<%BANNER%>

Dissociating Components of Cognitive Control Using High-Density Event-Related Potentials: Implementation of Control, Con...


PAGE 1

DISSOCIATING COMPONENTS OF CO GNITIVE CONTROL USING HIGHDENSITY EVENT-RELATED POTENTIALS: IMPLEMENTATION OF CONTROL, CONFLICT PROCESSING, AND ERROR MONITORING By MICHAEL JAMES LARSON 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 2004

PAGE 2

Copyright 2004 by Michael James Larson

PAGE 3

ACKNOWLEDGMENTS I thank my advisor, William M. Perlstein, as well as my collaborators Vonetta Jones, Grant Webber, and Bao-thuy Hoang. I also thank my parents and my wife for their support. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT.......................................................................................................................ix CHAPTER 1 INTRODUCTION........................................................................................................1 Cognitive Control.........................................................................................................1 Regulative/Strategic Processes..............................................................................1 Evaluative Processes.............................................................................................3 Neural Correlates of Stroop Performance.............................................................3 Dissociation of Cognitive Control Processes........................................................4 Event-related Potentials................................................................................................7 ERPs and Cognitive Control.........................................................................................9 Electrophysiological Correlates of Evaluative Processes......................................9 Electrophysiological Correlates of Regulative Processes...................................12 Dissociation of Cognitive Control Component Processes using ERPs...............13 Predictions..................................................................................................................14 Behavioral Data...................................................................................................14 ERP Data.............................................................................................................15 Regulative Processes....................................................................................15 Evaluative Processes....................................................................................15 2 METHOD...................................................................................................................17 Participants.................................................................................................................17 Materials and Procedure.............................................................................................17 EEG Acquisition and Reduction.................................................................................19 EEG Acquisition..................................................................................................19 EEG Data Reduction...........................................................................................19 Statistical Analyses.....................................................................................................21 Behavioral Data Analyses...................................................................................21 ERP Data Analyses..............................................................................................22 iv

PAGE 5

3 RESULTS...................................................................................................................23 Behavioral Analyses...................................................................................................23 Error Rate Analyses.............................................................................................23 RT Analyses........................................................................................................24 ERP Analyses.............................................................................................................26 Instruction-related Activity.................................................................................26 Stimulus-related Activity.....................................................................................27 Response-related Activity...................................................................................27 4 DISCUSSION.............................................................................................................36 Behavioral Data..........................................................................................................36 ERP Data....................................................................................................................37 Regulative Processes...........................................................................................37 Evaluative Processes...........................................................................................39 Alternative Explanations and Potential Limitations...................................................40 Future Directions........................................................................................................42 Summary.....................................................................................................................43 LIST OF REFERENCES...................................................................................................44 BIOGRAPHICAL SKETCH.............................................................................................50 v

PAGE 6

LIST OF TABLES Table page 2-1. Demographic information...........................................................................................17 3-1. Means and standard errors of error rates (%) in the single-trial Stroop.....................24 3-2. Means and standard errors of median RT (ms) in the single-trial Stroop..................25 vi

PAGE 7

LIST OF FIGURES Figure page 1-1. Schema of the single-trial Stroop task. Participants are first presented with an instructional cue (Color or Word), followed by a delay (1 or 5 seconds) and Stroop stimulus (congruent, neutral, incongruent).....................................................5 1-2. Schemata of cognitive control functions and subsequent ERP manifestations predicted in the current study...................................................................................16 2-1. Sensor layout of 64-channel geodesic sensor net. Slow-wave activity was quantified at scalp sites #2 and #4 (blue); N450 activity at scalp site #55 (red); and, ERN activity at scalp site #5 (green). See ERP Data Analyses in text for details.........20 3-1. Error rates by task and congruency conditions collapsed across delay. Error bars represent standard errors..........................................................................................24 3-2. Median RTs by task and congruency conditions collapsed across delay. Error bars represent standard errors..........................................................................................26 3-3. Top view of spherical-spline interpolated voltage maps representing instruction-related activity of the CN task, WR task, and the difference between the two tasks (CN WR, bottom). ERP waveforms of the grand average differential slow-wave activity between the CN and WR tasks at scalp site #2 (top)..................................28 3-4. Grand average instruction-locked ERPs for all scalp sites of the CN (red) and WR (blue) task presentation............................................................................................29 3-5. Top view of spherical-spline interpolated voltage maps representing instruction-related activity of subsequent correct and incorrect trials, and the difference (correct incorrect, bottom). ERP waveforms of the grand average differential slow-wave activity between subsequent correct and incorrect trials collapsed across conditions at scalp site #4 (top).................................................................................................30 3-6. Grand average instruction-locked ERPs for all scalp sites of correct (red) and incorrect (blue) task presentation collapsed across conditions................................31 3-7. Top view of spherical-spline interpolated voltage maps representing stimulus-related activity of the congruent CN condition, the incongruent CN condition, and the difference between the conditions (congruent incongruent, bottom). ERP vii

PAGE 8

waveforms reflect the grand average N450 to the incongruent CN trials at scalp site #55 (top)...................................................................................................................32 3-8. Grand average stimulus-locked ERPs for all scalp sites of the congruent (red) and incongruent (blue) CN task......................................................................................33 3-9. Top view of spherical-spline interpolated voltage maps representing response-related activity of correct and incorrect trials (collapsed across conditions), and the difference (correct incorrect, bottom). ERP waveforms of the grand average ERN deflection to the incorrect trials as compared to the correct trials at scalp site #5 (top)..........................................................................................................................34 3-10. Grand average response-locked ERPs for all scalp sites of correct and incorrect responses collapsed across conditions......................................................................35 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 DISSOCIATING COMPONENTS OF COGNITIVE CONTROL USING HIGH-DENSITY EVENT-RELATED POTENTIALS: IMPLEMENTATION OF CONTROL, CONFLICT PROCESSING, AND ERROR MONITORING By Michael James Larson May 2004 Chair: William M. Perlstein Major Department: Clinical and Health Psychology Recent theories suggest that cognitive control is a dynamic process instantiated within a fronto-cortical network that implements regulative or strategic control over top-down processes, monitors and detects processing conflicts, and signals for adjustments in performance when necessary. We examined this complement of cognitive control processes behaviorally and neurally. To do this, we acquired high-density brain event-related potentials (ERPs) while 24 neurologically-normal participants performed a cued, single-trial Stroop task that temporally dissociated instruction-related regulative processes (i.e., representing and maintaining the attentional demands of the task), from evaluative processes (i.e., conflict processing, error monitoring). Implementation of control was reflected in a right frontal instruction-related slow-wave associated with the more attentionally demanding color-naming task. A mid-to-lateral frontal stimulus-related conflict N450 was elicited by the incongruent color-naming task condition. ix

PAGE 10

Response-locked ERPs to incorrect responses, collapsed across task conditions, revealed a mid-fronto-central error-related negativity (ERN). Behaviorally, response times and error rates were greatest in the incongruent color-naming task condition, indicating Stroop response time and error-rate interference. Overall, ERPs are determined to be an effective methodology for examining component processes of cognitive control. Furthermore, findings are consistent with previous research indicating two distinct neural systems in cognitive control, one for regulative processes involved in the implementation of control and one involved in the evaluative processes of conflict detection and error monitoring. x

PAGE 11

CHAPTER 1 INTRODUCTION Cognitive Control Cognitive control refers to the ability to guide thought and action in accord with internal intentions (Botvinick, Carter, Braver, Barch, & Cohen, 2001; Cohen, Botvinick, & Carter, 2000; Miller, 2000; Miller & Cohen, 2001) and encompasses those processes necessary for controlled information processing and coordinated actions. Current cognitive neuroscience theories distinguish between at least two important components of cognitive control: a regulative or strategic component responsible for activation and implementation of control processes, and an evaluative component responsible for monitoring the need for control and signaling when adjustments in control are necessary (Botvinick et al., 2001; Kerns et al., 2004). Regulative/Strategic Processes Regulative processes are those involved in the top-down control of cognition and include functions such as representing and maintaining task context and goals (e.g., working memory), the allocation of limited attentional resources, as well as preparing to execute cognitive tasks and override prepotent response tendencies (Cohen, Barch, Carter, & Servan-Schreiber, 1999). Perhaps a prototypical cognitive task that requires the implementation of cognitive control is the Stroop color-word task (MacLeod, 1991; Stroop, 1935). Although there are many different versions of the original Stroop task (MacLeod, 1991), the basic paradigm requires participants to either read words or name the color in which the words are written. To perform this task successfully, participants 1

PAGE 12

2 must selectively attend to one stimulus attribute (i.e., word or printed color). This is particularly true when naming the color of an incongruent or conflict stimulus (e.g., the word GREEN printed in blue) because there is a strong automatic tendency to read the word (GREEN), which competes with the less automatic instruction to name the color (blue). Typically, participants show robust Stroop interference effects, wherein there is increased reaction times (RT) or error rates when participants are required to name the color of the word when the word name and word color are incongruent (conflict condition). The ability to successfully complete this task and overcome the conflict illustrates selective allocation of attentional resources and the ability to select a weaker, task-relevant response in the face of competition from an otherwise more automatic, but task-irrelevant option (Miller & Cohen, 2001); a fundamental aspect of cognitive control. Regulative processes of cognitive control also require the maintenance of task-relevant context and goals. Cohen and colleagues (1999) suggest that maintenance of context representations is critical to adequate performance on the Stroop task. For example, in order to respond to the appropriate dimension of the stimulus, participants must hold in mind the instruction for the trial, providing the necessary context for interpreting the stimulus and generating the correct response. In the card or Golden version of the Stroop task (Golden, 1978) frequently used in neuropsychological testing, task conditions are blocked, wherein all stimuli for each condition are presented as lists on their respective cards. This arrangement consistently reinforces the proper context (i.e., task instruction) and, therefore, places minimal demands on the representation and maintenance of context. In contrast, a single-trial computerized cued version of the Stroop task devised by Cohen et al. (1999) presents trials individually and randomly varies task instruction (color-naming, CN, word-reading, WR). To complete this task,

PAGE 13

3 participants must maintain the context of the task instruction (CN or WR) prior to stimulus presentation and employ these context representations to provide the correct response. Evaluative Processes A second set of processes essential for cognitive control are those involved in the evaluation of performance and include functions such as detection of processing conflicts and performance monitoring. These evaluative processes are believed to play a crucial role in signaling for adjustments of top-down control needed to adapt to constantly changing task demands (Kerns et al., 2004). For example, in the card version of the Stroop task described above, participants show greater interference on the initial one or two trials in each block than on subsequent trials in a series (Botvinick et al., 2001). Additionally, participants completing a modified Stroop task show less interference on incongruent trials if the incongruent trials are frequent relative to congruent trials than if they are rare (Lindsay & Jacoby, 1994). The results of these studies indicate detection of conflict and subsequent adjustment in control processes to more efficiently perform the task as well as increased conflict when there is decreased expectancy of a conflict stimulus. Thus, cognitive control is a dynamic process that is most reliably invoked during tasks involving conflicts in information processing where modification of performance is required to successfully complete the task (Botvinick et al., 2001; Norman & Shallice, 1986). Neural Correlates of Stroop Performance As noted in the examples above, one experimental paradigm that consistently evokes response conflict conditions where cognitive control is necessary is the Stroop task. The Stroop task has been utilized by numerous cognitive scientists to examine a

PAGE 14

4 myriad of cognitive functions, including: automatic and controlled cognitive processes (Posner & Snyder, 1975), selective attention (Rebai, Bernard, & Lannou, 1997), and disturbances in these processes due to psychiatric disorders (Cohen et al., 1999; Perlstein, Carter, Barch, & Baird, 1998). In addition, recent technological advances in hemodynamicand electrophysiologically-based neuroimaging methods have led to increased insight into the neural correlates of the Stroop task, as well as the functional neural bases of cognitive control more generally. Traditionally, the Stroop task has been employed as an instrument used to measure pre-potent response inhibition (Miyake, Friedman, Emerson, Witzki, & Howerter, 2000), a function often attributed to the frontal lobes (Kiefer, Marzinzik, Weisbrod, Scherg, & Spitzer, 1998). Hemodynamic-based neuroimaging research has described many frontal sites as critical to the ability to overcome pre-potent tendencies, including: left inferior lateral cortex (Taylor, Kornblum, Lauber, Minoshima, & Koeppe, 1997), left superomedial cortex (Pardo, Pardo, Janer, & Raichle, 1990), right frontal polar cortex (Bench et al., 1993), and bilateral anterior cingulate cortex (Bench et al., 1993). More recently, the Stroop task has been used to examine not only the neural correlates of the ability to overcome pre-potent response tendencies, but also the neural correlates of the regulative and evaluative components of cognitive control. Dissociation of Cognitive Control Processes To distinguish among component processes of cognitive control, a modified version of the Stroop paradigm (Figure 1-1) has been introduced that allows one to dissociate the regulative and evaluative processes required to successfully complete the Stroop task (Cohen et al., 1999). In this modified Stroop task, participants are given an instruction before each trial indicating whether to read the word (a more automatic

PAGE 15

5 response) or name the color (a condition requiring an increased amount of control due to the need to override the more automatic tendency to read the word). Following a brief delay, the Stroop color-word stimulus is presented and the participant responds. Thus, the task temporally separates the instruction-related regulative processes (representing the context/goal of the task in the CN or WR trials) from the response-related evaluative processes (the detection of incongruencies and errors and signaling for adjustments in control processes). This modified version of the Stroop task can be contrasted with the traditional Golden Stroop task, where participants respond in blocks to trials of the same type, thus temporally confounding regulative and evaluative processes. Figure 1-1. Schema of the single-trial Stroop task. Participants are first presented with an instructional cue (Color or Word), followed by a delay (1 or 5 seconds) and Stroop stimulus (congruent, neutral, incongruent). Using the modified Stroop paradigm described above and event-related functional magnetic resonance imaging (fMRI) techniques, MacDonald et al. (2000) demonstrated a double dissociation of the roles of the dorsolateral pre-frontal cortex (dlPFC) and anterior cingulate cortex (ACC) in the regulative and evaluative component processes of cognitive control. Specifically, they found that during maintenance of the task context

PAGE 16

6 (CN, WR), the left dlPFC was more active following instructions to perform the CN task than the WR task, consistent with a role of the dlPFC in preparation to execute the more demanding color-naming task. In contrast, ACC activity was increased upon presentation of incongruent color-word stimuli as compared to congruent color-word stimuli, consistent with a role in detection of response conflict. Based on these data, the authors suggested that the network necessary for cognitive control is dissociable into either the regulative or evaluative processes necessary for completion of separate task aspects. This study, however, did not allow for the specific examination of conflict processing and error detecting aspects of the evaluative component of cognitive control as the stimulus-related activity did not differentiate error trials from correct trials. Thus, the evaluative processes of conflict processing and error detection are potentially confounded. Despite the limitation, this study is of great importance to the cognitive control literature as it demonstrates a functional dissociation of not only the behavioral aspects of cognitive control, but also the neuroanatomical bases of cognitive control; elucidating the dlPFC mediated regulative control processes and the ACC mediated evaluative processes. This study also emphasizes the reality that conventional behavioral methodologies do not permit cognitive psychological processes and representations to be assessed directly. Furthermore, different neuroimaging methods have different strengths in the examination of cognitive and neural processes. For example, fMRI results represent blood flow-based hemodynamic response mechanisms to specific stimuli that are not particularly sensitive to the temporal course of neural activity. In the MacDonald study, the hemodynamic-response signal was examined over five 2.5-second increments per trial, while verbal responses to Stroop stimuli tend to occur in less than one second. Thus, results of this fMRI study reflect a spatially sensitive common neural output, rather than a direct

PAGE 17

7 reflection of neural processes. In contrast to hemodynamic-based measures of brain activity, event-related potentials (ERPs) provide direct measurement of neuronal electrical activity with sub-millisecond resolution. Moreover, under some circumstances, ERPs may be able to detect neural activity that is unaccompanied by secondary phenomena such as changes in regional blood flow or local metabolic activity (Gaetz & Bernstein, 2001). However, ERPs have limited spatial resolution, potentially leading to ambiguous spatial localization. Therefore, a convergence of information based on multiple neuroimaging methods, including scalp-recorded brain ERPs, can provide additional clarity into the nature of the cognitive and neuronal processes of cognitive control. Event-related Potentials In order to have a thorough understanding of the contributions of ERPs to theories of cognitive control, one must understand the assumptions behind ERP neuroimaging methods. Briefly, two major assumptions of ERP neuroimaging research are, first, that the distribution of electrical activity across the scalp indicates the activities of underlying neural structures, and, second, that this neural activity corresponds with specific cognitive states and processes. To the extent these assumptions are valid, 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. This overall background, or ongoing electrical activity, however, is not the interest of the present study; rather, the time(event-) locked averaged electrical activity associated with the presentation of specific events is of

PAGE 18

8 interest. Initially, the event-related signal associated with the presentation of a stimulus is embedded in the noise of the background EEG activity. Extracting the signal associated with a specific cognitive activity from the noise (background activity and measurement error) is accomplished by averaging multiple samples of the EEG that are time-locked to repeated occurrences of the event (i.e., stimulus or response) of interest. The logic of averaging is that the signal does not change from trial to trial, while the noise is random, thus, the signal is enhanced by a factor proportional to the square-root of the number of trials, while the noise is reduced essentially to zero (Fabiani, Gratton, & Coles, 2000). Due to the direct measure of electrical brain activity associated with specific cognitive events, ERPs are currently considered the gold standard in terms of temporal resolution among noninvasive imaging methods (Fabiani et al., 2000). ERP waveforms typically consist of a series of discrete deflections (i.e., peaks and troughs), often followed by so-called slow-wave potentials 1 Characteristics of ERP waveforms usually include descriptors of polarity (positive or negative) and latency (in milliseconds). For example, P300 refers to an ERP with a positive peak that has an approximate latency of 300 milliseconds. Another similar labeling system involves a descriptor of polarity followed by a number representing the ordinal latency of the component. Using these labeling criteria, P3 refers to the third positive peak in the ERP waveform. Other descriptors, such as scalp location at which the component is maximal (e.g., frontal P3) are also used. Cortically, the activity of neurons associated with ERP activity is attributed primarily to post-synaptic potentials (Williamson & Kaufman, 1990). As an example, 1 A slow-wave potential is a temporally extended change in the ERP waveform, rather than a distinct or punctate deflection.

PAGE 19

9 consider the case of an excitatory post-synaptic potential (although similar activities occur for inhibitory post-synaptic potentials). While at rest, neurons contain a lesser concentration of sodium ions (Na + ) and a greater concentration of potassium ions (K + ) inside the cell. When the dendrites of a neuron receive an excitatory signal from an adjacent neuron, the resulting change in the cell membrane allows Na + to flow into the cell. This results in a reduction of positive ions in the extracellular spacemaking the space more negative. This negativity in the extracellular space is known as the current sink. The positive Na + ions that entered the cell repel like-charged ions, and create a current that sends the K + ions toward the cell body. This buildup of positive charge near the cell body is known as the current source. The current source repels like-charged ions in the extracellular space, which are then attracted back to the sink, producing a dipolar extracellular current. It is this extracellular current that produces the electrical potentials. ERPs, then, represent the net activities of a large population of neurons that must be synchronously active and configured in such a way they produce dipolar electromagnetic fields that can be measured at the scalp. Such a synchronously activated configuration of neurons is known as an open field, and usually involves the alignment of neurons in an orientation parallel to the scalp (Coles & Rugg, 1995) 2 ERPs and Cognitive Control Electrophysiological Correlates of Evaluative Processes In cognitive control research, ERPs can serve an important purpose because they are particularly sensitive to the temporal course of neural activity and, by extension, the concomitant underlying sensory, motor, and cognitive processes. Of particular interest to 2 Information on the physiological bases of ERPs taken from Coles & Rugg, 1995; Fabiani et al., 2000; Simmons, 1998. Please see these sources for additional information.

PAGE 20

10 the current study is the idea that ERPs can be used to temporally dissociate the component processes of cognitive control by facilitating inferences regarding the timing, level of processing, and, roughly, the anatomical location of neural mechanisms supporting these processes. The preponderance of research on component processes of cognitive control using ERPs has focused on the evaluative components of conflict detection and error monitoring. For example, using tasks that reliably invoke processing conflicts between simultaneously active task-relevant representations (e.g., Stroop or Eriksen flanker tasks 3 ), investigators have found a reliably evoked late fronto-central ERP signature referred to as the N450 or N2 component (van Veen & Carter, 2002a, 2002b; West, 2003). These components both reflect conflict detection processes and differ only based on the stimulus paradigm presented (e.g., Stroop vs. Eriksen flanker tasks). Conflict detection in the Stroop paradigm has been associated with the negative deflection between 400ms and 500ms known as the N450 (Liotti, Woldorff, Perez, & Mayberg, 2000; West & Alain, 1999), while conflict associated with the incongruent condition of the Eriksen flanker has been associated with a slight negative deflection in the ERP waveform known as the N2 (van Veen & Carter, 2002b). These ERP components are largest under conditions in which response conflict is high, such as the incongruent condition of the Stroop color-naming task (Grapperon, Vidal, & Leni, 1988; Liotti et al., 2000; Rebai et al., 1997). Increased amplitude of the N450 has been observed following unexpected or rare incongruent trial presentation (West & Alain, 2000b), consistent with the hemodynamic-based neuroimaging results presented 3 The Eriksen Flanker task (Eriksen & Eriksen, 1974) consists of a target central stimulus flanked by either congruent or incongruent stimuli. For example, participants may be instructed to press a button with their left hand if a central target arrow points left (e.g., <) or a button with their right hand if a central target arrow points right (e.g., >). For congruent trials the flankers are the same as the specified target (e.g., <<<<<), while incongruent trials have flankers that are opposite the target (e.g., <<><<).

PAGE 21

11 previously of MacDonald et al. (2000) implicating the role of the ACC in conflict detection. In support of this hypothesis, source localization algorithms have also localized regions of the ACC as the neural generators of the N450 and N2 components (van Veen & Carter, 2002a, 2002b; West, 2003). These results implicate the N450 and N2 components as neurobiological indices of conflict detection, and support the role of ACC as a conflict detection mechanism. Detection of errors is another component of the evaluative processes in cognitive control that, due to the exquisite temporal sensitivity of electrophysiological-based neuroimaging methods, has been widely investigated using ERPs. The detection of errors has been associated with a midline fronto-central negative deflection in response-locked ERPs that occurs within 100ms of committing an error. This negative deflection is known as the error-related negativity (ERN), and is the first identified neurobiological index of performance monitoring (Dahaene, Posner, & Tucker, 1994; Falkenstein, Hoormann, Christ, & Hohnsbein, 2000; Gehring, Goss, Coles, Meyer, & Donchin, 1993; Luu, Flaisch, & Tucker, 2000). The precise function responsible for the ERN continues to be debated (Cohen et al., 2000b; Gehring & Knight, 2000); however, the ERN has typically been referred to as a representation of an error/conflict monitoring system that operates across various stimulus and response modalities. Specifically, the ERN has been found with visual, auditory, and movement responses (Bernstein, Scheffers, & Coles, 1995; Holroyd, Dien, & Coles, 1998), is present following errors of omission as well as errors of commission (Falkenstein et al., 2000), and is greatest in amplitude when participants are aware that an error has been made (Luu, Collins, & Tucker, 2000; Scheffers & Coles, 2000)--consistent with the action of an error detector/performance monitor in cognitive control. Furthermore, the ERN is observed when participants make

PAGE 22

12 partial errors (begin to make an error but spontaneously correct themselves), and is greater in amplitude on the trials with partial errors than errors where the conflict is not produced in time to spontaneously correct the response (Gehring et al., 1993). These results implicate the ERN as an on-line index of error detection or performance monitoring. Such an index is critical because an important function of the human brain is to monitor behavior and prevent undesirable actions. Like the N450 and N2 components, dipole-modeling techniques have localized the neural activity associated with the ERN to the ACC (Holroyd et al., 1998; van Veen & Carter, 2002b). In fact, van Veen and Carter (2002a) reviewed multiple studies of the N2 and ERN components and concluded that both components are representations of similar conflict detection processes. Specifically, they concluded that the N2 reflects pre-response detection of conflict between competing response tendencies, while the ERN reflects post-response detection of incompatible responses following error trials. The similarities between the N2/N450 and the ERN provide additional support for a model of cognitive control with ACC mediated evaluative processes detecting incongruities and possibly signaling for changes in the top-down control of cognition. Electrophysiological Correlates of Regulative Processes Regulation of cognitive control has also been examined using ERPs. The allocation of attentional resources under challenging task conditions and the active maintenance of goal/task-representations have been shown to be reflected in differential ERP slow-wave activity between tasks requiring differing levels of attentional demands (Curtin & Fairchild, 2003). Under the appropriate task conditions, investigators have observed a slow-wave in the ERP that appears to be associated with the implementation of cognitive control, perhaps reflecting an active biasing of processing in favor of the

PAGE 23

13 more attentionally-demanding aspect of the task (Curtin & Fairchild, 2003; West, 2003). This slow-wave has also been used to distinguished between subsequent correct and incorrect responses on a version of the cued-Stroop task, with greater slow-wave activity preceding correct than incorrect responses (West, 2003). The slow-wave activity that differentiates trials where there is a dysregulation of control processes (incorrect trials) and trials where control is adequately implemented (correct trials) also supports the contention that slow-wave activity reflects the implementation of control processes. Dissociation of Cognitive Control Component Processes using ERPs Cognitive control component processes have also been dissociated temporally using ERPs. Using a variation of the modified Stroop described above (Cohen et al., 1999; MacDonald, Cohen, Stenger, & Carter, 2000), West (2003) reported a temporal dissociation between the regulative and evaluative component processes of cognitive control. Following presentation of task instruction, implementation of regulative processes was exhibited by an occipital-parietal slow-wave that differentiated correct (goal-compatible) and incorrect (goal incompatible) response trials. In addition, implementation of control was associated with frontal slow-wave ERP modulation that differentiated CN trials, more attentionally demanding trials where participants prepare to override the more automatic WR response, from less demanding WR trials. These findings are consistent with the allocation of increased attentional resources during tasks requiring increased cognitive processes, as well as increased implementation of control on correct trials. Conflict detection was associated with a fronto-central N450 with greater amplitude for incongruent than congruent trials. Consistent with a role in conflict monitoring, the N450 amplitude was reflected equally for incongruent trials in both CN and WR conditions. In addition, West reported a frontal slow-wave component 500ms to

PAGE 24

14 600ms following incongruent stimulus presentation. This component, known as the conflict slow-wave potential (conflict SP), was interpreted to reflect the allocation of increased attentional resources in preparation for future incongruent trials during the incongruent condition of both CN and WR trials. These findings taken together indicate that the different neural processes reflecting the regulative and evaluative of cognitive control can be dissociated in time using ERPs. Predictions Given the sensitivity of ERPs in examining the temporal course of neural activity, the goal of the present study was to use the modified Stroop task developed by Cohen and colleagues to give added support to the temporal dissociation of the regulative and evaluative components of cognitive control using high-density event-related potentials. Based on previous research the following predictions are offered: Behavioral Data It is predicted that standard Stroop effects for both RT and error-rates will be manifest by increased RTs and error rates to incongruent CN and WR conditions. In addition, it is predicted that interference effects will be disproportionately greater in the CN as compared to the WR task. It is also predicted that facilitation effects (faster RTs to congruent than neutral conditions) for both the CN and WR tasks will be manifest. For delay conditions, it is hypothesized that greater time between instruction and stimulus presentation will allow for increased implementation of regulative processes; therefore, fewer errors and faster RTs are predicted for trials following the 5s delay, as compared to trials following the 1s delay.

PAGE 25

15 ERP Data Since the behavioral data alone cannot address the potential mechanism(s) underlying task performance and allocation/implementation of cognitive control resources, ERPs are necessary to further examine the electrophysiological signatures of the following possibilities (Figure 1-2) 4 : Regulative Processes Regulative processes associated with the preparation to override more automatic response tendencies and the maintenance of task-relevant representations are predicted to be manifest in frontal slow-wave activity locked to the task instructional cue that will differentiate preparation to engage in the more attentionally demanding CN task as compared to the more automatic WR task. In addition, it is hypothesized that lapses in regulation of cognitive control processes leads to increased errors. This will be examined in instruction-locked slow-wave ERP activity collapsed across conditions for error and correct trials. It is predicted that task instruction-related slow-wave activity will differentiate subsequent correct and incorrect trials. Evaluative Processes Evaluative processes associated with conflict detection are predicted to be reflected in an N450 ERP deflection that has increased amplitude to stimulus-locked incongruent vs. congruent CN stimuli. Evaluative processes associated with error processing will be reflected in response-locked ERPs that exhibit an increased ERN following incorrect trials that is less pronounced during correct trials--indicative of the detection of incompatibilities (conflicts) between the response given and the accurate response. 4 Examination of the effect of delay on stimulus-locked ERP reflections of task (CN, WR) effects is currently underway. These analyses, however, are incomplete and will not be discussed. Additionally, delay x congruency effects for each task will not be examined since the number of trials per condition is low (< 28, see Methods section below) resulting in inadequate signal-to-noise ratio for examining these effects.

PAGE 26

16 Figure 1-2. Schemata of cognitive control functions and subsequent ERP manifestations predicted in the current study.

PAGE 27

CHAPTER 2 METHOD Participants Twenty-four right-handed individuals (14 female) between the ages of 18 and 24 participated in the study in exchange for course credit. All participants provided written informed consent according to procedures established by the University of Florida Health Science Center Institutional Review Board. Participants were screened for depressive or anxious symptoms that may influence results using the Beck Depression Inventory (Beck & Steer, 1987; Beck, Steer, & Garbing, 1988) and Beck Anxiety Inventory (Beck & Steer, 1990), respectively. Descriptive information regarding participants age, education, and depression and anxiety levels is reported in Table 1. Participants were excluded if they reported previous neurological insults, traumatic brain injury (TBI), psychiatric diagnosis, current psychotropic medication use, or color-blindness (color-blindness was assessed using the Ishihara pseudo-isochromatic color plates, Clark, 1924). All participants reported normal or corrected-to-normal vision. Table 2-1. Demographic information Mean Standard Deviation Range Age (years) 19 1.4 18 24 Education (years) 13.4 1.0 12 16 Beck Depression Inventory score 5.5 5.2 0 25 Beck Anxiety Inventory score 5.2 4.8 0 19 Materials and Procedure Participants performed a computerized single-trial, cued version of the Stroop task (see Figure 1-1), originally developed by Cohen et al. (1999). In this task, each trial began with the computer presentation of an auditory instructional cue (the word color 17

PAGE 28

18 or word), followed, after a short delay, by a visual stimulus, which remained on the screen until participants response. Stimuli comprised the same three colors and color-words (red, green, and blue) used in the card Stroop (Golden, 1978), and commonly employed in clinical neuropsychological settings. Participants were instructed to respond verbally to the stimulus as quickly and accurately as possible. RTs were determined by a voice-activated relay connected to the computer, and the examiner manually coded response accuracy. Participants performed the color-naming and word-reading tasks, each comprising three congruency conditions. Congruent stimuli were words printed in the same color (e.g., RED printed in red), incongruent stimuli were words printed in a different color (e.g., RED printed in blue), and neutral stimuli were animal names printed in red, green, or blue (e.g., BEAR printed in red) for the color-naming trials and words displayed in white for word-reading trials. The context provided by the task instruction (e.g., color) must be used to override the influence of the stronger dimension (i.e., word) when the task is to respond to the less prepotent or automatic dimension (i.e., color). Additionally, reliance on context was increased by (a) varying the task to be performed on each trial and (b) introducing a delay between the task instruction for each trial and the stimulus to be responded to. For the current task, we used two delays (stimulus onset asynchrony, SOA, of 1s and 5s). Visual stimuli were presented in the center of a visual display, and delivered using an Apple Macintosh computer using PsyScope software (Cohen, MacWhinney, Flatt, & Provost, 1993). A total of 336 experimental trials were distributed equally across task and congruency, resulting in 28 trials of each type. Prior to acquisition of electrophysiological data, participants completed a practice block consisting of random presentation of 12 trials, one of each stimulus type. On these

PAGE 29

19 trials, if the participants RT was over 1000ms, an auditory beep was presented, indicating to the participant a need to respond more quickly. EEG Acquisition and Reduction EEG 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 250Hz with a 16-bit analog-to-digital converter. A right posterior electrode served as common ground. The impedance of all electrodes was kept below 50 k, consistent with procedures suggested by the manufacturer. EEG Data Reduction Due to the volume conducting nature of the brain, no one site on the head can be considered an inactive reference site (Tucker, Liotti, Potts, Russell, & Posner, 1994); therefore, data was mathematically rereferenced against an average reference (Bertrand, Perrin, & Pernier, 1985). 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; (Scherg, 1990).

PAGE 30

20 2 4 5 55 Figure 2-1. Sensor layout of 64-channel geodesic sensor net. Slow-wave activity was quantified at scalp sites #2 and #4 (blue); N450 activity at scalp site #55 (red); and, ERN activity at scalp site #5 (green). See ERP Data Analyses in text for details. Individual-subject ERP averages were divided into three categories and included a pre-stimulus baseline period: auditory task instruction-related activity, visual stimulus-presentation activity, and response-related activity (see Figure 1-2). Instruction-locked epochs, associated with the implementation of control, were extracted from 100ms prior to instruction presentation to 1000ms post instruction presentation. Instruction-locked epochs were calculated for correct trials of both the presentation of the color-naming and word-reading tasks. Similarly, instruction-locked epochs were extracted separately for correct and incorrect responses from 100ms prior to instruction presentation to 1000ms post instruction presentation, collapsed across color-naming and word-reading tasks.

PAGE 31

21 Stimulus-locked epochs, associated with the detection of conflict, were extracted with a duration of 100ms prior to stimulus presentation and 750ms post-stimulus presentation. Individual subject averages were calculated for the correct-trial congruent, neutral, and incongruent stimulus-locked conditions. Response-locked averages were created separately for correct and incorrect responses, collapsed across color-naming and word-reading tasks, as well as congruencies. Collapsing across conditions was necessary because of insufficient numbers of incorrect responses to conduct specific error x congruency analyses. In addition, two-participants did not make any errors, their data, therefore, were not included in the analyses of response-related activity. Response-locked activity was extracted with a duration of 400ms pre-stimulus presentation and 400ms post-stimulus presentation. All averaged ERP epochs were baseline corrected using a 100ms window prior to stimulus or response onset and digitally filtered using a 30 Hz low-pass filter and a .5 Hz high-pass filter. Statistical Analyses Behavioral Data Analyses For analysis of behavioral data, correct-trial RTs and overall error rates were analyzed separately. For each trial type and participant, we calculated the median RT for correct responses as well as proportion of errors by subjecting them to 2-Task x 2-Delay x 3-Congruency repeated-measures analysis of variance analyses (ANOVA). Tests of simple effects were used to decompose interaction effects. For error rates, due to the high probability of no errors in several conditions (e.g., congruent word-reading condition), raw data were normalized prior to analysis using the arcsine transformation (Neter, Wasserman, & Kutner, 1985).

PAGE 32

22 ERP Data Analyses Analysis of ERP waveforms focused on instruction-related, stimulus-related, and response-related activity as indicated in Figure 1-2. Statistical analyses of ERP waveforms were performed on mean voltages over specified temporal windows extracted from individual electrode sites. Scoring windows and electrode positions for each condition of interest were determined by examination of grand-average ERP waveforms and spline-interpolated scalp voltage distribution plots (Perrin, Pernier, Bertrand, & Echallier, 1989). Instruction-related activity was quantified at electrode site #2 (see Figure 2-1) and was examined over the period of 800ms to 1000ms post-stimulus presentation. Paired t-tests were conducted between the CN and WR instruction-related slow-wave activity. Additionally, instruction-related activity for subsequent correct and incorrect trials, collapsed across conditions, was quantified at electrode site #4 (see Figure 2-1) and was examined for the period of 700 to 1,000ms post-instruction presentation. Stimulus-related activity was examined over the period of 440ms to 500ms post-stimulus presentation, and was quantified at electrode site #55 (see Figure 2-1). Paired t-tests examined conflict detection through comparison of congruent and incongruent CN stimulus presentation. Response-related activity was examined over the period of 32ms to 72ms post-response, and quantified at electrode site #5 (see Figure 2-1). Paired t-tests were used to compare mean amplitudes of correct and incorrect responses. Previous research has shown the ERN and the N450 to be phasic components of the ERP waveform (i.e., brief deflections, Rebai et al., 1997; West & Alain, 1999), and durations for examination were chosen to reflect their short-term nature. Slow-wave activity is tonic in nature, therefore, a broader window for averaging was chosen.

PAGE 33

CHAPTER 3 RESULTS Behavioral Analyses Initial analyses examined the possibility of a speed/accuracy trade-off by correlating RT and error-rates. Analyses revealed that RT and error-rate data were negatively correlated (r = -.30); however, the correlation was not significant (p>.10), indicating a slight, but non-significant speed/accuracy trade-off. Error Rate Analyses Mean error rates and standard errors as a function of task condition, congruency, and delay are provided in Table 3-1. Analyses revealed the standard Stroop effects: A main effect of task, F(1,23 ) = 10.51, p<.01, with WR more accurate than CN; a main effect of congruency, F(2,46) = 53.2, p<.01, with significantly more errors in the incongruent condition than congruent condition, F(1,23) = 59.53, p<.01, and the neutral condition, F(1,23) = 55.51, p<.01. There was also a Task x Congruency interaction, F(2,46) = 7.02, p<.01, with increased errors in the incongruent condition of the CN task compared to the congruent condition, F(1,23) = 10.86, p<.01, and the neutral condition, F(1,23) = 6.40, p<.01, of the CN and WR tasks (i.e., error rate interference), but no significant difference in error facilitation (neutral condition errors congruent condition errors), F(1,23) = .46, p>.40 (Figure 3-1). Of note is the finding that there was not a significant main effect of delay, F(1,23) = 1.01, p>.50, and delay did not significantly interact with any task, F(2,46) = 2.55, p>.10, or congruency conditions, F(2,46) = .62, p>.50. 23

PAGE 34

24 Table 3-1. Means and standard errors of error rates (%) in the single-trial Stroop M SE Color-naming Short Delay Congruent .31 .20 Neutral 2.01 .70 Incongruent 10.20 2.70 Long Delay Congruent 1.39 .40 Neutral 1.54 .40 Incongruent 11.90 3.10 Word-reading Short Delay Congruent .31 .20 Neutral 1.08 .40 Incongruent 4.32 .90 Long Delay Congruent .62 .30 Neutral .93 .40 Incongruent 3.40 .70 Figure 3-1. Error rates by task and congruency conditions collapsed across delay. Error bars represent standard errors. RT Analyses Means and standard errors of median RTs as a function of task condition, congruency, and delay are provided in Table 3-2. As with the accuracy data, analyses of

PAGE 35

25 RT data revealed the standard Stroop effects: A main effect of task, F(1,23) = 47.31, p<.01, with WR faster than CN; a main effect of congruency, F(2,46) = 53.2, p<.01, with significantly slower RTs in the incongruent (i.e., interference) condition as compared to the neutral condition, F(1,23) = 51.20, p<.01, and congruent condition, F(1,23) = 64.71, p<.01, and a facilitation effect as evidenced by significantly faster RTs to congruent conditions compared to neutral conditions, F(1,23) = 6.68, p<.05. There was a significant Task x Congruency interaction, F(2,46) = 12.6, p<.01, with stronger interference effects in the CN task compared to the WR task, F(1,23) = 17.99, p<.01 (see Figure 3-2). Similar to the error rate analyses, there was not a significant main effect of delay, F(1,23) = .13, p>.70, and delay did not significantly interact with task, F(1,23) = .16, p>.60, or congruency conditions F(2,46) = 1.25, p>.25. Table 3-2. Means and standard errors of median RT (ms) in the single-trial Stroop M SE Color-naming Short Delay Congruent 735.88 25.73 Neutral 799.67 23.19 Incongruent 895.65 32.79 Long Delay Congruent 743.65 23.59 Neutral 777.46 22.71 Incongruent 896.58 29.67 Word-reading Short Delay Congruent 724.50 25.72 Neutral 721.90 23.88 Incongruent 800.73 33.07 Long Delay Congruent 723.42 27.58 Neutral 717.83 27.42 Incongruent 805.65 34.96

PAGE 36

26 Figure 3-2. Median RTs by task and congruency conditions collapsed across delay. Error bars represent standard errors. ERP Analyses Instruction-related Activity Spline-interpolated scalp voltage maps for instruction-related ERP activity showed a right inferior frontal difference between the CN and WR instruction presentation (Figure 3-3). ERPs to the task instructional cues (CN or WR) showed differential slow-wave activity quantified at scalp electrode site #2 following instructions to engage in the CN and WR tasks t(23) = 1.73, p<.05 (Figure 3-4). These findings are consistent with the deployment of regulative processes associated with the implementation of cognitive control to the more attentionally-demanding CN task (e.g., implementation of control reflected in preparation to override the prepotent WR response tendency). Additionally, inspection of spline-interpolated scalp voltage maps revealed a medial frontal difference between instruction-related activity of subsequent correct and incorrect responses (Figure 3-5). ERP slow-wave activity quantified at scalp site #4 revealed slow-wave activity that differentiated subsequent execution of correct and incorrect responses collapsed across conditions, t(23) = 6.3, p<.01 (Figure 3-6).

PAGE 37

27 Stimulus-related Activity The behavioral data (RTs, error rates) did not differ as a function of delay; therefore, stimulus-locked ERPs were collapsed across delay conditions. In addition, increased conflict was reflected in the significantly longer RTs to the incongruent CN condition as compared to the incongruent WR condition; therefore, ERP analyses focused on the congruent and incongruent CN conditions, collapsed across delay. Consistent with previous research (Liotti et al., 2000; West, 2003; West & Alain, 1999), spline-interpolated scalp voltage maps for stimulus-related ERP activity showed a mid to right-lateral frontal difference between the congruent and incongruent conditions of the CN task (Figure 3-7). Stimulus-locked ERPs showed significantly greater N450 deflection to the incongruent than congruent stimuli of the CN task quantified at scalp site #55 approximately 450ms post-stimulus presentation, t(23) = 3.28, p<.01 (Figure 3-8). Similar findings have been interpreted as the neural representation of conflict detection in the incongruent condition as compared to decreased conflict in the congruent condition (Liotti et al., 2000; West & Alain, 2000b). Response-related Activity Spline-interpolated scalp voltage maps for response-related ERP activity revealed a medial-frontal difference between the correct and incorrect responses, collapsed across task and congruency due to insufficient error trials (Figure 3-9). Response-locked ERPs to correct and incorrect responses exhibited a significant negative deflection (ERN) quantified at scalp site #5 occurring approximately 50ms following incorrect responses that was not present following correct responses, t(21) = 2.11, p<.01 (Figure 3-10). These results are consistent with previous research and have been previously interpreted

PAGE 38

28 to represent the detection of conflict in task performance (Falkenstein et al., 2000; Gehring et al., 1993; van Veen & Carter, 2002a). Figure 3-3. Top view of spherical-spline interpolated voltage maps representing instruction-related activity of the CN task, WR task, and the difference between the two tasks (CN WR, bottom). ERP waveforms of the grand average differential slow-wave activity between the CN and WR tasks at scalp site #2 (top).

PAGE 39

29 Figure 3-4. Grand average instruction-locked ERPs for all scalp sites of the CN (red) and WR (blue) task presentation.

PAGE 40

30 Figure 3-5. Top view of spherical-spline interpolated voltage maps representing instruction-related activity of subsequent correct and incorrect trials, and the difference (correct incorrect, bottom). ERP waveforms of the grand average differential slow-wave activity between subsequent correct and incorrect trials collapsed across conditions at scalp site #4 (top).

PAGE 41

31 Figure 3-6. Grand average instruction-locked ERPs for all scalp sites of correct (red) and incorrect (blue) task presentation collapsed across conditions.

PAGE 42

32 Figure 3-7. Top view of spherical-spline interpolated voltage maps representing stimulus-related activity of the congruent CN condition, the incongruent CN condition, and the difference between the conditions (congruent incongruent, bottom). ERP waveforms reflect the grand average N450 to the incongruent CN trials at scalp site #55 (top).

PAGE 43

33 Figure 3-8. Grand average stimulus-locked ERPs for all scalp sites of the congruent (red) and incongruent (blue) CN task.

PAGE 44

34 Figure 3-9. Top view of spherical-spline interpolated voltage maps representing response-related activity of correct and incorrect trials (collapsed across conditions), and the difference (correct incorrect, bottom). ERP waveforms of the grand average ERN deflection to the incorrect trials as compared to the correct trials at scalp site #5 (top).

PAGE 45

35 Figure 3-10. Grand average response-locked ERPs for all scalp sites of correct and incorrect responses collapsed across conditions.

PAGE 46

CHAPTER 4 DISCUSSION Current theories of cognitive control recognize two features as essential to negotiating everyday cognitive tasks: an evaluative component responsible for monitoring the need for internal adjustments in control and signaling when such adjustments are necessary and a regulative component responsible for activation and implementation of control processes. The current study utilized ERPs and a modified version of the Stroop task to temporally dissociate the electrophysiological signatures of the regulative and evaluative processes of cognitive control. Behavioral Data Participants displayed the typical Stroop interference effects, with increased error rates and RTs on incongruent CN trials. These results reflect the increased influence of word-reading over color-naming on the incongruent CN trials and the need to override the more automatic tendency to read the word. Delay did not significantly affect RT or error-rate performance in any task or congruency conditions. The absence of delay effects may be attributed to the demographics of the sample used in this study. The sample consisted of educated college students who had intact working memory skills and had little difficulty maintaining the representation of the CN or WR tasks; therefore, performance on the long and short delay conditions was nearly equivalent. These results can be contrasted with those of Seignourel et al. (in preparation) who found in healthy, slightly older, control participants decreased error-rates following the 5s delay condition, but no differences in RTs. Cohen et al. (1999) found increased facilitation of RTs (faster 36

PAGE 47

37 RTs to congruent than neutral stimuli) in the 5s delay condition as compared to the 1s delay, also in slightly older controls. The RT facilitation and decreased error rates in the 5s condition were attributed to the increased time allowed to engage regulative control mechanisms and prepare to override prepotent response tendencies. Interestingly, Seignourel et al. found that moderate-to-severely traumatic brain injured participants did not display differential error rate performances by delay and Cohen et al. found no within, or between subjects error-rate differences in patients diagnosed with schizophrenia and normal controls. Based on this information, current theories hypothesizing that increased delay between instruction-cue and stimulus presentation facilitates the implementation of regulative control processes remain ambiguous. Nonetheless, computational modeling studies do suggest that, in healthy participants using a different cognitive control task (i.e., AX-CPT, (Braver, Barch, & Cohen, 1999), the biasing of control requires some time to achieve full strength. Furthermore, functional neuroimaging studies have shown that this delay-related effect is mediated, at least in part, by the dlPFC (Barch et al., 2001). More studies are necessary to determine the effects of delay between instruction-cue and stimulus presentation on the implementation of regulative processes in cognitive control. ERP Data Regulative Processes As predicted, regulative processes reflecting the implementation of control were shown in frontal slow-wave activity, more specifically, instruction-related slow-wave activity that differentiated the CN from the WR tasks. These results are consistent with an increased requirement for top-down control and increased allocation of attentional resources to the CN task in preparation to override the more automatic tendency to read

PAGE 48

38 the word. These results are similar to the fMRI results of MacDonald et al. (2000) who attributed increased left dlPFC activity to the CN task as compared to the WR task to the implementation of increased cognitive control processes in preparation for the more demanding CN task. Reasons for the differences in lateralization are currently unclear; however, due to the volume-conducting nature of the brain, the lack of spatial sensitivity associated with ERPs, and the failure of the current investigation to provide consistent source localization results using more advanced dipole modeling techniques, speculation about specific anatomical locations of control processes is considered tentative. The results and interpretations of the current study and those of MacDonald et al. can be contrasted with those of West (2003). Using a similar modified version of the Stroop task as that used by the current study and MacDonald et al., West also found instruction-locked slow-wave activity that differentiated the CN and WR tasks. The slow-wave activity found by West, however, was reflected primarily in a slow negativity over the occipital-parietal regions and positivity over the frontal-central region. West also described an instruction-locked slow-wave that differentiated correct and incorrect responses. This slow-wave activity reflected greater negativity for incorrect relative to correct responses and was specifically related to whether or not a correct response was made, rather than any particular aspects of the presented stimulus. Dipole-based source localization methods used by West determined that the best dipole model for the slow-wave activity occurred using mirrored dipoles in the occipital-parietal region, with a single dipole in the left dlPFC. The dipole in the left dlPFC contributed only moderately to the overall fit of the model. Based on these findings, West concluded that instruction cue-related slow-wave activity over the occipito-parietal region supported the processing

PAGE 49

39 of goal-compatible responses (correct responses) rather than preferential processing of a more attention demanding stimulus attribute. To examine the hypothesis that occipital-parietal slow-wave activity is associated with correct and incorrect responses, we collapsed CN and WR trials by accuracy. Our findings revealed a mid frontal-central slow-wave that differentiated subsequent correct and incorrect responses. However, in contrast to the findings of West, our slow-wave activity reflected greater negativity for correct relative to incorrect responses. Based on the assumption that errors are committed when there is a break down in the regulative processes of cognitive control, these findings could represent a possible dysregulation of control processes during error trials as compared to correct trials. Due to the fact, however, that trials were collapsed across conditions, no conjectures about specific stimulus attributes associated with the correct and incorrect trials are offered on the basis of the current study. Evaluative Processes Evaluative processes were reflected in a mid to right-lateral frontal N450 with greater negativity to the incongruent CN trials than congruent CN trials. These results are consistent with previous studies of the Stroop task (Liotti et al., 2000; West, 2003; West & Alain, 1999) that suggest the N450 is an electrophysiological reflection of conflict detection. In support of this interpretation, West (2003) showed increased negativity to the incongruent trials of both the CN and WR tasks, allowing him to conclude that the N450 is a neural representation of a cognitive process with a role in conflict detection independent of task condition. Previous dipole source localization research on the N450 has localized possible neural generation of the N450 to the ACC (West, 2003; West & Alain, 1999). These results are consistent with fMRI results found

PAGE 50

40 by MacDonald et al. and several PET studies (Carter, Mintun, & Cohen, 1995; Zysset, Muller, Lohmann, & von Cramon, 2001) indicating greater activity in the ACC during the incongruent, compared to congruent, CN trials. Evaluative processes were further explored in the current study by examining response-related ERP activity following correct and incorrect responses. The ERN, a significantly greater negativity following error trials compared to correct trials, has been suggested to reflect an error detection mechanism that plays a role in signaling for adjustments in performance following the occurrence of errors. Results of the current study replicate previous findings discussed previously of the ERN following error trials. Previous dipole source localization of the ERN has also implicated the ACC as the most likely neural generator (Falkenstein et al., 2000; Gehring et al., 1993; van Veen & Carter, 2002a, 2002b). Overall, convergence of evidence from N450 and ERN components of electrophysiological-based ERP methods, as well as results of hemodynamic-based neuroimaging studies, emphasizes an evaluative mechanism of cognitive control that is neurally distinct from the regulative components of cognitive control. Alternative Explanations and Potential Limitations While we propose that the observed ERP results reflect the regulative and evaluative components of cognitive control, other alternatives and potential limitations must be considered. First, the precise function responsible for the ERN has been subject to multiple interpretations (Cohen, Botvinick, & Carter, 2000a; Gehring & Knight, 2000). Research has not clarified whether the ERN is caused by the error itself, or by processing conflicts that produce uncertainty and predispose to error. Additionally, previous research has been incongruous on performance adjustments following error trials, with some research correlating post-error slowing with the ERN and some failing to document

PAGE 51

41 such a slowing effect (Hajcak, McDonald, & Simons, 2003); therefore, it remains unclear how a monitoring function such as the ERN relates to mechanisms of cognitive control. In addition, there is considerable evidence that attentive and motivational factors modulate ERN magnitude. For example, when participants are motivated to make accurate rather than speedy responses, the ERN is larger (Falkenstein et al., 2000). Similarly, larger ERNs occur when participants are certain they have made an error (Gehring et al., 1993). Continued research is required to further elucidate the precise cognitive focus reflected in the ERN. Second, the current study employed ERP methodologies to provide direct measurement of the temporal course of neurological reflections of cognitive control processes. Under some circumstances, ERPs may be able to detect neural activity that is unaccompanied by changes in regional blood flow or metabolic activity. However, ERPs have limited spatial resolution, potentially leading to ambiguous spatial localization. Nonetheless, relatively recent developments resulting in high-density spatial sampling of EEG (i.e., 64, 128, 256 channels), combined with sophisticated source localization algorithms, have been used to provide greater confidence in source localization (Pascual-Marqui, Michel, & Lehmann, 1994; Scherg, 1990). In the current study, source localization methods were not employed due to artifact introduced into the EEG recordings by concerns of contamination due to the vocal response mechanism. Vocalization-related cortical potentials (VRCPs) are one potential contaminant and consist of a movement-related cortical potential preceding vocalization and an auditory-evoked negative potential that immediately follows vocalization (Masaki, Tanaka, Takasawa, & Yamazaki, 2001). Additionally, following each trial response, electromyographic activity was observed throughout the EEG that could not be

PAGE 52

42 completely removed using current data-correction algorithms. Due to this variability associated with vocal response introduced into the EEG data, it was felt that source localization methods would not be reliable estimates of the potential neural generators of cognitive control processes. Nonetheless, our finding of N450 modulation over medial-frontal regions is consistent with several studies that employed manual responses (West, 2003; West & Alain, 1999, 2000b) and both manual and verbal responses (Liotti et al., 2000). Future Directions The present study lends continued credence to the use of a single-trial version of the Stroop task that specifically examines the component processes of cognitive control. Previous research using the Golden version of the Stroop task has not had the sensitivity of the single-trial version of the Stroop used in this study due to the temporal confound between the regulative and evaluative processes of cognitive control. This is of particular relevance in the study of functional impairments in groups who demonstrate putative cognitive control deficits on traditional measures (e.g., blocked Stroop task), as well as groups whose underlying deficits have been proposed to be related to impaired goal maintenance or conflict detection processes, for example, schizophrenia (Cohen et al., 1999; Perlstein, Dixit, Carter, Noll, & Cohen, 2003), aging (West, 1999; West & Alain, 2000a), and traumatic brain injury (Seignourel et al., in preparation) The current study lays the foundation for examining the electrophysiological representations of cognitive control in these groups. Future studies using ERPs and hemodynamic-based neuroimaging methods will examine the specific components of goal maintenance and conflict processing deficits in these populations where dysfunctions in broader cognitive control processes have been hypothesized to account for disease-specific impairments.

PAGE 53

43 Specifically, future studies are underway in aging, ADHD, and traumatic brain injury (TBI; Larson, Kelly, & Perlstein, 2003) that will further elucidate the specific nature of deficits and the underlying neural substrates of cognitive control dysfunction. Summary Current cognitive neuroscience theories distinguish between at least two important components of cognitive control: an evaluative component responsible for monitoring the need for control and signaling when adjustments in control are necessary, and a regulative component responsible for activation and implementation of control processes. The current study used ERPs in an effort to temporally dissociate these regulative and evaluative components of cognitive control. The findings of this research are consistent with the hypothesis that regulative and evaluative components of cognitive control are dissociable with ERP methods. This conclusion lays the foundation for future studies that will provide increased clarity into specific cognitive control deficits in clinical populations (e.g., TBI).

PAGE 54

LIST OF REFERENCES Barch, D. M., Carter, C. S., Braver, T. S., Sabb, F. W., MacDonald, A. W., Noll, D., et al. (2001). Selective deficits in prefrontal cortex function in medication-naive patients with schizophrenia. Archives of General Psychiatry, 58, 280-288. Beck, A. T., & Steer, R. A. (1987). The Beck Depression Inventory Manual. San Antonio, TX: Psychological Corporation. Beck, A. T., & Steer, R. A. (1990). The Beck Anxiety Inventory Manual. San Antonio, TX: The Psychological Corporation. Beck, A. T., Steer, R. A., & Garbing, M. G. (1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77-100. Bench, C. J., Frith, C. D., Grasby, P. M., Friston, K. J., Paulesu, E., & Frackowiack, R. J. S. (1993). Investigation of the functional neuroanatomy of attention using the Stroop task. Neuropsychologia, 31, 907-922. Bernstein, P. S., Scheffers, M. K., & Coles, M. G. H. (1995). "Where did I go wrong?" A psychophysiological analysis of error detection. Journal of Experimental Psychology: Human Perception and Performance, 21, 1312-1322. Bertrand, O., Perrin, F., & Pernier, J. (1985). A theoretical justification of the average-reference in topographic evoked potential studies. Electroencephalography and Clinical Neurophysiology, 62, 462-464. Botvinick, M. W., Carter, C. S., Braver, T. S., Barch, D. M., & Cohen, J. D. (2001). Conflict monitoring and cognitive control. Psychological Review, 108, 624-652. Braver, T. S., Barch, D. M., & Cohen, J. D. (1999). Cognition and control in schizophrenia: a computational model of dopamine and prefrontal function. Biological Psychiatry, 46, 312-328. Carter, C. S., Mintun, M., & Cohen, J. D. (1995). Interference and facilitation effects during selective attention: an H 2 15-O PET study of Stroop task performance. NeuroImage, 2, 264-272. Clark, J. H. (1924). The Ishihara test for color blindness. American Journal of Physiological Optics, 5, 269-276. 44

PAGE 55

45 Cohen, J. D., Barch, D. M., Carter, C. S., & Servan-Schreiber, D. (1999). Context-processing deficits in schizophrenia: converging evidence from three theoretically motivated cognitive tasks. Journal of Abnormal Psychology, 108(1), 120-133. Cohen, J. D., Botvinick, M., & Carter, C. S. (2000). Anterior cingulate and prefrontal cortex: Who's in control? Nature Neuroscience, 3, 421-423. Cohen, J. D., MacWhinney, B., Flatt, M. R., & Provost, J. (1993). PsyScope: An interactive graphic system for designing and controlling experiments in the psychology laboratory using Macintosh computers. Behavioral Research Methods: Instruments and Computers, 25, 257-271. 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. Curtin, J. J., & Fairchild, B. A. (2003). Alcohol and cognitive control: Implications for regulation of behavior during response conflict. Journal of Abnormal Psychology, 112, 424-436. Dahaene, S., Posner, M. I., & Tucker, D. M. (1994). Localization of a neural system for error detection and compensation. Psychological Science, 5, 303-305. Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters upon the identification of a target letter in a nonsearch task. Perception & Psychophysics, 16, 143-149. Fabiani, M., Gratton, G., & Coles, M. G. H. (2000). Event-related potentials. In J. T. Cacioppo, L. G. Tassinary & G. G. Berntson (Eds.), Handbook of Psychophysiology (2nd ed.). Cambridge, England: Cambridge University Press. Falkenstein, M., Hoormann, J., Christ, S., & Hohnsbein, J. (2000). ERP components on reaction errors and their functional significance: A tutorial. Biological Psychology, 51, 87-107. Gaetz, M., & Bernstein, D. M. (2001). The current status of electrophysiological procedures for the assessment of mild traumatic brain injury. Journal of Head Trauma Rehabilitation, 16, 386-405. Gehring, W. J., Goss, B., Coles, M. G. H., Meyer, D. E., & Donchin, E. (1993). A neural system for error detection and compensation. Psychological Science, 4, 385-390. Gehring, W. J., & Knight, R. T. (2000). Prefrontal-cingulate interactions in action monitoring. Nature Neuroscience, 3, 516-520. Golden, C. J. (1978). Stroop Color and Word Test. Chicago: Stoelting.

PAGE 56

46 Grapperon, J., Vidal, F., & Leni, P. (1988). The contribution of cognitive evoked potentials to knowledge mechanisms of the Stroop color-word interference effect. Neuropsychologia, 38, 701-711. Hajcak, G., McDonald, N., & Simons, R. F. (2003). To err is autonomic: Error-related brain potentials, ANS activity, and post-error compensatory behavior. Psychophysiology, 40, 895-903. Holroyd, C. B., Dien, J., & Coles, M. G. H. (1998). Error-related scalp potentials elicited by hand and foot movements: Evidence for an output-independent error-processing system in humans. Neuroscience Letters, 242, 65-68. Kerns, J. G., Cohen, J. D., MacDonald, A. W., Cho, R. Y., Stenger, V. A., & Carter, C. S. (2004). Anterior cingulate conflict monitoring and adjustments in control. Science, 303, 1023-1026. Kiefer, M., Marzinzik, F., Weisbrod, M., Scherg, M., & Spitzer, M. (1998). The time course of brain activations during response inhibition: Evidence from event-related potentials in a go/no go task. Neuroreport, 9, 765-770. Kutas, M., & Dale, A. (1995). Electrical and magnetic readings of mental functions. In M. D. Rugg (Ed.), Cognitive Neuroscience (pp. 197-242). Cambridge, MA: MIT Press. Larson, M. J., Kelly, K. G., & Perlstein, W. M. (2003). Functional neuroimaging of executive dysfunction in closed head injury: A cognitive neuroscience perspective. Special Interest Division 2, American Speech and Hearing Association, 13, 20-28. Lindsay, D. S., & Jacoby, L. L. (1994). Stroop process dissociations: the relationship between facilitation and interference. Journal of Experimental Psychology: Human Perception and Performance, 20, 219-234. Liotti, M., Woldorff, M. G., Perez, R., & Mayberg, H. S. (2000). An ERP study of the temporal course of the Stroop color-word interference effect. Neuropsychologia, 38, 701-711. Luu, P., Collins, P., & Tucker, D. M. (2000). Mood, personality, and self-monitoring: Negative affect and emotionality in relation to frontal lobe mechanisms of error monitoring. Journal of Experimental Psychology: General, 129, 43-60. Luu, P., Flaisch, T., & Tucker, D. M. (2000). Medial frontal cortex in action monitoring. Journal of Neuroscience, 20, 464-469. MacDonald, A. W., Cohen, J. D., Stenger, V. A., & Carter, C. S. (2000). Dissociating the role of the dorsolateral prefrontal cortex in cognitive control. Science, 288, 1835-1838.

PAGE 57

47 MacLeod, C. M. (1991). Half a century of research on the Stroop effect: an integrative review. Psychological Bulletin, 109, 163-203. Masaki, H., Tanaka, H., Takasawa, N., & Yamazaki, K. (2001). Error-related brain potentials elicited by vocal errors. Neuroreport, 12, 1851-1855. Miller, E. K. (2000). The prefrontal cortex and cognitive control. Nature Reviews: Neuroscience, 1, 59-66. Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex. Annual Review of Neuroscience, 24, 167-202. Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., & Howerter, A. (2000). The unity and diversity of executive functions and their contributions to complex "frontal lobe" tasks: A latent variable analysis. Cognitive Psychology, 41, 49-100. Neter, J., Wasserman, W., & Kutner, M. H. (1985). Applied linear statistical models: Regression, analysis of variance, and experimental designs (2nd ed.). Homewood, Ill.: R.D. Irwin. Norman, D. A., & Shallice, T. (1986). Attention to action: Willed and automatic control of behavior. In R. J. Davidson, G. E. Schwartz & D. Shapiro (Eds.), Consciousness and self-regulation (Vol. 4). New York: Plenum Press. Pardo, J. V., Pardo, J. P., Janer, K. W., & Raichle, M. E. (1990). The anterior cingulate cortex mediates processing selection in the Stroop attentional conflict paradigm. Proceedings of the National Academy of Sciences, 87, 256-259. Pascual-Marqui, R. D., Michel, C. M., & Lehmann, D. (1994). Low-resolution electromagnetic tomography: A new method for localizing electrical activity in the brain. International Journal of Psychophysiology, 18, 49-65. Perlstein, W. M., Carter, C. S., Barch, D. M., & Baird, J. (1998). The Stroop task and attentional deficits in schizophrenia: A critical evaluation of card and single-trial Stroop methodologies. Neuropsychology, 12, 414-425. Perlstein, W. M., Dixit, N. K., Carter, C. S., Noll, D., & Cohen, J. D. (2003). Prefrontal cortex dysfunction mediates deficits in working memory and prepotent responding in schizophrenia. Biological Psychiatry(53), 25-38. Perrin, F., Pernier, J., Bertrand, O., & Echallier, J. F. (1989). Spherical splines for scalp potential and current density mapping. Electroencephalography and Clinical Neurophysiology, 72, 184-187. Posner, M. I., & Snyder, C. (1975). Attention and cognitive control. In R. L. Solos (Ed.), Information Processing and Cognition (Vol. 1, pp. 55-85). Hillsdale, N.J.: Erlbaum.

PAGE 58

48 Rebai, M., Bernard, C., & Lannou, J. (1997). The Stroop test evokes a negative brain potential, the N400. International Journal of Neuroscience, 91, 85-94. Scheffers, M. K., & Coles, M. G. H. (2000). Performance monitoring in a confusing world: Error-related brain activity, judgements or response accuracy, and types or errors. Journal of Experimental Psychology: Human Perception and Performance, 26, 141-151. 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. Seignourel, P. J., Robbins, D., Demery, J., Larson, M. J., Cole, M., & Perlstein, W. M. (in preparation). A Tale of Two Stroops: Cognitive control dysfunction in CHI. Simmons, M. (1998). Brain activity correlates of the Stroop effect: An event-related potentials investigation. University of Florida, Gainesville, FL. Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18, 643-662. Taylor, S. F., Kornblum, S., Lauber, E. J., Minoshima, S., & Koeppe, R. A. (1997). Isolation of specific interference processing in the Stroop task: PET activation studies. NeuroImage, 6, 81-92. 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. van Veen, V., & Carter, C. S. (2002a). The anterior cingulate as a conflict monitor: fMRI and ERP studies. Physiology and Behavior, 77, 477-482. van Veen, V., & Carter, C. S. (2002b). The timing of action-monitoring processes in the anterior cingulate cortex. Journal of Cognitive Neuroscience, 14, 593-602. West, R. (1999). Age differences in lapses of intention in the Stroop task. Journal of Gerontology: Psychological Sciences, 54, 34-43. West, R. (2003). Neural correlates of cognitive control and conflict detection inthe Stroop and digit-location tasks. Neuropsychologia, 41, 1122-1135. West, R., & Alain, C. (1999). Event-related neural activity associated with the Stroop task. Cognitive Brain Research, 8, 157-164. West, R., & Alain, C. (2000a). Age-related decline in inhibitory control contributes ot the increased Stroop effect observed in older adults. Psychophysiology, 37, 179-189.

PAGE 59

49 West, R., & Alain, C. (2000b). Effect of task context and fluctuations of attention on neural activity supporting performance of the Stroop task. Brain Research, 873, 102-111. Williamson, S. J., & Kaufman, L. (1990). Theory of neuroelectric and neuromagnetic fields. In F. Grandori, M. Hoke & G. L. Romani (Eds.), Auditory Evoked Magnetic Fields and Electric Potentials. Advances in Audiology (Vol. 6, pp. 1-39). Basel: Karger. Zysset, S., Muller, K., Lohmann, G., & von Cramon, D. Y. (2001). Color-word matching Stroop task: Separating interference and response conflict. NeuroImage, 13, 29-36.

PAGE 60

BIOGRAPHICAL SKETCH Michael James Larson received a Bachelor of Science degree in psychology from Brigham Young University in 2002, and will receive his Master of Science degree from the University of Florida in 2004. He plans to continue his studies in clinical-cognitive neuroscience and neuropsychology and receive his Ph.D. 50


Permanent Link: http://ufdc.ufl.edu/UFE0004801/00001

Material Information

Title: Dissociating Components of Cognitive Control Using High-Density Event-Related Potentials: Implementation of Control, Conflict Processing, and Error Monitoring
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: UFE0004801:00001

Permanent Link: http://ufdc.ufl.edu/UFE0004801/00001

Material Information

Title: Dissociating Components of Cognitive Control Using High-Density Event-Related Potentials: Implementation of Control, Conflict Processing, and Error Monitoring
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: UFE0004801:00001


This item has the following downloads:


Full Text












DISSOCIATING COMPONENTS OF COGNITIVE CONTROL USING HIGH-
DENSITY EVENT-RELATED POTENTIALS: IMPLEMENTATION OF CONTROL,
CONFLICT PROCESSING, AND ERROR MONITORING















By

MICHAEL JAMES LARSON


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


2004

































Copyright 2004

by

Michael James Larson















ACKNOWLEDGMENTS

I thank my advisor, William M. Perlstein, as well as my collaborators Vonetta

Jones, Grant Webber, and Bao-thuy Hoang. I also thank my parents and my wife for

their 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

L IST O F F IG U R E S .... ...... ................................................ .. .. ..... .............. vii

ABSTRACT ........ .............. ............. ...... ...................... ix

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

C ognitiv e C control ................. ...... ............................ ...... ........ .. ...........
R egulative/Strategic Processes.................................... .................. .... ........... ..
Evaluative Processes ................................. .............. .. .. ........ ........ 3
N eural Correlates of Stroop Perform ance ........................................ ..................3
Dissociation of Cognitive Control Processes ....................... ................ ........... 4
Event-related Potentials ............................................. .. ...... ............... ...
ERPs and Cognitive Control.................... ......... ......................... ........ ............... 9
Electrophysiological Correlates of Evaluative Processes............... .......... 9
Electrophysiological Correlates of Regulative Processes ...................................12
Dissociation of Cognitive Control Component Processes using ERPs ..............13
P reduction s ............................................................................ 14
B behavioral D ata .................................................................................. 14
E R P D ata ....................................................... 15
R egulative Processes ............................................... ....... ... ............... 15
Evaluative Processes .......................... ...... ................ ...... .. ........ .. 15

2 M E T H O D .............................................................................17

P a rtic ip a n ts ........................................................................................................... 1 7
M materials and P procedure ........... ..................................................... .......................17
EEG Acquisition and Reduction.................................................... .. ............... 19
EEG A acquisition ......... .... .... .................... ........ ........ .... .......... .... 19
E E G D ata R edu action ............................................. ........................................ 19
Statistical A analyses ........... .... .......... .. ...... ............................ 21
Behavioral D ata Analyses ............................................................................21
E R P D ata A naly ses........... ...... .................................................. .. .... ..... .. 22











3 R E SU L T S ....................................................... 2 3

B behavioral A analyses ....................................... .......... ..... .............. 23
E rror R ate A naly ses............ ... .................................................... .. .. .... .... ... 23
R T A n aly ses ................................................................2 4
ERP Analyses ....................................................... 26
Instruction-related A activity ........................................... ........... ............... 26
Stim ulu s-related A activity ................................................................ ...............27
Response-related A activity ............................................................................27

4 D ISC U S SIO N ............................................................................... 36

B eh av ioral D ata ................................................................36
E R P D ata ................................................................3 7
R egu lativ e P ro cesses ..................................................................................... 3 7
Evaluative Processes .................... ............ ......... 39
Alternative Explanations and Potential Limitations ............................ ...........40
Future D directions ...................................... ...................... ... .... 42
S u m m a ry ..................................................................................................................... 4 3
Summary...i ........... .. ........... ........................43

L IST O F R EFER EN CE S ........................ .. ............................................ ............... 44

B IO G R A PH IC A L SK E TCH ..................................................................... ..................50





























v
















LIST OF TABLES

Table p

2-1. D em graphic inform ation................................................. .............................. 17

3-1. Means and standard errors of error rates (%) in the single-trial Stroop ...................24

3-2. Means and standard errors of median RT (ms) in the single-trial Stroop ..................25















LIST OF FIGURES


Figure page

1-1. Schema of the single-trial Stroop task. Participants are first presented with an
instructional cue ("Color" or "Word"), followed by a delay (1 or 5 seconds) and
Stroop stimulus (congruent, neutral, incongruent) .................................................5

1-2. Schemata of cognitive control functions and subsequent ERP manifestations
predicted in the current study. ..... ......................................................................16

2-1. Sensor layout of 64-channel geodesic sensor net. Slow-wave activity was quantified
at scalp sites #2 and #4 (blue); N450 activity at scalp site #55 (red); and, ERN
activity at scalp site #5 (green). See 'ERP Data Analyses' in text for details.........20

3-1. Error rates by task and congruency conditions collapsed across delay. Error bars
represent standard errors. ............................................... .............................. 24

3-2. Median RTs by task and congruency conditions collapsed across delay. Error bars
represent standard errors. ............................................... .............................. 26

3-3. Top view of spherical-spline interpolated voltage maps representing instruction-
related activity of the CN task, WR task, and the difference between the two tasks
(CN WR, bottom). ERP waveforms of the grand average differential slow-wave
activity between the CN and WR tasks at scalp site #2 (top). ................................28

3-4. Grand average instruction-locked ERPs for all scalp sites of the CN (red) and WR
(blue) task presentation. ........................... .... ................ ............ .... ...... ...... 29

3-5. Top view of spherical-spline interpolated voltage maps representing instruction-
related activity of subsequent correct and incorrect trials, and the difference (correct
incorrect, bottom). ERP waveforms of the grand average differential slow-wave
activity between subsequent correct and incorrect trials collapsed across conditions
at scalp site #4 (top). ........................... ....... ... .. ....... ........... 30

3-6. Grand average instruction-locked ERPs for all scalp sites of correct (red) and
incorrect (blue) task presentation collapsed across conditions. ............................31

3-7. Top view of spherical-spline interpolated voltage maps representing stimulus-
related activity of the congruent CN condition, the incongruent CN condition, and
the difference between the conditions (congruent incongruent, bottom). ERP









waveforms reflect the grand average N450 to the incongruent CN trials at scalp site
#55 (top). ........................................................................... 32

3-8. Grand average stimulus-locked ERPs for all scalp sites of the congruent (red) and
incongruent (blue) CN task. ............................................ ............................. 33

3-9. Top view of spherical-spline interpolated voltage maps representing response-
related activity of correct and incorrect trials (collapsed across conditions), and the
difference (correct incorrect, bottom). ERP waveforms of the grand average ERN
deflection to the incorrect trials as compared to the correct trials at scalp site #5
(to p ). ............................................................. ................ 3 4

3-10. Grand average response-locked ERPs for all scalp sites of correct and incorrect
responses collapsed across conditions...... .................. ...............35















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

DISSOCIATING COMPONENTS OF COGNITIVE CONTROL USING HIGH-
DENSITY EVENT-RELATED POTENTIALS: IMPLEMENTATION OF CONTROL,
CONFLICT PROCESSING, AND ERROR MONITORING

By

Michael James Larson

May 2004



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

Recent theories suggest that cognitive control is a dynamic process instantiated

within a fronto-cortical network that implements regulative or strategic control over top-

down processes, monitors and detects processing conflicts, and signals for adjustments in

performance when necessary. We examined this complement of cognitive control

processes behaviorally and neurally. To do this, we acquired high-density brain event-

related potentials (ERPs) while 24 neurologically-normal participants performed a cued,

single-trial Stroop task that temporally dissociated instruction-related regulative

processes (i.e., representing and maintaining the attentional demands of the task), from

evaluative processes (i.e., conflict processing, error monitoring). Implementation of

control was reflected in a right frontal instruction-related slow-wave associated with the

more attentionally demanding color-naming task. A mid-to-lateral frontal stimulus-

related conflict N450 was elicited by the incongruent color-naming task condition.










Response-locked ERPs to incorrect responses, collapsed across task conditions, revealed

a mid-fronto-central error-related negativity (ERN). Behaviorally, response times and

error rates were greatest in the incongruent color-naming task condition, indicating

Stroop response time and error-rate interference. Overall, ERPs are determined to be an

effective methodology for examining component processes of cognitive control.

Furthermore, findings are consistent with previous research indicating two distinct neural

systems in cognitive control, one for regulative processes involved in the implementation

of control and one involved in the evaluative processes of conflict detection and error

monitoring.














CHAPTER 1
INTRODUCTION

Cognitive Control

Cognitive control refers to the ability to guide thought and action in accord with

internal intentions (Botvinick, Carter, Braver, Barch, & Cohen, 2001; Cohen, Botvinick,

& Carter, 2000; Miller, 2000; Miller & Cohen, 2001) and encompasses those processes

necessary for controlled information processing and coordinated actions. Current

cognitive neuroscience theories distinguish between at least two important components of

cognitive control: a regulative or strategic component responsible for activation and

implementation of control processes, and an evaluative component responsible for

monitoring the need for control and signaling when adjustments in control are necessary

(Botvinick et al., 2001; Kerns et al., 2004).

Regulative/Strategic Processes

Regulative processes are those involved in the top-down control of cognition and

include functions such as representing and maintaining task context and goals (e.g.,

working memory), the allocation of limited attentional resources, as well as preparing to

execute cognitive tasks and override prepotent response tendencies (Cohen, Barch,

Carter, & Servan-Schreiber, 1999). Perhaps a prototypical cognitive task that requires

the implementation of cognitive control is the Stroop color-word task (MacLeod, 1991;

Stroop, 1935). Although there are many different versions of the original Stroop task

(MacLeod, 1991), the basic paradigm requires participants to either read words or name

the color in which the words are written. To perform this task successfully, participants









must selectively attend to one stimulus attribute (i.e., word or printed color). This is

particularly true when naming the color of an incongruent or conflict stimulus (e.g., the

word GREEN printed in blue) because there is a strong automatic tendency to read the

word (GREEN), which competes with the less automatic instruction to name the color

(blue). Typically, participants show robust Stroop interference effects, wherein there is

increased reaction times (RT) or error rates when participants are required to name the

color of the word when the word name and word color are incongruent (conflict

condition). The ability to successfully complete this task and overcome the "conflict"

illustrates selective allocation of attentional resources and the ability to select a weaker,

task-relevant response in the face of competition from an otherwise more automatic, but

task-irrelevant option (Miller & Cohen, 2001); a fundamental aspect of cognitive control.

Regulative processes of cognitive control also require the maintenance of task-

relevant context and goals. Cohen and colleagues (1999) suggest that maintenance of

context representations is critical to adequate performance on the Stroop task. For

example, in order to respond to the appropriate dimension of the stimulus, participants

must hold in mind the instruction for the trial, providing the necessary context for

interpreting the stimulus and generating the correct response. In the card or Golden

version of the Stroop task (Golden, 1978) frequently used in neuropsychological testing,

task conditions are blocked, wherein all stimuli for each condition are presented as lists

on their respective cards. This arrangement consistently reinforces the proper context

(i.e., task instruction) and, therefore, places minimal demands on the representation and

maintenance of context. In contrast, a single-trial computerized cued version of the

Stroop task devised by Cohen et al. (1999) presents trials individually and randomly

varies task instruction (color-naming, CN, word-reading, WR). To complete this task,









participants must maintain the context of the task instruction (CN or WR) prior to

stimulus presentation and employ these context representations to provide the correct

response.

Evaluative Processes

A second set of processes essential for cognitive control are those involved in the

evaluation of performance and include functions such as detection of processing conflicts

and performance monitoring. These evaluative processes are believed to play a crucial

role in signaling for adjustments of top-down control needed to adapt to constantly

changing task demands (Kerns et al., 2004). For example, in the card version of the

Stroop task described above, participants show greater interference on the initial one or

two trials in each block than on subsequent trials in a series (Botvinick et al., 2001).

Additionally, participants completing a modified Stroop task show less interference on

incongruent trials if the incongruent trials are frequent relative to congruent trials than if

they are rare (Lindsay & Jacoby, 1994). The results of these studies indicate detection of

conflict and subsequent adjustment in control processes to more efficiently perform the

task as well as increased conflict when there is decreased expectancy of a conflict

stimulus. Thus, cognitive control is a dynamic process that is most reliably invoked

during tasks involving conflicts in information processing where modification of

performance is required to successfully complete the task (Botvinick et al., 2001; Norman

& Shallice, 1986).

Neural Correlates of Stroop Performance

As noted in the examples above, one experimental paradigm that consistently

evokes response conflict conditions where cognitive control is necessary is the Stroop

task. The Stroop task has been utilized by numerous cognitive scientists to examine a









myriad of cognitive functions, including: automatic and controlled cognitive processes

(Posner & Snyder, 1975), selective attention (Rebai, Bernard, & Lannou, 1997), and

disturbances in these processes due to psychiatric disorders (Cohen et al., 1999; Perlstein,

Carter, Barch, & Baird, 1998). In addition, recent technological advances in

hemodynamic- and electrophysiologically-based neuroimaging methods have led to

increased insight into the neural correlates of the Stroop task, as well as the functional

neural bases of cognitive control more generally. Traditionally, the Stroop task has been

employed as an instrument used to measure pre-potent response inhibition (Miyake,

Friedman, Emerson, Witzki, & Howerter, 2000), a function often attributed to the frontal

lobes (Kiefer, Marzinzik, Weisbrod, Scherg, & Spitzer, 1998). Hemodynamic-based

neuroimaging research has described many frontal sites as critical to the ability to

overcome pre-potent tendencies, including: left inferior lateral cortex (Taylor, Kornblum,

Lauber, Minoshima, & Koeppe, 1997), left superomedial cortex (Pardo, Pardo, Janer, &

Raichle, 1990), right frontal polar cortex (Bench et al., 1993), and bilateral anterior

cingulate cortex (Bench et al., 1993). More recently, the Stroop task has been used to

examine not only the neural correlates of the ability to overcome pre-potent response

tendencies, but also the neural correlates of the regulative and evaluative components of

cognitive control.

Dissociation of Cognitive Control Processes

To distinguish among component processes of cognitive control, a modified

version of the Stroop paradigm (Figure 1-1) has been introduced that allows one to

dissociate the regulative and evaluative processes required to successfully complete the

Stroop task (Cohen et al., 1999). In this modified Stroop task, participants are given an

instruction before each trial indicating whether to read the word (a more automatic









response) or name the color (a condition requiring an increased amount of control due to

the need to override the more automatic tendency to read the word). Following a brief

delay, the Stroop color-word stimulus is presented and the participant responds. Thus,

the task temporally separates the instruction-related regulative processes (representing the

context/goal of the task in the CN or WR trials) from the response-related evaluative

processes (the detection of incongruencies and errors and signaling for adjustments in

control processes). This modified version of the Stroop task can be contrasted with the

traditional Golden Stroop task, where participants respond in blocks to trials of the same

type, thus temporally confounding regulative and evaluative processes.


Single-Trial Cued Stroop
I ---io ~"Color"
Instruction

Short (Is)
Delay
Long (5s)
,, /- Congnrent
CongaIency Neutral
"t i i jL,. h i j t _aJll It

Figure 1-1. Schema of the single-trial Stroop task. Participants are first presented with
an instructional cue ("Color" or "Word"), followed by a delay (1 or 5 seconds)
and Stroop stimulus (congruent, neutral, incongruent).


Using the modified Stroop paradigm described above and event-related functional

magnetic resonance imaging (fMRI) techniques, MacDonald et al. (2000) demonstrated a

double dissociation of the roles of the dorsolateral pre-frontal cortex (dlPFC) and anterior

cingulate cortex (ACC) in the regulative and evaluative component processes of

cognitive control. Specifically, they found that during maintenance of the task context









(CN, WR), the left dlPFC was more active following instructions to perform the CN task

than the WR task, consistent with a role of the dlPFC in preparation to execute the more

demanding color-naming task. In contrast, ACC activity was increased upon presentation

of incongruent color-word stimuli as compared to congruent color-word stimuli,

consistent with a role in detection of response conflict. Based on these data, the authors

suggested that the network necessary for cognitive control is dissociable into either the

regulative or evaluative processes necessary for completion of separate task aspects. This

study, however, did not allow for the specific examination of conflict processing and

error detecting aspects of the evaluative component of cognitive control as the stimulus-

related activity did not differentiate error trials from correct trials. Thus, the evaluative

processes of conflict processing and error detection are potentially confounded. Despite

the limitation, this study is of great importance to the cognitive control literature as it

demonstrates a functional dissociation of not only the behavioral aspects of cognitive

control, but also the neuroanatomical bases of cognitive control; elucidating the dlPFC

mediated regulative control processes and the ACC mediated evaluative processes. This

study also emphasizes the reality that conventional behavioral methodologies do not

permit cognitive psychological processes and representations to be assessed directly.

Furthermore, different neuroimaging methods have different strengths in the examination

of cognitive and neural processes. For example, fMRI results represent blood flow-based

hemodynamic response mechanisms to specific stimuli that are not particularly sensitive

to the temporal course of neural activity. In the MacDonald study, the hemodynamic-

response signal was examined over five 2.5-second increments per trial, while verbal

responses to Stroop stimuli tend to occur in less than one second. Thus, results of this

fMRI study reflect a spatially sensitive common neural output, rather than a direct









reflection of neural processes.

In contrast to hemodynamic-based measures of brain activity, event-related

potentials (ERPs) provide direct measurement of neuronal electrical activity with sub-

millisecond resolution. Moreover, under some circumstances, ERPs may be able to detect

neural activity that is unaccompanied by secondary phenomena such as changes in

regional blood flow or local metabolic activity (Gaetz & Bernstein, 2001). However,

ERPs have limited spatial resolution, potentially leading to ambiguous spatial

localization. Therefore, a convergence of information based on multiple neuroimaging

methods, including scalp-recorded brain ERPs, can provide additional clarity into the

nature of the cognitive and neuronal processes of cognitive control.

Event-related Potentials

In order to have a thorough understanding of the contributions of ERPs to theories

of cognitive control, one must understand the assumptions behind ERP neuroimaging

methods. Briefly, two major assumptions of ERP neuroimaging research are, first, that

the distribution of electrical activity across the scalp indicates the activities of underlying

neural structures, and, second, that this neural activity corresponds with specific cognitive

states and processes. To the extent these assumptions are valid, 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. This overall background, or ongoing electrical activity,

however, is not the interest of the present study; rather, the time- (event-) locked

averaged electrical activity associated with the presentation of specific events is of









interest. Initially, the event-related signal associated with the presentation of a stimulus

is embedded in the noise of the background EEG activity. Extracting the signal

associated with a specific cognitive activity from the "noise" (background activity and

measurement error) is accomplished by averaging multiple samples of the EEG that are

time-locked to repeated occurrences of the event (i.e., stimulus or response) of interest.

The logic of averaging is that the signal does not change from trial to trial, while the

noise is random, thus, the signal is enhanced by a factor proportional to the square-root of

the number of trials, while the noise is reduced essentially to zero (Fabiani, Gratton, &

Coles, 2000). Due to the direct measure of electrical brain activity associated with

specific cognitive events, ERPs are currently considered the "gold standard" in terms of

temporal resolution among noninvasive imaging methods (Fabiani et al., 2000).

ERP waveforms typically consist of a series of discrete deflections (i.e., peaks and

troughs), often followed by so-called slow-wave potentials1. Characteristics of ERP

waveforms usually include descriptors of polarity (positive or negative) and latency (in

milliseconds). For example, "P300" refers to an ERP with a positive peak that has an

approximate latency of 300 milliseconds. Another similar labeling system involves a

descriptor of polarity followed by a number representing the ordinal latency of the

component. Using these labeling criteria, "P3" refers to the third positive peak in the

ERP waveform. Other descriptors, such as scalp location at which the component is

maximal (e.g., frontal P3) are also used.

Cortically, the activity of neurons associated with ERP activity is attributed

primarily to post-synaptic potentials (Williamson & Kaufman, 1990). As an example,


1 A slow-wave potential is a temporally extended change in the ERP waveform, rather than a distinct or
punctate deflection.









consider the case of an excitatory post-synaptic potential (although similar activities

occur for inhibitory post-synaptic potentials). While at rest, neurons contain a lesser

concentration of sodium ions (Na+) and a greater concentration of potassium ions (K+)

inside the cell. When the dendrites of a neuron receive an excitatory signal from an

adjacent neuron, the resulting change in the cell membrane allows Na+ to flow into the

cell. This results in a reduction of positive ions in the extracellular space-making the

space more negative. This negativity in the extracellular space is known as the current

sink. The positive Na+ ions that entered the cell repel like-charged ions, and create a

current that sends the K+ ions toward the cell body. This buildup of positive charge near

the cell body is known as the current source. The current source repels like-charged ions

in the extracellular space, which are then attracted back to the sink, producing a dipolar

extracellular current. It is this extracellular current that produces the electrical potentials.

ERPs, then, represent the net activities of a large population of neurons that must be

synchronously active and configured in such a way they produce dipolar electromagnetic

fields that can be measured at the scalp. Such a synchronously activated configuration of

neurons is known as an "open field," and usually involves the alignment of neurons in an

orientation parallel to the scalp (Coles & Rugg, 1995)2.

ERPs and Cognitive Control

Electrophysiological Correlates of Evaluative Processes

In cognitive control research, ERPs can serve an important purpose because they

are particularly sensitive to the temporal course of neural activity and, by extension, the

concomitant underlying sensory, motor, and cognitive processes. Of particular interest to


2 Information on the physiological bases of ERPs taken from Coles & Rugg, 1995; Fabiani et al., 2000;
Simmons, 1998. Please see these sources for additional information.









the current study is the idea that ERPs can be used to temporally dissociate the

component processes of cognitive control by facilitating inferences regarding the timing,

level of processing, and, roughly, the anatomical location of neural mechanisms

supporting these processes. The preponderance of research on component processes of

cognitive control using ERPs has focused on the evaluative components of conflict

detection and error monitoring. For example, using tasks that reliably invoke processing

conflicts between simultaneously active task-relevant representations (e.g., Stroop or

Eriksen flanker tasks3), investigators have found a reliably evoked late fronto-central

ERP signature referred to as the N450 or N2 component (van Veen & Carter, 2002a,

2002b; West, 2003). These components both reflect conflict detection processes and

differ only based on the stimulus paradigm presented (e.g., Stroop vs. Eriksen flanker

tasks). Conflict detection in the Stroop paradigm has been associated with the negative

deflection between 400ms and 500ms known as the N450 (Liotti, Woldorff, Perez, &

Mayberg, 2000; West & Alain, 1999), while conflict associated with the incongruent

condition of the Eriksen flanker has been associated with a slight negative deflection in

the ERP waveform known as the N2 (van Veen & Carter, 2002b). These ERP

components are largest under conditions in which response conflict is high, such as the

incongruent condition of the Stroop color-naming task (Grapperon, Vidal, & Leni, 1988;

Liotti et al., 2000; Rebai et al., 1997). Increased amplitude of the N450 has been

observed following unexpected or rare incongruent trial presentation (West & Alain,

2000b), consistent with the hemodynamic-based neuroimaging results presented

3 The Eriksen Flanker task (Eriksen & Eriksen, 1974) consists of a target central stimulus "flanked" by
either congruent or incongruent stimuli. For example, participants may be instructed to press a button with
their left hand if a central target arrow points left (e.g., <) or a button with their right hand if a central target
arrow points right (e.g., >). For congruent trials the flankers are the same as the specified target (e.g.,
<<<<<), while incongruent trials have flankers that are opposite the target (e.g., <><<).










previously of MacDonald et al. (2000) implicating the role of the ACC in conflict

detection. In support of this hypothesis, source localization algorithms have also

localized regions of the ACC as the neural generators of the N450 and N2 components

(van Veen & Carter, 2002a, 2002b; West, 2003). These results implicate the N450 and

N2 components as neurobiological indices of conflict detection, and support the role of

ACC as a conflict detection mechanism.

Detection of errors is another component of the evaluative processes in cognitive

control that, due to the exquisite temporal sensitivity of electrophysiological-based

neuroimaging methods, has been widely investigated using ERPs. The detection of

errors has been associated with a midline fronto-central negative deflection in response-

locked ERPs that occurs within 100ms of committing an error. This negative deflection

is known as the error-related negativity (ERN), and is the first identified neurobiological

index of performance monitoring (Dahaene, Posner, & Tucker, 1994; Falkenstein,

Hoormann, Christ, & Hohnsbein, 2000; Gehring, Goss, Coles, Meyer, & Donchin, 1993;

Luu, Flaisch, & Tucker, 2000). The precise function responsible for the ERN continues

to be debated (Cohen et al., 2000b; Gehring & Knight, 2000); however, the ERN has

typically been referred to as a representation of an error/conflict monitoring system that

operates across various stimulus and response modalities. Specifically, the ERN has been

found with visual, auditory, and movement responses (Bernstein, Scheffers, & Coles,

1995; Holroyd, Dien, & Coles, 1998), is present following errors of omission as well as

errors of commission (Falkenstein et al., 2000), and is greatest in amplitude when

participants are aware that an error has been made (Luu, Collins, & Tucker, 2000;

Scheffers & Coles, 2000)--consistent with the action of an error detector/performance

monitor in cognitive control. Furthermore, the ERN is observed when participants make









"partial" errors (begin to make an error but spontaneously correct themselves), and is

greater in amplitude on the trials with "partial" errors than errors where the conflict is not

produced in time to spontaneously correct the response (Gehring et al., 1993). These

results implicate the ERN as an on-line index of error detection or performance

monitoring. Such an index is critical because an important function of the human brain is

to monitor behavior and prevent undesirable actions. Like the N450 and N2 components,

dipole-modeling techniques have localized the neural activity associated with the ERN to

the ACC (Holroyd et al., 1998; van Veen & Carter, 2002b). In fact, van Veen and Carter

(2002a) reviewed multiple studies of the N2 and ERN components and concluded that

both components are representations of similar conflict detection processes. Specifically,

they concluded that the N2 reflects pre-response detection of conflict between competing

response tendencies, while the ERN reflects post-response detection of incompatible

responses following error trials. The similarities between the N2/N450 and the ERN

provide additional support for a model of cognitive control with ACC mediated

evaluative processes detecting incongruities and possibly signaling for changes in the

top-down control of cognition.

Electrophysiological Correlates of Regulative Processes

Regulation of cognitive control has also been examined using ERPs. The

allocation of attentional resources under challenging task conditions and the active

maintenance of goal/task-representations have been shown to be reflected in differential

ERP slow-wave activity between tasks requiring differing levels of attentional demands

(Curtin & Fairchild, 2003). Under the appropriate task conditions, investigators have

observed a slow-wave in the ERP that appears to be associated with the implementation

of cognitive control, perhaps reflecting an active biasing of processing in favor of the









more attentionally-demanding aspect of the task (Curtin & Fairchild, 2003; West, 2003).

This slow-wave has also been used to distinguished between subsequent correct and

incorrect responses on a version of the cued-Stroop task, with greater slow-wave activity

preceding correct than incorrect responses (West, 2003). The slow-wave activity that

differentiates trials where there is a "dysregulation" of control processes (incorrect trials)

and trials where control is adequately implemented (correct trials) also supports the

contention that slow-wave activity reflects the implementation of control processes.

Dissociation of Cognitive Control Component Processes using ERPs

Cognitive control component processes have also been dissociated temporally

using ERPs. Using a variation of the modified Stroop described above (Cohen et al.,

1999; MacDonald, Cohen, Stenger, & Carter, 2000), West (2003) reported a temporal

dissociation between the regulative and evaluative component processes of cognitive

control. Following presentation of task instruction, implementation of regulative

processes was exhibited by an occipital-parietal slow-wave that differentiated correct

("goal-compatible") and incorrect ("goal incompatible") response trials. In addition,

implementation of control was associated with frontal slow-wave ERP modulation that

differentiated CN trials, more attentionally demanding trials where participants prepare to

override the more automatic WR response, from less demanding WR trials. These

findings are consistent with the allocation of increased attentional resources during tasks

requiring increased cognitive processes, as well as increased implementation of control

on correct trials. Conflict detection was associated with a fronto-central N450 with

greater amplitude for incongruent than congruent trials. Consistent with a role in conflict

monitoring, the N450 amplitude was reflected equally for incongruent trials in both CN

and WR conditions. In addition, West reported a frontal slow-wave component 500ms to









600ms following incongruent stimulus presentation. This component, known as the

conflict slow-wave potential (conflict SP), was interpreted to reflect the allocation of

increased attentional resources in preparation for future incongruent trials during the

incongruent condition of both CN and WR trials. These findings taken together indicate

that the different neural processes reflecting the regulative and evaluative of cognitive

control can be dissociated in time using ERPs.

Predictions

Given the sensitivity of ERPs in examining the temporal course of neural activity,

the goal of the present study was to use the modified Stroop task developed by Cohen and

colleagues to give added support to the temporal dissociation of the regulative and

evaluative components of cognitive control using high-density event-related potentials.

Based on previous research the following predictions are offered:

Behavioral Data

It is predicted that standard Stroop effects for both RT and error-rates will be

manifest by increased RTs and error rates to incongruent CN and WR conditions. In

addition, it is predicted that interference effects will be disproportionately greater in the

CN as compared to the WR task. It is also predicted that facilitation effects (faster RTs to

congruent than neutral conditions) for both the CN and WR tasks will be manifest. For

delay conditions, it is hypothesized that greater time between instruction and stimulus

presentation will allow for increased implementation of regulative processes; therefore,

fewer errors and faster RTs are predicted for trials following the 5s delay, as compared to

trials following the Is delay.









ERP Data

Since the behavioral data alone cannot address the potential mechanisms)

underlying task performance and allocation/implementation of cognitive control

resources, ERPs are necessary to further examine the electrophysiological signatures of

the following possibilities (Figure 1-2)4:

Regulative Processes

* Regulative processes associated with the preparation to override more automatic
response tendencies and the maintenance of task-relevant representations are
predicted to be manifest in frontal slow-wave activity locked to the task
instructional cue that will differentiate preparation to engage in the more
attentionally demanding CN task as compared to the more automatic WR task.

* In addition, it is hypothesized that lapses in regulation of cognitive control
processes leads to increased errors. This will be examined in instruction-locked
slow-wave ERP activity collapsed across conditions for error and correct trials. It
is predicted that task instruction-related slow-wave activity will differentiate
subsequent correct and incorrect trials.

Evaluative Processes

* Evaluative processes associated with conflict detection are predicted to be reflected
in an N450 ERP deflection that has increased amplitude to stimulus-locked
incongruent vs. congruent CN stimuli.


* Evaluative processes associated with error processing will be reflected in response-
locked ERPs that exhibit an increased ERN following incorrect trials that is less
pronounced during correct trials--indicative of the detection of incompatibilities
(conflicts) between the response given and the accurate response.









4 Examination of the effect of delay on stimulus-locked ERP reflections of task (CN, WR) effects is
currently underway. These analyses, however, are incomplete and will not be discussed. Additionally,
delay x congruency effects for each task will not be examined since the number of trials per condition is
low (< 28, see Methods section below) resulting in inadequate signal-to-noise ratio for examining these
effects.











Instruction Locked Stimulus Locked


Response Locked


cognitivee Implementation Performance
Conflict Detection .
Process of Control Conflict Det Monitoring



ERP Signature Post-instruction Post-stimulus Post-response
slow-wave N450 ERN


Time-in-Task Trial


Figure 1-2. Schemata of cognitive control functions and subsequent ERP manifestations
predicted in the current study.















CHAPTER 2
METHOD

Participants

Twenty-four right-handed individuals (14 female) between the ages of 18 and 24

participated in the study in exchange for course credit. All participants provided written

informed consent according to procedures established by the University of Florida Health

Science Center Institutional Review Board. Participants were screened for depressive or

anxious symptoms that may influence results using the Beck Depression Inventory (Beck

& Steer, 1987; Beck, Steer, & Garbing, 1988) and Beck Anxiety Inventory (Beck &

Steer, 1990), respectively. Descriptive information regarding participants' age,

education, and depression and anxiety levels is reported in Table 1. Participants were

excluded if they reported previous neurological insults, traumatic brain injury (TBI),

psychiatric diagnosis, current psychotropic medication use, or color-blindness (color-

blindness was assessed using the Ishihara pseudo-isochromatic color plates, Clark, 1924).

All participants reported normal or corrected-to-normal vision.

Table 2-1. Demographic information
Mean Standard Deviation Range
Age (years) 19 1.4 18-24
Education (years) 13.4 1.0 12 16
Beck Depression Inventory score 5.5 5.2 0 25
Beck Anxiety Inventory score 5.2 4.8 0 19

Materials and Procedure

Participants performed a computerized single-trial, cued version of the Stroop task

(see Figure 1-1), originally developed by Cohen et al. (1999). In this task, each trial

began with the computer presentation of an auditory instructional cue (the word "color"

17









or "word"), followed, after a short delay, by a visual stimulus, which remained on the

screen until participant's response. Stimuli comprised the same three colors and color-

words (red, green, and blue) used in the card Stroop (Golden, 1978), and commonly

employed in clinical neuropsychological settings. Participants were instructed to respond

verbally to the stimulus as quickly and accurately as possible. RTs were determined by a

voice-activated relay connected to the computer, and the examiner manually coded

response accuracy. Participants performed the color-naming and word-reading tasks,

each comprising three congruency conditions. Congruent stimuli were words printed in

the same color (e.g., "RED" printed in red), incongruent stimuli were words printed in a

different color (e.g., "RED" printed in blue), and neutral stimuli were animal names

printed in red, green, or blue (e.g., "BEAR" printed in red) for the color-naming trials and

words displayed in white for word-reading trials. The context provided by the task

instruction (e.g., color) must be used to override the influence of the stronger dimension

(i.e., word) when the task is to respond to the less prepotent or automatic dimension (i.e.,

color). Additionally, reliance on context was increased by (a) varying the task to be

performed on each trial and (b) introducing a delay between the task instruction for each

trial and the stimulus to be responded to. For the current task, we used two delays

(stimulus onset asynchrony, SOA, of Is and 5s). Visual stimuli were presented in the

center of a visual display, and delivered using an Apple Macintosh computer using

PsyScope software (Cohen, MacWhinney, Flatt, & Provost, 1993). A total of 336

experimental trials were distributed equally across task and congruency, resulting in 28

trials of each type.

Prior to acquisition of electrophysiological data, participants completed a practice

block consisting of random presentation of 12 trials, one of each stimulus type. On these









trials, if the participants RT was over 1000ms, an auditory beep was presented, indicating

to the participant a need to respond more quickly.

EEG Acquisition and Reduction

EEG 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

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

common ground. The impedance of all electrodes was kept below 50 kM, consistent with

procedures suggested by the manufacturer.

EEG Data Reduction

Due to the volume conducting nature of the brain, no one site on the head can be

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

therefore, data was mathematically rereferenced against an average reference (Bertrand,

Perrin, & Pemier, 1985). 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; (Scherg, 1990).










In
64 63
11 6

14 7 1
12

19 15 8 3 61 60
19 60
13 62

23 16 9 8 57 59
20 56
17 54
Left ear 21 53 Right ear
Ref
Ift ear 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
35 36 44 Corn
39




Figure 2-1. Sensor layout of 64-channel geodesic sensor net. Slow-wave activity was
quantified at scalp sites #2 and #4 (blue); N450 activity at scalp site #55 (red);
and, ERN activity at scalp site #5 (green). See 'ERP Data Analyses' in text
for details.

Individual-subject ERP averages were divided into three categories and included a

pre-stimulus baseline period: auditory task instruction-related activity, visual stimulus-

presentation activity, and response-related activity (see Figure 1-2). Instruction-locked

epochs, associated with the implementation of control, were extracted from 100ms prior

to instruction presentation to 1000ms post instruction presentation. Instruction-locked

epochs were calculated for correct trials of both the presentation of the color-naming and

word-reading tasks. Similarly, instruction-locked epochs were extracted separately for

correct and incorrect responses from 100ms prior to instruction presentation to 1000ms

post instruction presentation, collapsed across color-naming and word-reading tasks.









Stimulus-locked epochs, associated with the detection of conflict, were extracted with a

duration of 100ms prior to stimulus presentation and 750ms post-stimulus presentation.

Individual subject averages were calculated for the correct-trial congruent, neutral, and

incongruent stimulus-locked conditions. Response-locked averages were created

separately for correct and incorrect responses, collapsed across color-naming and word-

reading tasks, as well as congruencies. Collapsing across conditions was necessary

because of insufficient numbers of incorrect responses to conduct specific error x

congruency analyses. In addition, two-participants did not make any errors, their data,

therefore, were not included in the analyses of response-related activity. Response-

locked activity was extracted with a duration of 400ms pre-stimulus presentation and

400ms post-stimulus presentation. All averaged ERP epochs were baseline corrected

using a 100ms window prior to stimulus or response onset and digitally filtered using a

30 Hz low-pass filter and a .5 Hz high-pass filter.

Statistical Analyses

Behavioral Data Analyses

For analysis of behavioral data, correct-trial RTs and overall error rates were

analyzed separately. For each trial type and participant, we calculated the median RT for

correct responses as well as proportion of errors by subjecting them to 2-Task x 2-Delay

x 3-Congruency repeated-measures analysis of variance analyses (ANOVA). Tests of

simple effects were used to decompose interaction effects. For error rates, due to the

high probability of no errors in several conditions (e.g., congruent word-reading

condition), raw data were normalized prior to analysis using the arcsine transformation

(Neter, Wasserman, & Kutner, 1985).









ERP Data Analyses

Analysis of ERP waveforms focused on instruction-related, stimulus-related, and

response-related activity as indicated in Figure 1-2. Statistical analyses of ERP

waveforms were performed on mean voltages over specified temporal windows extracted

from individual electrode sites. Scoring windows and electrode positions for each

condition of interest were determined by examination of grand-average ERP waveforms

and spline-interpolated scalp voltage distribution plots (Perrin, Pernier, Bertrand, &

Echallier, 1989). Instruction-related activity was quantified at electrode site #2 (see

Figure 2-1) and was examined over the period of 800ms to 1000ms post-stimulus

presentation. Paired t-tests were conducted between the CN and WR instruction-related

slow-wave activity. Additionally, instruction-related activity for subsequent correct and

incorrect trials, collapsed across conditions, was quantified at electrode site #4 (see

Figure 2-1) and was examined for the period of 700 to 1,000ms post-instruction

presentation. Stimulus-related activity was examined over the period of 440ms to 500ms

post-stimulus presentation, and was quantified at electrode site #55 (see Figure 2-1).

Paired t-tests examined conflict detection through comparison of congruent and

incongruent CN stimulus presentation. Response-related activity was examined over the

period of 32ms to 72ms post-response, and quantified at electrode site #5 (see Figure 2-

1). Paired t-tests were used to compare mean amplitudes of correct and incorrect

responses. Previous research has shown the ERN and the N450 to be phasic components

of the ERP waveform (i.e., brief deflections, Rebai et al., 1997; West & Alain, 1999), and

durations for examination were chosen to reflect their short-term nature. Slow-wave

activity is tonic in nature, therefore, a broader window for averaging was chosen.














CHAPTER 3
RESULTS

Behavioral Analyses

Initial analyses examined the possibility of a speed/accuracy trade-off by

correlating RT and error-rates. Analyses revealed that RT and error-rate data were

negatively correlated (r = -.30); however, the correlation was not significant (p>. 10),

indicating a slight, but non-significant speed/accuracy trade-off

Error Rate Analyses

Mean error rates and standard errors as a function of task condition, congruency,

and delay are provided in Table 3-1. Analyses revealed the standard Stroop effects: A

main effect of task, F(1,23 ) = 10.5 1, p<.01, with WR more accurate than CN; a main

effect of congruency, F(2,46) = 53.2, p<.01, with significantly more errors in the

incongruent condition than congruent condition, F(1,23) = 59.53, p<.01, and the neutral

condition, F(1,23) = 55.51, p<.01. There was also a Task x Congruency interaction,

F(2,46) = 7.02, p<.01, with increased errors in the incongruent condition of the CN task

compared to the congruent condition, F(1,23) = 10.86, p<.01, and the neutral condition,

F(1,23) = 6.40, p<.01, of the CN and WR tasks (i.e., error rate interference), but no

significant difference in error facilitation (neutral condition errors congruent condition

errors), F(1,23) = .46, p>.40 (Figure 3-1). Of note is the finding that there was not a

significant main effect of delay, F(1,23) = 1.01, p>.50, and delay did not significantly

interact with any task, F(2,46) = 2.55, p>. 10, or congruency conditions, F(2,46) = .62,

p>.50.











Table 3-1. Means and standard errors of error rates (%) in the single-trial Stroop
M SE
Color-naming
Short Delay
Congruent .31 .20
Neutral 2.01 .70
Incongruent 10.20 2.70
Long Delay
Congruent 1.39 .40
Neutral 1.54 .40
Incongruent 11.90 3.10
Word-reading
Short Delay
Congruent .31 .20
Neutral 1.08 .40
Incongruent 4.32 .90
Long Delay
Congruent .62 .30
Neutral .93 .40
Incongruent 3.40 .70


Figure 3-1. Error rates by task and congruency conditions collapsed across delay. Error
bars represent standard errors.


RT Analyses

Means and standard errors of median RTs as a function of task condition,

congruency, and delay are provided in Table 3-2. As with the accuracy data, analyses of


Proportion of Errors

0 14
o 012
WU 0.1
4a-
0 0.08 Color-narming
.0 0.06 Word-reading
o 0.04
o 0 02

Congruent Neutral Incongruent


ngruent Neutral Incongruent









RT data revealed the standard Stroop effects: A main effect of task, F(1,23) = 47.31,

p<.01, with WR faster than CN; a main effect of congruency, F(2,46) = 53.2, p<.01, with

significantly slower RTs in the incongruent (i.e., interference) condition as compared to

the neutral condition, F(1,23) = 51.20, p<.01, and congruent condition, F(1,23) = 64.71,

p<.01, and a facilitation effect as evidenced by significantly faster RTs to congruent

conditions compared to neutral conditions, F(1,23) = 6.68, p<.05. There was a

significant Task x Congruency interaction, F(2,46) = 12.6, p<.01, with stronger

interference effects in the CN task compared to the WR task, F(1,23) = 17.99, p<.01 (see

Figure 3-2). Similar to the error rate analyses, there was not a significant main effect of

delay, F(1,23) = .13, p>.70, and delay did not significantly interact with task, F(1,23) =

.16, p>.60, or congruency conditions F(2,46) = 1.25, p>.25.


Table 3-2. Means and standard errors of median RT (ms) in the single-trial Stroop
M SE
Color-naming
Short Delay
Congruent 735.88 25.73
Neutral 799.67 23.19
Incongruent 895.65 32.79
Long Delay
Congruent 743.65 23.59
Neutral 777.46 22.71
Incongruent 896.58 29.67
Word-reading
Short Delay
Congruent 724.50 25.72
Neutral 721.90 23.88
Incongruent 800.73 33.07
Long Delay
Congruent 723.42 27.58
Neutral 717.83 27.42
Incongruent 805.65 34.96











Median RT
950
90 900
oo
r 850

750
700
650 Color-naming
650 Word-readinq
E 600
= 550
500
Congruent Neutral Incongruent

Figure 3-2. Median RTs by task and congruency conditions collapsed across delay.
Error bars represent standard errors.

ERP Analyses

Instruction-related Activity

Spline-interpolated scalp voltage maps for instruction-related ERP activity

showed a right inferior frontal difference between the CN and WR instruction

presentation (Figure 3-3). ERPs to the task instructional cues (CN or WR) showed

differential slow-wave activity quantified at scalp electrode site #2 following instructions

to engage in the CN and WR tasks t(23) = 1.73, p<.05 (Figure 3-4). These findings are

consistent with the deployment of regulative processes associated with the

implementation of cognitive control to the more attentionally-demanding CN task (e.g.,

implementation of control reflected in preparation to override the prepotent WR response

tendency). Additionally, inspection of spline-interpolated scalp voltage maps revealed a

medial frontal difference between instruction-related activity of subsequent correct and

incorrect responses (Figure 3-5). ERP slow-wave activity quantified at scalp site #4

revealed slow-wave activity that differentiated subsequent execution of correct and

incorrect responses collapsed across conditions, t(23) = 6.3, p<.01 (Figure 3-6).









Stimulus-related Activity

The behavioral data (RTs, error rates) did not differ as a function of delay;

therefore, stimulus-locked ERPs were collapsed across delay conditions. In addition,

increased conflict was reflected in the significantly longer RTs to the incongruent CN

condition as compared to the incongruent WR condition; therefore, ERP analyses focused

on the congruent and incongruent CN conditions, collapsed across delay. Consistent with

previous research (Liotti et al., 2000; West, 2003; West & Alain, 1999), spline-

interpolated scalp voltage maps for stimulus-related ERP activity showed a mid to right-

lateral frontal difference between the congruent and incongruent conditions of the CN

task (Figure 3-7). Stimulus-locked ERPs showed significantly greater N450 deflection to

the incongruent than congruent stimuli of the CN task quantified at scalp site #55

approximately 450ms post-stimulus presentation, t(23) = 3.28, p<.01 (Figure 3-8).

Similar findings have been interpreted as the neural representation of conflict detection in

the incongruent condition as compared to decreased conflict in the congruent condition

(Liotti et al., 2000; West & Alain, 2000b).

Response-related Activity

Spline-interpolated scalp voltage maps for response-related ERP activity revealed a

medial-frontal difference between the correct and incorrect responses, collapsed across

task and congruency due to insufficient error trials (Figure 3-9). Response-locked ERPs

to correct and incorrect responses exhibited a significant negative deflection (ERN)

quantified at scalp site #5 occurring approximately 50ms following incorrect responses

that was not present following correct responses, t(21) = 2.11, p<.01 (Figure 3-10).

These results are consistent with previous research and have been previously interpreted











to represent the detection of conflict in task performance (Falkenstein et al., 2000;

Gehring et al., 1993; van Veen & Carter, 2002a).




























CN ITask WVVR Task Difference
Used color versus vaue of potential

-3 -2 -1 0 1 2 3
Potential [VY]


Figure 3-3. Top view of spherical-spline interpolated voltage maps representing
instruction-related activity of the CN task, WR task, and the difference
between the two tasks (CN WR, bottom). ERP waveforms of the grand
average differential slow-wave activity between the CN and WR tasks at scalp
site #2 (top).











Kr~;~j~


Front
jriSWe-:


12-j4
I -ow- wc 4
C- c


4V c
4 OcN
Ba~ck


.-,d ^- N taskh







t'









S






2 li00
T^a----- ,

4t&
-4vr-.-4-- r



0 1000
9nms


Figure 3-4. Grand average instruction-locked ERPs for all scalp sites of the CN (red) and
WR (blue) task presentation.





































I co ,rro
Incorrect


Sieff re nce


Used colorversus value of potential

-4 -3 -2 -1 0 1 2 3 4
Potential [pV]


Figure 3-5. Top view of spherical-spline interpolated voltage maps representing
instruction-related activity of subsequent correct and incorrect trials, and the
difference (correct incorrect, bottom). ERP waveforms of the grand average
differential slow-wave activity between subsequent correct and incorrect trials
collapsed across conditions at scalp site #4 (top).


o0 0o
Co Qr o

Co rrcot









Front
4s-ds>-


- CI~orrssc
-Inco~rrcct


12 2 I C=

Th...A & -~c...B (V6\--


Left .,X -O,'L c &4t-\t4IV< J -- ... Flight




4^ ... 9 2 .," "



4j4. 3 V.1


Va -2.4
'G a .- : [ __



t i 0 vllq.1

Figure 3-6. Grand average instruction-locked ERPs for all scalp sites of correct (red) and
incorrect (blue) task presentation collapsed across conditions.


~c---r;






































Congruent Incongruent DifferrIence
Usedcolorversusvalue of potential

-4 -3 -2 -1 0 1 2 3 4
Potential [yV]


Figure 3-7. Top view of spherical-spline interpolated voltage maps representing
stimulus-related activity of the congruent CN condition, the incongruent CN
condition, and the difference between the conditions (congruent -
incongruent, bottom). ERP waveforms reflect the grand average N450 to the
incongruent CN trials at scalp site #55 (top).










_j4 FrOnt -----. r
S... Inzorngrucnt Cj N








L-eft;;51Tr t--qiA-"-,--
S*Fi Right


I v

c k ... i l 8_
41^ T --:--'... I L







^> V^ ^., -.



Figure 3-8. Grand average stimulus-locked ERPs for all scalp sites of the congruent (red)ed and incongruent (blue) CN task.


ed)
and incongruent (blue) CN task.





































Correct


Incorrect D ifferencre


Usedcolorversus value of potential

-2 -1 0 1 2
Pot ental [YV]


Figure 3-9. Top view of spherical-spline interpolated voltage maps representing
response-related activity of correct and incorrect trials (collapsed across
conditions), and the difference (correct incorrect, bottom). ERP waveforms
of the grand average ERN deflection to the incorrect trials as compared to the
correct trials at scalp site #5 (top).











.Front ,C-rrp Hqpqn



*I rrt p in ^






L eft- ,,. _I

















incorrect responses collapsed across conditions.
^^ 'S4^^ i! ^.^ .-.-
^ ^ ^ ^ ) __ ) __
B'aIPck Tim

Fiur 310.Gad vrg rsos-locke E fr alsap ie fcorc n
icret rsose olasdaroscndtos














CHAPTER 4
DISCUSSION

Current theories of cognitive control recognize two features as essential to

negotiating everyday cognitive tasks: an evaluative component responsible for

monitoring the need for internal adjustments in control and signaling when such

adjustments are necessary and a regulative component responsible for activation and

implementation of control processes. The current study utilized ERPs and a modified

version of the Stroop task to temporally dissociate the electrophysiological signatures of

the regulative and evaluative processes of cognitive control.

Behavioral Data

Participants displayed the typical Stroop interference effects, with increased error

rates and RTs on incongruent CN trials. These results reflect the increased influence of

word-reading over color-naming on the incongruent CN trials and the need to override

the more automatic tendency to read the word. Delay did not significantly affect RT or

error-rate performance in any task or congruency conditions. The absence of delay

effects may be attributed to the demographics of the sample used in this study. The

sample consisted of educated college students who had intact working memory skills and

had little difficulty maintaining the representation of the CN or WR tasks; therefore,

performance on the long and short delay conditions was nearly equivalent. These results

can be contrasted with those of Seignourel et al. (in preparation) who found in healthy,

slightly older, control participants decreased error-rates following the 5s delay condition,

but no differences in RTs. Cohen et al. (1999) found increased facilitation of RTs (faster









RTs to congruent than neutral stimuli) in the 5s delay condition as compared to the Is

delay, also in slightly older controls. The RT facilitation and decreased error rates in the

5s condition were attributed to the increased time allowed to engage regulative control

mechanisms and prepare to override prepotent response tendencies. Interestingly,

Seignourel et al. found that moderate-to-severely traumatic brain injured participants did

not display differential error rate performances by delay and Cohen et al. found no

within, or between subjects error-rate differences in patients diagnosed with

schizophrenia and normal controls. Based on this information, current theories

hypothesizing that increased delay between instruction-cue and stimulus presentation

facilitates the implementation of regulative control processes remain ambiguous.

Nonetheless, computational modeling studies do suggest that, in healthy participants

using a different cognitive control task (i.e., AX-CPT, (Braver, Barch, & Cohen, 1999),

the biasing of control requires some time to achieve full strength. Furthermore,

functional neuroimaging studies have shown that this delay-related effect is mediated, at

least in part, by the dlPFC (Barch et al., 2001). More studies are necessary to determine

the effects of delay between instruction-cue and stimulus presentation on the

implementation of regulative processes in cognitive control.

ERP Data

Regulative Processes

As predicted, regulative processes reflecting the implementation of control were

shown in frontal slow-wave activity, more specifically, instruction-related slow-wave

activity that differentiated the CN from the WR tasks. These results are consistent with

an increased requirement for top-down control and increased allocation of attentional

resources to the CN task in preparation to override the more automatic tendency to read









the word. These results are similar to the fMRI results of MacDonald et al. (2000) who

attributed increased left dlPFC activity to the CN task as compared to the WR task to the

implementation of increased cognitive control processes in preparation for the more

demanding CN task. Reasons for the differences in lateralization are currently unclear;

however, due to the volume-conducting nature of the brain, the lack of spatial sensitivity

associated with ERPs, and the failure of the current investigation to provide consistent

source localization results using more advanced dipole modeling techniques, speculation

about specific anatomical locations of control processes is considered tentative. The

results and interpretations of the current study and those of MacDonald et al. can be

contrasted with those of West (2003). Using a similar modified version of the Stroop

task as that used by the current study and MacDonald et al., West also found instruction-

locked slow-wave activity that differentiated the CN and WR tasks. The slow-wave

activity found by West, however, was reflected primarily in a slow negativity over the

occipital-parietal regions and positivity over the frontal-central region. West also

described an instruction-locked slow-wave that differentiated correct and incorrect

responses. This slow-wave activity reflected greater negativity for incorrect relative to

correct responses and was specifically related to whether or not a correct response was

made, rather than any particular aspects of the presented stimulus. Dipole-based source

localization methods used by West determined that the best dipole model for the slow-

wave activity occurred using mirrored dipoles in the occipital-parietal region, with a

single dipole in the left dlPFC. The dipole in the left dlPFC contributed only moderately

to the overall fit of the model. Based on these findings, West concluded that instruction

cue-related slow-wave activity over the occipito-parietal region supported the processing









of "goal-compatible responses" (correct responses) rather than preferential processing of

a more attention demanding stimulus attribute.

To examine the hypothesis that occipital-parietal slow-wave activity is associated

with correct and incorrect responses, we collapsed CN and WR trials by accuracy. Our

findings revealed a mid frontal-central slow-wave that differentiated subsequent correct

and incorrect responses. However, in contrast to the findings of West, our slow-wave

activity reflected greater negativity for correct relative to incorrect responses. Based on

the assumption that errors are committed when there is a break down in the regulative

processes of cognitive control, these findings could represent a possible "dysregulation"

of control processes during error trials as compared to correct trials. Due to the fact,

however, that trials were collapsed across conditions, no conjectures about specific

stimulus attributes associated with the correct and incorrect trials are offered on the basis

of the current study.

Evaluative Processes

Evaluative processes were reflected in a mid to right-lateral frontal N450 with

greater negativity to the incongruent CN trials than congruent CN trials. These results

are consistent with previous studies of the Stroop task (Liotti et al., 2000; West, 2003;

West & Alain, 1999) that suggest the N450 is an electrophysiological reflection of

conflict detection. In support of this interpretation, West (2003) showed increased

negativity to the incongruent trials of both the CN and WR tasks, allowing him to

conclude that the N450 is a neural representation of a cognitive process with a role in

conflict detection independent of task condition. Previous dipole source localization

research on the N450 has localized possible neural generation of the N450 to the ACC

(West, 2003; West & Alain, 1999). These results are consistent with fMRI results found










by MacDonald et al. and several PET studies (Carter, Mintun, & Cohen, 1995; Zysset,

Muller, Lohmann, & von Cramon, 2001) indicating greater activity in the ACC during

the incongruent, compared to congruent, CN trials.

Evaluative processes were further explored in the current study by examining

response-related ERP activity following correct and incorrect responses. The ERN, a

significantly greater negativity following error trials compared to correct trials, has been

suggested to reflect an error detection mechanism that plays a role in signaling for

adjustments in performance following the occurrence of errors. Results of the current

study replicate previous findings discussed previously of the ERN following error trials.

Previous dipole source localization of the ERN has also implicated the ACC as the most

likely neural generator (Falkenstein et al., 2000; Gehring et al., 1993; van Veen & Carter,

2002a, 2002b). Overall, convergence of evidence from N450 and ERN components of

electrophysiological-based ERP methods, as well as results of hemodynamic-based

neuroimaging studies, emphasizes an evaluative mechanism of cognitive control that is

neurally distinct from the regulative components of cognitive control.

Alternative Explanations and Potential Limitations

While we propose that the observed ERP results reflect the regulative and

evaluative components of cognitive control, other alternatives and potential limitations

must be considered. First, the precise function responsible for the ERN has been subject

to multiple interpretations (Cohen, Botvinick, & Carter, 2000a; Gehring & Knight, 2000).

Research has not clarified whether the ERN is caused by the error itself, or by processing

conflicts that produce uncertainty and predispose to error. Additionally, previous

research has been incongruous on performance adjustments following error trials, with

some research correlating post-error slowing with the ERN and some failing to document









such a slowing effect (Hajcak, McDonald, & Simons, 2003); therefore, it remains unclear

how a monitoring function such as the ERN relates to mechanisms of cognitive control.

In addition, there is considerable evidence that attentive and motivational factors

modulate ERN magnitude. For example, when participants are motivated to make

accurate rather than speedy responses, the ERN is larger (Falkenstein et al., 2000).

Similarly, larger ERNs occur when participants are certain they have made an error

(Gehring et al., 1993). Continued research is required to further elucidate the precise

cognitive focus reflected in the ERN.

Second, the current study employed ERP methodologies to provide direct

measurement of the temporal course of neurological reflections of cognitive control

processes. Under some circumstances, ERPs may be able to detect neural activity that is

unaccompanied by changes in regional blood flow or metabolic activity. However, ERPs

have limited spatial resolution, potentially leading to ambiguous spatial localization.

Nonetheless, relatively recent developments resulting in high-density spatial sampling of

EEG (i.e., 64, 128, 256 channels), combined with sophisticated source localization

algorithms, have been used to provide greater confidence in source localization (Pascual-

Marqui, Michel, & Lehmann, 1994; Scherg, 1990). In the current study, source

localization methods were not employed due to artifact introduced into the EEG

recordings by concerns of contamination due to the vocal response mechanism.

Vocalization-related cortical potentials (VRCPs) are one potential contaminant and

consist of a movement-related cortical potential preceding vocalization and an auditory-

evoked negative potential that immediately follows vocalization (Masaki, Tanaka,

Takasawa, & Yamazaki, 2001). Additionally, following each trial response,

electromyographic activity was observed throughout the EEG that could not be









completely removed using current data-correction algorithms. Due to this variability

associated with vocal response introduced into the EEG data, it was felt that source

localization methods would not be reliable estimates of the potential neural generators of

cognitive control processes. Nonetheless, our finding of N450 modulation over medial-

frontal regions is consistent with several studies that employed manual responses (West,

2003; West & Alain, 1999, 2000b) and both manual and verbal responses (Liotti et al.,

2000).

Future Directions

The present study lends continued credence to the use of a single-trial version of

the Stroop task that specifically examines the component processes of cognitive control.

Previous research using the Golden version of the Stroop task has not had the sensitivity

of the single-trial version of the Stroop used in this study due to the temporal confound

between the regulative and evaluative processes of cognitive control. This is of particular

relevance in the study of functional impairments in groups who demonstrate putative

cognitive control deficits on traditional measures (e.g., blocked Stroop task), as well as

groups whose underlying deficits have been proposed to be related to impaired goal

maintenance or conflict detection processes, for example, schizophrenia (Cohen et al.,

1999; Perlstein, Dixit, Carter, Noll, & Cohen, 2003), aging (West, 1999; West & Alain,

2000a), and traumatic brain injury (Seignourel et al., in preparation) The current study

lays the foundation for examining the electrophysiological representations of cognitive

control in these groups. Future studies using ERPs and hemodynamic-based

neuroimaging methods will examine the specific components of goal maintenance and

conflict processing deficits in these populations where dysfunctions in broader cognitive

control processes have been hypothesized to account for disease-specific impairments.









Specifically, future studies are underway in aging, ADHD, and traumatic brain injury

(TBI; Larson, Kelly, & Perlstein, 2003) that will further elucidate the specific nature of

deficits and the underlying neural substrates of cognitive control dysfunction.

Summary

Current cognitive neuroscience theories distinguish between at least two important

components of cognitive control: an evaluative component responsible for monitoring the

need for control and signaling when adjustments in control are necessary, and a

regulative component responsible for activation and implementation of control processes.

The current study used ERPs in an effort to temporally dissociate these regulative and

evaluative components of cognitive control. The findings of this research are consistent

with the hypothesis that regulative and evaluative components of cognitive control are

dissociable with ERP methods. This conclusion lays the foundation for future studies

that will provide increased clarity into specific cognitive control deficits in clinical

populations (e.g., TBI).















LIST OF REFERENCES


Barch, D. M., Carter, C. S., Braver, T. S., Sabb, F. W., MacDonald, A. W., Noll, D., et al.
(2001). Selective deficits in prefrontal cortex function in medication-naive patients
with schizophrenia. Archives of General Psychiatry, 58, 280-288.

Beck, A. T., & Steer, R. A. (1987). The BeckDepression Inventory Manual. San Antonio,
TX: Psychological Corporation.

Beck, A. T., & Steer, R. A. (1990). The Beck Anxiety Inventory Manual. San Antonio,
TX: The Psychological Corporation.

Beck, A. T., Steer, R. A., & Garbing, M. G. (1988). Psychometric properties of the Beck
Depression Inventory: Twenty-five years of evaluation. Clinical Psychology
Review, 8, 77-100.

Bench, C. J., Frith, C. D., Grasby, P. M., Friston, K. J., Paulesu, E., & Frackowiack, R. J.
S. (1993). Investigation of the functional neuroanatomy of attention using the
Stroop task. Neuropsychologia, 31, 907-922.

Bernstein, P. S., Scheffers, M. K., & Coles, M. G. H. (1995). "Where did I go wrong?" A
psychophysiological analysis of error detection. Journal ofExperimental
Psychology: Human Perception and Performance, 21, 1312-1322.

Bertrand, O., Perrin, F., & Pernier, J. (1985). A theoretical justification of the average-
reference in topographic evoked potential studies. Electroencephalography and
Clinical Neurophysiology, 62, 462-464.

Botvinick, M. W., Carter, C. S., Braver, T. S., Barch, D. M., & Cohen, J. D. (2001).
Conflict monitoring and cognitive control. Psychological Review, 108, 624-652.

Braver, T. S., Barch, D. M., & Cohen, J. D. (1999). Cognition and control in
schizophrenia: a computational model of dopamine and prefrontal function.
Biological Psychiatry, 46, 312-328.

Carter, C. S., Mintun, M., & Cohen, J. D. (1995). Interference and facilitation effects
during selective attention: an H215-0 PET study of Stroop task performance.
Neurolmage, 2, 264-272.

Clark, J. H. (1924). The Ishihara test for color blindness. American Journal of
Physiological Optics, 5, 269-276.










Cohen, J. D., Barch, D. M., Carter, C. S., & Servan-Schreiber, D. (1999). Context-
processing deficits in schizophrenia: converging evidence from three theoretically
motivated cognitive tasks. Journal ofAbnormal Psychology, 108(1), 120-133.

Cohen, J. D., Botvinick, M., & Carter, C. S. (2000). Anterior cingulate and prefrontal
cortex: Who's in control? Nature Neuroscience, 3, 421-423.

Cohen, J. D., MacWhinney, B., Flatt, M. R., & Provost, J. (1993). PsyScope: An
interactive graphic system for designing and controlling experiments in the
psychology laboratory using Macintosh computers. Behavioral Research Methods:
Instruments and Computers, 25, 257-271.

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.

Curtin, J. J., & Fairchild, B. A. (2003). Alcohol and cognitive control: Implications for
regulation of behavior during response conflict. Journal ofAbnormal Psychology,
112, 424-436.

Dahaene, S., Posner, M. I., & Tucker, D. M. (1994). Localization of a neural system for
error detection and compensation. Psychological Science, 5, 303-305.

Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters upon the identification of
a target letter in a nonsearch task. Perception & Psychophysics, 16, 143-149.

Fabiani, M., Gratton, G., & Coles, M. G. H. (2000). Event-related potentials. In J. T.
Cacioppo, L. G. Tassinary & G. G. Berntson (Eds.), Handbook of
Psychophysiology (2nd ed.). Cambridge, England: Cambridge University Press.

Falkenstein, M., Hoormann, J., Christ, S., & Hohnsbein, J. (2000). ERP components on
reaction errors and their functional significance: A tutorial. Biological Psychology,
51, 87-107.

Gaetz, M., & Bernstein, D. M. (2001). The current status of electrophysiological
procedures for the assessment of mild traumatic brain injury. Journal of Head
Trauma Rehabilitation, 16, 386-405.

Gehring, W. J., Goss, B., Coles, M. G. H., Meyer, D. E., & Donchin, E. (1993). A neural
system for error detection and compensation. Psychological Science, 4, 385-390.

Gehring, W. J., & Knight, R. T. (2000). Prefrontal-cingulate interactions in action
monitoring. Nature Neuroscience, 3, 516-520.


Golden, C. J. (1978). Stroop Color and Word Test. Chicago: Stoelting.









Grapperon, J., Vidal, F., & Leni, P. (1988). The contribution of cognitive evoked
potentials to knowledge mechanisms of the Stroop color-word interference effect.
Neuropsychologia, 38, 701-711.

Hajcak, G., McDonald, N., & Simons, R. F. (2003). To err is autonomic: Error-related
brain potentials, ANS activity, and post-error compensatory behavior.
Psychophysiology, 40, 895-903.

Holroyd, C. B., Dien, J., & Coles, M. G. H. (1998). Error-related scalp potentials elicited
by hand and foot movements: Evidence for an output-independent error-processing
system in humans. Neuroscience Letters, 242, 65-68.

Kerns, J. G., Cohen, J. D., MacDonald, A. W., Cho, R. Y., Stenger, V. A., & Carter, C. S.
(2004). Anterior cingulate conflict monitoring and adjustments in control. Science,
303, 1023-1026.

Kiefer, M., Marzinzik, F., Weisbrod, M., Scherg, M., & Spitzer, M. (1998). The time
course of brain activations during response inhibition: Evidence from event-related
potentials in a go/no go task. Neuroreport, 9, 765-770.

Kutas, M., & Dale, A. (1995). Electrical and magnetic readings of mental functions. In
M. D. Rugg (Ed.), Cognitive Neuroscience (pp. 197-242). Cambridge, MA: MIT
Press.

Larson, M. J., Kelly, K. G., & Perlstein, W. M. (2003). Functional neuroimaging of
executive dysfunction in closed head injury: A cognitive neuroscience perspective.
Special Interest Division 2, American Speech and Hearing Association, 13, 20-28.

Lindsay, D. S., & Jacoby, L. L. (1994). Stroop process dissociations: the relationship
between facilitation and interference. Journal of Experimental Psychology: Human
Perception and Performance, 20, 219-234.

Liotti, M., Woldorff, M. G., Perez, R., & Mayberg, H. S. (2000). An ERP study of the
temporal course of the Stroop color-word interference effect. Neuropsychologia,
38, 701-711.

Luu, P., Collins, P., & Tucker, D. M. (2000). Mood, personality, and self-monitoring:
Negative affect and emotionality in relation to frontal lobe mechanisms of error
monitoring. Journal ofExperimental Psychology: General, 129, 43-60.

Luu, P., Flaisch, T., & Tucker, D. M. (2000). Medial frontal cortex in action monitoring.
Journal ofNeuroscience, 20, 464-469.

MacDonald, A. W., Cohen, J. D., Stenger, V. A., & Carter, C. S. (2000). Dissociating the
role of the dorsolateral prefrontal cortex in cognitive control. Science, 288, 1835-
1838.









MacLeod, C. M. (1991). Half a century of research on the Stroop effect: an integrative
review. Psychological Bulletin, 109, 163-203.

Masaki, H., Tanaka, H., Takasawa, N., & Yamazaki, K. (2001). Error-related brain
potentials elicited by vocal errors. Neuroreport, 12, 1851-1855.

Miller, E. K. (2000). The prefrontal cortex and cognitive control. Nature Reviews:
Neuroscience, 1, 59-66.

Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex. Annual
Review ofNeuroscience, 24, 167-202.

Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., & Howerter, A. (2000). The
unity and diversity of executive functions and their contributions to complex
"frontal lobe" tasks: A latent variable analysis. Cognitive Psychology, 41, 49-100.

Neter, J., Wasserman, W., & Kutner, M. H. (1985). Applied linear statistical models:
Regression, analysis of variance, and experimental designs (2nd ed.). Homewood,
Ill.: R.D. Irwin.

Norman, D. A., & Shallice, T. (1986). Attention to action: Willed and automatic control
of behavior. In R. J. Davidson, G. E. Schwartz & D. Shapiro (Eds.), Consciousness
and self-regulation (Vol. 4). New York: Plenum Press.

Pardo, J. V., Pardo, J. P., Janer, K. W., & Raichle, M. E. (1990). The anterior cingulate
cortex mediates processing selection in the Stroop attentional conflict paradigm.
Proceedings of the National Academy ofSciences, 87, 256-259.

Pascual-Marqui, R. D., Michel, C. M., & Lehmann, D. (1994). Low-resolution
electromagnetic tomography: A new method for localizing electrical activity in the
brain. International Journal ofPsychophysiology, 18, 49-65.

Perlstein, W. M., Carter, C. S., Barch, D. M., & Baird, J. (1998). The Stroop task and
attentional deficits in schizophrenia: A critical evaluation of card and single-trial
Stroop methodologies. Neuropsychology, 12, 414-425.

Perlstein, W. M., Dixit, N. K., Carter, C. S., Noll, D., & Cohen, J. D. (2003). Prefrontal
cortex dysfunction mediates deficits in working memory and prepotent responding
in schizophrenia. Biological Psychiatry(53), 25-38.

Perrin, F., Pernier, J., Bertrand, O., & Echallier, J. F. (1989). Spherical splines for scalp
potential and current density mapping. Electroencephalography and Clinical
Neurophysiology, 72, 184-187.

Posner, M. I., & Snyder, C. (1975). Attention and cognitive control. In R. L. Solos (Ed.),
Information Processing and Cognition (Vol. 1, pp. 55-85). Hillsdale, N.J.:
Erlbaum.









Rebai, M., Bernard, C., & Lannou, J. (1997). The Stroop test evokes a negative brain
potential, the N400. International Journal ofNeuroscience, 91, 85-94.

Scheffers, M. K., & Coles, M. G. H. (2000). Performance monitoring in a confusing
world: Error-related brain activity, judgements or response accuracy, and types or
errors. Journal ofExperimental Psychology: Human Perception and Performance,
26, 141-151.

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.

Seignourel, P. J., Robbins, D., Demery, J., Larson, M. J., Cole, M., & Perlstein, W. M. (in
preparation). A Tale of Two Stroops: Cognitive control dysfunction in CHI.

Simmons, M. (1998). Brain activity correlates of the Stroop effect: An event-related
potentials investigation. University of Florida, Gainesville, FL.

Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of
Experimental Psychology, 18, 643-662.

Taylor, S. F., Komblum, S., Lauber, E. J., Minoshima, S., & Koeppe, R. A. (1997).
Isolation of specific interference processing in the Stroop task: PET activation
studies. Neurolmage, 6, 81-92.

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.

van Veen, V., & Carter, C. S. (2002a). The anterior cingulate as a conflict monitor: fMRI
and ERP studies. Physiology and Behavior, 77, 477-482.

van Veen, V., & Carter, C. S. (2002b). The timing of action-monitoring processes in the
anterior cingulate cortex. Journal of Cognitive Neuroscience, 14, 593-602.

West, R. (1999). Age differences in lapses of intention in the Stroop task. Journal of
Gerontology: Psychological Sciences, 54, 34-43.

West, R. (2003). Neural correlates of cognitive control and conflict detection inthe Stroop
and digit-location tasks. Neuropsychologia, 41, 1122-1135.

West, R., & Alain, C. (1999). Event-related neural activity associated with the Stroop
task. Cognitive Brain Research, 8, 157-164.

West, R., & Alain, C. (2000a). Age-related decline in inhibitory control contributes ot the
increased Stroop effect observed in older adults. Psychophysiology, 37, 179-189.






49


West, R., & Alain, C. (2000b). Effect of task context and fluctuations of attention on
neural activity supporting performance of the Stroop task. Brain Research, 873,
102-111.

Williamson, S. J., & Kaufman, L. (1990). Theory of neuroelectric and neuromagnetic
fields. In F. Grandori, M. Hoke & G. L. Romani (Eds.), Auditory EvokedMagnetic
Fields and Electric Potentials. Advances in Audiology (Vol. 6, pp. 1-39). Basel:
Karger.

Zysset, S., Muller, K., Lohmann, G., & von Cramon, D. Y. (2001). Color-word matching
Stroop task: Separating interference and response conflict. Neurolmage, 13, 29-36.















BIOGRAPHICAL SKETCH

Michael James Larson received a Bachelor of Science degree in psychology from

Brigham Young University in 2002, and will receive his Master of Science degree from

the University of Florida in 2004. He plans to continue his studies in clinical-cognitive

neuroscience and neuropsychology and receive his Ph.D.