecholamine modulation of prefrontal cortical cognitive function. Trends in Cognitive Sciences, 2 (11), 436-447. Ashby, F.G., Isen, A.M., & Turken, A.U. (1999). A neuropsychological theory of positive affect and its influence on cognition. Psychological Review, 106 (3), 529-550. Baas, J.M.P., Kenemans, J.L., Bocker, K.B.E., & Verbaten, M.N. (2002). Threatinduced cortical processing and startle potentiation. Neuroreport, 13, 133-137. Baddeley, A.D. (1986). Working memory. Oxford: Oxford University Press. Beck, A.T., Steer, R.A., & Brown, G.K. (1996). Manual for the BDI-II. San Antonio, TX: The Psychological Coorperation. Bertrand, O., Perrin, F., & Pernier, J. (1985). A theoretical justification of the averagereference in topographic evoked potential studies. Electroencephalography and Clinical Neurophysiology, 62, 462-464. Brown, T. A., DiNardo, P. A., & Barlow, D. H. (1994). Anxiety disorders interview schedule for DSM-IV (ADIS-IV). Albany, NY: Graywind Publications. Carlsson, K., Petersson, K.M., Lundqvist, D., Karlsson, A., Ingvar, M. & Ohman, A. (2004). Fear and the amygdala: Manipulation of awareness generates differential cerebral responses to phobic and fear-relevant (but not feared) stimuli. Emotion, 4 (4), 340-353. 59

PAGE 70

60 Cohen, J.D., Perlstein, W.M., Braver, T.S., Nystrom, L.E., Noll, D.C., Jonides, J. & Smith, E.E. (1997). Temporal dynamics of brain activation during a working memory task. Nature, 386, 604-608. Coles, M.G.H. & Rugg, M.D. (1995). Event-related brain potentials: an introduction. In M.G.H. Coles & M.D. Rugg (Eds.), Electrophysiology of mind: Event-related potentials and cognition (pp. 1-35). Oxford, England: Oxford University Press. Compton, R.J. & Banich, M.T. (2003). Paying attention to emotion: An fMRI investigation of cognitive and emotional Stroop tasks. Cognitive, Affective, & Behavioral Neuroscience, 3 (2), 81-96. Curtis, C.E. & DEsposito, M. (2003). Persistent activity in the prefrontal cortex during working memory. Trends in Cognitive Sciences, 7 (9), 415-423. Cuthbert, B.N., Schupp, H.T., Bradley, M.M., Birbaumer, N., & Lang, P.J. (2000). Brain potentials in affective picture processing: Covariation with autonomic arousal and affective report. Biological Psychology, 52, 95-111. Dolan, R.J. (2002). Emotion, cognition, and behavior. Science, 298, 1191-1194. Donchin, E. & Coles, M.G. (1988). Is the P300 component a manifestation of context updating? Behavioral and Brain Sciences, 11, 357-427. Fabiani, M., Gratton, G., & Coles, M.G.H. (2000). Event-related potentials. In J.T. Cacioppo, L.G. Tassinary & G.G. Bernston (Eds.), Handbook of psychphysiology (2 nd ed.). Cambridge, England: Cambridge University Press. Fox, E. (1993). Allocation of visual attention and anxiety. Cognition and Emotion, 8, 165-195. Fox, E., Russo, R., Bowles, R., & Dutton, K. (2001). Do threatening stimuli draw or hold visual attention in subclinical anxiety? Journal of Experimental Psychology: General, 130 (4), 681-700. Fuster, J.M. (1997). The prefrontal cortex: Anatomy, physiology, and neuropsychology of the frontal lobe (3 rd ed.). New York: Raven. 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. Goldman-Rakic, P.S. (1987). Circuitry of primate prefrontal cortex and regulation of behavior by representational memory. In F. Plum (Ed.), Handbook of physiology; The nervous system (pp. 373-414). Bathesda, MD: American Physiological Society.

PAGE 71

61 Goldman-Rakic, P.S. (1996). Regional and cellular fractionation of working memory. Proceedings of the National Academy of Sciences, 93, 13473-13480. Hermans, D., Vansteenwegen, D., & Eelen, P. (1999). Eye movement registration as a continuous index of attention deployment: Data from a group of spider anxious students. Cognition and Emotion, 13, 419-434. Huynh, H. & Feldt, L.S. (1976). Estimation of the box correction for degrees of freedom from sample data in the randomized block and split plot designs. Journal of Educational Statistics, 1, 69 -82. Ikeda, M., Iwanaga, M., & Seiwa, H. (1996). Test anxiety and working memory system. Perceptual Motor Skills, 82, 1223-1231. Johnson, V.S., Miller, D.R., & Burleson, M.H. (1986). Multiple P3s to emotional stimuli and their theoretical significance. Psychophysiology, 23, 684-693. Keil, A., Bradley, M. M., Hauk, O., Rockstroh, B., Elbert, T., & Lang, P. J. (2002). Large-scale neural correlates of affective picture processing. Psychophysiology, 39, 641-649. Keppel, G. (1982). Design and analysis: A researchers handbook, (2 nd ed.). Englewood Cliffs, NJ: Prentice-Hall. Kindt, M. & Brosschot, J.F. (1997). Phobia-related cognitive bias for pictorial and linguistic stimuli. Journal of Abnormal Psychology, 106 (4), 644-648. Klorman, R., Hastings, J.E., Weerts, T.C., Melamed, B.G., & Lang, P.J. (1974). Psychometric description of some specific-fear questionnaires. Behavior Therapy, 5, 401-409. Kutas, M. & Dale, A. (1995). Electrical and magnetic readings of mental functions. In M.D. Rugg (Ed.), Cognitive Neuroscience (pp 187-242). Cambridge, MA: MIT Press. Lane, R.D., Reiman, E.M., Bradley, M.M., Lang, P.J., Ahern, G.L., Davidson, R.J., & Schwartz, G.E. (1997). Neuroanatomical correlates of pleasant and unpleasant emotion. Neuropsychologia, 35, 1437-1444. Lang, P. J. (1980). Behavioral treatment and bio-behavioral assessment: Computer applications. In J. B. Sidowski, J. H. Johnson, & T. A. Williams (Eds.), Technology in mental health care delivery systems (pp. 119-137). Norwood, NJ: Ablex.

PAGE 72

62 Lang, P.J., Bradley, M.M., & Cuthbert, B.N. (1997). Motivated attention: Affect, activation, and action. In P. Lang, R.F. Simmons, & M. Balaban (Eds.), Attention and orienting: Sensory and motivational processes: Hillsdale, NJ: Erlbaum Associates. Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1998). International affective picture system (IAPS): Technical manual and affective ratings. Gainesville, FL: The Center for Research in Psychophysiology, University of Florida. Lavric, A., Rippon, G., & Gray, J.R. (2003). Threat-evoked anxiety disrupts spatial working memory performance: An attentional account. Cognitive Therapy and Research, 27 (5), 489-504. Ledoux, J. E. (1996). The emotional brain. New York, NY: Simon & Schuster. Low, A., Rockstroh, B., Harsch, S., Berg, P., & Cohen, R. (2000). Event-related potentials in a working-memory task in schizophrenics and controls. Schizophrenia Research, 46, 175-186. Luck, S.J., Woodman, G.F., & Vogel, E.K. (2000). Event-related potential studies of attention. Trends in Cognitive Sciences, 4 (11), 432-440. MacLeod, C.M. (1991). Half a century of research on the Stroop effect: An integrative review. Psychological Bulletin, 190, 163-203. MacLeod, C.M, Matthews, A., & Tata, P. (1986). Attentional bias in emotional disorders. Journal of Abnormal Psychology, 95, 15-20. Miller, E.K. & Cohen, J.D. (2001). An integrative theory of prefrontal cortex function. Annual Review of Neuroscience, 24, 167-202. Morris, J.S., Frith, C.D., Perrett, D.O., Rowland, D., Young, A.W., Calder, A.J., & Dolan, R.J. (1996). A differential neural response in the human amygdala to fearful and happy facial expressions. Nature, 383 (6603), 812-815. Ohman, A., Flykt, A., & Esteves, F. (2001). Emotion drives attention: Detecting the snake in the grass. Journal of Experimental Psychology: General, 130 (3), 466478. Ohman, A. & Mineka, S. (2001). Fear, phobias, and preparedness: Toward an evolved module of fear and fear learning. Psychological Review, 108, 483-522. Ohman, A. & Mineka, S. (2003). The malicious serpent: Snakes as a prototypical stimulus for an evolved module of fear. Current Directions in Psychological Science, 12 (1), 5-9.

PAGE 73

63 Ohman, A. & Soares, J.J.F. (1998). Emotional conditioning to masked stimuli: Expectancies for aversive outcomes following nonrecognized fear-relevant stimuli. Journal of Experimental Psychology: General, 127, 69-82. Ohman, A., & Soares, J.J.F. (1993). On the automatic nature of phobic fear: Conditioned electrodermal responses to masked fear-relevant stimuli. Journal of Abnormal Psychology, 102. 121-132. Ohman, A., & Soares, J.J.F. (1994). Unconscious anxiety: Phobic responses to masked stimuli. Journal of Abnormal Psychology, 103, 231-240. Perlstein, W.M., Elbert, T., & Stenger, V.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. Posner, M.I., Inhoff, A.W., Friedrich, F.J., & Cohen, A. (1987). Isolating attentional systems: A cognitive-anatomical analysis. Psychobiology, 15, 107-121. Ratcliff, R. (1993). Methods for dealing with reaction time outliers. Psychological Bulletin, 114, 510-532. Runchkin, D.S., Johnson, R., Mahaffey, D., & Sutton, S. (1988). Toward a functional categorization of slow waves. Psychophysiology, 25, 339-353. Sadock, B.J. & Sadock, V.A. (2003). Kaplan and Sadocks synopsis of psychiatry: Behavioral sciences/clinical psychiatry (9 th ed.). Philadelphia, PA: Lippincott Williams & Wilkins. Scherg, M. (1990). Fundamentals of dipole source potential analysis. In F. Grandori & M. Hoke (Eds.), Auditory evoked magnetic fields and electric potentials. Advances in audiology (Vol. 6, pp. 65-78). Basel: Karger. Schimmack, U. (2005). Attentional interference effects of emotional pictures: Threat, negativity, or arousal? Emotion, 5 (1), 55-66. Schupp, H.T., Cuthbert, B.N., Bradley, M.M., Birbaumer, N., & Lang, P.J. (1997). Probe P3 and blinks: Two measures of affective startle modulation. Psychophysiology, 34, 1-6. Schupp, H.T., Junhhofer, M., Weike, A.I., & Hamm, A.O. (2003). Emotional facilitation of sensory processing in the visual cortex. Psychological Science, 14 (1), 7-13. Schupp, H.T., Junhhofer, M., Weike, A.I., & Hamm, A.O. (2004). The selective processing of briefly presented affective pictures: An ERP analysis. Psychophysiology, 41, 441-449.

PAGE 74

64 Smith, E.E. & Jonides, J. (1999). Storage and executive processes in the frontal lobes. Science, 283, 1657-1661. Spielberger, C. S., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the state trait anxiety inventory. Palo Alto, CA: Consulting Psychologists Press. Tucker, D.M., Liotti, M., Potts, G.F., Russell, G.S., & Posner, M.I. (1994). Spatiotemporal analysis of brain electrical fields. Human Brain Mapping, 1, 134152. Vrana, S., Roodman, A., & Beckham, J. (1995). Selective processing of trauma-relevant words in post-traumatic stress disorder. Journal of Anxiety Disorders, 9, 515-530. Williams, J.M.G., Watts, F.N., Macleod, C., & Matthews, A. (1997). Cognitive psychology and emotional disorders, (2 nd ed.). Chichester, England: Wiley. Williamson, S.J. & Kaufman, L. (1990). Theory of neuroelectric and neuromagnetic fields. In F. Grandor, M. Hokle, & G.L. Romani (Eds.), Auditory Evoked Magnetic Fields and Electric Potentials. Advances in Audiology (Vol. 6, pp. 139). Basel, Switzerland: Karger Publications. Yamasaki, H., LaBar, K.S., & McCarthy, G. (2002). Dissociable prefrontal brain systems for attention and emotion. Proceedings of the National Academy of Sciences, 99, 11447-11451. Yerkes, R.M. & Dodson, J.D. (1908). The relation of strength of stimulus to rapidity of habit-formation. Journal of Comparative Neurology and Psychology, 18, 459-482.

PAGE 75

BIOGRAPHICAL SKETCH David Stigge-Kaufman received his B.A. degree from Bethel College in North Newton, Kansas, in 1998, majoring in biology and psychology. He plans to receive his M.S. from the University of Florida in 2005, and then plans to continue his doctoral study in clinical psychology, concentrating his clinical and research training in neuropsychology. 65


Permanent Link: http://ufdc.ufl.edu/UFE0010550/00001

Material Information

Title: Interference Effects of Anxiety and Affective Processing on Working Memory: Behavioral and Electrophysiological Accounts
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0010550:00001

Permanent Link: http://ufdc.ufl.edu/UFE0010550/00001

Material Information

Title: Interference Effects of Anxiety and Affective Processing on Working Memory: Behavioral and Electrophysiological Accounts
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0010550:00001


This item has the following downloads:


Full Text












INTERFERENCE EFFECTS OF ANXIETY AND AFFECTIVE PROCESSING ON
WORKING MEMORY: BEHAVIORAL AND ELECTROPHYSIOLOGICAL
ACCOUNTS














By

DAVID ANDREW STIGGE-KAUFMAN


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2005





























Copyright 2005

by

DAVID ANDREW STIGGE-KAUFMAN















ACKNOWLEDGMENTS

I thank my advisor, William M. Perlstein, for his excellent guidance and support

during this project. I also thank the other members of the Clinical-Cognitive

Neuroscience Lab for their assistance. Lastly, I thank my wife and family for their love

and support.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ................................................................................................. iii

LIST OF TABLES ................................ .. .......... .................................... vi

LIST OF FIGURES ............................. .. .......... .................................. vii

A B S T R A C T ........................................................................................................ ........... ix

CHAPTER

1 IN TR O D U C T IO N ........ .. ......................................... ..........................................1.

P hobias and F ear .......................................... .... ... ....... ....... ..................... 1
Attentional and Perceptual Components of Emotional Processing ............................. 2
Emotional Connections with Working Memory............... .....................................5...
E v ent-R elated P otentials ...........................8.......... .. ....................... ..................... 8
Neural Correlates of Cognitive and Affective Picture Processing .......................... 10
Summary and Rationale for the Current Study............... .....................................12
P re d ictio n s .............................................................................................................. . 14
W M T ask P perform ance...................................................................... ............... 14
ERP Reflections of Task-Irrelevant Interference Stimulus Processing ............ 15

2 M E T H O D ................................................................................................................. .. 16

P a rtic ip a n ts ................................................................................................................ 1 6
M materials and P procedure .................................................................... ............... 18
E E G A acquisition and R eduction............................................................ ................ 21
E E G D ata A acquisition ......................................... ........................ ................ 2 1
E E G D ata R education ................................................................... ................ 22
Statistical Analyses .................. .. ............ ............................... 23
W M P perform ance D ata ........................................................................................24
P picture R response D ata ......................................... ........................ ................ 24
E R P D ata ............................................................................................................. 2 4

3 R E S U L T S ................................................................................................................. .. 2 6

T ask P perform an ce .................................................. ............................................. 2 6
E effects of W M L oad .............................................. ......................... ............... 2 6









Effects of Interference on WM Task Accuracy................................................ 27
Effects of Interference on WM Task Reaction Time ....................................28
P picture R response D ata ..................................................................... ... ......... 3 1
Interfering Picture D election D ata.................................................. ................ 31
View ing Tim es During Picture Rating ........................................... ................ 32
Manipulation Check: Picture Rating Data............... .....................................33
V alence ratings .................................................................. ............... 33
A rou sal rating s ........................................................ ................ ... . ........ 34
Event-R elated Potential (ERP) D ata...................................................... ................ 36
E effects of W M L oad .............................................. ......................... ............... 36
Effects of Picture Interference.................. .................................................. 40

4 D IS C U S SIO N ................................................................................................... 4 8

Interference Effects on WM Task Performance ...................................................... 48
Picture R response D ata ................... ............................................... 50
ERP Reflections of Interference-Stimulus Processing .........................................51
WM Load Effects on Interference-Related Activity ....................51
Valence-Related Effects Interference Activity ..............................................52
Emerging Patterns................. .... ..... ............. ...................... 53
Alternative Explanations and Possible Limitations...............................................54
F u tu re D direction s ........................................................................................................ 57
S u m m a ry .................................................................................................................. .. 5 8

L IST O F R EFE R E N C E S .... ....................................................................... ................ 59

BIO GR APH ICAL SK ETCH .................................................................... ................ 65

























v















LIST OF TABLES


Table page

2-1. Demographic and emotional functioning data for all participants ............................. 18

2-2. Intercorelations between measures of anxiety and depression .................................. 18

3-1. ANOVA statistics from the WM task accuracy data.............................................. 28

3-2. Valence effects in the WM task accuracy data....................................................... 28

3-3. ANOVA statistics from the probe reaction time data...........................................29

3-4. Valence effects in the probe reaction time data....................................................29

3-5. ANOVA statistics from the ERP components......................................................42

3-6. V alence effects in the ERP com ponents............................................... ................ 43















LIST OF FIGURES


Figure page

1-1. Extraction of the ERP waveform from ongoing EEG..........................................9...

2-1. Overview of each trial, showing the time course of each cue, picture
interference, and probe .......................................... .. ....................... ............... 20

2-2. Sensor layout of the 64-channel geodesic sensor net..........................................22

3-1. WM task performance for each WM load during no-interference trials.............. 27

3-2. Error rates by interference category, and fear group...........................................29

3-3. Probe reaction time by interference category, and fear group ...............................31

3-4. Mean picture detection reaction times by valence and fear group ........................32

3-5. Mean RTs for picture viewing during the rating procedure at the end of the
ex p e rim e n t............................................................................................................... 3 3

3-6. Subjective ratings for picture valence ................................................. ................ 34

3-7. Subjective ratings for picture arousal ................................................. ................ 35

3-8. Grand-averaged ERPs for all scalp sites during interference picture processing in
the low (blue) and high (red) WM load conditions.............................................37

3-9. Spherical-spline interpolated scalp voltage maps representing the differences in
neural processing between the high and low WM loads for the early and late
L P P s an d slow w av e ................................................................................................. 3 8

3-10. Grand-averaged ERPs for site #34 during interference picture processing in the
low (blue) and high (red) W M load conditions...................................... ............... 38

3-11. Mean ERP amplitudes for low and high WM load conditions for the early and
late L PP and slow w ave.. .. ............................................................... ................ 39

3-12. Grand-averaged ERPs for all scalp sites during interference picture processing in
the neutral (blue), pleasant (black), unpleasant (green), and threat (red)
c o n d itio n s. ............................................................................................................... 4 1









3-13. Spherical-spline interpolated scalp voltage maps representing the differences in
neural processing between affective and neutral pictures for the early and late
LPPs and slow w aves of high load trials............................................. ................ 42

3-14. Grand-averaged ERPs for site #34 during interference picture processing in the
neutral (blue), pleasant (black), unpleasant (green), and evolutionary threat (red)
conditions at high WM loads, collapsed across groups....................43

3-15. Mean ERP amplitudes for the early and LPP and slow wave for controls ..............44

3-16. Mean ERP amplitudes for the early and LPP and slow wave for high-fear
p articip an ts. .............................................................................................................. 4 5

3-17. Grand-averaged ERPs for controls and high-fear participants during threat and
unpleasant interference picture processing in the at high WM load ................ 45















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

INTERFERENCE EFFECTS OF ANXIETY AND AFFECTIVE PROCESSING ON
WORKING MEMORY: BEHAVIORAL AND ELECTROPHYSIOLOGICAL
ACCOUNTS

By

David Andrew Stigge-Kaufman

May, 2005

Chair: William M. Perlstein
Major Department: Clinical and Health Psychology

Emotionally salient information is often given preferential processing in attention

and perception. While this offers many benefits, it can become problematic in disorders

of anxiety when threat-related processing is excessive and dysfunctional. Affective

processing has been shown to modulate higher-level cognitive functions like working

memory, although the nature of these effects is poorly understood for both anxious and

healthy individuals alike. This study examined whether affective interference impairs

representations in working memory (WM) and whether this impairment is greater in

response to interfering emotional stimuli that have intrinsic clinical or evolutionary

relevance. High-fear participants and low-fear controls performed a novel delayed

matching-to-sample (DMS) task involving high and low WM "loads" and the

presentation of interference pictures of varying valence (pleasant, neutral, unpleasant, and

clinically/evolutionarily-relevant threat) during the delay. Error rates, reaction time (RT),









and high-density event-related potentials (ERP) were acquired while participants

performed the task. Performance data from WM task revealed that WM performance is

disrupted by affective interference and that arousal and anxiety play key roles in that

interference. ERP data suggest that increased WM load and the processing of task-

irrelevant affective interference compromise neural resources of attention. Deficits in the

neural processes involved in sustained attention appear to be associated with enhanced

threat-based interference of WM in high-fear participants, which may have important

implications for broader behavioral symptoms associated with anxiety. In the end, these

findings offer new insights into the complex interplay between competing cognitive and

affective demands on attention, and suggest ways that anxiety disrupts higher-level

cognitive processes involved in WM.














CHAPTER 1
INTRODUCTION

Phobias and Fear

Recent epidemiological studies estimate that phobias are the single most common

mental disorder in the United States, affecting approximately 5 to 10 percent of the

population (Sadock & Sadock, 2003). Distress associated with phobias often results in a

restricted lifestyle and can lead to many other psychiatric complications, including other

anxiety disorders, major depressive disorder, and substance-related disorders (American

Psychiatric Association [APA], 2000). Although phobias are extremely common, many

persons with phobias do not seek help to overcome their phobias or are misdiagnosed

when given medical or psychiatric attention (Sadock & Sadock, 2003). Specific phobics

have been shown to exhibit cognitive biases for threat-related pictures and words (Kindt

& Brosschot, 1997), which likely play a role in the maintenance of fear. Additionally,

individuals with subclinical levels of anxiety exhibit notable deficits in the cognitive

processing of threatening information (Fox, Russo, Bowles, & Dutton, 2001; Ohman &

Soares, 1998). Animals that pose evolutionary threat (e.g., snakes, spiders) are common

objects of intense fear, reported by as many as 38% of females and 12% of males

sampled (Ohman & Mineka, 2003).

Given their high prevalence, the functional consequences of phobias and high

levels of fear are worthy of considerable attention. Previous investigations have

examined the perceptual and attentional processing of emotional information, while other

lines of research have begun to focus on connections between emotional processing and









higher levels of cognition, including working memory. The combined study of affective

and cognitive processes is complex, but previous findings now lay the foundation for a

better understanding of the neural processes involved in affective influences on cognition,

which makes it possible to more carefully study the functional impact of fear and other

emotional factors on higher cognitive processes like working memory.

Attentional and Perceptual Components of Emotional Processing

Examinations of emotions from an evolutionary perspective are quick to highlight

the importance of allocating preferential perceptual processing to certain environmental

stimuli (Dolan, 2002). For millions of years, species' survival has depended on the

ability to efficiently process the wide variety of sensory cues that exist in the natural

world (Ledoux, 1996). Emotional stimuli that occur naturally in the environment (such

as snakes and spiders) are detected faster than non-emotional stimuli, suggesting that

evolutionarily relevant threatening stimuli are effective in capturing visual attention

quickly and consistently (Ohman, Flykt, & Esteves, 2001). Enhanced attention has even

been shown to spread to non-emotional stimuli that merely appear within spatial

proximity to emotional cues (Williams, Watts, MacLeod, & Matthews, 1997).

Participants of spatial orienting tasks respond faster to non-emotional targets appearing

on the same side or same location as an emotional cue (e.g., positive or negative faces),

while a slower response results for targets appearing on the opposite side or different

location (Armony & Dolan, 2002), suggesting preferential processing for emotionally

salient information.

Despite the obvious benefits that accompany enhanced perceptual processing of

emotional stimuli, excessive levels of attention may give rise to the emotional

dysfunction present in certain psychological disorders. It has been suggested that at least









some anxiety disorders are caused by automatic biases of attention that cause excessive

engagement to threat-related stimuli (Ledoux, 1996; Williams et al., 1997). These claims

have been fueled in part by results from tasks that measure the reaction time of

effectively modulated attentional processing. In studies involving the "dot-probe"

paradigm, for example, anxious participants detected a target dot more quickly after it

appeared in the location previously occupied by a threat-related word (MacLeod,

Matthews & Tata, 1986; Fox, 1993). Other studies employing the emotional Stroop

paradigm have found that color naming was slower on anxiety-related words compared to

neutral words (MacLeod, 1991, Vrana, Roodman, & Beckham, 1995), suggesting that an

interference effect can be facilitated by emotional processing. However, not all

researchers attribute anxiety to hypervigilence of the attentional system.

Fox and colleagues (2001) demonstrated that threatening words and faces may not

serve to attract attention, but instead disrupt the disengagement of attention by using a

modified version of the exogenous cueing paradigm (Posner, Inhoff, Friedrich, & Cohen,

1987). They concluded that threatening cues increase the amount of time that attention is

allocated to a stimulus by making it difficult to disengage attention from that stimulus.

While initial disengagement may constitute an early stage of the attentional bias for

threatening stimuli, research in spider phobics has suggested that high levels of anxiety

can also cause rapid disengagement at later processing stages (Hermans, Vansteenwegen,

& Eelen, 1999). Although the debate over the mechanism underlying these findings has

not been resolved, it seems likely that emotionally-motivated attention is mediated by a

complex interplay between multiple, overlapping neural systems that mediate different

kinds of psychological processing (Compton & Banich, 2003).









Capturing the attentional system does not appear to be the only way that

emotional stimuli influence perception. In situations where attention has been

systematically limited, emotional stimuli continue to show preferential processing.

Visual backward masking paradigms tap the automatic regulation of attention by

presenting an threatening target stimulus briefly (30 milliseconds) that is then masked by

an immediately following second stimulus. Even when the emotional target (e.g.,

pictures of snakes or spiders) is not consciously perceived, physiological processing (as

measured by skin conductance responses) reveals differential autonomic processing in

individuals who are fearful of those stimuli (Ohman & Soares, 1994). Furthermore, other

studies have found similar effects in non-fearful subjects who were aversively

conditioned to fear the masked stimuli (Ohman & Soares, 1993). From these findings, it

appears that certain threat-provoking stimuli are selectively processed by autonomic and

central nervous systems even in the absence of conscious awareness.

A brain structure that plays a key role in automatic processing of threat is the

amygdala, a subcortical structure of the limbic system located deep in the anterior medial

temporal lobe (Armony & LeDoux, 1997). Neuroimaging studies consistently find the

amygdala to be active during the evaluation of fearful stimuli, such as fearful facial

expressions (Morris et al., 1996) and affective pictures (Lane et al., 1997). Patients with

amygdala lesions show significant deficits in recognition of fearful facial expressions

(Adolphs et al., 2005). Anatomical analyses reveal that the amygdala receives inputs

from the cortex as well as direct sensory inputs from the thalamus, suggesting that the

amygdala can process emotional information independently of conscious processing

streams (Ledoux, 1996). Rapid processing of emotional information is highly adaptive









when the brain needs to evaluate threatening information; however, this automaticity

comes at the cost of more in-depth cortical processing of stimuli (Armony & LeDoux,

1997). With its anatomical connections and its role in evaluating fearful stimuli, the

amygdala appears to serve the evolutionary role of the brain's "threat detector" (Ohman

& Mineka, 2001).

All in all, there is strong evidence that humans give preferential attention to

emotional stimuli. This bias in attention allows for rapid adaptation to environmental

cues that promotes species survival. However, humans possess higher levels of cognitive

control that allow for more flexible adaptation of behavior to the contextual demands of

our environments (Miller & Cohen, 2001). Are these higher levels of cognitive

functioning also affected by threat-provoking cues that are present in the environment, or

is affective modulation of cognition limited only to the basic levels of attention and

perception?

Emotional Connections with Working Memory

The ability to internally store and manipulate information provides additional

survival benefits beyond those afforded by perception and attention alone. These benefits

have been provided by a fundamental set of cognitive processes known as working

memory. Working memory (WM) involves the active maintenance and manipulation of

selective symbolic information while inhibiting other information, utilizing two

components: short-term storage and executive processes (Baddeley, 1986). Through its

dependence on processing symbolic representations, WM serves to free an organism from

a dependence on the presence of environmental cues and is critical for behavioral

flexibility, internal monitoring, and guidance of contextually appropriate action (Miller &

Cohen, 2001). In experimental tasks that measure WM, participants are required to









actively maintain representations or manipulate information over the course of a delay.

Anatomical analyses in primates have concluded that WM is mediated by a network of

brain structures that includes the prefrontal cortex (e.g., Goldman-Rakic, 1987).

Functional neuroimaging research in humans has shown that multiple frontal regions are

active during short-term WM storage (including Broca's area and motor areas), while the

dorsolateral prefrontal cortex (dlPFC) is involved in active maintenance of

representations (Cohen, et al., 1997; Fuster, 1997; Smith & Jonides, 1999). Recent

findings suggest that the dlPFC aids in the maintenance of information by directing

attention toward internal representations of sensory and motor information that are stored

in more posterior processing centers of the brain (Curtis & D'Esposito, 2003).

Understanding the neurochemical functioning of the PFC provides clues as to

how WM may be integrated with affective processing. Research has shown that normal

PFC activity is dependent on dopamine (DA), a catecholanime neurotransmitter that is

modulated by the amygdala during times of stress (Arnsten, 1998). Both stress and

pleasant affect can trigger the release of DA, which also increases the activity of the

amygdala, in turn triggering more production of DA in the PFC (Ashby, Isen, & Turken,

1999). Unfortunately, when Dl receptors receive too much DA in the PFC, cognitive

dysfunction such as poor attention, impaired response inhibition, and disruptions in WM

often result (Arnsten, 1998, Goldman-Rakic, 1996). Paradoxically, mild levels of

positive affect have been shown to enhance performance on WM tasks (Ashby et al.,

1999), while anxiety can serve to impair both visuospatial (Lavric, Rippon, & Gray,

2003) and verbal WM (Ikeda, Iwanaga, and Seiwa, 1996). The end result of this appears

to be that "both too little and too much dopamine Dl stimulation are detrimental to









prefrontal function" (Goldman-Rakic, 1996, p. 13478), which may serve to take the

dlPFC offlinee" (Arnsten, 1998).

If the dlPFC is taken offlinee," subcortical structures like the amygdala may have

the opportunity to guide behavior with more automatic and reflexive affective processing

and responding. Recent evidence has found that task-irrelevant pictures that are

emotionally arousing cause greater interference effects in cognitive tasks (e.g., solving

math problems) than emotional pictures that represent evolutionary threat or general

negativity in low-fear individuals (Schimmack, 2005). This suggests that emotional

arousal may be the key ingredient for interference of cognitive functioning; however,

little is known about the underlying neural processes that are involved with affective

interference on higher cognitive functions. It is also not clear if these effects would

generalize to high-fear individuals, who show more susceptibility for threat-evoked

decreases in dlPFC activity than controls (Carlsson et al., 2004).

Further clues about the dissociations between affective and cognitive processing

have been studied using fMRI in attention- and WM-related tasks. Recent research has

found that the middle frontal gyrus (i.e., dorsolateral region) of the PFC is actively

involved in the processing of cognitive targets but deactivated during the detection of

novel, non-target emotional stimuli (Yamasaki, LaBar, & McCarthy, 2002). Perlstein,

Elbert, and Stenger (2002) found that WM-related dlPFC activity was influenced by the

emotional characteristics of task-relevant stimuli, but only when brought "on-line" by

task demands that required active maintenance. Furthermore, they found that unpleasant

affective content reduced WM-related brain activation in the dlPFC relative to neutral

and pleasant content. Other researchers have sought to examine the effects of emotional









states, finding that pleasant states enhance verbal WM and decrease spatial WM

performance, while unpleasant states enhance spatial WM and decrease verbal WM

performance (Gray, Braver, & Raichle, 2002). Overall, the lateral PFC was highly

correlated with these behavioral changes, suggesting that this region of the brain is

important for emotional-cognitive integration. This research has yielded many insights

into the interactions between cognition and emotion, but a key disadvantage of fMRI

research is that it offers less sensitivity to rapid changes in the brain compared to direct

measures of neural activity, such as electrophysiological techniques like event-related

potentials.

Event-Related Potentials

Before examining the contributions of event-related potential (ERP) methodology

to the study of emotion, one must understand the basic assumptions behind ERP methods.

One assumption is that the distribution of electrical activity across the scalp is reflective

of the activities of underlying neural structures. A second assumption is that this neural

activity corresponds with specific cognitive states and processes. As an extension of

these assumptions, electrical potentials then represent information regarding cognitive

states and processes (Kutas & Dale, 1995).

The electrical activity of the brain can be measured non-invasively across the scalp

using electrodes. The electroencephalogram (EEG) is the record of the volume-

conducted electrical activity of the brain. EEGs can be used to observe ongoing electrical

brain activity. Alternatively, electrical activity can be averaged in association with the

presentation of specific events of interest. Initially, the event-related response associated

with the presentation of a stimulus is embedded in the ongoing EEG activity. Extracting

an ERP waveform associated with a specific stimulus is accomplished by averaging









multiple samples of the EEG that are time-locked to repeated occurrences of the stimulus

(see Figure 1-1). The benefit of averaging is that the ERPs should remain somewhat

consistent from trial to trial, while the ongoing background EEG is random and will be

averaged out.


(a) Stimulus 1 Stimulus 2... -

I I

EEG On nVV I", -W,--




Stimulus 2
_--- ---------- -- ---- -------

(b) ------------I

Stimulus 2


Stimuls -------- --V^-- -- g



Stimulus N
.-------------


Stimulus N A2 V


Stimulus N







Averaged ERP
Averaged ERP


L-- -- Time (ms)

Figure 1-1. Extraction of the ERP waveform from ongoing EEG. (a) Stimuli (1.. .N) are
presented while the EEG is being recorded, but the specific response to each
stimulus is too small to be seen in the much larger EEG. (b) To isolate the
ERP from the ongoing EEG, the EEG segments following each stimulus are
extracted and averaged together to create the averaged ERP waveform. Taken
from Luck, Woodman, & Vogel (2000).

ERPs are highly sensitive to changes in neural activity on the level of

milliseconds (ms), making them the "gold standard" among noninvasive imaging

methods in terms of temporal resolution (Fabiani, Gratton, & Coles, 2000). ERP

waveforms usually consist of discrete voltage deflections that can be positive or negative,

which are often followed by longer lasting potentials that are called "slow-waves."









Specific components of ERP waveforms are usually named in accordance with their

polarity (positive or negative) and latency (in ms). A common example is "P300," which

refers to an ERP component with a positive peak that has latency of approximately 300

ms. Sometimes it is more appropriate for descriptors to use broader latency terminology,

referring to deflections as "early" (100-200 ms) or "late" potentials (300-600 ms). Other

descriptors may incorporate scalp location at which the component is maximal, such as

"late centro-parietal potentials."

The neural activity associated with ERP activity is attributed primarily to post-

synaptic potentials in pyramidal neurons of the cortex (Williamson & Kaufman, 1990).

Neurons and the extracellular space that surrounds them are filled with charged ions like

sodium (Na ) and potassium (K ). While neurons are at rest, Na+ ions are more plentiful

in the extracellular space than in the neurons, but this changes when action potentials

cause the post-synaptic membrane to depolarize. When dendrites of a neuron depolarize,

Na+ ions flow into the cell, making the extracellular space more negative (known as the

current sink). These Na+ ions that enter the cell repel other positive ions (like K ) away

from the dendrites, creating a current that sends them toward the cell body. The

accumulation of positively charged ions in the cell body (known now as the current

source) repels positively charged ions away from the surrounding region in the

extracellular space, sending those extracellular positive ions back toward the current sink.

This process creates a dipolar extracellular current that accumulates with large

populations of neurons, giving rise to detectable scalp potentials (Coles & Rugg, 1995).

Neural Correlates of Cognitive and Affective Picture Processing

Since phobia and specific fear often have strong visual components, much can be

learned about their neural correlates by examining the neural processes underlying









affective picture processing. Past investigations have demonstrated that a variety of ERP

components have larger amplitudes to emotionally salient or arousing pictures (pleasant

or unpleasant) compared to neutral pictures. This finding has been consistently

associated with enhanced amplitudes of late positive potentials (including P300) and slow

wave components (Cuthbert, Schupp, Bradley, Birbaumer, & Lang, 2000; Johnson,

Miller, & Burleson, 1986). Some researchers have claimed that these effects in picture

viewing paradigms have theoretical links to motivated attention, in which attentional

resources are aroused and directed by motivationally relevant stimuli (Lang, Bradley, &

Cuthbert, 1997).

Research has shown that affective pictures elicit larger amplitudes for late

positive potentials (LPPs) that typically begin 300-400 ms after picture onset (Cuthbert et

al., 2000). These LPPs show greater amplitude differences when the pictures are more

arousing and emotionally salient, even when the pictures are only presented briefly

(Schupp et al., 2004). Visual processing of pictures associated with threat has evoked

larger deflections in early LPPs (Baas, Kenemans, Bocker, & Verbaten, 2001). Early

components of LPPs have been interpreted as a response of capturing attention,

recognition, and stimulus evaluation (Donchin & Coles, 1988). As a result, ERPs can be

used to investigate the amount of resources available for attentionally-dependent

processing. Schupp and colleagues (1997) found that the attentional resources engaged

while viewing affective pictures can serve to detract from the attentional resources that

are available for other forms of processing, leading to a diminished probe startle ERP to a

burst of noise. From this finding, it appears that the resources of attention that are

available for a given task can be limited by other forms of attentional processing, even









across different sensory domains. A provocative question that remains is how the

attentional demands of higher-order cognitive processes, like WM, could potentially limit

the attentional processing available for processing affective content.

LPPs are not the only ERP components that are sensitive to affective stimuli.

Enhanced ERP positivity from affective pictures may last up to 5 seconds after the

picture onset in the form of slow waves (Cuthbert et al., 2000). Ruchkin and colleagues

(1988) have proposed that positive slow waves vary with amount of memory storage

required for processing. Positive slow potentials have been interpreted as reflecting

increased sustained attention (Cuthbert et al., 2000) and possible roles in memory storage

(Donchin & Coles, 1988). Given their apparent functional significance and capacity for

lengthy durations, positive slow waves are likely to have meaningful effects subsequent

objects of attention or motivation.

Summary and Rationale for the Current Study

Previous research has paved the way for a more comprehensive understanding of

the complex interplay between cognitive and affective processing systems of the brain

and the additional role that anxiety plays in these interactions. Emotional stimuli receive

preferential processing on the perceptual and attentional level, and can even be processed

automatically without conscious awareness. Emotional stimuli activate the amygdala,

which modulates PFC concentrations of dopamine and may serve to disrupt prefrontally-

mediated WM functioning in times of heightened emotional arousal. The dlPFC is

involved in active maintenance of representations in WM, yet recent studies have also

shown that its activity can be modulated by affective or emotional context, as well.

Additionally, threat has been shown to modulate WM performance in both visuospatial

and verbal domains.









ERP methods can be used to detect rapid neural changes in the affective

processing in order to map out a time-course of brain responses that are difficult to assess

using other techniques. The method of ERP is well suited for measuring changes in

neural processing that reflect changes in attentional processing, yet little research has

used this technique to study cognitive-emotional interactions in this regard. Certain

forms of affective processing (e.g., picture viewing) can serve to modulate the amount of

attentional resources available for other forms of emotional behavior (e.g., startle), as

evidenced by LPPs. However, much still needs to be learned about the role that anxiety

and other forms of emotion play in modulating neural and cognitive behavior.

The current study examined the potential role of task-irrelevant emotionally

evocative interference stimulation on PFC-mediated WM processes. The primary

objective was to determine the effects of task-irrelevant affective (pleasant and

unpleasant) stimuli presented during the delay period of a task requiring active

maintenance of stimulus representation in WM. Key questions regarding this issue

include: 1) does affective interference "capture" attention and, thereby, lead to a

degradation of the strength of task-relevant stimulus representations in WM, resulting in

impaired performance; 2) does the emotional valence or arousal of these interfering

stimuli differ in the extent to which they draw resources away from the active

maintenance of task-relevant representations; 3) do evolutionarily-relevant threat stimuli

(i.e., snakes, spiders) disproportionately capture attention and impair WM performance;

and 4) does active maintenance of task-relevant representations in WM draw resources

away from the ability to process task-irrelevant interfering stimuli? A secondary goal

was to determine if individuals with high fear of these evolutionarily-relevant threat









stimuli (e.g., specific phobics/sub-clinical phobics) evidence a disproportionate

attentional capture by these stimuli and a consequent increase in WM impairment.

To address these aims, high- and controls performed high and low "load"

conditions of a novel visual delayed matching-to-sample (DMS) task in which task-

irrelevant pictorial interfering stimuli of different valence categories (pleasant, neutral,

unpleasant, evolutionary threat) were presented during the delay or retention period.

Task performance was measured by error rate and reaction time (RT); high-density event-

related potentials (ERPs) indexed the extent to which task-irrelevant interfering stimuli

were processed.

Predictions

WM Task Performance

It is predicted that error rates and probe RTs during the WM task will be

modulated by the emotional valence of the interfering pictures. Emotionally arousing

interference pictures (pleasant, unpleasant, evolutionary threat) are expected to cause

greater task performance decrements than neutral pictures. Furthermore, it is predicted

that high-fear participants will show a disproportionate decrement in task performance to

the fear specific (i.e., evolutionary threat) interference stimuli than controls, and that this

dissociation will be most evident at the high WM load condition when active

maintenance of task-relevant stimulus representations is greatest. Finally, it is not

anticipated that high-fear subjects will show deficits in "baseline" WM task performance,

as there is little evidence to suggest that under non-arousing conditions they have

impaired dlPFC/WM function.









ERP Reflections of Task-Irrelevant Interference Stimulus Processing

If, as predicted, the active maintenance of task-relevant representations in WM

draws resources away from the ability to process task-irrelevant interfering stimuli, it is

expected that there will be load-dependent effects on the LPP components and slow

waves evoked during the presentation of interference pictures. That is, the LPP

components will be smaller in amplitude during the high- than low-load task conditions.

Furthermore, it is predicted that the LPPs and slow wave amplitudes evoked by the

interfering pictures will be greater in amplitude for emotionally-arousing than neutral

interference stimuli, indicative of a greater capture-of-attention effect by emotionally-

arousing interference stimuli. Finally, it is also predicted that high-fear participants will

show greater LPP and slow wave amplitudes to the evolutionary threat interference

compared to interference of other valence categories and compared to controls.














CHAPTER 2
METHOD

Participants

Thirty individuals (16 female) between the ages of 19 and 45 participated in the

study in exchange for course credit or financial compensation. All participants provided

written informed consent in accordance with procedures of the University of Florida

Health Science Center Institutional Review Board. All participants were given the

"Specific Phobia" section of the Anxiety Disorders Interview Schedule for the DSM-IV

(ADIS-IV; Brown et al., 1994), along with other relevant sections necessary to assess

comorbid psychopathology. The ADIS-IV provides diagnostic information pertaining to

all anxiety disorder categories and the full range of mood, substance-related, and

somatoform disorders.

Based on the results of the ADIS-IV, 14 of the participants were judged to have

high levels of fear to either snakes or spiders, ranging from "moderate" to "very severe."

Six of the high fear participants also met diagnostic criteria for other anxiety disorders,

including Panic Disorder without Agoraphobia, Generalized Anxiety Disorder, and

Social Phobia. Gender makeup was not significantly different between the high-fear and

control groups [X2(1, N= 30) = 1.27, p > .25], although 64% of the high-fear participants

were female, compared to 44% of the controls. Age and educational background were

also equated across fear group, as shown in Table 2-1. All of the participants completed

the interview and WM task, yet one of the controls had to be excluded from the ERP

analysis due to technical problems. This subject is included in the sample statistics for









demographic information, emotional assessment, and WM task performance (N= 30),

while the ERP data does not include this subject (N= 29).

Although the ADIS-IV provided the diagnostic information necessary to form the

two participant groups, several other assessments were used to quantify each participant's

level of emotional functioning and affective health. The Snake Anxiety Questionnaire

and Spider Questionnaire (SNAQ and SPQ; Klorman, Hastings, Weerts, Melamed, &

Lang, 1974) further assessed each participant's fear of snakes or spiders. The State-Trait

Anxiety Inventory (STAI; Spielberger, Gorsuch, & Lushene, 1970) was also used to

provide broader information regarding the participants' general levels of anxiety as

manifest in temporary states of distress and more long-term personality traits. Finally the

Beck Depression Inventory Second Edition, (BDI-II; Beck, Steer, & Brown, 1996) was

used to assess for elevated levels of depressive symptoms. Participants were excluded if

they reported depressive symptoms sufficient for a current diagnosis of major depressive

disorder, previous neurological disease, traumatic brain injury (TBI), or current

psychotropic medication use. A total of seven participants were excluded from

participation due to excessive symptoms of depression or substance abuse. As shown in

Table 2-1, participants in the high fear group exhibited significantly higher levels of

specific fear, state anxiety, trait anxiety, and depressive symptoms compared to the

controls. Levels of specific fear, state anxiety, trait anxiety, and depressive symptoms

were highly correlated with each other across all subjects, as shown in Table 2-2. Mean

BDI scores, while different between groups, were below clinical cut-off levels considered

to reflect depression (e.g., 14 for mild depression; Beck et al., 1996).










Table 2-1. Demographic and emotional functioning data for all participants
Mean (SD)
High Fear (n=14) Low Fear (n=16) t-statistic
Age 25.3 (7.5) 25.0(5.8) .117ns
Education 15.6 (1.7) 15.7 (2.1) -.165 n
SNAQ/SPQ 10.6 (2.6) 2.7(2.9) 7.93**
STAI-State 37.7(10.1) 29.1 (11.4) 2.17*
STAI-Trait 42.9 (11.3) 29.7 (6.7) 3.96**
BDI 10.2 (5.1) 5.0 (5.4) 2.70*
p> .85. *p < .05. p < .001

Table 2-2. Intercorelations between measures of anxiety and depression
STAI-State STAI-Trait BDI
SNAQ/SPQ .391 .536* .538"
STAI-State --- .572 .539*
STAI-Trait --- --- .719***
p < .05. **p<.01. *p<.001. df= 28.

Materials and Procedure

Participants performed a visual delayed match to sample (DMS) task, using

stimuli modified from Low, Rockstroh, Harsch, Berg, & Cohen (2000). Participants

were seated in front of a computer screen (approximately one meter from screen to nose)

and a keyboard. In this task, geometric shapes were presented as cue-probe pairs with

pictorial interference occurring during the delay period (Figure 2-1). Participants gave

forced-choice responses indicating if they saw the probe stimuli in the previously

presented cue. Each participant completed a total of 400 trials. Each trial began with the

computer presentation of a visual cue, which remained on the screen for 2500 ms. The

cues consisted of two difficulty levels, including one diamond (low load) and three

diamonds (high load). The high and low load trials were presented randomly with equal

probability. The diamonds varied by size and orientation, and participants were

instructed to remember each diamond in the cue by its individual combination of size and

orientation features.









A task-irrelevant interfering picture was presented in 80% of the trials. This

interfering picture was presented 1000 ms following cue offset and remained on the

screen for 750 ms, during which the participants were instructed to press a button to

ensure that the picture was detected and encoded. The pictures were taken from the

International Affective Picture System (IAPS; Lang, Bradley, & Cuthbert, 1998), and

consist of four distinct valences, or emotional categories: pleasant, neutral, unpleasant,

and evolutionary threat (snakes or spiders). Each of the four valence categories consisted

of 25 pictures that were presented multiple times in a pseudo-randomized fashion over

the course of the experiment.1 Because of limited numbers of IAPS pictures that consist

of snakes or spiders, this category of pictures were supplemented with several non-IAPS

pictures. IAPS pictures have standardized ratings for pleasantness (valence) and arousal

on a 1 to 9 scale (9 equals most pleasant and most arousing). The ratings for the pleasant,

neutral, and unpleasant pictures used in this experiment can be seen in Table 2-3. With

regard to pleasantness, pleasant pictures were rated significantly higher than neutral, t(48)

= 18.8, p < .001, which were rated significantly higher than unpleasant, t(48) = 22.0, p <

.001. With regard to arousal, pleasant and unpleasant pictures did not differ, t(24) = .22,

p > .80, but these pictures were rated significantly higher than neutral, ts(24) > 10.2, ps <

.001. In 20% of the trials, participants viewed a continuous fixation point in place of the

interfering picture in order to create a "no interference" condition.



1 IAPS identification numbers for pleasant pictures are 1440, 1460, 1463, 1604, 1710, 2040, 2070, 2071,
4599, 4608, 4623, 4680, 5470, 5621, 5623, 7230, 7260, 7330, 8030, 8370, 8380, 8470, 8501, 8510; for
neutral pictures: 2191, 2235, 2372, 2383, 2393, 2394, 2515, 2745.1, 2840, 2850, 2980, 5390, 5740, 7036,
7100, 7140, 7180, 7217, 7234, 7285, 7491, 7496, 7595, 7710, 8311; for unpleasant pictures: 2095, 2205,
2375.1, 2683, 2800, 6213, 6550, 6570, 6831, 6834, 6838, 8485, 9050, 9220, 9280, 9342, 9400, 9410, 9470,
9471, 9520, 9561, 9830, 9910, 9921; for evolutionary threat pictures: 1019, 1022, 1050, 1051, 1052, 1070,
1080, 1090, 1101, 1110, 1111, 1112, 1113, 1114, 1120, 1201, 1220.









Table 2-3. Mean ratings of each group of IAPS pictures (from Lang et al., 1998)
Mean Ratings (SD)
Pleasant Neutral Unpleasant
Valence 7.59 (.40) 5.25 (.47) 2.44 (.43)
Arousal 5.51 (.92) 3.29 (.58) 5.56 (.88)


A memory probe (single diamond in a box) was presented on the screen 2750 ms

following the offset of the original cue. The probe remained on the screen for 1500 ms,

during which the participants were asked to determine if the diamond shown was an exact

match in both size and orientation to one of the previously shown diamonds for that trial.

Participants pressed "m" to indicate that the probe diamond was a match to the cue and

"n" to indicate that the probe diamond was not a match. 1000 ms following the offset of

the probe, the next trial began with a new cue. Participants were instructed to look at a

fixation point (+) during the time intervals between the cue, picture, and probe.


1000 ms Response



200ms 1000 ms----
OR -



2500 ms

Figure 2-1. Overview of each trial, showing the time course of each cue, picture
interference, and probe. Since the probe diamond in this example matches
one of the cue diamonds exactly in its size and orientation, the correct
response for the trial shown here is "m".


100c; MS

1500 MS









After completing the experimental task, participants then rated each picture that

they viewed during the course of the experiment in order to provide a manipulation check

on effects of valence and arousal. These ratings were computerized, using the Self-

Assessment Manikin (SAM) rating system developed by Lang (1980). Participants were

instructed to view each picture for as long as they needed in order to be able to rate the

pictures, and they self-terminated their viewing of each picture with a button press.

Participants then rated each picture separately for valence and arousal, using the same 1

to 9 scale used in the standardized IAPS ratings (9 equals most pleasant and most

arousing).

EEG Acquisition and Reduction


EEG Data Acquisition

EEG was recorded from 64 scalp sites using a 64-channel geodesic sensor net

(Figure 2-1) and amplified at 20K using an Electrical Geodesics Incorporated (EGI)

amplifier system (nominal bandpass .10 100Hz). Electrode placements enabled

recording vertical and horizontal eye movements reflected in electro-oculographic (EOG)

activity: one placed above and below each eye and centered around the pupil to record

vertical eye movements; the others placed at the outer canthus of each eye for recording

horizontal eye movements. EEG was referenced to Cz and was digitized continuously at

250 Hz with a 16-bit analog-to-digital converter. A right posterior electrode served as

common ground. The impedance of all electrodes was maintained below 50 kQ,

consistent with procedures suggested by the manufacturer.











in
64 63
11 6

14 7 1
12 2

15 8 3 61
13 62
4
23 16 9 5B 57 59
20 56
5 55
17 54
Left ear 21 53 Right ear
Ref
I aft ar 24 18 43 52 Right ear
25 30 50
26 29 42 51
27 28 (34) 46 49
33 41
38
31 32 45 48

37 40
55 36 44 Com
59




Figure 2-2. Sensor layout of the 64-channel geodesic sensor net. Electrode #34
(outlined) was used for measurement of picture-related ERPs.



EEG Data Reduction

Due to the volume-conducting nature of the brain, no single scalp site can be


considered an "inactive" reference site (Tucker, Liotti, Potts, Russell, & Posner, 1994);


therefore, data were mathematically re-referenced against an average reference (Bertrand,


Perrin, & Pernier, 1989). In this procedure, the activity of each electrode site is reflected


as the difference between itself and the average of all the other recording sites. Editing of


the EEG for movement, electromyographic muscle artifact, electro-ocular eye movement,


and blink artifacts was performed by computer algorithm in Brain Electrical Source


Analysis software (BESA version 5.0; Scherg, 1990). EEG during which voltage


exceeded 150 [V and with point-to-point transitions exceeded 125 [V were excluded


from averaging.









Individual-subject event-related potentials (ERPs) were extracted and averaged

together from the ongoing EEG recording in discrete temporal windows that coincided

with the onset of each stimulus (refer back to Figure 1-1). ERP averages from each

subject were divided into three categories: cue activity, interfering picture activity, and

probe activity. Stimulus-locked epochs were extracted with a duration of 100ms prior to

stimulus presentation (constituting a 100ms pre-stimulus baseline period) and 1000ms

post-stimulus presentation. Collapsing across certain conditions was necessary because

of insufficient numbers of responses to conduct specific analyses. All averaged ERP

epochs were baseline corrected using a 100 ms window prior to stimulus onset and

digitally filtered at 15 Hz low-pass and a .5 Hz high-pass. Although ERPs were acquired

for stimulus-related activity for the cues, probes, and interfering pictures, only the ERPs

for interfering pictures will be discussed in this study.

Statistical Analyses

WM task performance, picture responses, and ERP amplitudes were analyzed

using repeated measures analyses of variance (ANOVAs). Because of the a priori

hypotheses regarding group differences on these measures, planned contrasts were used

to decompose interaction effects within the ANOVAs that were performed. When

applicable, these contrasts utilized orthogonal comparisons of different levels of affective

picture categories: 1) affective vs. neutral, 2) pleasant vs. negative (unpleasant and

threat), and 3) unpleasant vs. threat. In addition to these planned contrasts, follow-up

contrasts were employed and, where appropriate, adjusted for multiple comparisons using

the modified Bonferroni method (Keppel, 1982). For ANOVAs where there were more

than two levels of a within-subject factor, the Huynh-Feldt epsilon adjustment (Huynh &

Feldt, 1976) was used; uncorrected degrees of freedom and corrected p-values are









reported. In addition, two-tailed Pearson product-moment correlations were performed to

determine associations between continuous variables of interest.

WM Performance Data

For each participant, error rates and probe RTs were calculated for each WM load

and interfering picture valence category. For all RT measurements, median RTs were

initially calculated for each participant in order to better accommodate outliers (Ratcliff,

1993), and then means of these individual RTs were compared in subsequent inferential

analyses. Error rates and probe RTs for the no-interference condition were then

examined for an effect of WM load, using a 2-Group x 2 Load repeated measures

ANOVA. After detecting a significant effect of WM load on task performance, error

rates and probe RTs were analyzed by subjecting them both to a 2-Group x 4-Valence

ANOVA.

Picture Response Data

A 2-Group x 4-Valence repeated measures ANOVA was performed on the RTs of

the interfering picture detection responses, as well as the mean viewing times during the

picture rating procedure. Post-experimental picture ratings were subjected to 2-Group x

4-Valence repeated measures ANOVAs separately for ratings of pleasantness and

arousal.

ERP Data

Analysis of ERP waveforms focused on activity reflecting processing of the

interfering pictures. Statistical analyses of ERP waveforms assessed the mean voltages

over specified temporal windows (epochs) of individual subject ERPs extracted from

individual electrode sites for correct trials. Scoring windows and electrode positions for

each condition of interest were determined by examination of grand-averaged ERP









waveforms and the grand mean global power of the voltage obtained over all electrode

sites. Picture-related ERP activity was quantified at electrode site #34 (Figure 2-2) and

was examined over three different time periods believed to represent different aspects of

visual processing. The first time period was 296 ms to 356 ms, which corresponds to the

early portion of late positive potentials (LPP) associated with affective picture

processing, while the second time period was 452 ms to 512 ms, corresponding to the late

portion of LPP (Schupp et. al., 2003). The third time period was 660 ms to 760 ms,

corresponding to the slow wave component of visual processing (Keil, et al., 2002).

Thus, three ERP components associated with processing interfering pictures were

examined: early LPP, late LPP, and slow wave. A 2-Group x 2 Load repeated measures

ANOVA was performed on the amplitudes of each ERP component to determine the

overall effects of WM load on ERPs, collapsing across different picture types to

determine the effects of WM load on ERPs to the interfering pictures. Subsequently,

amplitudes of each ERP component were analyzed in 2-Group x 4-Valence repeated

measures ANOVAs to determine the effects of interfering picture valence on ERPs to the

interfering pictures.














CHAPTER 3
RESULTS

Task Performance

Initial analyses were performed to examine the possibility of a speed/accuracy

trade-off by correlating error rates on the WM task and reaction times to the probe.

These analyses revealed that probe reaction time (RT) and error-rate were not

significantly correlated, r(27) = .145, p > .40, indicating that speed-accuracy trade-off

does not account for the behavioral findings reported below.

Effects of WM Load

An important aspect of the present research was to demonstrate that the WM load

manipulation was indeed associated with alterations in behavioral performance. To

examine effects of WM load in the absence of interfering pictures, 2-Group x 2 Load

repeated measures ANOVAs were performed on error rates and probe RTs for the no-

interference trials. As expected, participants as a whole exhibited more errors, F(1,28) =

106.26, p < .001, r2 = .791, and longer probe RT, F(1,28) = 181.46, p < .001, r2 = .866,

in high WM load trials compared to low load trials, as shown in Figure 3-1. Fear group

did not exert a significant main effect on error rate, F(1,28) = .01, p > .90, qr2 < .001, or

probe RT, F(1,28) = 3.44, p > .07, qr2 = .109. Furthermore, examination of Group x

Load interactions revealed that high-fear participants did not significantly differ from

controls in their load-related WM performance with regard to error rates, F(1,28) = .13, p

> .70, qr2 = .005, or probe RTs, F(1,28)=2.92,p > .10, r2 = .094, for no-interference

trials.











Effects of WM Load on Accuracy Effects of WM Load on Probe RT

0.4 1400
0.35 1200
0.3
2 1000
W 0.25
E 800- 0 Low Load
a 0.2 Ea High Load
o E 600
S0.15 I-
00.1 400--
0.05 200--
0 0
Controls High-Fear Controls High-Fear
Group Group

Figure 3-1. WM task performance for each WM load during no-interference trials.
Error bars represent standard errors.

These findings suggest that high WM load trials required more active

maintenance of the stimulus representations (three diamonds) in WM than low load trials

(one diamond), and that high levels of fear did not impair WM performance. Since the

pictures presented during the delay were predicted to interfere with the active

maintenance of WM representations, it is likely that the interference will have its

maximal effect on the high WM load trials. As a result, subsequent analyses will focus

on high WM load trials only.

Effects of Interference on WM Task Accuracy

A 2-Group x 4-Valence repeated measures ANOVA was performed on the error

rates for high WM load trials to determine the effects of the interfering pictures. Planned

orthogonal contrasts were used to test a priori hypotheses about the effects of interfering

picture valence. Statistics for the ANOVA main effects and interactions can be found in

Table 3-1, while results from the contrasts can be found in Table 3-2.









Table 3-1. ANOVA statistics from the WM task accuracy data
Effect F-Ratio rj2
Group (G)a 1.25 .043
Valence (V)b 3.19* .102
G x Vb 1.60 .054
adf= 1,28. bdf= 3,84. *p < .05.

Table 3-2. Valence effects in the WM task accuracy data
Valence Valence x Group
(Across Group) (Across Load)
Valence Contrast" F-Ratio qr2 F-Ratio r|2
Affective vs. Neutral 9.49 .253 1.56 .053
Negative vs. Pleasant .53 .018 3.11t .100
Threat vs. Unpleasant .02 .001 .28 .010
adf= 1,28. tp< .10. *p <.01.

Means and standard errors of the error rates as a function of fear group and

interference category can be seen in Figure 3-2. A main effect of picture valence was

present for the error rates of the high WM load interference trials. Planned contrasts

revealed that participants as a whole made more errors in each of the three affective

picture trials compared to neutral. A trend-level Group x Valence interaction emerged,

with high-fear participants making more errors than controls during negative (unpleasant

and threat) picture trials relative to pleasant. Follow-up group-wise contrasts using

modified Bonferroni-corrected comparisons at each valence level (critical p = .0188)

revealed that controls exhibited more errors during pleasant picture trials than neutral (p

< .0188). Error rates for high-fear participants did not significantly differ among the

interfering picture valences or in comparison with those of the controls.

Effects of Interference on WM Task Reaction Time

A 2-Group x 4-Valence repeated measures ANOVA was performed on the RTs

for high WM load trials to determine the effects of the interfering pictures. Planned

orthogonal contrasts were used to test a priori hypotheses about the effects of interfering









picture valence. Statistics for the ANOVA main effects and interactions can be found in

Table 3-3, while results from the contrasts can be found in Table 3-4.

Effects of Interference on Accuracy


0 Control
U High-Fear


Neutral Pleas. Unpleas. Threat
Interference Condition

Figure 3-2. Error rates by interference category, and fear group. Error bars represent
standard errors.

Table 3-3. ANOVA statistics from the probe reaction time data
Effect F-Ratio q2
Group (G)a 1.58 .053
Valence (V)b 3.76* .118
G x Vb 1.78 .060
adf= 1,28. bdf= 3,84. p < .05.

Table 3-4. Valence effects in the probe reaction time data
Valence Valence x Group
(Across Group) (Across Load)
Valence Contrasta F-Ratio q2 F-Ratio q2
Affective vs. Neutral .76 .026 5.60* .167
Negative vs. Pleasant 8.27** .228 .17 .006
Threat vs. Unpleasant 1.32 .045 .02 .001
adf= 1,28. *p< .05. **p<.01.

A main effect of picture valence was present for the probe RTs of the high WM

load interference trials, and planned contrasts revealed that participants took longer to


T T

T T









respond to the probe during negative (unpleasant and threat) picture trials compared to

pleasant. Means and standard errors of the reaction times to the probes as a function of

fear group and interference category can be seen in Figure 3-3. Follow-up contrasts

clarified that this negative picture effect was actually due to the threat pictures, as

interfering threat picture trials resulting in significantly longer probe RTs than pleasant,

(p < .0188), while unpleasant picture trials did not (p = .05). Follow-up group-wise

contrasts using modified Bonferroni-corrected comparisons revealed that high-fear

participants took longer to respond to the probe during threat picture trials than neutral (p

< .0188), while controls failed to show this pattern (p > .200). Instead, controls exhibited

faster probe RTs during pleasant picture trials compared to neutral (p < .0188). Faster

probe RTs for pleasant picture trials in controls were not significantly correlated with

increased error rates, r(27) = .013, p > .95, suggesting that they did not exhibit a

speed/accuracy tradeoff for pleasant picture trials. Although high-fear participants took

longer to respond in every category of interfering picture trial, their probe RTs did not

significantly differ from controls.

To summarize the WM task performance data, WM load had very strong effects

on both accuracy and RT, with participants making more errors and taking longer to

respond to the probe during the high WM load trials. Participants as a whole made more

errors during trials involving affective picture interference compared to those involving

neutral pictures. Controls made more errors and took less time to respond to the probe

during pleasant picture trials compared to neutral. High-fear participants exhibited a

trend for making more errors in the negatively-valenced (unpleasant and threat) picture

trials compared to pleasant and exhibited significantly longer probe RTs to threat picture










trials compared to neutral, showing that they demonstrated the expected pattern of WM

impairment from the threat picture interference.


Effects of Interference on Probe RT


1250

1200

1150

1100

1050

1000

950

900

850


Figure 3-3.


O Control
* High-Fear


Neutral Pleas. Unpleas. Threat
Interference Condition

Probe reaction time by interference category, and fear group. Error bars
represent standard errors.

Picture Response Data


Interfering Picture Detection Data

Timed responses were measured as participants indicated that they detected the

interfering pictures during the WM task. While the primary goal of requiring participants

to button-press to the presentation of interfering pictures was to increase the likelihood

that they did, indeed, actually view the pictures, RTs to the presentation of interfering

pictures could potentially shed light on aspects of picture processing. Thus, a 2-Group x

4-Valence repeated measures ANOVA was performed on the mean picture detection

RTs. Means and standard errors of the reaction times to the pictures as a function of fear

group and valence can be seen in Figure 3-4. A significant main effect of picture valence

was seen in the interference picture detection, F(3,81) = 13.07, p < .001, r2 = .326.










Participants were significantly faster in responding to interfering threat pictures compared

to neutral, pleasant, and unpleasant pictures (ps < .0188). A main effect of fear group

was also seen in these RTs, F(1,27) = 5.25, p < .05, r2 = .163, with high-fear participants

taking significantly longer to detect the interfering pictures as a whole than controls.


Mean Picture Detection RT
620
600
580
560
E
) 540 T
rE 0 Control
520-
U Figh-Fear
500--
480--
460
Neutral Pleas. Unpleas. Threat
Picture Category
Figure 3-4. Mean picture detection reaction times by valence and fear group. Error bars
represent standard errors.

Viewing Times During Picture Rating

Timed responses to pictures were measured as participants viewed the self-

terminating pictures prior to making their subjective ratings at the end of the experiment.

Means and standard errors of the picture viewing time as a function of fear group and

valence can be seen in Figure 3-5. A significant main effect of valence emerged, F(3,84)

= 7.48, p < .01, r2 = .211, as participants as a whole chose to view unpleasant pictures

longer than pleasant (p < .0188) and threat (p < .0188) pictures. Although high-fear

participants took more time than controls to view the pictures as a whole before rating

them, this difference was not significant, F(1,28) = 1.72, p = .20, r2 = .058.










Mean Picture Viewing Time

5100 -
50 Ce o Controlr
4600 H High-Fear

4100 -

E 3600

S3100 -

2600

2100 -

1600
Neutral Pleas. Unpleas. Threat
Picture Category

Figure 3-5. Mean RTs for picture viewing during the rating procedure at the
end of the experiment. Error bars represent standard errors.

Manipulation Check: Picture Rating Data

An important premise of the current research with respect to modulation of WM

by emotionally interfering stimuli, and examination of evolutionarily-relevant threat

stimuli in low- and high-fear participants, is that the interfering stimuli were, indeed,

perceived as both emotionally-arousing and as producing the intended emotional valence.

These "manipulation checks" are examined below.

Valence ratings

Significant main effects were found for participants' ratings of the pictures'

pleasantness, F(3,84) = 174.94, p < .001, r2 = .862, which can be seen in Figure 3-6. As

expected, participants rated pleasant pictures as more "pleasant" compared to ratings of

neutral, unpleasant, and threat pictures (ps < .0188). Also as expected, participants rated

unpleasant pictures as more "unpleasant" than neutral and pleasant pictures (ps < .0188).

However, this main effect of valence was qualified by a Group x Valence interaction,

F(3,84) = 5.86, p < .01, r2 = .173, as high-fear participants rated threat pictures as more









"unpleasant" than controls (p < .0188). Controls rated threat pictures as less "unpleasant"

than the unpleasant pictures (p < .0188), while high-fear participants rated threat and

unpleasant pictures as equally "unpleasant" (p > .95).


Picture Valence Ratings
9-
.* | 8 OControl
n 7 High-Fear
c 6
5
S 4--




Neutral Pleas. Unpleas. Threat

Picture Category
Figure 3-6. Subjective ratings for picture valence. Error bars represent standard
errors. p < .001

Arousal ratings

Significant main effects were found for participants' ratings of the pictures' level

of arousal, F(3,84) = 19.83,p < .001, q2 = .415, which can be seen in Figure 3-7.

Participants rated the pleasant, unpleasant, and threat pictures as more arousing than the

neutral (ps < .05). However, this main effect of arousal was qualified by a Group x

Valence interaction, F(3,84) = 8.98, p < .001, q2 = .243, as controls rated pleasant

pictures as more arousing than high-fear participants (p < .0188), while high-fear

participants rated threat pictures as significantly more arousing than controls (p < .0188).

Additionally, high-fear participants rated threat pictures as more arousing than neutral

pictures (p < .0188), while controls did not (p > .35), but instead rated pleasant pictures









as more arousing than neutral (p < .0188). There was no significant difference between

high-fear participants' arousal ratings of threat pictures and controls' arousing pictures of

pleasant pictures, t(28) = 1.138, p > .26, although Levene's test was significant, F(1,28)

= 11.05, p < .01, indicating that these two sets of ratings did not show equal variances

and this result should be interpreted with caution.


Picture Arousal Ratings

9 n Control
8 High-Fear
S 7


5
4
0
3
02-j


Neutral Pleas. Unpleas. Threat
Picture Category

Figure 3-7. Subjective ratings for picture arousal. Error bars represent standard
errors. *p < .05, **p < .001

To summarize the picture response data, participants detected interfering threat

pictures during the WM task faster than the other categories, and high-fear participants

took longer than controls to signify their detection of interfering pictures. Picture SAM

ratings, which provide a manipulation check on participant's judgments of valence and

arousal, demonstrated that the interfering pictures indeed were effective in producing the

intended arousal and valence effects. Participants' ratings of valence and arousal

differentiated neutral, pleasant, and unpleasant pictures in ways that were generally

expected. High-fear participants rated threat pictures as significantly less pleasant than









the neutral pictures, giving the threat pictures more arousing ratings than controls.

Controls surprisingly rated pleasant pictures as significantly more arousing than high-fear

participants. These arousal ratings show interesting correspondence with the task

performance data, as high-fear participants made exhibited errors and longer probe RTs

during threat picture trials, while controls exhibited more errors and shorter RTs during

pleasant picture trials. Finally, during SAM ratings, participants viewed unpleasant

pictures the longest prior to making their ratings.


Event-Related Potential (ERP) Data

Effects of WM Load

ERP responses to interfering pictures were acquired from all scalp electrodes, and

converted into grand averaged waveforms of picture interference by load (Figure 3-8).

Electrode #34 showed clearly defined LPP and slow wave components. Spherical-spline

interpolated scalp voltage maps showed that a centro-parietal region, encompassing

electrode #34, was sensitive to WM load-related effects in the early and late LPP epochs

(Figure 3-9). Thus, for simplicity, and because a primary aim of the research was to

examine affective interference effects on active maintenance in WM, electrode #34 was

used in all of the WM-related ERP analyses that follow.

Differential load-related activity can clearly be seen over the centro-parietal region of

the scalp in the earlier stages of picture processing (LPP), yet this effect is absent in the

slow wave (see Figure 3-9). Figure 3-10 illustrates the grand-averaged ERP waveforms

for the low and high WM load trials. 2-Group x 2 Load repeated measures ANOVAs

were performed to determine the overall effects of WM load on ERPs in the early LPP,

late LPP, and slow wave epochs from the interfering pictures, collapsing across different










picture types. As shown in Figure 3-11, voltage amplitudes measured from electrode #34

were found to be more positive during low WM load trials compared to the high load for

the LPP time windows, yet no load effect was seen in the slow wave, F(1,27) = 1.484, p

> .20, r2 = .052. The WM load effects on LPP were stronger for the early, F(1,27) =

28.367, p < .001, r2 = .512, than late, F(1,27) = 7.239, p < .05, qr2 = .211, time window.

Controls and high-fear participants were equally affected by these WM load effects, as no

significant Group x Load interaction emerged, Fs(1,27) < .423, ps > .50, qr2 < .015.


Z_ Front 3


1i-


- Low Load
High Load


Left






r




both 3cor-vec avr,




Figure 3-8.


0r .. 0 1000
Back Time [ms]

Grand-averaged ERPs for all scalp sites during interference picture
processing in the low (blue) and high (red) WM load conditions.


1
9,4


7










Early LPP



Front: .


Late LPP


~rr.
* 4.
It.


Side:
_.,: ,, ,, !


.2




AW7N.
,V0/ --=
'r -*/^


Back:


Slow Wave
, ,, .. ".
;*:.. ^ : ;1,,*J


'4
A
it .,








II~


La-

6::



V


326.0 ms reference free
EEG -Voltage 0.20 pVf step


482.0 ms
EEG Voltage


reference free
1 f 0.20 pV/step


710.0 ms + reference free
EEG Voltage T IITiIITM 0.20 IpV step


Spherical-spline interpolated scalp voltage maps representing the differences
in neural processing between the high and low WM loads for the early and
late LPPs and slow wave. Voltage difference maps were calculated by
subtracting high load activity from low load, and were calculated at 326ms
(early LPP), 482ms (late LPP), and 710ms (slow wave) after interference
picture onset. Positive voltage differences are indicated in red, showing
strongest load-related differences in the early LPP and the late LPP.


1000


0 250 500 750
Time [ms]


Figure 3-10.


Grand-averaged ERPs for site #34 during interference picture processing
in the low (blue) and high (red) WM load conditions.


Figure 3-9.











Effects of WM Load on ERP Amplitudes
5-

4



U Low Load
E High Load



0-

-1 Early LPP Late LPP Slow Wave
ERP Component
Figure 3-11. Mean ERP amplitudes for low and high WM load conditions for the early
and late LPP and slow wave. Error bars represent standard errors.

Figure 3-11 clearly shows that early stages (LPP) of interference picture

processing show greater effect of WM load than later stages (slow wave), which fits the

scalp distribution maps of the load-related differences (Figure 3-9). In order to compare

the effect of WM load on these different time-locked ERP components, a 2-Group x 2-

Load x 3-Epoch repeated measures ANOVA was performed. Significant main effects of

load, F(1,27) = 12.49, p = .001, r2 = .316, and epoch, F(1.64, 44.13) = 33.45, p < .001,

2 = .546, were found, but these were qualified by a significant Load x Epoch interaction,

F(1.95, 52.54) = 11.05,p <.001, r2= .290. Planned contrasts revealed that the effect of

WM load on slow wave amplitudes was significantly smaller than it was for the two LPP

epochs, F(1,27) = 13.42, p .001, r2= .332. Of the two LPP time windows, the early LPP

was modulated more by WM load than the late LPP, F(1,27) = 8.46, p < .01, q2= .239.

These load effects did not differ between controls and high-fear participants, as there was









no main effect of group or group interactions with regard to load or epoch effects, Fs <

.80,ps > .45, 2<.030.

Overall, the effects of WM load on the ERP components of interest support the

prediction that fewer attentional resources are available for processing interfering picture

stimuli during the delay period of the high WM load condition relative to the low load.

That is, less neural activity is engaged in the capture of attention for the pictures while

participants actively maintain greater stimulus representations during the high load

conditions. Since the pictures and WM representations are both competing for similar

reservoirs of neural processing, this finding suggests that more neural resources are

allocated to the active maintenance of high load representations (three diamonds)

compared to the low load (one diamond). As the high WM load engages more neural

resources, the opportunity is optimal for interference effects caused by the pictures. As a

result, only ERP components taken from high WM load trials will be examined in the

analyses that follow.

Effects of Picture Interference

ERP responses to interfering pictures were acquired from all scalp electrodes, and

converted into grand averaged waveforms of picture interference for each picture

category (Figure 3-12). Spherical-spline interpolated scalp voltage maps showed that a

central region (including electrode #34) was differentially sensitive to affective pictures

relative to neutral in the early LPP epoch; however, this region of highest differential

sensitivity shifted more centro-frontally during the late LPP and slow wave time windows

(Figure 3-13). In determining the electrode site to use for evaluating the effects of picture

interference, consideration was given to prior research involving affective pictures and










the earlier findings of this study relating to WM-related effects. Since previous research

has found centro-parietal regions to show maximal positive potentials in response to

affective pictures and this region also showed sensitivity to load-related effects in this

study, electrode #34 was used for the picture-related analyses that follow.




44 Front 3
1.0

1 d Neutral
1:4 1: -Pleasant
... Unpleasant
jA -. --'-\ Threat
1:5 di



1 3 j d '
1:16 '---- 43.J- .




Left Right
4both inor-i c aE r -2



-,----,-i---0 -, "----







0 1000



processing in the neutral (blue), pleasant (black), unpleasant (green), and
4-,----.1



both line -with the WM task performance data, 2-Group x 4-Valence repeated
,5 0---_4-__ 6 t~, ~
,. .. 0|. ,1.0.00.
Back- Time [ms]


Figure 3-12. Grand-averaged ERPs for all scalp sites during interference picture
processing in the neutral (blue), pleasant (black), unpleasant (green), and
threat (red) conditions.

In line with the WM task performance data, 2-Group x 4-Valence repeated

measures ANOVAs were performed on mean amplitudes taken from electrode #34 for

the early LPP, late LPP, and slow wave time periods in order to determine the effects of









the interfering pictures at the high WM load. Planned orthogonal contrasts were used to

test a priori hypotheses regarding effects of interfering picture valence. Statistics for the

ANOVA main effects and interactions can be found in Table 3-5, while results from the

contrasts can be found in Table 3-6.


Early LPP



i- L "


Late LPP


Front:


Slow Wave


iT2/


Side: ,'i
..'j_'., "' i'"' W'


Back:


326.0 ms reference free
EEG- Voflage 0.11l V I step


482.0 ms + reference free
EEG Voltage f -- 0.11 pV step


710.0 ms + reference free
EEG Voltage 0.11 V f step


Figure 3-13.


Spherical-spline interpolated scalp voltage maps representing the
differences in neural processing between affective and neutral pictures for
the early and late LPPs and slow waves of high load trials. Voltage
difference maps were calculated by subtracting neutral from affective
picture activity, and were calculated at 326ms (early LPP), 482ms (late
LPP), and 710ms (slow wave) after picture onset. Positive voltage
differences are indicated in red, showing strongest affective-related
differences in the early LPP and the late LPP.


Table 3-5. ANOVA statistics from the ERP components
Early LPP Late LPP Slow Wave
Effect F-Ratio r 2 F-Ratio r2 F-Ratio r2
Group (G)a .025 .001 .194 .007 .24 .009
Valence (V)b 6.70*** .199 6.67*** .198 12.64** .319
G x Vb .173 .006 .619 .022 1.62 .056
adf= 1,27. bdf= 3,81. *p <.001.


7~7









Table 3-6. Valence effects in the ERP components


Valence Contrasts"
Early LPP
Affective vs. Neutral
Negative vs. Pleasant
Threat vs. Unpleasant

Late LPP
Affective vs. Neutral
Negative vs. Pleasant
Threat vs. Unpleasant

Slow Wave
Affective vs. Neutral
Negative vs. Pleasant
Threat vs. Unpleasant


Valence
(Across Group)
F-Ratio qr2

4.26* .136
16.47** .379
2.20 .075


4.743* .149
15.37*** .363
.056 .814


.021 .001
11.05* .290
32.96*** .550


adf= 1,28. p < .05. 'p < .01. j'p < .001.

Early
LPP


p


Valence x Group
(Across Load)
F-Ratio q2

.001 .001
.641 .023
.009 .001


.337
.591
.886


.08
.756
5.38*


Late
LPP

'"\


.012
.021
.032


.003
.027
.166


Neutral
Pleasant
Unpleasant
Threat


Slow
Wave


0 250 500 750 1000
Time [ms]
Figure 3-14. Grand-averaged ERPs for site #34 during interference picture processing
in the neutral (blue), pleasant (black), unpleasant (green), and evolutionary
threat (red) conditions at high WM loads, collapsed across groups.

Figure 3-14 illustrates the grand-averaged ERP waveforms as a function of

interfering picture valence category during high WM load trials. As can be seen, valence

category of the interfering pictures modulated the amplitudes of all ERP epochs. Planned

contrasts revealed that participants as a whole exhibited greater positive deflections in

both LPP time windows for each of the three affective picture types compared to neutral.


/r


-------------------------------------- ------- ------- ---


v









Both LPPs and slow waves had elevated positive amplitudes for negative (unpleasant and

threat) compared to pleasant interfering pictures. Additionally, slow waves were more

positive for unpleasant compared to threat pictures. This slow wave effect was qualified

by planned contrasts that revealed a significant Group x Valence interaction, with

controls exhibiting significantly more negative slow waves than high-fear participants for

threat relative to unpleasant pictures. Put another way, this indicates that the slow waves

of high-fear participants were more positive than controls with respect to unpleasant slow

waves. Figures 3-15 and 3-16 show the overall effects of all picture valence categories

on ERP amplitudes for controls and high-fear participants, while Figure 3-17 shows

grand-averaged ERP waveforms that reveal the Group x Valence interaction where high-

fear participants show more positive slow wave for threat pictures (relative to unpleasant)

than controls.

Effects of Picture Valence on ERPs:
Controls
5-

4-

0 Neutral
2--
-o2 Pleas.

.1 M Unpleas.
E U Threat

w -1
-2
Early LPP Late LPP Slow wave
-3
ERP Epoch
Figure 3-15. Mean ERP amplitudes for the early and LPP and slow wave for controls.
Error bars represent standard errors.










Effects of Picture Valence on ERPs:
High-Fear


4-

3-

2-

1 -

0

-1


O Neutral
O Pleas.
* Unpleas.
* Threat


-2 Early LPP Late LPP Slow wave
ERP Epoch
Figure 3-16. Mean ERP amplitudes for the early and LPP and slow wave for high-fear
participants. Error bars represent standard errors.


Early
LPP


-Control Threat
- High-Fear Threat
-Control Unpleas.
- High-Fear Unpleas.


0 250 500 750 1000
Time [ms]
Figure 3-17. Grand-averaged ERPs for controls and high-fear participants during threat
and unpleasant interference picture processing in the at high WM load.

The enhanced positivity exhibited in high-fear participants' slow waves for the

threat pictures is intriguing because it may account for group differences in sustained

attention to the threat picture interference. The slow wave time window used for the









present analysis was taken from 660-760 ms following picture onset, but as Figure 3-17

shows, the group difference in threat picture slow waves increases dramatically beyond

this point. If the time window for slow wave analysis was extended to 1000 ms

following picture onset, the group difference found in the slow waves would likely be

much more powerful than the present analyses indicate. Unfortunately, activity recorded

after 750 ms is confounded by the fact that a fixation point replaced the pictures at this

time, making it difficult to interpret this activity as purely reflecting sustained attention

and not some degree of response to the picture offset.

One final analysis was conducted to determine if the increased positivity seen in

the fear participants' slow waves for the threat pictures was associated with behavioral

symptomatology reported by the participants in the self-report measures. To accomplish

this, mean slow wave amplitudes for threat pictures were correlated with mean scores on

the SNAQ and SPQ, STAI-trait, STAI-state, and BDI. Correlations revealed that slow

wave responses to the threat pictures were moderately correlated to phobic symptoms

reported in the SNAQ/SPQ, r(27) = .363, p = .053, and the state anxiety reported in the

STAI, r(27) = .408, p < .05. Correlations further revealed associations between threat

slow waves and trait anxiety reported in the STAI, [r(27) = .417, p < .05 and depressive

symptoms reported in the BDI, r(27) = .402, p < .05. These findings suggest that

sustained attentional processing to interfering threat pictures is associated with behavioral

symptoms of negative affect (including specific fear), which may account for group

differences in WM impairment caused by threat picture interference.

To summarize the ERP results, WM load manipulation had a strong effect on the

neural activity devoted to processing the interfering pictures. This effect mirrors that









seen in the WM task data. Low WM load trials evoked more positive early and late LPPs

compared to the high load, suggesting that less attentional resources were available to

process the pictures during high load trials. Differential load-related activity was centro-

parietal in its scalp distribution; however, this effect diminished over time and was not

significant in the slow wave components. With regard to the valence effects of the

interfering pictures, LPPs had larger positive amplitudes for emotionally salient

compared to neutral interfering pictures. This also reflects effects that were seen in the

accuracy and RT data, suggesting that the greater the positivity of the interfering picture

response, the greater the disruption in the maintenance of the WM representation. All

ERP components showed greater centro-parietal positivity for negative pictures

(unpleasant and threat) compared to neutral. Slow waves were most positive for

unpleasant pictures in the centro-parietal region, yet controls and high-fear participants

differed in their slow waves such that high-fear participants showed more positive

activity for the interfering threat pictures than controls. The enhanced slow wave

positivity for threat pictures was associated with self-reported behavioral symptoms of

negative affect, including specific fear of the threatening content.














CHAPTER 4
DISCUSSION

This study investigated the effects of affective interference on visual working

memory (WM). The main objectives were: 1) to determine if emotionally-arousing

interference degrades the active maintenance of task-relevant representations in WM and,

if so, 2) whether this degradation is disproportionately greater to interfering

evolutionarily-relevant threat stimuli. Additionally, this study sought to determine if

high-fear participants showed evidence of greater WM-related disruption to clinically-

relevant interfering stimuli. Evidence from WM task performance and

electrophysiological data converge to suggest that the representations in WM were

differentially modulated by task-irrelevant affective interference. Changes in task

performance attributable to both WM load and affective interference appear to be

associated with rapidly changing neural processes. Taken together, these findings raise

new questions about interactions between affect and higher-level cognitive processes like

WM, and provide insight into the disruptive nature of fear and anxiety.

Interference Effects on WM Task Performance

WM load had very strong effects on both accuracy and RT, with participants

making more errors and taking longer to respond to the probe during the high- than low-

load conditions. This indicates that participants had more difficulty with the more

challenging 3-diamond cues and engaged more WM resources to actively maintain

representations of the higher load stimuli. Valence of the interfering picture had different

group-related effects on accuracy and probe RT, as controls made more errors and took









less time to respond to the probe during trials that involved pleasant compared to neutral

picture interference; high-fear participants took more time to respond to the probe during

threat picture trials and showed a trend-level increase compared to controls in errors

during the threat and unpleasant relative to pleasant picture interference.

The finding that controls made more errors and took less time to respond during

pleasant picture trials is somewhat surprising at first; however, controls rated pleasant

pictures as the most arousing category of pictures, significantly higher than neutral.

Correlations between error rates and probe RTs for the pleasant picture trials were not

significant, suggesting that they did not exhibit a speed-accuracy trade-off for the

pleasant picture trials. Therefore, it appears instead that the controls' WM task

performance was influenced by arousal in such a way as to cause greater interference in

WM representations while at the same time facilitating faster processing of the probe.

Ashby (1999) proposed that moderate levels of positive affect enhance WM, but the

pleasant pictures used in this study may have been too arousing for this benefit to be

realized. This finding fits a pattern similar to that of the classic Yerkes-Dodson law

(Yerkes & Dodson, 1908) that performance is related to arousal in an inverted U-shaped

pattern such that too low or high arousal leads to impairment in cognitive performance.

In this light, the controls' arousal ratings may suggest that the pictures were too arousing

and may have overwhelmed the PFC with excessive levels of dopamine (Arnsten, 1998;

Goldman-Rakic, 1996), thereby disrupting WM function and leading to more impulsive

responding.

Task-related findings for high-fear participants fit some of the predictions that

their WM performance would be most impaired during trials that involved threat pictures.









Increased reaction time in probe responses may be indicative of a reduced cognitive

capacity to actively maintain the representation of the cue stimuli and/or execute the

recognition decision and response. Since high-fear participants rated threat pictures as

the most arousing, this pattern also appears to fit the pattern of the Yerkes-Dodson law.

However, it is interesting that this effect is opposite to that seen in controls where

pleasant pictures actually facilitated faster response to the probe. This may suggest that

the underlying neural processing involved in threat-based affective interference in high-

fear participants may be functionally different from those involved in pleasant affective

interference in controls. It seems plausible that this threat-based effect may be linked to a

greater level of disengaging attention bias invoked by the threatening stimuli (Fox et al.,

2001), which may not occur for pleasant pictures. This would account for the ERP

findings of more positive slow waves to threat pictures in high-fear participants.

Nevertheless, the dissociation between different valence-specific arousal effects cannot

be fully explained with these data alone.

Picture Response Data

Participants detected interfering threat pictures during the WM task faster than the

other categories, and high-fear participants took longer than controls to detect interfering

pictures as a whole. This is an interesting finding, as it shows that threat pictures

successfully captured the participants' attention better than any other type of picture -

regardless of the their level of fear and that higher levels of anxiety has a cost on overall

efficiency of attentional processing. This supports previous work that has found that

threat cues evoke automatic attention biases (Ohman, Flykt, & Esteves, 2001), and may

be explained by neural pathways that are present for both healthy and anxious individuals

that route threatening information directly to the amygdala for automatic processing









(Ledoux, 1996). It appears that the picture detection response (which involved a

keypress) may have made an impact on the late LPP component, as the peak latencies for

each valence group seem to coincide somewhat with the picture detection RTs.

While making their ratings at the end of the experiment, participants chose to

view unpleasant pictures the longest prior to making their ratings. One likely explanation

for this is that the unpleasant pictures were more complex and required more sustained

attention in order to rate them. Instead of simple pictures of snakes, smiling babies, or

non-affective items such as an ironing board, unpleasant pictures included some complex

scenes such as accidents, natural disasters, and saddening events. It is also possible that

they depicted more novelty and provoked more curiosity in the participants. Participants'

ratings of valence and arousal differentiated neutral, pleasant, and unpleasant pictures in

line with the expected patterns, revealing that the key manipulation of picture valence

was successfully executed. High-fear participants rated threat pictures as significantly

more arousing and less pleasant than controls, while controls rated pleasant pictures as

significantly more arousing than other picture categories.

ERP Reflections of Interference Stimulus Processing

WM Load Effects on Interference-Related Activity

As predicted, there was a clear load-related effect on ERPs to the task-irrelevant

interfering stimuli. Both the early and late LPP was smaller during the high- than low-

load condition. This finding suggests that fewer resources were available during the

high-load maintenance interval to process interfering stimuli and extends previous work

that observed a similar compromise of attentional resources with regard to picture

processing and startle modulation (Schupp et al., 1997). However, the WM load

modulation of picture-related LPPs decayed over time and did not affect the slow wave









ERP component. Load-related WM modulation of interference picture processing was

most pronounced for the earlier stages of attentional processing and,

electrophysiologically, was less reflected during later stages of more sustained attention.

Valence-Related Effects Interference Activity

As predicted, the valence characteristics of the interfering stimuli influenced the

amplitudes of the early and late LPP and slow wave. LPPs were larger for emotionally-

arousing than neutral interference pictures, which supports previous research that

affective pictures selectively engage attention based on motivational importance

(Cuthbert et al., 2000; Schupp et al., 2003, Baas et al., 2001). Additionally, all three ERP

components were larger in amplitude to negative (unpleasant and threat) than pleasant

pictures. This finding is consistent with the hypothesis that the unpleasant and threat

pictures recruited more attentional processing than did pleasant pictures. Slow waves

were most positive for unpleasant pictures, suggesting that unpleasant pictures engaged

more sustained attention than the other picture categories. Slow waves were very

responsive to differences in picture valence category, but they appeared to be cut short by

the offset of the picture; that is, there is a clear offset potential evoked by picture

termination, and this offset potential disrupted the slow wave making it difficult to

evaluate the duration of differential processing associated with emotional valence. A

significant Group x Valence interaction revealed that controls and high-fear participants

differed in their slow waves. High-fear participants showed more positive activity than

controls for threat pictures relative to unpleasant. Unpleasant pictures still evoked the

most positive slow waves in both groups, but high-fear participants exhibited

dramatically more sustained processing of the threat pictures compared to controls. This

finding may help to explain the increased WM impairment for high-fear participants









caused by the threat picture interference. Surprisingly, no ERP effects seemed to

correlate with the controls' WM impairments caused by the pleasant picture interference.

Interestingly, the scalp-voltage distribution maps shown in Figure 3-13 suggest

that emotionally arousing versus neutral scalp-voltage difference maps became

progressively more frontal over time. While this effect has not yet been quantitatively

examined, there are several potentially important implications. First, the choice of

electrode site for measuring valence-related effects may not have been optimal for

sensitively detecting differences in the different components. Second, the progressively

more frontal distribution over time may be consistent with the hypothesis that

emotionally-arousing interference stimuli both disrupt active maintenance of task-

relevant representations in WM and, furthermore, that representations of these task-

irrelevant interfering stimuli become preferentially processed over representations of the

task-relevant stimuli. That is, it is possible that representations of the task-irrelevant

interference stimuli may actually become task relevant and displace those of the task-

relevant stimuli in the dlPFC.

Emerging Patterns

The data suggest that affective interference disrupts the active maintenance of

representations in WM and that this impairment is associated with the arousal

characteristics of the interfering stimuli. High-fear participants exhibited greater

allocation of sustained attention for threat pictures than controls, which corresponded

with a longer RT to make recognition decisions during the WM task. This may suggest

that high-fear participants have difficulty disengaging their attention from clinically-

relevant threat stimuli, which may lead to maintenance of their fear. Controls also

showed task-related patterns of WM impairment from interfering pleasant pictures









(which they rated as highly arousing), but these impairments did not show the neural

signature that was detected for an overall attention bias for threatening pictures or the

threat-induced WM impairments in high-fear participants.

Taken with the task performance data, the ERP results from this study suggest

several important findings. First, high- and low-fear individuals both show selective

processing of effectively arousing stimuli, even in the context of a challenging cognitive

task. Second, clinically-relevant threat interference has the capacity to disrupt neural

processes of sustained attention that are needed for higher-level cognitive processes such

as WM, and these changes in sustained attention are significantly associated with

behavioral symptoms of negative affect, including specific fear. Third, when attentional

resources are devoted to a challenging visual WM task, they are effectively subtracted

from the attentional capacity that is available for the processing of incoming affective

pictures, indicating that WM can diminish the degree of affective processing resources

that are available. Taken together, these ERP effects provide novel insights into neural

processes involved in attention, WM, and emotion that have not been previously

examined. Future studies are needed to clarify the nature of these cognitive-affective

interactions and expand these findings to other domains of cognitive and affective

processing.

Alternative Explanations and Possible Limitations

The results of this study offer new hope into better understanding nature of

affective interference of WM and the potential role that anxiety plays on higher-level

cognition. However, alternative explanations and possible limitations need to be

considered. The operational definitions associated with any form of emotion research are

complex. Lang (1998) has postulated that emotions can be measured on two dimensions:









valence and arousal. As a result, this study sought to define affective stimuli along these

two dimensions. However, it is possible that the pictures varied on other criteria that

were not controlled (e.g., brightness, complexity, human content) and may have affected

the outcome. It is also possible that the sample selected for participation may have had

life experiences that were not assessed that may have influenced their reactions to the

pictures. Group differences may have also been influenced by demand characteristics

that were placed on the high-fear participants (i.e., special attention given to them during

recruitment and interview because of their higher levels of fear). Alternatively, it is also

possible that the "high-fear" participants were not fearful enough for robust group effects

to emerge, as other studies have found individuals with spider or snake phobias score

much higher on the SNAQ/SPQ than the high-fear participants in this study. The mean

SNAQ/SPQ scores for high-fear participants in this study was 10.6, while snake and

spider phobics have obtained mean scores over 23 in other studies (e.g., Fredrikson,

1983).

Although this study repeated pictures during the WM task, it did not have a

sufficient number of trials to examine potential habituation effects; that is, repeated

exposure to each of the pictures may reasonably be assumed to be associated with

decreased arousal to subsequent exposures. However, given the limited number of trials

per condition, there was not adequate signal-to-noise ratio in the ERP data to examine

effects of repeated exposure. An additional concern is that the baseline accuracy

condition for comparing groups was measured as performance no-interference trials, yet

this condition was randomly mixed with the four categories of picture interference,

making the no-interference condition a bit of an "odd-ball" that may not truly represent









baseline WM performance. Lastly, this study had a small sample size that unfortunately

could not be completely balanced on gender in each group due to the low numbers of

available high-fear individuals.

Several additional potential limitations of the ERP data must be considered. First,

terms like "late positive potential" have somewhat limited interpretational value aside

from discussions of general attentional resources. Part of the reason for this is that

volume conduction of electrical signals in the brain makes it very difficult to map sources

of brain activity with precise spatial resolution, although new technologies are emerging

that make source localization more feasible and effective. A considerable amount of ERP

variability exists from participant to participant, and some participants exhibit the same

ERP components at slightly different latencies. Grand-averaging subjects together in

certain cases might actually average out a key component of interest. Also, due to the

number of trials that were available, no ERP analyses were performed on incorrect trials,

meaning that accuracy could not be considered as a variable in the ERP analysis.

Another concern emerged after analyzing the data taken from the centro-parietal

electrode, as the sensitivity to affective picture characteristics shifted more frontally and

may have been more pronounced if the slow wave had been analyzed from a more frontal

electrode site. It appears that 750 ms may not have not been a long enough picture

presentation to assess full slow wave effects. Finally, since analysis of the cue and probe

waveforms went beyond the scope of this project, it is difficult to completely interpret the

picture processing ERPs without a more complete understanding of the neural responses

to the task-relevant stimuli associated with the WM task.









Future Directions

Despite its limitations, this study provides important new information about the

interfering effects of emotional arousal on active maintenance of task-relevant

representations within WM. For the sake of complexity, and to address the primary aims

of the current research, task-relevant stimulus-evoked activity (i.e., cues and probes) was

not examined. These data will be examined in the future to more fully elucidate the role

of affective interference on WM-related processes. Additionally, as the arousal level

ii li/hi/ a valence category has been shown to modulate startle probe and LPP amplitudes

during emotional picture viewing (i.e., larger LPP and startle with higher arousal ratings

within category; Schupp et al., 1997), it would be useful to examine arousal-related

effects more parametrically. By necessity, in the current study, arousal and emotional

versus neutral stimuli are confounded. That is, pleasant and unpleasant stimuli were both

rated as high in arousal, while neutral stimuli were rated as low in arousal.

While localization of the neural sources giving rise to the observed interference

effects was not a primary aim of the study, such information will certainly be useful.

Findings by Perlstein et al. (2002) and Gray et al. (2002) have indeed shown that WM-

related dlPFC activity is modulated by emotionally arousing contexts. Future efforts will

be aimed at applying source localization routines to the current data using Brain Electric

Source Analysis (BESA; Scherg, 1990) to determine if the effects observed in the current

study are similarly localizable to the dlPFC. Also, the paradigm used for the present

study is currently being adapted for the functional MRI context, in order to facilitate

examination of brain activity that offers a higher degree of spatial resolution. Future

extensions of the project could also include measuring different types of specific fear or

anxiety (e.g., needle phobia, PTSD, etc.), to learn whether the findings in the current









study generalize to other clinical populations. Furthermore, this paradigm could

potentially be used to assess behavioral and electrophysiological effects of therapy that

focuses on reduction of fear symptoms in phobic individuals.

Summary

This study sought to investigate the effects of affective interference on visual

working memory (WM), testing whether affective interference impairs representations in

working memory (WM) and the degree to which this interference is due to effects of

arousal brought about by the clinical or evolutionary relevance of threat in participants

with high levels of fear. Performance data from the WM task revealed that WM

performance is modulated by affective interference and that arousal plays a key role in

that interference, causing high-fear participants to respond more slowly as a result of

threat interference and controls to make more errors and respond more quickly as a result

of pleasant picture interference. Subtle differences in sustained attention appear to be the

only source of group differences in the neural effects associated with threat-based

interference of WM, but no neural effects appear to explain the effects of pleasant picture

WM interference in controls. As a whole, participants demonstrated selective detection

of threatening stimuli, which corresponded to enhanced positivity in late positive

potentials measured at the centro-parietal region of the scalp. Importantly, ERP results

showed that increased WM load and the processing of affective interference

compromised the neural resources of attention. In the end, these findings help to better

illustrate the complexity of cognitive-affective interactions, and indicate that anxiety can

play a key role in disrupting higher-level cognitive processes involved in WM.














LIST OF REFERENCES


Adolphs, R., Gosselin, F., Buchanan, T.W., Tranel, D., Schyns, P., & Damasio, A.R.
(2005). A mechanism for impaired fear recognition after amygdala damage.
Nature, 433 (7021), 68-72.

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental
disorders (4th ed., text revision). Washington, DC: Author.

Armony, J.L. & Dolan, R.J. (2002). Modulation of spatial attention by fear-conditioned
stimuli: An event-related fMRI study. Neuropsychologia, 40, 817-826.

Armony J.L. & Ledoux, J.E. (1997). How the brain processes emotional information.
Annals of the New York Academy of Sciences, 821, 259-270.

Arnsten, A.F.T. (1998). Catecholamine modulation of prefrontal cortical cognitive
function. Trends in Cognitive Sciences, 2 (11), 436-447.

Ashby, F.G., Isen, A.M., & Turken, A.U. (1999). A neuropsychological theory of
positive affect and its influence on cognition. Psychological Review, 106 (3),
529-550.

Baas, J.M.P., Kenemans, J.L., Bocker, K.B.E., & Verbaten, M.N. (2002). Threat-
induced cortical processing and startle potentiation. Neuroreport, 13, 133-137.

Baddeley, A.D. (1986). Working memory. Oxford: Oxford University Press.

Beck, A.T., Steer, R.A., & Brown, G.K. (1996). Manual for the BDI-II. San Antonio,
TX: The Psychological Coorperation.

Bertrand, 0., Perrin, F., & Pernier, J. (1985). A theoretical justification of the average-
reference in topographic evoked potential studies. Electroencephalography and
Clinical Neurophysiology, 62, 462-464.

Brown, T. A., DiNardo, P. A., & Barlow, D. H. (1994). Anxiety disorders interview
schedule for DSM-IV (ADIS-IV). Albany, NY: Graywind Publications.

Carlsson, K., Petersson, K.M., Lundqvist, D., Karlsson, A., Ingvar, M. & Ohman, A.
(2004). Fear and the amygdala: Manipulation of awareness generates differential
cerebral responses to phobic and fear-relevant (but not feared) stimuli. Emotion,
4 (4), 340-353.









Cohen, J.D., Perlstein, W.M., Braver, T.S., Nystrom, L.E., Noll, D.C., Jonides, J. &
Smith, E.E. (1997). Temporal dynamics of brain activation during a working
memory task. Nature, 386, 604-608.

Coles, M.G.H. & Rugg, M.D. (1995). Event-related brain potentials: an introduction. In
M.G.H. Coles & M.D. Rugg (Eds.), Electrophysiology of mind. Event-related
potentials and cognition (pp. 1-35). Oxford, England: Oxford University Press.

Compton, R.J. & Banich, M.T. (2003). Paying attention to emotion: An fMRI
investigation of cognitive and emotional Stroop tasks. Cognitive, Affective, &
Behavioral Neuroscience, 3 (2), 81-96.

Curtis, C.E. & D'Esposito, M. (2003). Persistent activity in the prefrontal cortex during
working memory. Trends in Cognitive Sciences, 7 (9), 415-423.

Cuthbert, B.N., Schupp, H.T., Bradley, M.M., Birbaumer, N., & Lang, P.J. (2000).
Brain potentials in affective picture processing: Covariation with autonomic
arousal and affective report. Biological Psychology, 52, 95-111.

Dolan, R.J. (2002). Emotion, cognition, and behavior. Science, 298, 1191-1194.

Donchin, E. & Coles, M.G. (1988). Is the P300 component a manifestation of context
updating? Behavioral and Brain Sciences, 11, 357-427.

Fabiani, M., Gratton, G., & Coles, M.G.H. (2000). Event-related potentials. In J.T.
Cacioppo, L.G. Tassinary & G.G. Bemston (Eds.), Handbook ofpsychphysiology
(2nd ed.). Cambridge, England: Cambridge University Press.

Fox, E. (1993). Allocation of visual attention and anxiety. Cognition and Emotion, 8,
165-195.

Fox, E., Russo, R., Bowles, R., & Dutton, K. (2001). Do threatening stimuli draw or
hold visual attention in subclinical anxiety? Journal of Experimental Psychology:
General, 130 (4), 681-700.

Fuster, J.M. (1997). The prefrontal cortex: Anatomy, physiology, and neuropsychology
of the frontal lobe (3rd ed.). New York: Raven.

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.

Goldman-Rakic, P.S. (1987). Circuitry of primate prefrontal cortex and regulation of
behavior by representational memory. In F. Plum (Ed.), Handbook of physiology;
The nervous system (pp. 373-414). Bathesda, MD: American Physiological
Society.









Goldman-Rakic, P.S. (1996). Regional and cellular fractionation of working memory.
Proceedings of the National Academy of Sciences, 93, 13473-13480.

Hermans, D., Vansteenwegen, D., & Eelen, P. (1999). Eye movement registration as a
continuous index of attention deployment: Data from a group of spider anxious
students. Cognition and Emotion, 13, 419-434.

Huynh, H. & Feldt, L.S. (1976). Estimation of the box correction for degrees of
freedom from sample data in the randomized block and split plot designs.
Journal of Educational Statistics, 1, 69 -82.

Ikeda, M., Iwanaga, M., & Seiwa, H. (1996). Test anxiety and working memory system.
Perceptual Motor \/Aill, 82, 1223-1231.

Johnson, V.S., Miller, D.R., & Burleson, M.H. (1986). Multiple P3s to emotional
stimuli and their theoretical significance. Psychophysiology, 23, 684-693.

Keil, A., Bradley, M. M., Hauk, 0., Rockstroh, B., Elbert, T., & Lang, P. J. (2002).
Large-scale neural correlates of affective picture processing. Psychophysiology,
39, 641-649.

Keppel, G. (1982). Design and analysis: A researcher's handbook, (2nd ed.). Englewood
Cliffs, NJ: Prentice-Hall.

Kindt, M. & Brosschot, J.F. (1997). Phobia-related cognitive bias for pictorial and
linguistic stimuli. Journal ofAbnormal Psychology, 106 (4), 644-648.

Klorman, R., Hastings, J.E., Weerts, T.C., Melamed, B.G., & Lang, P.J. (1974).
Psychometric description of some specific-fear questionnaires. Behavior
Therapy, 5, 401-409.

Kutas, M. & Dale, A. (1995). Electrical and magnetic readings of mental functions. In
M.D. Rugg (Ed.), Cognitive Neuroscience (pp 187-242). Cambridge, MA: MIT
Press.

Lane, R.D., Reiman, E.M., Bradley, M.M., Lang, P.J., Ahem, G.L., Davidson, R.J., &
Schwartz, G.E. (1997). Neuroanatomical correlates of pleasant and unpleasant
emotion. Neuropsychologia, 35, 1437-1444.

Lang, P. J. (1980). Behavioral treatment and bio-behavioral assessment: Computer
applications. In J. B. Sidowski, J. H. Johnson, & T. A. Williams (Eds.),
Technology in mental health care delivery systems (pp. 119-137). Norwood, NJ:
Ablex.









Lang, P.J., Bradley, M.M., & Cuthbert, B.N. (1997). Motivated attention: Affect,
activation, and action. In P. Lang, R.F. Simmons, & M. Balaban (Eds.), Attention
and orienting: Sensory and motivational processes: Hillsdale, NJ: Erlbaum
Associates.

Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1998). International affective picture
system (IAPS): Technical manual and affective ratings. Gainesville, FL: The
Center for Research in Psychophysiology, University of Florida.

Lavric, A., Rippon, G., & Gray, J.R. (2003). Threat-evoked anxiety disrupts spatial
working memory performance: An attentional account. Cognitive Therapy and
Research, 27 (5), 489-504.

Ledoux, J. E. (1996). The emotional brain. New York, NY: Simon & Schuster.

Low, A., Rockstroh, B., Harsch, S., Berg, P., & Cohen, R. (2000). Event-related
potentials in a working-memory task in schizophrenics and controls.
Schizophrenia Research, 46, 175-186.

Luck, S.J., Woodman, G.F., & Vogel, E.K. (2000). Event-related potential studies of
attention. Trends in Cognitive Sciences, 4 (11), 432-440.

MacLeod, C.M. (1991). Half a century of research on the Stroop effect: An integrative
review. Psychological Bulletin, 190, 163-203.

MacLeod, C.M, Matthews, A., & Tata, P. (1986). Attentional bias in emotional
disorders. Journal ofAbnormal Psychology, 95, 15-20.

Miller, E.K. & Cohen, J.D. (2001). An integrative theory of prefrontal cortex function.
Annual Review ofNeuroscience, 24, 167-202.

Morris, J.S., Frith, C.D., Perrett, D.O., Rowland, D., Young, A.W., Calder, A.J., &
Dolan, R.J. (1996). A differential neural response in the human amygdala to
fearful and happy facial expressions. Nature, 383 (6603), 812-815.

Ohman, A., Flykt, A., & Esteves, F. (2001). Emotion drives attention: Detecting the
snake in the grass. Journal of Experimental Psychology: General, 130 (3), 466-
478.

Ohman, A. & Mineka, S. (2001). Fear, phobias, and preparedness: Toward an evolved
module of fear and fear learning. Psychological Review, 108, 483-522.

Ohman, A. & Mineka, S. (2003). The malicious serpent: Snakes as a prototypical
stimulus for an evolved module of fear. Current Directions in Psychological
Science, 12 (1), 5-9.









Ohman, A. & Soares, J.J.F. (1998). Emotional conditioning to masked stimuli:
Expectancies for aversive outcomes following nonrecognized fear-relevant
stimuli. Journal of Experimental Psychology: General, 127, 69-82.

Ohman, A., & Soares, J.J.F. (1993). On the automatic nature of phobic fear:
Conditioned electrodermal responses to masked fear-relevant stimuli. Journal of
Abnormal Psychology, 102. 121-132.

Ohman, A., & Soares, J.J.F. (1994). Unconscious anxiety: Phobic responses to masked
stimuli. Journal ofAbnormal Psychology, 103, 231-240.

Perlstein, W.M., Elbert, T., & Stenger, V.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.

Posner, M.I., Inhoff, A.W., Friedrich, F.J., & Cohen, A. (1987). Isolating attentional
systems: A cognitive-anatomical analysis. Psychobiology, 15, 107-121.

Ratcliff, R. (1993). Methods for dealing with reaction time outliers. Psychological
Bulletin, 114, 510-532.

Runchkin, D.S., Johnson, R., Mahaffey, D., & Sutton, S. (1988). Toward a functional
categorization of slow waves. Psychophysiology, 25, 339-353.

Sadock, B.J. & Sadock, V.A. (2003). Kaplan and Sadock's synopsis of psychiatry:
Behavioral sciences/clinical psychiatry (9th ed.). Philadelphia, PA: Lippincott
Williams & Wilkins.

Scherg, M. (1990). Fundamentals of dipole source potential analysis. In F. Grandori &
M. Hoke (Eds.), Auditory evoked magnetic fields and electric potentials.
Advances in audiology (Vol. 6, pp. 65-78). Basel: Karger.

Schimmack, U. (2005). Attentional interference effects of emotional pictures: Threat,
negativity, or arousal? Emotion, 5 (1), 55-66.

Schupp, H.T., Cuthbert, B.N., Bradley, M.M., Birbaumer, N., & Lang, P.J. (1997).
Probe P3 and blinks: Two measures of affective startle modulation.
Psychophysiology, 34, 1-6.

Schupp, H.T., Junhhofer, M., Weike, A.I., & Hamm, A.O. (2003). Emotional
facilitation of sensory processing in the visual cortex. Psychological Science, 14
(1),7-13.

Schupp, H.T., Junhhofer, M., Weike, A.I., & Hamm, A.O. (2004). The selective
processing of briefly presented affective pictures: An ERP analysis.
Psychophysiology, 41, 441-449.









Smith, E.E. & Jonides, J. (1999). Storage and executive processes in the frontal lobes.
Science, 283, 1657-1661.

Spielberger, C. S., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the state trait
anxiety inventory. Palo Alto, CA: Consulting Psychologists Press.

Tucker, D.M., Liotti, M., Potts, G.F., Russell, G.S., & Posner, M.I. (1994).
Spatiotemporal analysis of brain electrical fields. Human Brain Mapping, 1, 134-
152.

Vrana, S., Roodman, A., & Beckham, J. (1995). Selective processing of trauma-relevant
words in post-traumatic stress disorder. Journal ofAnxiety Disorders, 9, 515-530.

Williams, J.M.G., Watts, F.N., Macleod, C., & Matthews, A. (1997). Cognitive
psychology and emotional disorders, (2nd ed.). Chichester, England: Wiley.

Williamson, S.J. & Kaufman, L. (1990). Theory of neuroelectric and neuromagnetic
fields. In F. Grandor, M. Hokle, & G.L. Romani (Eds.), Auditory Evoked
Magnetic Fields and Electric Potentials. Advances in Audiology (Vol. 6, pp. 1-
39). Basel, Switzerland: Karger Publications.

Yamasaki, H., LaBar, K.S., & McCarthy, G. (2002). Dissociable prefrontal brain
systems for attention and emotion. Proceedings of the National Academy of
Sciences, 99, 11447-11451.

Yerkes, R.M. & Dodson, J.D. (1908). The relation of strength of stimulus to rapidity of
habit-formation. Journal of Comparative Neurology and Psychology, 18, 459-
482.















BIOGRAPHICAL SKETCH

David Stigge-Kaufman received his B.A. degree from Bethel College in North

Newton, Kansas, in 1998, majoring in biology and psychology. He plans to receive his

M.S. from the University of Florida in 2005, and then plans to continue his doctoral study

in clinical psychology, concentrating his clinical and research training in

neuropsychology.