<%BANNER%>

Developing a cognitive-affective neuroimaging probe to explore differential diagnosis of traumatic brain injury (TBI) an...

Permanent Link: http://ufdc.ufl.edu/UFE0041654/00001

Material Information

Title: Developing a cognitive-affective neuroimaging probe to explore differential diagnosis of traumatic brain injury (TBI) and posttraumatic stress disorder (PTSD) with functional magnetic resonance imaging (fMRI) A preliminary study
Physical Description: 1 online resource (65 p.)
Language: english
Creator: Tyner, Callie
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Traumatic brain injury (TBI) survivors often experience persistent post-concussive symptoms (PCS), including cognitive, affective, and somatic complaints. These symptoms reflect problematic overlap with those of posttraumatic stress disorder (PTSD), complicating differential diagnosis and potentially hampering tailoring of treatment. This investigation aims to develop a functional magnetic resonance imaging (fMRI) probe of rostral anterior cingulate cortex (rACC) and other emotion-critical brain regions using a cognitive-affective task the Cued Emotional Counting Stroop modeled after one used in this lab to probe caudal anterior cingulate cortex (cACC) in TBI. Our larger aim is to develop an fMRI probe to differentiate PCS in TBI and PTSD. Four neurologically-normal adults completed an event-related fMRI paradigm using this task, which includes an instructional-cue followed by a stimulus-probe containing pleasant, neutral, and unpleasant words. We predicted cue-related activity in dorsolateral prefrontal cortex (dlPFC), emotional probe-related activity in rACC and amygdala, and increased response latency and error rates to emotionally salient words. No significant differences in reaction times (RT) or error-rates by probe-stimulus valence were found. Neuroimaging revealed task-related activity to emotionally salient words in bilateral amygdala and rACC, and frontal activation, including dlPFC, during context maintenance. Results support further exploration of this task as a reliable probe of rACC and cognitive-affective brain circuits important in both PTSD and TBI. Further application to clinical populations, in conjunction with measures of symptomatology, is encouraged.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Callie Tyner.
Thesis: Thesis (M.S.)--University of Florida, 2010.
Local: Adviser: Perlstein, William.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041654:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041654/00001

Material Information

Title: Developing a cognitive-affective neuroimaging probe to explore differential diagnosis of traumatic brain injury (TBI) and posttraumatic stress disorder (PTSD) with functional magnetic resonance imaging (fMRI) A preliminary study
Physical Description: 1 online resource (65 p.)
Language: english
Creator: Tyner, Callie
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Traumatic brain injury (TBI) survivors often experience persistent post-concussive symptoms (PCS), including cognitive, affective, and somatic complaints. These symptoms reflect problematic overlap with those of posttraumatic stress disorder (PTSD), complicating differential diagnosis and potentially hampering tailoring of treatment. This investigation aims to develop a functional magnetic resonance imaging (fMRI) probe of rostral anterior cingulate cortex (rACC) and other emotion-critical brain regions using a cognitive-affective task the Cued Emotional Counting Stroop modeled after one used in this lab to probe caudal anterior cingulate cortex (cACC) in TBI. Our larger aim is to develop an fMRI probe to differentiate PCS in TBI and PTSD. Four neurologically-normal adults completed an event-related fMRI paradigm using this task, which includes an instructional-cue followed by a stimulus-probe containing pleasant, neutral, and unpleasant words. We predicted cue-related activity in dorsolateral prefrontal cortex (dlPFC), emotional probe-related activity in rACC and amygdala, and increased response latency and error rates to emotionally salient words. No significant differences in reaction times (RT) or error-rates by probe-stimulus valence were found. Neuroimaging revealed task-related activity to emotionally salient words in bilateral amygdala and rACC, and frontal activation, including dlPFC, during context maintenance. Results support further exploration of this task as a reliable probe of rACC and cognitive-affective brain circuits important in both PTSD and TBI. Further application to clinical populations, in conjunction with measures of symptomatology, is encouraged.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Callie Tyner.
Thesis: Thesis (M.S.)--University of Florida, 2010.
Local: Adviser: Perlstein, William.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041654:00001


This item has the following downloads:


Full Text

PAGE 1

DEVELOPING A COGNITIVE-AF FECTI VE NEUROIMAGING PROBE TO EXPLORE DIFFERENTIAL DIAGNOSIS OF TRAU MATIC BRAIN INJURY (TBI) AND POSTTRAUMATIC STRESS DISORDER (PTS D) WITH FUNCTIONAL MAGNETIC RESONANCE IMAGING (FMRI): A PRELIMINARY STUDY By CALLIE ELIZABETH TYNER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORID A IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2010 1

PAGE 2

2010 Callie Elizabeth Tyner 2

PAGE 3

To Curtis 3

PAGE 4

ACK NOWLEDGMENTS I thank my mentor, Dr. William M. Perl stein, for his help and guidance as a compassionate and respectful advisor. I must also thank Christopher N. Sozda who provided continual intellectual and moral support throughout this project. I would also like to acknowledge the time and contributions of the study participants. I am also grateful for the input and dedi cation of my committee mem bers: Dr. Glenn Ashkanazi, Dr. Russell Bauer, and Dr. Stephen Boggs. 4

PAGE 5

TABL E OF CONTENTS page ACKNOWLEDGMENTS ..................................................................................................4LIST OF TABLES ............................................................................................................7LIST OF FIGURES ..........................................................................................................8ABSTRACT .....................................................................................................................9 CHA PTER 1 INTRODUC TION ....................................................................................................11Overview of the Problem .........................................................................................11Background .............................................................................................................11Traumatic Brain Injury (TBI) and Po st Concussive Symptoms (PCS) ..............11Posttraumatic Stress Disorder (PTSD) .............................................................14Comorbid Mild TBI and PTSD ..........................................................................15Combat specific exposures ........................................................................17Differential diagnosis ..................................................................................20Neuroanatomy Relevant to PCS and PTSD .....................................................22Neuroanatomy of cognitive impairment in TBI ...........................................22Neuroanatomy of attention and executive function ....................................23Functional neuroanatomy of emotion .........................................................24Anterior cingulate cortex (ACC) .................................................................25Emotional Stroop Tasks ...................................................................................26Study Aims and Predictions ....................................................................................29Behavioral Data ................................................................................................30Functional Magnetic Resonance Imaging (fMRI) Data .....................................302 METHOD S..............................................................................................................31Participants .............................................................................................................31Materials and Procedures .......................................................................................31FMRI Cognitive Task: Cued Emotional Counting Stroop ..................................31Stimuli ...............................................................................................................34Image Acquisition .............................................................................................35Data Analysis ..........................................................................................................36Behavioral Data Analysis ..................................................................................36Functional Image Data Reduction and Pre-Processing ....................................37Functional Imaging Data Analysis ....................................................................373 RESULT S...............................................................................................................39Behavioral Results ..................................................................................................39 5

PAGE 6

Self-Assessment Manikin (SAM) Ratings Analyses .........................................39Valence ratings ..........................................................................................39Arousal ratings ...........................................................................................40Task Performance Analyses .............................................................................42Reaction time (RT) analysis .......................................................................42Error rate analysis ......................................................................................43FMRI Results ..........................................................................................................44Head Movement ...............................................................................................44Probe-Related Activation ..................................................................................44Cue-Related Activation .....................................................................................454 DISCU SSION.........................................................................................................50Review of Findings .................................................................................................50Behavioral Data ................................................................................................50FMRI Data ........................................................................................................51Probe stimuli related activation ..................................................................51Instructional cue related activation .............................................................52Implications .............................................................................................................53Limitations ...............................................................................................................54Limitations of FMRI ..........................................................................................54Sample Size Limitations ...................................................................................55Future Directions ....................................................................................................56Summary ................................................................................................................57APPENDIX: ST IMULI....................................................................................................58LIST OF REFERENCES ...............................................................................................59BIOGRAPHICAL SKETCH ............................................................................................65 6

PAGE 7

LIST OF TABLES Table page 1-1 Post concussive symptoms (P CS): Domains of dysfunction ..............................131-2 Symptoms accompanying posttraumatic stress disorder (PTSD): Domains of dysfunction .........................................................................................................151-3 Overlapping symptoms of PTSD and PCS .........................................................162-1 Valence and arousal ratings of stimuli words based on normative data .............343-1 Participant Self-Assessment Ma nikin (SAM) ratings: Mean valence and arousal by category ............................................................................................393-2 Cued Emotional Counting Stroop: Behavioral results .........................................433-3 Average head motion by valence category .........................................................443-4 Areas of significant probe-related ac tivity for emotional ly arousing stimuli .........453-5 Areas of significant cue-related ac tivity for emotionally arousing stimuli ............46A-1 Stimuli used in Cued Em otional Counting Stroop task .......................................58 7

PAGE 8

LIST OF FIGURES Figure page 2-1 Diagram of Cued Emot ional Counting Stroop task .............................................322-2 Graph of Affective Norms for Englis h Words (ANEW) stimuli: Valence by arousal ................................................................................................................353-1 Participant Self-Assessment Ma nikin (SAM) ratings: Mean valence ..................403-2 Participant SAM ratings: Mean arousal ..............................................................423-3 Image of bilateral anterior cingulat e cortex (ACC) functional magnetic resonance imaging (fMRI) activation ..................................................................473-4 Image of bilateral precuneus fMRI activation ......................................................483-5 Image of bilateral amygdala fMRI activation .......................................................49 8

PAGE 9

Abstract of Thesis Pres ented to the Graduate School of the University of Florida in Partial Fulf illment of the Requirements for t he Degree of Master of Science DEVELOPING A COGNITIVE-AFFECTI VE NEUROIMAGING PROBE TO EXPLORE DIFFERENTIAL DIAGNOSIS OF TRAU MATIC BRAIN INJURY (TBI) AND POSTTRAUMATIC STRESS DISORDER (PTS D) WITH FUNCTIONAL MAGNETIC RESONANCE IMAGING (FMRI): A PRELIMINARY STUDY By Callie Elizabeth Tyner May 2010 Chair: William M. Perlstein Major: Psychology Traumatic brain injury (TBI) survivors o ften experience persistent post-concussive symptoms (PCS), including cognitive, a ffective, and somatic complaints. These symptoms reflect problematic overlap with those of posttraumat ic stress disorder (PTSD), complicating differential diagnosis and potentially hamper ing tailoring of treatment. This investigation aims to dev elop a functional magnetic resonance imaging (fMRI) probe of rostral anter ior cingulate cortex (rACC) and ot her emotion-critical brain regions using a cognitive-affective ta skthe Cued Emotional Counting Stroop modeled after one used in this lab to probe caudal anterior cingulate cortex (cACC) in TBI. Our larger aim is to develop an fMRI probe to differentiate PCS in TBI and PTSD. Four neurologically-normal adults comp leted an event-relat ed fMRI paradigm using this task, which includes an inst ructional-cue followed by a stimulus-probe containing pleasant, neutral, and unpleasant word s. We predicted cue -related activity in dorsolateral prefrontal cortex (dlPFC), emotional probe-related activity in rACC and amygdala, and increased response latency and error rates to emotionally salient words. 9

PAGE 10

10 No significant differences in reaction time s (RT) or error-rates by probe-stimulus valence were found. Neuroimaging revealed ta sk-related activity to emotionally salient words in bilateral amygdala and rACC, and fr ontal activation, including dlPFC, during context maintenance. Results support further ex ploration of this task as a reliable probe of rACC and cognitive-affective brain circuits important in both PTSD and TBI. Further application to clinical populations, in conjunc tion with measures of symptomatology, is encouraged.

PAGE 11

CHA PTER 1 INTRODUCTION Overview of the Problem The incidence of traumatic brain injury (TBI) is staggering (Kraus & Chu, 2005), yet understanding of the causes of f unctional impairments and best methods for treatment are currently incomplete. In a s ubset of patients, t hese impairments can be either confused with, or magnified by, posttraumatic stress disorder (PTSD). The neuroanatomical underpinnings of PTSD overl ap with the brain areas where dysfunction often presents in TBI. The present study aims to study regions of the brain implicated in both TBI and PTSD, using a task of attent ion and executive function that uses emotionally arousing words as stimuli, to determine if such a task could have utility in explicating the neurocognitive dysf unction seen in both TBI and PTSD. Background Traumatic Brain Injury (TBI) and Post Concussive Symptoms (PCS) TBI is a major public health problem, wit h more than 1.5 million people sustaining a TBI each year in the United States (US), and treatment costs reaching more than $48 billion per year (Kraus & C hu, 2005). Approximately 2% of the US population lives with disabilities from TBI (Kraus & Chu, 2005, Stein & McAllister, 2009). After sustaining a TBI of any severity, the areas of the brai n most often damaged are the orbitofrontal cortex and the ventral surface of the temporal lobe where t he forces of acceleration and deceleration of the brain (coup and contre-coup forces) cause lesions at the point(s) of contact with the bony protuberanc es on the vault of the sku ll (Kennedy, Jaffee, Leskin, Stokes, Leal, & Fitzpatrick, 2007; Maas, Sto cchetti, & Bullock, 2008; Rao & Lyketsos, 2000; Stein & McAllister, 2009; Vasterling, Ve rfaellie, & Sullivan, 2009). In addition to 11

PAGE 12

these primary injuries, the brain can undergo secondary damage due to chemical forces including, for example, neurotransmitter disruptions and hypoxia (Kennedy et al., 2007; Maas et al., 2008; Rao & Lyketsos, 2000). TBI is diagnosed bas ed on presence and dura tion of clinical symptoms (Maas et al., 2008). Although definitions vary widely, a TBI is generally considered mild when at least one of the following is present: (1) mental status is altered at the time of the accident, (2) there is a loss of consciousne ss with duration less than 30 minutes, (3) there is memory loss of events before or after the accident that does not exceed 24 hours, or (4) the presence of any focal neurological deficits (Hoge et al., 2008; Kay et al., 1993; McAllister, 2005; Stein & McA llister, 2009). Mild TBI, also known as concussion, represents the majority of all TBIs sustained, accounting for approximately 80% of all case s (Kraus & Chu, 2005; Levin, Katz, Dade, & Black, 2002). The prognosis for mild TBI is often relati vely favorable, compared with moderate and severe TBI, and most individuals recover quickly with the majority of notable impairments subsiding within wee ks (Kay et al., 1992; Vasterli ng et al., 2009). Moderate and severe TBI are defined as head injuries with demonstrable structural brain abnormalities typically accompanied by longer periods of unconsciousness and/or more severe impairments in mental status, and correspond with worse prognosis (Kraus & Chu, 2005). Post concussive symptoms (PCS) are a group of symptoms that can appear in the period immediately following TBI of any seve rity, that broadly affect multiple areas of functioning, including physical, behaviora l, cognitive, and em otional symptoms (See Table 1-1; Kay, Newman, Cavallo, Ezrach i, & Resnick, 1992; Kennedy et al., 2007; 12

PAGE 13

McAllister, 2005; Rao & Lyketsos, 2000; Stei n & McAllis ter, 2009; Vasterling et al., 2009). For the vast majority of individuals who incur a mild TBI, symptoms from these domains will dissipate within three to six months (Rao & Lyketsos, 2000). However, approximately 10-15% of those with mild TBI report impa irments that persist for longer than one year (Rao & Lyketsos, 2000; Stein & McAllister, 2009; Vasterling et al., 2009). Headache is the most frequently reported sym ptom for individuals reporting lingering symptoms at one year, but a surprising num ber of individuals experience lasting problems in cognitive and affective domains as well (see Table 1-1; McAllister, 2005). Table 1-1. Post concussive symptoms: Domains of dysfunction Physical/Behavioral Cognitive Emotional Coordination Attention Anger Dizziness Concentration Anxiety Fatigue Executive function Depression Headache Memory Irritability Noise sensitivity Problem solving Reduced self-esteem Pain Processing speed Sleep (Kay et al., 1992; Kennedy et al., 2007; McAllister, 2005; Rao & Lyketsos, 2000; Stein & McAllister, 2009; Vasterling et al., 2009) It is important to distinguish, however, t he fact that PCS is not a term applied solely to mild TBI, and can appear after TBI of any severity; it is a mistake to equate PCS solely with mild TBI (McAllister, 2005) Yet, the causes of these persistent symptoms in seemingly mild head injuri es are poorly understood. Neurologic exams and neuroimaging studies of patients evin cing persistent PCS can be completely unremarkable (Hoge et al., 2008; Rao & Lyketsos, 2000). However, modern neuroimaging techniques have lead some to believe that when these persistent, multisystem deficits appear in mild TBI they are a product of diffuse brain damage to frontal and temporal lobe white matter through what is known as diffuse axonal injury (DAI), which is caused by shearing and tensile forces that result from sudden deceleration and 13

PAGE 14

rotation of the head, resulting in micro-hemorrhages or lesions to white matter tracts (Maas et al ., 2008; McAllister, 2005; Rao & Lyketsos, 2000; Vasterling et al., 2009). Although this type of white matter damage is common in both the frontal and temporal lobes in mild TBI (Stein & McAllister, 2009; Vasterling et al., 2009), it can be difficult to detect on the computed tomography (CT) scans used immediately after TBI to assess extent of brain damage (Bigler, 2005). Posttraumatic Stress Disorder (PTSD) There is considerable overlap between PCS and those symptom s seen in patients suffering with PTSD, an anxiety disorder that can develop after an individual has been exposed to a trauma that was experienced wit h fear, helplessness, or horror. Diagnostic symptoms of PTSD include recurr ent and intrusive recollections or reexperiencing of the trauma, avoidance of associated stimuli and emotional numbing, and symptoms of persistent physiological arousal including night mares (American Psychiatric Association, 2005). These symptoms must be present for at least one month and must cause impairment in important areas of functioning. However, beyond the emotional symptoms that are diagnostic of PTSD, patients with this disorder are often affected by symptoms in broad areas of functioning including physica l, cognitive, and behavioral domains (see Table 1-2; Kennedy et al, 2007). Current estimates indicate that approx imately 8% of the adult US population will be diagnosed with PTSD at so me point during their lifetime (American Psychiatric Association, 2005). If one takes into account subclinical pres entations of the disorder, lifetime prevalence for the general population may be as high as 23% (Sadock & Sadock, 2007). Whats more, for certain high-risk groups, researchers have estimated lifetime prevalence rates for the disorder as high as 75% (Sadock & Sadock, 2007). 14

PAGE 15

Table 1-2. Symptoms accompany ing PTSD: Domains of dysfunction Physical Cognitive Behavioral Cardiovascular symptoms Attention Alienation Exhaustion Confusion Impaired school performance Gastrointestinal problems Forgetfulness Impaired work performance Headaches Impaired concentration Increased relational conflict Hyperarousal Impaired decision making Reduced relational intimacy Insomnia Impaired learning Social withdrawal Musculoskeletal disorders Memory impairment Startle response Slowed processing speed (Kennedy et al., 2007) Neuroscience research has been conducted examining various cognitive processes in patients with PTSD and other anxie ty disorders (Bremner et al., 2004; Shin et al., 2001) and using animal models of PTSD (Nemeroff et al, 2006). Broadly, neuroimaging research, including functi onal neuroimaging studies with positron emission tomography (PET) and volumetric analyses with magnetic resonance imaging (MRI), suggests dysfunction of the anterior cingulate, dorsolate ral and medial prefrontal cortex, hippocampus, amygdala, thalamus, and insula in pat ients with PTSD, with these areas implicated in both the diagnostic sympt oms of the disorder and the cognitive deficits often reported (Bremner et al., 2004; Nemeroff et al., 2006; Vasterling et al., 2009). Comorbid Mild TBI and PTSD There is widespread belief that PTSD and TBI can co-occur; concomitant diagnosis can be appropriate (Bombardier et al., 2006; Kennedy et al., 2007). Both disorders result from trauma experiences a nd there is striking ov erlap in the symptoms of PTSD and those considered PCS (see Tabl e 1-3; Kay et al, 1992; Kennedy et al., 2007; Parker, 2002; Vasterling et al., 2009). Ho wever, some research indicates that concomitant diagnosis of PTSD is rare after mild TBI, but may be more likely to develop after a physical trauma of a personally s ensitive nature (i.e., mugging, combat; 15

PAGE 16

Bombardier et al, 2006; Kay et al, 1992). Ne vertheless, symptoms of mood disorders are common after TBI (e.g., depression, mania) and have been reported for centuries (Rao & Lyketsos, 2000). Additionally, cert ain behaviors are associated with both disorders, such as substance abuse, suic idal behav ior, and pain conditions (Stein & McAllister, 2009; Vasterling et al., 2009). There are also similarities in the risk factors for both conditions, for example low education, feeling terrified, helpless, of being intoxicated during the traumat ic event (Bombardier et al ., 2006; Stein & McAllister, 2009). Individuals with TBI and PTSD are also more likely to suffer with comorbid mental health diagnosis, such as major depression (Hoge et al., 2008; Sadock & Sadock, 2007; Stein & McAllister, 2009). Recent research indicates that somewhere between 11 to 30% patients with TBI meet cr iteria for PTSD, depending on the sample (Bombardier et al., 2006). Table 1-3. Overlapping symptoms of PTSD and PCS Physical/Behavioral Cognitive Emotional Fatigue/exhaustion Attention problems Anger Headaches Difficulty concentrating Anxiety Noise/startle sensitivity Memory problems Depersonalization Sleep disruption/insomnia Processing speed Depression Irritability (Kay et al, 1992; Kennedy et al., 2007; Parker, 2002; Vasterling et al., 2009) While there has been some debate in the past over whether complete memory for an event is required for PTSD to develop, it appears that even if an individual has a temporary loss of consciousness or post-traum atic amnesia (PTA) after TBI, they can still develop PTSD (Warden & Labbate, 2005). There are several proposed mechanisms by which PTSD can develop in the absence of a complete memory for trauma, including, for example, the presence of tr aumatic memories from before or after the incident, unpleasant experiences during hospita lization or treatment, or being told the 16

PAGE 17

details by others (Kennedy et al., 2007; Warden & Labbate, 2005). In fact, there appears to be a protective effect of PTA th at is more common in moderate and severe TBI, which serves as an explanation as to why PTSD is actually more likely to develop after mild TBI (Bombardier et al., 2006; Kennedy et al., 2007; Vasterling et al., 2009). Howev er, the nature of the comorbidity of PTSD and mild TBI is not completely understood; the DSM-IV-TR does not address how to approach diagnosis when there are overlapping symptoms from both PTSD and post-concussive profiles (Parker, 2002). Additionally, there are currently no clin ical guidelines for treating patients with comorbid TBI and PTSD (Kennedy et al., 2007). There is clear overlap in the symptoms and functional impairments associated with PCS after mild TBI and PTSD that needs to be better understood (Stein & McAllister, 2009). Combat specific exposures The connection between PTSD and TBI has been brought to the forefront of clinical and research interest in the last decade due to the military conflicts in Iraq and Afghanistan (Stein & McAllister, 2009). In fact, some consider TBI and PTSD the signature injuries of Operation Enduring Fr eedom (OEF) in Afghanistan and Operation Iraqi Freedom (OIF) in Iraq (Hoge et al., 2008; Stein & McAllister, 2009). Indeed, the number of TBI cases worldwide has increas ed dramatically since these wars began. A primary reason for the surge in the number of TBIs is due to the astounding advances in medical care for head injuries that allows a la rger number of soldiers to survive injuries that would have been fatal in previous militar y conflicts (Hoge et al., 2008). In fact, TBI is the most frequently reported physical injury sustained by combatants of OEF and OIF (Stein & McAllister, 2009). 17

PAGE 18

Regarding epidemiological estim ates, out of the injured service members who are evacuated for medical treatment, about 25% have sustained a serious head or neck injury (Hoge et al., 2008; Kennedy et al., 2007). Unfortunately, diagnosis of non-severe TBI can sometimes be particularly difficult because mild TBIs are often not formally documented or treated since the immediat e symptoms are typica lly brief loss of consciousness or alteration in mental state, and lasting PTA is rare (Vasterling et al., 2009). Therefore, estimating t he incidence and prevalence of mild TBI in OEF/OIF combatants is difficult. Nevertheless, us ing documented loss of consciousness or altered mental state as criteria, one study estimated that mild TBI occurs in 15% of soldiers returning from duty (Hoge et al ., 2008). Other researchers have proposed prevalence rates of TBI ranging from 5% to 23% in OEF/OIF veterans (Vasterling et al., 2009). Currently, the Department of Defense and the Departm ent of Veterans Affairs are working to collect more accurate data regarding rates of mild TBI in returning veterans, so future estimates should reflect more thorough and accurate documentation (Hoge et al., 2008). PTSD in veterans has also been a long-documented phenomenon, with this syndrome of symptoms appearing in veterans of every modern war (Nemeroff et al., 2006; Sadock & Sadock, 2007). Strikingly, PTSD is appearing at alarming rates in those currently returning from duty, with approxim ately 16-17% of OEF/OIF veterans being diagnosed with PTSD (Sadock & Sadock, 2007; Stein & McAllister, 2009). There appears to be a positive correlation between sustaining a combat injury and developing PTSD (Kennedy et al., 2007). Whats mo re, experiencing a mild TBI is uniquely associated with PTSD, even after controlling fo r the effect of type of combat exposure 18

PAGE 19

and the effect of sustaining a combat in jury of another type (Hoge et al., 2008; Vasterling et al., 2009). Specifically, in those returning from duty who report experienc ing a loss of consciousness, researcher s estimate that 44% also meet criteria for PTSD (Hoge et al., 2008). One confounding factor unique to the c onflicts in Iraq and Afghanistan are the striking rates of TBI with blast injuries as the primary cause. Blast injury due to improvised explosive devices (IEDs) is now seen as a unique and not previously considered mechanism of TBI, with a unique profile of damage to the brain (Maas et al., 2008). In fact, IEDs are reported as the most frequent cause of TBI in Iraq and Afghanistan (Vasterling et al., 2009). The mechanism of damage to the brain due to blast is incompletely understood, yet pathology is assumed to result from the dramatic changes in air pressure as the explosion sh ock wave passes through the brain, causing small vessels in the frontal and temporal lobes to rupture leading to early brain swelling, hemorrhage, and vasospasm (Maas et al., 2008; Vasterling et al., 2009). Regardless, in blast and non-blast TBI, the areas of brain dam aged are typically the same, and in fact, these regions overlap significantly with those regions found to have structural and functional abnormalities in PTSD. These blast-related TBIs are particularly difficult to document because there is often no observable physical contact with the head. In addition these injuries sometimes occur repeatedly within a short time span, making extent of TBI exposure nearly impossible to document. Thus, when symptoms present in a soldier who has been exposed to combat and IED explosion, differential diagnosis of mild TBI and PTSD is particularly tricky. Some preliminary res earch suggests that mild TBI due to blast 19

PAGE 20

injury may be associat ed with worse PTSD symptomatology when compared with nonblast related mild TBI (Vasterling et al., 2009). Differential diagnosis Retuning now to the problems associated wi th differential diagnosis in civilian populations, there are other confounding diagnostic factors, in addition to the overlapping symptom presentat ions, that often render diffe rential diagnosis of PTSD and PCS difficult. For one, motor vehicle a ccidents and assaults are two of the most frequent causes of the trauma causative of both TBI and PTSD (Maas et al., 2008; Nemeroff et al., 2006; Stein & McAllister, 2009). There are also biologic al similarities in the etiologies of PTSD and post concussi ve symptoms; for ex ample, both are associated with decreased function of t he pituitary gland and dysregulation of neurotransmitters (Kennedy et al., 2007). Both disorders evince signs of acute and chronic stress in terms of changes in autonomic arousal and immune responses (Hoge et al., 2008; Parker, 2002; Warden & Labbate 2005), especially in terms of the hypothalamic-pituitary-adrenal axis (Hog e et al., 2008; Kennedy et al., 2007). Researchers have found repeatedly that the presence of PTSD se ems to magnify the cognitive deficits associated with PCS and c an also serve as a hindrance to recovery (Kennedy et al., 2007; Vasterling et al., 2009). Additionally, the brain areas implicated in PTSD are some of the brain areas most vu lnerable to damage in TBI (Kennedy et al., 2007; Vasterling et al., 2009), a point that will be discussed in more depth shortly. Unfortunately, due to the sometimes signifi cant difficulty encountered when trying to differentiate these two symptom profiles, clinicians experience a lack of a precise way to appropriately diagnose all patients. In tu rn, the primary impacts of misdiagnosis are the slowing of treatment in itiation and appropriateness of type rehabilitation selected. 20

PAGE 21

This inability to tailor tr eatment appropriately influenc es quality of care and speed of service delivery, which is critical when considering the recommendat ion that cognitive rehabilitation for TBI begin within 6 months after injury (Rao & Lyketsos, 2000). The speed with which patients can re integrate into their communities is also an important factor. In fact, community ex clusion or feelings of det achment from others can be a problem for both PTSD and TBI patients, which is probably compounded when these disorders occur co-morbidly (Kennedy et al., 2007). Whats mo re, the quandary of differentiating PCS in TBI from PTSD is an instance where so-called symptom management is inappropriate and treatment designed for the specific disorder is required (Rao & Lyketsos, 2000). Clearly, differ ential diagnosis is a critical concern for patient care, beyond the interests of res earchers wishing to document these disease processes. Current methods for differential diagnos is are based on ex amining symptom presentation. Unfortunately, both PTSD and PCS after mild TBI can show mild deficiencies in neuropsychological testing of attention, learning, memory, and executive functioning (Kennedy et al., 2007; Stein & Mc Allister, 2009; Vasterling et al., 2009). Promising recent research, however, is beginning to reveal similarities in the neural basis of these two disorders, with pattern s of deficits shown for both disorders in orbitofrontal cortex, prefr ontal cortex, hippocampus, and anterior cingulate (Stein & McAllister, 2009). An important aspect in understanding the shared neurological correlates of PTSD and PCS is rejection of the notion that PTSD is caused by purely psychological trauma while TBI results from purely physical damage to the brain; conceptualization of PTSD and TBI in te rms of neuropsychologic al and neurological 21

PAGE 22

damage and changes emphasizes the fact t hat cognitio n and affect can both be impacted by biomechanical an d psychological trauma (Stein & McAllister, 2009). The various distributed brain regions implicated in PCS and PTSD, and the neural basis of these disorders and their shared symptoms, is the area on whic h the current research is focused. Specifically, to the extent that so me brain regions show differential impairment in the two disorders, the present research aims to develop a functional neuroimaging approach for differentiating brain activity that might be more impaired in one disorder compared with another. Thus, this research is intended to reflect a first step in developing a paradigm for bet ter understanding PCS and PTSD. Neuroanatomy Relevant to PCS and PTSD Neuroanatomy of cognitive impairment in TBI As mentioned previously, the areas of t he brain most likely to sustain damage in TBI are orbitofrontal cortex, dorsolateral pref rontal cortex, the po les of the temporal lobes, and hippocampus (Stein & McAllister, 2009). These patterns of damage to the frontal and temporal lobes most often affect systems of cognition, behavior, and affect that can result in the appearance of cogniti ve sequelae, personality change, and mood disorders (Rao & Lyketsos, 2000). Changes in patterns of functional brain activation (Perlstein et al., 2004; Sozda, 2009) and in lesion localization (Levin, Williams, Eisenberg, High, & Guinto, 1992) in thes e regions have been documented post TBI. Attention and executive functions of the frontal and temporal lobes are frequently examined as a domain particularly sensitive to mild TBI (Malojcic, Mubrin, Coric, Susnic, & Spilich, 2009). 22

PAGE 23

Neuroanat omy of attention and executive function The neuroanatomical underpinn ings of attention and executive function have been studied in human and animal models for ma ny decades, and the literature broadly implicates the frontal lobes (Alvarez & Emory, 2006; Smith & Jonides, 1999). One task of particular salience in this domain in the Stroop task (Stroop, 1935; for review see MacLeod, 1991). Research using the original color-word Stroop and modified versions of the task (i.e., task switching Stroop) in controls have show n that the anterior cingulate cortex (ACC) and a distributed neural network of other frontal and temporal lobe structures, including the dorsola teral prefrontal cortex (dlPFC ), are critical for accurate performance in the incongruent, or more c ognitively taxing, condition of the task (Alvarez & Emory, 2006; Bush et al., 1998; Gruber, Rogowska, Holcomb, Soraci, & Yurgelun-Todd, 2002; MacDonald, Cohen, Stenger, & Carter, 2000; Markela-Lerenc et al., 2004;). The role of these r egions of the brain in maintaining attention in the face of distraction, also known as the executive control of attention, is well documented (Compton et al., 2003; Kompus, Hugdahl, Oh man, Marklund, & Nyberg, 2009; Stuss, Shallice, Alexander & Picton, 1995). Previous research in this lab has focu sed on the deficits in cognitive control present in patients with moderate and severe TBI (Larson, Perlstein, Demery, & StiggeKaufman, 2006; Perlstein, Larson, Dotson, & Kelly, 2006; Seignourel et al., 2005). Specifically, a variant of the traditional color-word Str oop has been shown to elicit differential activation of the ACC and dlPFC in patients versus healthy controls (Sozda, 2009). This task, the Task-Switching Cued Str oop requires participants to either read the word (red, green, or blue) or name the color of ink used to print it (red, green, or blue). This paradigm instructs participants which task to perform using a cue that 23

PAGE 24

appears several seconds before the stimuli appear s. This cue requires participants to engage working memory systems actively in order to maintain the experimental context The purpose of using this instructional cue is to increase the cognitive demands of the task and to elicit medial-prefrontal brai n activation in terms of ACC and dlPFC (Perlstein, Dixit, Carter, Noll, & Cohen, 2003). Functional neuroanatomy of emotion Research on the neurocircuitry of PTSD tends to focus on limbic structures, including the thalamus, hippocampus, amygdal a, and on the frontal cortex, especially the medial prefrontal cortex regions of anterior cingulate and orbitofrontal cortex (Bremner et al., 2004; Nemeroff et al., 2006). The amygdala has particularly been touted as being important in emotion, in term s of the role emotion plays in learning, memory, attention, perception, and info rmation processing (Phelps & LeDoux, 2005). This brain region is also important in eval uation of stimuli in te rms of emotional and motivational significance (Mohanty et al., 2005). Functional neuroimaging in PTSD has demonstrated altered patterns of responsivity of the amygdala, frontal cortex, anterior cingulat e, and hippocampus during fear processing (Kennedy et al., 2007; Nemero ff et al., 2006; Stein & McAllister, 2009; Vasterling et al., 2009), with some areas sho wing increased activity (e.g., amygdala) and others showing decreased activity (e.g., prefrontal cortex). These changes associated with PTSD or chronic stress are also supported by anatomic volumetric analysis in humans and in animal lesion model s (Kennedy et al., 2007; Nemeroff et al., 2006; Shin et al., 2001; Vasterling et al., 2009). Behaviorally, patients with PTSD are robustly susceptible to an attentional bi as toward threat based information and emotionally salient information (Williams, Mathews, & MacLeod, 1996; Vasterling et al., 24

PAGE 25

2009). When PTSD patients perform a functiona l neuroimaging tas k requiring executive control of attention in the face of emotionally relev ant stimuli, an atypical pattern of brain activation in the anterior cingulate is often not ed, which is assumed to reflect a failure of this region to inhibit a hyper-responsive amy gdala (Bremner et al., 2004; Kennedy et al., 2007; Shin et al., 2001; Stein & McAllister, 2009; Vasterling et al., 2009; Whalen et al., 1998). This decrease in medio-frontal activity in PTSD is typically coupled with an increased in activation of the amygdala (Ken nedy et al., 2007; Stein & McAllister, 2009; Vasterling et al., 2009; Whalen et al., 1998). C onversely, in control subjects, successful performance on this type of task is associ ated with an increase in activation of the prefrontal cortex and anterior cingulate and a decrease in amygdala activation (Kompus et al., 2009). Anterior cingulate cortex (ACC) Taking the neural basis of attention, executive functions, and emotion into consideration in light of t he neural changes seen in mild TBI and PTSD, the ACC stands out as an area of common interest. Anatom ically, the ACC wraps around the frontal portion of the corpus callosum, anterior to the rostrum and super ior to the genu of the corpus callosum, in both hemispheres, and is considered part of frontal cortex (Devinsky, Morrel, & Vogt, 1995). The ACC is strongly associated with processes requiring attention and/or concentration (Bremner et al., 2004; Bush et al., 1998; Stuss et al., 1995). However, the ACC does not have just one function in t he brain, and may be better considered as subdivided into different functional areas (D avis et al., 2005; Devinsky et al., 1995; Mayberg, 1997; Vogt, Finch, & Olson, 1992; Whalen et al., 1998). The rostral ACC (rACC), defined as Brodmanns areas (BA) 33 and rostral area 24, seems to be 25

PAGE 26

particularly involv ed in em otion-based processing and responding, and the caudal ACC (cACC), defined as BA 32 and caudal area 24, is implicated in performance of cognitive tasks (Devinsky et al., 1995). These divi sions are supported by the anatomical projections of the regions, with the affective or rACC having di rect afferent and efferent connections with the amygdala and the hypothalamus, and the cognitive or cACC having many connections with the prefrontal cortex and motor systems (Bush et al., 1998; Devinsky et al., 1995; Vogt et al., 1992). This division is also supported by patterns of deficits after brain damage, lesion studies in animals, research using single cell recordings in humans, and by patterns documented with functional neuroimaging in patients and controls (Bush et al., 1998; Da vis et al., 2005; Devinsky et al., 1995; Mayberg, 1997; Mohanty et al., 2005; Whalen et al., 1998). The neuroimaging research that supports these divisions does so by demonstrating greater rACC activation durin g emotionally salient, compared with neutral, aspects of tasks requiring emotiona l expression, emotiona l learning, or with emotionally arousing stimuli (Whalen et al., 1998). On the other hand, the caudal division demonstrates unique activation dur ing tasks requiring conflict detection, response competition, and salience detection of non-emotionally ar ousing stimuli (Bush et al., 1998). The existence of these two s ubdivisions of the ACC is well accepted. Some consider the purpose of the ACC, theref ore, as a center for the interaction of affect and intellect in the brain (Bush, Luu, & Posner, 2000; Devinsky et al., 1995). Emotional Stroop Tasks Tasks to activate either division of the ACC in neuroimaging research often involve some form of decision making and conflict t hat require the subject to override a prepotent response tendency, for example, t he Stroop task, flanker tasks, and certain 26

PAGE 27

Go/No-Go tasks. Many tasks used to study th e rACC are variations of what is known as an emotional Stroop. While the basis of the interference in the traditional color-word Stroop is the antagonistic combination of colors of ink and color words, the interference in emotional Stroop tasks comes from usi ng emotionally arousing words as stimuli (Compton et al., 2003; Williams et al., 1996). Thus, the so-called emotional Stroop effect is the increased latency to name the color of ink used to print the stimuli when the words are emotionally salient (Williams et al., 1996). While there has been some debate about the appropriatenes s of considering these tasks Stroop tasks (see: Algom, Chajut, & Lev, 2004 and Dagliesh, 2005) there appears to be relatively strong consensus as to the utility and suitability of using these tasks for their stated purposes (Williams et al., 1996). Emotional Stroop tasks have a long history of use in popul ations of patients with anxiety disorders (Gotlib & McCann, 1984; Mathews & MacLeod, 1985; Watts, McKenna, Sharrock, & Trezise, 1986). Studies using emotional Str oop tasks to examine performance of patients with various disor ders have shown interference effects for depression, specific phobia, conduct disor der, PTSD, obsessive-compulsive disorder, as well as state and trait anxiety when the stim uli are targeted at the concerns of the participants (i.e., spider related words for spider phobics; Compton et al., 2003; PerezEdgar & Fox, 2003; Vythilingam et al., 2007; W illiams et al., 1996). Beyond the fact that the type of words that produce an interference effect can serve as a marker for the patients condition, the emot ional Stroop effect has also shown improvement with desensitization treatment for anxiety disorders (Williams et al., 1996). 27

PAGE 28

More recen tly, emotional Stroop tasks have been developed for use in functional neuroimaging of patient groups and controls (B remner et al., 2004; Bush et al., 1998; Compton et al., 2003; Malhi, Lagopoulos, Sachdev, Ivanovski, & Shnier, 2005; Shin et al., 2001; van den Heuvel et al., 2006; Whal en et al., 1998). In controls, these experiments demonstrate increased rACC acti vation to negative stimuli when compared with neutral stimuli (Compton et al., 2003; Whalen et al., 1998), along with increased activation of orbitofrontal cortex and amygdala (Compton et al., 2003). In patients with anxiety diso rders, the pattern of brain activation to disorderrelevant, emotionally negative stimuli is differ ent for that seen in controls. Specifically, when compared with controls, patients do not show the expected activation of rACC (Bremner et al., 2004; Shin et al., 2001). This pattern is often inte rpreted to signal a disruption in the regulatory mechanism of rACC in modulating the activity of the amygdala in response to the emotionally val ent stimuli (Bremner et al., 2004; Kennedy et al., 2007; Shin et al., 2001; Stein & McAllis ter, 2009; Vasterling et al., 2009; Whalen et al., 1998). However, increased amygdala acti vation to emotionally valent stimuli has not been observed in all studies (i.e., Bremne r et al., 2004). Recent behavioral studies using emotional Stroop tasks in TBI have dem onstrated a behavioral interference effect for trauma reminders in braininjured patients compared with controls (Coates, 2008). This indicates the feasibility of using this type of paradigm, even in neurologic patients without explicit memories of their tr auma experience, and again highlights the connection between TBI and PTSD (Coates, 2008). There is some debate as to whether pos itively valent words produce the same type of emotional Stroop effect as do negat ive words, although all studies compare 28

PAGE 29

performance to conditions using emotiona lly neutral words (McKenna & Sharma, 1995; Williams et al., 1996). Howev er, it appears that in healthy controls it is common to see no interference effect for positive and neutral st imuli, while an interference effect has been shown for positive words in patients with anxiety (Compton et al., 2003; McKenna & Sharma, 1995; Williams et al., 1996). So me studies do not examine the emotional Stroop effect in positive stimuli (e.g., Whal en et al., 1998). However, this point may be moot for the current purposes because t he small sample sizes used in functional neuroimaging often preclude detection of an interference effect on reaction time (Bremner et al., 2004; Whalen et al., 1998). Yet, in large behavioral studies using these tasks, interference effects are on the order of milliseconds and can range from 23 msec to 190 msec, for example, depending on the patient population under study (23 msec for mild depression: Gotlib & McCann, 1984; 190 msec for spider phobic: Watts, McKenna, Sharrock, & Trezise, 1986; Williams et al., 1996). In fact, Compton et al. note that the effect size for the interference seen for negative high-arousal words in another study was Cohens d = 0.33 and for positive high-arousal words was Cohens d = 0.24 (2003). Study Aims and Predictions This study was designed as a proof of c oncept for a task that could be used as a functional neuroimaging probe for the prefrontal-cingula te and amygdala-cingulate networks seen to been implicated in the def icits/symptoms of PCS in TBI and PTSD. Additionally, this neuroimaging probe might som eday have value as a tool for differential diagnosis of PTSD and TBI. This notion is bu ilt on the fact that there are documented patterns of functional activation of ACC in both PTSD and TBI using emotional Stroop (e.g., Shin et al., 2001) and cognitive Stroop (Sozda, 2009) tasks, respectively. 29

PAGE 30

30 Therefore, if the task used in this study, the Cued Emot ional Counting Stroop, can reliably activate rACC in controls, then it may provide a reliable method for differentiating PTSD from PCS based on patterns of neural activation. Behavioral Data In terms of behavioral performance during th is task, a pattern of slower reaction time and higher error rates to emotionally unpleasant words was predicted, based on the reasoning that these em otionally unpleasant words detract most from task performance. Such a finding has been reported in other studies using emotional Stroop tasks, though it is not particula rly robust in non-clinical populations (e.g., Coates, 2008; Whalen et al., 1998). Functional Magnetic Resonanc e Imaging (fMRI) Data Using functional neuroimaging to measure task-related brain activity, it was hypothesized that rACC activation would be pr esent during the component of the task requiring emotionally salient decision-making. Additionally, amygdala activation associated with the emot ionally salient probe stimuli was also expected. It was also hypothesized that the instruct ional cues would engage the dorso lateral prefrontal cortex, given the involvement of this region in wo rking memory and context maintenance. In conclusion, the primary aim of this study is to evaluate the patterns of brain activation when comparing pleasant and unpleasant emotionally arousing stimuli to neutral stimuli. A secondary aim is to evaluate the patterns of brain activation related to working memory or context maintenance requi red by the instructional cue.

PAGE 31

CHA PTER 2 METHODS Participants Four healthy, right handed, native Englis h-speaking adults we re recruited with flyers from the Gainesville community. Volunteers were screened for a history of traumatic brain injury and other neurolog ical problems, psychiatric illness, colorblindness, and MRI contraindications (e.g., ferromagnetic implant, claustrophobia). Female participants were screened for pregnancy. Participants were excluded if they endorsed a history of drug or alcohol abuse or were taking psychotropic medications. No subjects were excluded using these cr iteria. Three women participated along with one man and all participants were Cauc asian. Mean subject age was 26.3, SD = 4.6, range = 22-34, and average education co mpleted in years was 17.5, SD = 0.9, range = 16-18. All participants provided informed consent in accor dance with the guidelines of the Health Sciences Center Institutional Revi ew Board at the University of Florida. Materials and Procedures FMRI Cognitive Task: Cued Emotional Counting Stroop Participants completed a variant of t he emotional Stroop task (e.g., McKenna & Sharma, 1995; Williams et al., 1996). Em otional Stroop tasks have been designed as variants of the traditional St roop task (Stroop, 1935) but have used, in place of color words, emotionally pleasant, em otionally neutral, and emoti onally unpleasant words. In the task designed for the present study, stimuli wo rds were printed in red, green, or blue ink, and were shown one, two, or thr ee times on a screen. The task required participants either to count t he number of times the stimulus word was written or to name the color of ink with which the stimulus was printed. Participants were cued to 31

PAGE 32

which of the two tasks they should perform by an instruction screen appearing several seconds before the probe stimuli appeared. This instruction screen was either the word Color or Number and indic ated if the s ubject should count the number of times the word was printed or name the color of ink. The task is depicted in Figure 2-1. Figure 2-1. Diagram of Cued Emotional Co unting Stroop task. The task used in the present study consisted of a cue that in structed subjects to either name the color of ink or count the number of time s the word was printed. Stimuli were pleasant, unpleasant, and neutral words. Regarding timing parameters, this task required maintenance of task instructions during an 11-second delay. Subjects responded to a probe word usi ng a button press with the right hand. Probe stimuli were presented in red, gr een, or blue ink. An 8.5-second intertrial interval (ITI) was presented at the end of each trial and consisted of a fixation cross. Each trial comprised the following sequenc e: a cue word (number or color) presented for 1.5 seconds, 11 seconds of visual fixation, the stimulus or probe word for 1.5 seconds, and then 8.5 seconds of vi sual fixation. This trial sequence was 32

PAGE 33

repeated 18 times per block, and the entire ex periment was presented as six blocks, each separ ated by periods of rest. Partici pants were presented with the six blocks in random order and trials were presented in a fixed, predetermined random order within blocks. While fMRI data were acquired, participants completed this task using a button press to one of three buttons with the right hand. Behavioral data (i.e., reaction times and error rates) were collected during this task for subsequent eval uation of participant performance. Before beginn ing the experiment, participants were given a practice period outside of the MR environ ment to familiarize them with mapping the colors red, green, and blue and the num bers one, two, and three, to the fingers index, middle, and ring, respectively. The aspect of this task design that disti nguished it from other previously described emotional Stroop tasks was the presence of a task instruct ional cue. The goal of including this instructional cue was twofol d: (1) it allowed the two tasks (counting and color naming) to be intermixed, and thus required participants to be continually monitoring the experimental context, and (2) it introduced a working memory aspect to the task in that the instructions had to be ma intained in memory during the initial delay before the probe stimulus app eared. Both of these goals were relevant because previous experiments in this same lab have used a similarly designed cognitive task (i.e., Task-Switching Cued Stroop) to demonst rate patterns of functional activation in moderate to severe TBI patients and contro l participants (Sozda, 2009). Including the instructional cue increases the cognitive demand of this task, making it more sensitive to cognitive impairment than a typical emotional Stroop ta sk. Thus, the experiment 33

PAGE 34

designed for the present study was created with the intention of replicating the parameters used in previous inv est igations of TBI in this lab. Stimuli The 108 stimuli words used in this task we re taken from a standard list of words known as the Affective Norms for Englis h Words (ANEW; Bradley & Lang, 1999), and have been included as an Appendix to this manuscript. The words selected from the ANEW set represented an equal pr oportion from the three affe ctive valence categories: pleasant, neutral, and unpleasant. Thirty-six words were selected from each valence category, and the pleasant and unpleasant words were selected to be high in rated arousal. The mean valence and arousal ratings for the stimuli word s used are shown in Table 2-1 and a graph plotting the mean val ence and arousal ratings for the 108 words used in the present study is shown in Figure 2-2. Values were calculated based on the normative data provided with the AN EW set (Bradley & Lang, 1999). Table 2-1. Valence and arousal ratings of stimuli words based on normative data Valence Arousal Category M SD M SD Pleasant 7.90 0.57 6.34 0.60 Neutral 5.23 0.52 4.12 0.24 Unpleasant 2.09 0.39 6.46 0.63 These values were calculated from the normative values provided with the ANEW manual (Bradley & Lang, 1999). Words from the three valenc e categories were also chosen to be equal in terms of word frequency, as measured by Kucera-Fr ancis written word frequency (Kucera & Francis, 1967). A one-way analysis of variance (ANOVA) demonstrated no statistically significant difference in written word frequency for the three valence categories, F (2, 107) = .258, p = .773, using the values for Ku cera-Francis written word frequency provided with the ANEW normative information (Bradley & Lang, 1999). 34

PAGE 35

After compl eting the experiment and fMRI scanning session, participants rated each of the 108 words according to subjecti ve emotional valence and arousal using a computerized version of the Self Assessm ent Manikin (SAM; Bradley & Lang, 1994). The purpose of this rating process was as a manipulation check. Figure 2-2. Graph of ANEW stimuli: Valence by arousal. The words used as stimuli in the present study were taken from t he ANEW standardized word-set. This set included valence and arousal ratings by word. This graph shows a plot of the 108 words used in this study plotted by mean arousal and valence ratings. Values taken from ANEW database (Bradley & Lang, 1999) Image Acquisition MRI scanning was performed on the research -dedicated Phillips 3T MR scanner at the University of Florida McKnight Brain In stitute using a sensit ivity encoding (SENSE) radio frequency (RF) head coil. Prior to fu nctional scanning, a high-resolution T1weighted three-dimensional (3D) Magnetization Prepared Rapid Gradient Echo (MPRAGE) anatomical scan [Echo time (TE) of 4. 3 ms, repetition time (TR) = 2000 ms, field of view (FOV) = 24 cm, flip angle (FA) = 8, matrix size = 512 x 512 mm, 176 1-mm thick sagittal slices] was acquired to enable evaluation of structur al abnormalities and for transformation of functional data into standard reporting space (Talairach & 35

PAGE 36

Tournoux, 1988). High-order shimming was performed prior to functional image acquis ition in order to maximize field homogeneity. Functional images were acquired in 34 c ontiguous axial slices parallel to the anterior commissure-posterior commissure (AC-PC) line using a T2*-weighted echo planar imaging (EPI) pulse sequence (FOV = 24 cm, matrix = 64 x 64 at 3.75 mm3, TR = 2500 ms, TE = 30 ms, FA = 90). Scan acquisition was synchronized to each trialevent onset. Thus, the event-related fMRI ac quisition protocol acqu ired 9 volumes every 22.5 second trial, for a total of 972 functional volumes. Data Analysis Behavioral Data Analysis SAM ratings of valence and arousal were analyzed to determine the presence of a significant linear trend for valence ratings according to word category, and for the presence of a significant quadratic trend for arousal ratings according to category. This manipulation check would demonst rate that the participants were rating the valence and arousal of the words used in the experim ent according to the same parameters for which the words were selected. Task-performance error rates and reaction times (RT) on the Cued Emotional Counting Stroop task were analyzed separately using univariate mixed-model 2 3 analyses of variance (ANOVAs) to determine t he effects of task (2 : color, number) and valence (3: pleasant, neutral, unpleasant). M edian correct-trial RTs were analyzed to minimize RT dispersion (Ratcliff, 1993), and error rate s were analyzed by excluding trials with no response in order to minimize the positive skew of the distribution. 36

PAGE 37

Functional Image Data Reduction and Pre-Processin g Functional imaging data were processed using BrainVoyagerQX version 1.10.4 (Brain Innovation, Maastricht, The Net herlands; Goebel, Esposito, Formisano, 2006). Preprocessing of functional images consist ed of the following steps: rigid body 3D motion correction with tri-linear interpolati on, slice-scan time correction using sinc interpolation to account for timing variability across slice acquisitions, spatial smoothing with a 3D 8mm full-width half maximum (FWH M) Gaussian kernel, vo xel-wise linear detrending, and high-pass filtering of frequencies below 3 cycles per time course to remove low-frequency nonlinear drifts. Initial co-registration of functional im ages and high-resolution 3D anatomical volumes was completed using BrainVoyager QX software automatic co-registration alignment, and when necessary, manual alignment based on visual inspection was done. All images were spatially transformed into standard stereotactic Talairach space (Talairach & Tournoux, 1988) using a nine l andmark technique in order to allow for group-wise analysis of functional images. Functional Imaging Data Analysis Functional images were analyzed using a two-level mixed-model general linear modeling (GLM) approach (Friston et al., 1995). First, a separate fixed-effect GLM was specified for each participant, with separate pr edictors for each trial event (cue, probe), and condition (i.e., cue: color, number; probe: pleasant, neutral, unpleasant). Thus, a total of six predictors were used to examin e task-relevant effects. The hemodynamic response for each event was estimated by convolving each regressor with a standard gamma function (Boynton, Engel, Glover, & H eeger, 1996; Delta = 2.5; Tau = 1.25). For each voxel and each trial event, a parameter estimate ( ) was generated that indicated 37

PAGE 38

38 the strength of covariance between the data and the hemodynamic response function (HRF). Contrasts between parameter estimate s for different events of interest were calculated for each participant. Se cond, pair-wise contrasts between s for different events/conditions of interest were calcul ated for each participant and the results were submitted to group analyses that treated inte r-subject variability as a random effect, enabling generalization to population-level inferences. For all analyses, statistical maps were set to a threshold of p < .025 and a 3D cluster-size contiguity threshold of 20 voxels to control for type I error, providing a corres ponding image-wise false positive rate of p <.05, corrected for multiple com parisons (Forman et al., 1995). Planned statistical contrasts separately examined activity associated with the attentional component of contex t maintenance (i.e., cue-rela ted activity) and probe-word valence and arousal. Statistically defi ned region-of-interest analyses were also conducted on the resultant cluster-wise parameter estimates ( s) to determine the specific sources of co ndition-related effects.

PAGE 39

CHA PTER 3 RESULTS Behavioral Results Self-Assessment Manikin (S AM) Ratings Analyses SAM ratings were acquired as a manipulat ion check to determine that the words selected as stimuli were, indeed, effective at evoking the intended affective responses in the present sample of participants. Thus, SAM rating data were analyzed to determine that they conformed to the ex pected trends. Participant-rated mean valence and arousal ratings for words used in the pr esent study are provid ed in Table 3-1. Table 3-1. Participant SAM ratings: Mean valence and arousal by category Valence Ratings Arousal Ratings Category M SD M SD Pleasant 6.8a 1.6 3.7 2.4 Neutral 5.1b 0.9 1.7c 1.5 Unpleasant 2.2a,b 1.4 5.2c 2.5 Based on Bonferroni corrected post-hoc tests, these pairs of mean ratings differed significantly: a p = .002, b p = .017, c p = .017 Valence ratings Valence ratings were analyzed using a repeated measures mixed-model ANOVA to determine if mean valence rating varied according to the fixed effect valence category, accounting for the r andom effect of subject. A planned polynomial contrast was also conducted to determine if this re lationship was linear, such that the highest valence ratings were given to stimuli in the pleasant category and the lowest valence ratings to those in the unpleasant category. Bonferroni corrected post-hoc comparisons were also conducted in order to determine wh ich categories differed significant by mean valence rating. This analysis demonstrated a significant main effect of category on mean participant valence rating, F(2, 6) = 22.1, p = .002. Furthermo re, the polynomial contrast 39

PAGE 40

revealed that this effect was linear, t(6) = 6. 56, p = .001. Examinati on of the B onferroni corrected post-hoc tests revealed that unpleasant words were rated by participants as being of significantly lower valence than both neutral, t(7) = 4.21, p = .017, and pleasant words, t(7) = 6.57, p = 002, but that neutral and pleasant words were not rated as having significantly different valences, t(7) = 2.36, p = .170. Figure 3-1 illustrates this linear trend. These results generally confirm that the selection of words as pleasant, neutral, and unpleasant stimuli was consistent wit h the stated aims and that participants considered the valence level of the stim uli similarly to what was expected based on published normative data (Bradley & Lang, 1999). Thus, based on this manipulation check, the valences of the stimuli employed were different ially rated by the participants of the present study as a function of pleasant, neutral, and unpleasant valence. Figure 3-1. Participant SAM ratings: Mean va lence. The mean valence rating is linearly related to the word valence category. Arousal ratings Arousal ratings were also analyzed us ing a repeated measures mixed-model ANOVA to determine if mean arousal rating va ried according to the fixed effect of 40

PAGE 41

valence cat egory, accounting for the random effect of subject. A planned polynomial contrast was also conducted to determine if this relationship was quadratic, such that the highest arousal ratings were given to stimuli in the pleasant and unpleasant categories and the lowest arousal ratings to those in the neutral category. Bonferroni corrected post-hoc comparisons were also conducted in order to determine which categories differed significant by mean arousal rating. This analysis demonstrated a significant main effect of category on mean participant arousal rating, F (2,6) = 8.96, p = .016. Furthermore, t he polynomial contrast revealed that, as predicted, this effect was quadratic, t (6) = 3.79, p = .009. Examination of the Bonferroni corrected post-hoc tests re vealed that unpleasant words were rated by participants as being of significantly higher arousal than neutral words, t (7) = 4.22, p = .017, but that neutral and pleasant words, t (7) = 2.33, p = .176, and pleasant and unpleasant words, t (7) = 1.89, p = .322, were not rated as having significantly different arousal ratings. Figure 3-2 illustrates this quadr atic trend. These results similarly confirm that the selection of words as pleasant, neutral, and unpleasant stimuli was consistent with the stated aims and that participants considered the ar ousal level of the stimuli similarly to what was expected based on published normative data (Bradley & Lang, 1999). Thus, based on this manipulation check, the arousal ratings of the stimuli employed were rated by the participants of t he present study as a function of pleasant, neutral, and unpleasant valence, with pleasant and unpleasant words having higher arousal ratings. 41

PAGE 42

Figure 3-2. Participant SAM ratings: Mean ar ousal. The mean arousal rating is related to the word valence category by a quadr atic trend, such that the pleasant and unpleasant stimuli have the highest arousal ratings. Task Performance Analyses Participants reaction time (RT) and e rror rates were examined to determine the influence of task conditions on performance. Correct-trial RTs and error rates were evaluated separately using two repeated meas ures mixed-model 2 3 ANOVAs with task instruction (2: number, color) and probe word valence (3: pleasant, neutral, unpleasant) serving as within-subject fixed fa ctors, and controlling for the random effect of subject. Reaction time (RT) analysis A repeated measures mixed-model 2 3 ANOVA was conducted in order to determine if there was a significant main effect of task c ondition or valence category, controlling for the random effect of subject, on correct-trial RT. The results indicate a statistically significant main effect of task, F (1,15) = 25.1, p < .001, and a non-significant main effect of stimulus category, F (2,15) = .406, p = .673. The task valence category interaction was not significant, F (2,15) = 1.75, p = .207. These results indicate that RTs 42

PAGE 43

were significantly different across task condi tion, such that participants responded faster on average to the probe when t he instruction was to count the number of times the stimulus was printed, M = 792 msec, SD = 66.4 msec, compared to when the instruction was to name the color of ink used to print the stimulus, M = 941 msec, SD = 109 msec. These results also indicate that stimulus valence had no measurable effect on participant RT. Means and standard deviations of correct-trial mean RTs are presented according to stimulus valence in Table 3-2. Error rate anal ysis A second repeated measures mixed-model 2 3 ANOVA was conducted in order to determine if there were significant main ef fects of task condition or valence category, controlling for the random effect of subject, on error rates exclud ing errors due to nonresponse. The results indicate no statisti cally significant main effect of task, F (1,15) = .137, p = .717, or stimulus valence category, F (2,15) = 1.97, p = .174. The task valence category interaction was also not significant, F (2,15) = .341, p = .717. These results indicate that error rates did not di ffer significantly across task condition. These results also indicate that stimuli valence had no measurable effect on participant error rates. Means and standard deviations of error rates, excluding trials with no response, are presented according to stim ulus valence in Table 3-2. Table 3-2. Cued Emotional Count ing Stroop: Behavioral results Attribute Valence Mean SD Reaction Time* (msec) Pleasant 870.3 58.3 Neutral 858.8 65.6 Unpleasant 854.1 42.0 Error Rate (%) Pleasant 21.0 7.2 Neutral 14.6 7.6 Unpleasant 12.5 4.8 Note: No significant differences by valence category ( p > .05) *Correct trials only; Excluding trials with no response 43

PAGE 44

FMRI Results Head Movement Head movement in all six dimensions (x, y, z, pitch, roll, and yaw) was analyze d. Average motion in each direction was less than 1 voxel dimension (3.75mm or 1 degree). Two separate mixed-model 2 3 AN OVAs were conducted to determine if there was a significant difference in inter-s can displacement for both mean translation (x, y, and z) and mean rotation (pitch, roll, and yaw) according to task instruction condition and stimulus valence category, ta king into account the random effect of subject. These analyses revealed no statistically significant differences in translational movement by task, F (1,15) = .026, p = .874, or valence, F (2,15) = .524, p = .603, and no significant interaction, F (2,15) = 2.55, p = .112, as well as no significant differences in rotational movement by task, F (1,15) = .704, p = .415, or valence, F (2,15) = .421, p = .664, and no significant interaction, F (2,15) = 1.28, p = .308. Means and standard deviations of average movement in each direction are shown in Table 3-3. Table 3-3. Average head motion by valence category Translation (mm) Rotation (degrees) Valence Category Mean SD Mean SD Pleasant 0.013 0.005 0.015 0.002 Neutral 0.012 0.005 0.016 0.003 Unpleasant 0.012 0.003 0.015 0.001 Note: No significant differences in head motion by valence category. Translation consists of movement in the directions x, y, and z and is measured in mm. Rotation consists of the motions of pitch, roll, and yaw and is measured in degrees. Probe-Related Activation The clusters of brain voxels that showed statistically significant activation during the probe-word portion of the Cued Emotional Counting Stroop task are shown in Table 3-4. These clusters of activation were re vealed by contrasting pleasant and unpleasant stimuli versus neutral stimuli in the GLM. A cluster threshold of 20 contiguous voxels 44

PAGE 45

and a statistical threshold of p <.025 was used to identify ROIs. Areas of positive activation are produced by increased signa l contribution from the pleasant and unpleasant conditions. Three statistically def ined ROIs are shown graphically as fMRI overlay maps on an all-subject averaged brain. Figure 3-3 shows brain activation in bilateral ACC, specifically Brodmanns areas (BA) 32 and 9 (Talairach & Tournoux, 1988). Figure 3-4 shows activation in bilatera l precuneus, specifically BA 7, and Figure 3-4 shows bilateral amygdala activation. Table 3-4. Areas of significant probe-relate d activity for emotionally arousing stimuli Talairach Coordinates Area Hemisphere X Y Z Brodmann Area Amygdala R 22 -10 -12 Amygdala L -22 -15 -11 Anterior cingulate cortex (ACC) L -2 41 16 32/9 Hippocampus R 27 -8 -16 Inferior frontal gyrus (IFG) R 39 26 6 45 Inferior frontal gyrus (IFG) L -45 26 12 45 Insula R 29 17 -5 Insula L -33 15 -4 Medial precuneus L -2 -61 31 7 Middle frontal gyrus R 47 22 31 9 Superior frontal gyrus L -32 19 50 9 Supramarginal gyrus R 48 -65 32 39 Supramarginal gyrus L -48 -66 32 39 Note: R = right; L= left. X, Y, and Z are coordinates in standard stereotactic space (Talairach & Tournoux, 1988). Brodmann's areas are based on Talairach and Tournoux atlas maps (1988). This table shows the Talairach coordinates for the center of mass of the various volumes activated by the probe. These regions are significantly more active during the pleasant and unpleasant stimuli compared with the neutral. Cluster threshold was set for 20 contiguous voxels, p < .025. Cue-Related Activation The clusters of brain voxels that showed statistically significant activation during the task-instructional cue portion of the Cued Emotional Countin g Stroop task are shown in Table 3-5. No contrast betw een conditions was used; instead, the cue regressor for each condition was entered into the GLM. A cluster threshold of 20 contiguous voxels and a st atistical threshold of p < .005 were used to identify ROIs in 45

PAGE 46

the cue portion of the task. Most important w ith respect to the current study is that the task-instructional cue slide was associated wi th activation of brain regions typically associated with context maintenance and work ing memory, including the dorsolateral prefrontal and parietal cortices. Table 3-5. Areas of significant cue-relat ed activity for emotionally arousing stimuli Talairach Coordinates Area Hemisphere X Y Z Brodmann Area Inferior parietal lobule L 37 -49 51 40 Middle frontal gyrus L -34 30 30 46/9 Superior frontal gyrus R 21 33 32 9 Superior parietal lobule R 33 -51 52 7 Note: R = right; L= left. X, Y, and Z are coordinates in standard stereotactic space (Talairach & Tournoux, 1988). Brodmann's areas are based on Talairach and Tournoux atlas maps (1988). This table shows the Talairach coordinates for the center of mass of the various volumes activated by the task cue. Cluster threshold was set for 20 contiguous voxels, p < .005. 46

PAGE 47

Figure 3-3. Image of bilateral ACC fMRI activation. FMRI statistical overlay map showing significant clusters of activity in bilateral ACC associated with the pleasant and unpleasant probe stimuli condition versus neutral condition for all subjects (threshold: p < .025 and 20 contiguous voxels). 47

PAGE 48

Figure 3-4. Image of bilateral precuneus fMRI activation. FMRI statistical overlay map showing significant clusters of activi ty in bilateral precuneus associated with the pleasant and unpleasant probe stimuli condition versus neutral condition for all subjects (threshold: p < .025 and 20 contiguous voxels). 48

PAGE 49

49 Figure 3-5. Image of bilateral amygdala fMRI activation. FMRI stat istical overlay map showing significant clusters of activity in bilateral amygdala associated with the pleasant and unpleasant probe stimuli condition versus neutral condition for all subjects (threshold: p < .025 and 20 contiguous voxels).

PAGE 50

CHA PTER 4 DISCUSSION Emerging clinical research has highl ighted a unique need to understand the differences and similarities between the symptoms known as PCS and PTSD, in particular in veterans returning from the cu rrent wars where these profiles often appear concomitantly. Functional neuroimaging has begun to take up this cause; the current study aimed to determine the efficacy of the Cued Emotional C ounting Stroop task for activating regions of the brain implicated in both PCS and PTSD. The results indicate that this task was indeed successful at activating the ACC, bilateral amygdala, and other regions of the frontal and parietal lobes important in attention and emotion. Review of Findings Behavioral Data Considering performance on t he Cued Emotional Counting Stroop task in terms of RT and error rates, the results revealed no significant behavioral differences due to stimuli word valence. However, there was a statistically significant effect of task on RT, which reflects the fact that color naming is a more taxing task than simple counting. Nevertheless, results of the experimental task did not show the predicted increase in RT or error rates for more emotionally salient stimuli (i.e., pleasant and unpleasant words). While this pattern of interference has been documented repeatedly using patient groups in other versions of the emotional Stroop ta sk (e.g., Constans et al., 2004; Williams et al., 1996), it is not uncommon for other inve stigators using emotional Stroop tasks to find no significant differences in RT or erro r rates for control subjects (e.g., Coates, 2008; Whalen et al., 1998). The effect sizes noted by Compton et al. (2003) regarding the interference for unpleasant and pleasant words are d = 0.33 and d = 0.24, 50

PAGE 51

respectively, which suggests that in order to find a statistically signif icant interference effect in this study, the sample si ze would have had to be much larger. However, it is worth mentioning that t he pattern of valence-related RT and errorrate effects was not in the predicted direction (see Table 3-2). In fact, the responses were fastest to the unpleasant stimuli, althou gh the differences in RT by valence were not statistically significant. Due to lack of significant differences, there is no clear approach for interpreting these results. Howeve r, it remains plausi ble that with greater statistical power an interference effect or a fa cilitation effect of emotional valence might be observed. Both interference and facilit ation of emotional arousal have been documented in controls (e.g., McKenna & S harma, 1995; Perez-Edgar & Fox, 2003), yet there are no complete theories for why t hese patterns sometimes appear. Nevertheless, the current study precludes dr awing further conclusions on this point due to the absence of statistically signific ant differences in RT. FMRI Data Probe stimuli related activation As predicted, when examining probe stimulus related activation associated with correct trial performance, t here was greater medial pref rontal, including ACC, and amygdala activation associat ed with the pleasant and unpleasant probe conditions versus the neutral probe condition. This finding is consistent with numerous studies that show increased amygdalar (Isenberg et al ., 1999; Malhi et al., 2005) and medial prefrontal cortical (Compton et al., 2003; Kompus et al., 2009; Malhi et al., 2005) activation to emotionally arousing words. These results suggest that in order to successfully complete the Cued Emotional C ounting Stroop, medial prefrontal areas (i.e., ACC) are engaged to a greater degree to complete the color naming and counting 51

PAGE 52

tasks when the probe stimuli ar e emotionally arousing (i.e., pleasant and unpleasant) versus when the probe stimu li are emotionally neutral. Furthermore, these results suggest that both pleasant and unpleasant words reliably activate bilateral amygdala when compared to neutral words. Instructional cue related activation As predicted, when examining instructional cue related activation associated with correct trial performance, ther e was significant prefrontal and parietal cortical activation across task and valence condition. Activation of these regions is consistent with a host of functional imaging and human lesion studies demonstrating a critical role for these regions in working memory and context ma intenance (e.g., Braver, Barch, & Cohen, 1999; Miller & Cohen, 2001; Stuss et al., 1995). These results suggest that in order to maintain the task instructions and experimental context in memory successfully, during the delay before the probe stimulus appeared participants engaged prefrontal (i.e., dlPFC) and parietal cortices, regardless of ta sk instructional condit ion. Furthermore, these results suggest that both color naming and word counting, when presented intermixed, required active context main tenance in the Cued Emotional Counting Stroop; that is to say neit her task was automatic (i.e., requiring no cognitive control) because the two tasks were presented r andomly in mixed blocks. Based on the behavioral results, it appears that the word c ounting trials were easier (i.e., faster RT) when compared to the color naming trials. However, when examining the functional neuroimaging results, co llapsed across trial type, a signifi cant pattern of frontal and parietal signal was present. This indicat ed that both trial types were related with engagement of neural regions associated with working memory and cognitive processing, despite a clear RT effect. Toget her, these findings sug gest that there are 52

PAGE 53

medial prefrontal-amygdalar circ uits at wo rk during the Cued Emot ional Counting Stroop task, which indicates that this task could feasibly serve to illuminate neuroanatomical deficits seen in patients with PCS and PTSD. Implications Various emotional Stroop tasks used in functional neuroimaging studies have been shown to be effective for eliciting acti vation of brain areas important in both cognition and emotion (Bremner et al., 2004; Bush et al., 1998; Compton et al., 2003; Malhi et al., 2005; Shin et al., 2001; van den Heuvel et al., 2006; Whalen et al., 1998). However, none so far has incorporated an in structional cue component that requires active maintenance of task instructions in or der to complete the task successfully. The current study demonstrated that this Cued Emotional Counting Stroop task could be used in the context of fMRI to activate b ilateral amygdala, and ACC, and other regions of the frontal lobes important in attention and emotion. The unique attributes of this task are not trivial. Specif ically, using an instructional cue allows this task to be presented as mixed blocks, with color naming and word counting intermixed. This mixed design incr eases the processing demands of the task compared to traditional emoti onal Stroop tasks. Additionally the fact that the cue appears several seconds before the probe stimu li also requires engagement of working memory resources for active maintenance of task context. This allows the present task to be more cognitively demanding than a typi cal emotional Stroop task, which is arguably very important when studying PCS as the cognitive deficits in these patients are typically minor; deficits in attention and executive func tioning in PCS and PTSD may only appear with more highly demanding tasks. 53

PAGE 54

This task is also unique among other pr evious emotional Stroop experiments because the three emotional valences in the task (pleasant, neutral, and unple asant) were presented as a mixed paradigm dur ing functional neuroimaging. Whereas previous fMRI studies have predominantly used blocked paradigms for stimuli valence (i.e., blocks of several neutral trials followed by blocks of several unpleasant trials; e.g., Bremner et al., 2004; Compton et al., 2003; Malhi et al., 2005) the present study demonstrated statistically significant differences in brain activation according to pleasant and unpleasant emotional valenc e within a mixed paradigm. Taken as a whole, these neuroimaging results suggest that the C ued Emotional Countin g Stroop task is a potentially viable method for use in patients with PCS and PTSD in order to elucidate the hypothesized differential patterns of neural activation in these groups. Limitations Limitations of FMRI The viability of using fMRI as a correlate of neural activity is widely accepted (Bandettini, 2009; Friston, 2009). However, there remain criticisms of this methodology (Miller, 2008; Van Horn & Poldrack, 2009) a nd by all accounts, our understanding of the complex attributes of the hemodynamic re sponse in the brain is incompletely understood (Logothetis, 2008). However, certai n particular criticisms of previous research (e.g., that activation signals were due to mo tion artifact; Friston, Williams, Howard, Frackowiak, & Turner, 1996; Hajnal et al., 1994) have been well controlled for in the present study (no significant ta sk or valence differences in average head movement across experimental session), and therefore these resu lts are presented with confidence in their integrity. 54

PAGE 55

One criticism of this exper iment may be associated with how this task differs from more traditional emotional Str oop tasks. For example, it mi ght be said that conducting the Cued Emotional Counting Stroop as a mixed paradigms would produce bleed through activation as the trials vary from one affective valence to the next. This is in contrast with more traditional bloc ked paradi gms, which are believed to create mood states by lumping trials of one valence together, and some researchers have claimed that they can only elicit an emotional Stroop effect with a blocked design (Compton et al., 2003). However, even if effects of bleed th rough were present in the current study, any potential signal changes were not stro ng enough to outweigh the endogenous effect of each trial. Sample Size Limitations The small sample sized used for this study is a clear limitation. Published guidelines suggest that appr oximately 12 subjects ar e needed per group in fMRI research when data are analyzed using subjec t as a random effect (Desmond & Glover, 2002) in order to make population-level infer ences. Nonetheless, the sample size used in the present study appears to have been adequate to demonstrate the predicted effects at a suitably rigorous statistical criterion ( p < .05). Regardless, a larger sample size is indicated for future demonstration of the replicability of t hese findings. Along the same lines, it is worth noting that the inability of the present study to detect significant behavioral differences by word valence, in terms of RT or error rates, may be due to the small sample tested. However, the presence of significant interference effects has not been uniformly documented in non-clinical samples (Coates, 2008; Whalen et al., 1998). 55

PAGE 56

Future Directions The present study serves as a beginni ng step in a research endea vor aimed at using functional neuroimaging to document processes that may help differentiate PCS and PTSD. Future studies will work to demon strate the viability of the Cued Emotional Counting Stroop task in populatio ns of post brain-injury and anxiety disorder patients. These future lines of work will continue to hinge upon previous research in this lab that has demonstrated altered patterns of activa tion during tasks requiring cognitive control (i.e., Task Switching Cued Stro op) in patients with moderate and severe TBI (Larson et al., 2006; Perlstein et al., 2006; Seignourel et al., 2005; Sozda, 2009). The forthcoming projects using healthy control participants will use both the Task Switching Cued Stroop and the Cued Emotional Counting Stroop in the same individuals to demonstrate differential patterns of activation between thes e two tasks. If, as is hypothesized, these two tasks allow for a double dissociation of brain regions crit ical for cognitive and cognitive-affective neural ci rcuits, for example ACC and cACC, then these two tasks together should be well suited for use in diffe rentiating PCS and PTSD. To reiterate, the rACC has been implicated in emotion-based processing and responding, and the cACC has been implicated in performance of cognitive tasks, based on anatomical studies and functional imaging (Devinsky et al., 1995; V ogt et al., 1992). Based on patterns of functional neural activation in patient samp les that have been documented previously by other researchers (e.g., Bremner et al., 2004; Shin et al., 2001; Sozda, 2009), it is predicted that patients with PT SD would show disruption of activation in rACC and patients with PCS would show disruption of activation in cACC. It is worth noting that a similar double di ssociation project has been undertaken previously with healthy control participants with promising results (see: Bush et al., 1998 56

PAGE 57

57 and Whalen et al., 1998). However, the line of work by the Whalen and Bush groups did not employ a task-switching or instruct ional cue working memory component. This unique attribute, of tapping working memory in addition to emotional processing, allows broader aspects of cognitive-affective neural circuits to be explored with our Cued Emotional Counting Stroop task. For example, another line of future work will focus on dissociating the patterns of activation in thes e regions temporally, such that it may be found that dlPFC activity can predict activity in ACC and/or amygdala. It will also be explored whether there is a temporal corre lation between the activity in the ACC and amygdala, which would demonstr ate the presence of a network of activation that is currently assumed although not strictly empiri cally validated. These future endeavors in this realm are intended to hav e great clinical applicability. Yet at present, they depend upon a sequential, multi-step line of basic sci ence research, which is currently in its early stages in this and other laboratories. Summary The current study used fMRI to assess the feasibly of using the Cued Emotional Counting Stroop task as a probe for activati on of the amygdala, medial prefrontal regions including ACC, and lateral prefrontal cortical areas including the dlPFC. The results of this study suggest that this ta sk is indeed capable of activating these brain regions in a relatively small sample of healthy adult participants. These findings will be applied to a continuing line of research in this lab focused on better understanding the cognitive affective circuits implicated in the deficits shared by PCS and PTSD.

PAGE 58

APPENDIX STIMULI Table A-1. Stimuli used in C ued Emotional Count ing Stroop task Pleasant Neutral Unpleasant affection avenue abuse alert bathroom accident applause cellar afraid aroused circle anger baby clock angry bold coarse assault cash contents bomb champion custom cancer comedy detail cruel confident elbow crushed dancer elevator danger engaged engine destroy fame excuse devil glory fabric disaster handsome finger divorce happy habit execution holiday humble guilty hopeful journal hate improve lantern hatred joke medicine hurt kiss modest injury laughter muddy insane liberty nonsense killer loved nursery murderer lucky passage pain mighty phase poison miracle privacy prison passion shadow rejected pleasure sphere slaughter profit stomach slave proud teacher suicide romantic tower terrible success truck tragedy talent wagon troubled victory watch victim wedding writer violent All words are taken from the normative word set described in the Affective Norms for English Words (ANEW; Bradley & Lang, 1999). 58

PAGE 59

LIST OF REFERENCES Algom, D., Chajut, E., & Lev, S. (2004). A rational look at t he emotional Stroop phenomenon: A generic slowdown not a Stroop effect. Journal of Experim ental Psychology: General, 133 323-338. Alvarez, J. A., & Emory, E. (2006). Executive function and the frontal lobes: A metaanalytic review. Neuropsychology Review, 16 17-42. American Psychiatric Association (2005). Diagnostic and statistical manual of mental disorders (4th ed., text revi sion). Washington, DC: American Psychiatric Association. Bandettini, P. A. (2009). What's new in neuroimaging methods? Annals of the New York Academy of Sciences, 1156 260-293. Bigler, E. D., (2005). Structural imaging. In: J. M. Silver, T. W. McAllister, & Yudofsky, S. C. (Eds.), Textbook of traumatic brain injury (pp. 79-105). Arlington, VA: American Psychiatric Publishing, Inc. Bombardier, C. H., Fann, J. R., Temkin, N., Esselman, P. C., Pelzer, E., Keough, M., et al. (2006). Posttraumatic stress disorder symptoms during the firs t six months after traumatic brain injury. Journal of Neuropsychiatry and Clinical Neuroscience, 18 501-508. Boynton, G. M., Engel, S. A., Glover, G. H., & Heeger, D. J. (1996). Linear systems analysis of functional magnetic resonance imaging in human V1. Journal of Neuroscience, 16 4207-4221. Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: the Self-Assessment Manikin and the semantic differential. Journal of Behavioral Therapy and Experimental Psychiatry, 1 49-59. Bradley, M. M., & Lang, P. J. (1999). Affective norms for English words (ANEW): Stimuli, instruction ma nual and affective ratings Technical report C-1, Gainesville, FL. The Center for Research in Psyc hophysiology, University of Florida. 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. Bremner, J. D., Vermetten, E., Vythilingam, M., Afzal, N., Schm ahl, C., Elzinga, B., et al. (2004). Neural correlated of the classic color and emotional Stroop in women with abuse-related posttraumatic stress disorder. Biological Psychiatry, 55 612-620. Bush, G., Luu, P., & Posner, M. I. (2000). Co gnitive and emotional influences in anterior cingulated cortex. Trends in Cognitive Sciences, 4 215-222. Bush, G., Whalen, P. J., Rosen, B. R., Jenike, M. A., McIner ney, S. C., & Rauch, S. L. (1998). The counting Stroop: An interference task specialized for functional neuroimagingvalidation study with functional MRI. Human Brain Mapping, 6 270-282. 59

PAGE 60

Coates, R. C. (2008). Use of the emotional Stroop to a ssess psychological trauma following traumatic brain injury. Brain Injury, 22, 353-360. Compton, R. J., Banich, M. T. Mohanty, A., Milham, M. P., He rrington, J., Miller, G. A., et al. (2003). Paying attention to emotion: An fMRI investigation of cognitive and emotional Stroop tasks. Cognitive, Affective, and Behavioral Neuroscience, 3 8196. Dagliesh, T. (2005). Putting some feeli ng into itthe conceptual and empirical relationships between the classic and emotional Stroop tasks: Comment on Algom, Chajut, and Lev (2004). Journal of Experimental Psychology: General, 134 585-591. Davis, K. D., Taylor, K. S., Hutchinson, W. D., Dostrovsky, J. O., McAndrews, M. P., Richter, E. O., et al. (2005). Human anterior cingulate cortex neurons encode cognitive and emotional demands. The Journal of Neuroscience, 25 8402-8406. Desmond, J. E., & Glover, G. H. (2002). Estima ting sample size in functional MRI (fMRI) neuroimaging studies: Statistical power analyses. Journal of Neuroscience Methods, 188 115-128. Devinsky, O., Morrell, M. J., & Vogt, B. A. (1995). Contributions of the anterior cingulate cortex to behavior. Brain, 118 279-306. Forman, S. D., Cohen, J. D., Fitzgerald, M., Eddy, W. F., Mintun, M. A., & Noll, D. C. (1995). Improved assessment of significant activation in functional magnetic resonance imaging (fMRI): Use of a cluster-size threshold. Magnetic Resonance in Medicine, 33 636-647. Friston, K. J. (2009). M odalities, modes, and models in functional neuroimaging. Science, 326 399-403. Friston, K. J., Williams, S. Howard, R., Frackowiak, R. S., & Turner, R. (1995). Movement-related effects in fMRI time-series. Magnetic Resonance Medicine, 35 346-355. Goebel, R., Esposito, F., & Formisano, E. (2 006). Analysis of functional image analysis contest (FIAC) data with BrainVoyagerQX: From single-subject to cortically aligned group general linear model analysis and self-organizing group independent component analysis. Human Brain Mapping, 27 392-401. Gotlib, I. H., & McCann, C. D. (1984). Construct accessibility and depression: An examination of cognitive and affective factors. Journal of Personality and Social Psychology, 47 427-439. Gruber, S. A., Rogowska, J., Holcomb, P. Soraci, S., & Yurgelun-Todd, D. (2002). Stroop performance in normal cont rol subjects: An fMRI study. NeuroImage, 16 349-360. Hajnal, J. V., Myers, R., Oa tridge, A., Schwieso, J. E., Young, I. R., & Bydder, G. M. (1994). Artifacts due to stimulus correla ted motion in functional imaging of the brain. Magnetic Resonance Medicine, 31 283-291. 60

PAGE 61

Hoge, C. W., McGurk, D., Thom as, J. L., Cox, A. L., Enge l, C. C., & Castro, C. A. (2008). Mild traumatic brain injury in U.S. soldiers returning from Iraq. The New England Journal of Medicine, 358 453-463. Isenberg, N., Silbersweig, D., Engelien, A., Emmerich, S., Ma lavade, K., Beattie, B., et al. (1999). Linguistic threat activates the human amygdala. Proceedings of the National Ac ademy of Sciences, 96 10456-10459. Kay, T., Harrington, D. E., Adams, R., Ander son, T., Berrol, S., Cicerone, K., et al. (1993). Definition of mild traumatic br ain injury. Journal of Head Trauma and Rehabilitation, 8, 86-87. Kay, T., Newman, B., Cavall o, M., Ezrachi, O., & Resn ick, M. (1992). Toward a neuropsychological model of functional disabi lity after mild traumatic brain injury. Neuropsychology, 6 371-384. Kennedy, J. E., Jaffee, M. S., Leskin, G. A., Stokes, J. W., Leal, F. O., & Fitzpatrick, P. J. (2007). Posttraumatic stress disorder and posttraumatic stress disorder-like symptoms and mild traum atic brain injury. Journal of Rehabilit ation Research & Development, 44 894-920. Kompus, K., Hugdahl, K., Ohman, A., Marklund, P., & Nyberg, L. (2009). Distinct control networks for cognition and emotion in the prefrontal cortex. Neuroscience Letters, 467 76-80. Kraus, J. F., & Chu, L. D. ( 2005). Epidemiology. In: J. M. Silver, T. W. McAllister, & Yudofsky, S. C. (Eds.), Textbook of traumatic brain injury (pp. 3-26). Arlington, VA: American Psychiatric Publishing, Inc. Kucera, H., & Francis, W. N. (1967). Computational Analysis of Present-Day American English. Providence, RI: Brown University Press. Larson, M. J., Perlstein, W. M., Demery, J. A., & Stigge-Kaufman, D. A. (2006). Cognitive control impairments in traumatic brain injury. Journal of Clinical and Experimental Neuropsychology, 28 968-986. Levin, H. S., Williams, D. H., Eisenberg, H. M., High, W. M., & Gu into, F. C. (1992). Serial MRI and neurobehavioural findings after mild to moderate closed head injury. Journal of Neurology, Neurosurgery, and Psychiatry, 55 255-262. Logothetis, N. K. (2008). What we c an do and what we cannot do with fMRI. Nature, 453 869-878. Maas, A. I. R., Stocchetti, N., & Bullock, R. (2008). Moderat e and severe traumatic brain injury in adults. Lancet Neurology, 7 728-741. MacDonald, A. W., Cohen, J. D., Stenger, V. A., & Carter, C. S. (2000). Dissociating the role of the dorsolateral pref rontal and anterior ci ngulate cortex in cognitive control. Science, 288 1835-1838. MacLeod, C. M. (1991). Half a century of research on the St roop effect: An integrative review. Psychological Bulletin, 109 163-203. 61

PAGE 62

Malhi, G. S ., Lagopoulos, J., Sachdev, P. S., Ivanovski, B., & Shnier, R. (2005). An emotional Stroop functional MRI study of euthymic bipolar disorder. Bipolar Disorders, 7 58-69. Malojcic, B., Mubrin, Z., Coric, B., Susnic M., & Spilich, G. J. (2008). Consequences of mild traumatic brain injury on informati on processing assessed with attention and short-term memory tasks. Journal of Neurotrauma, 25 30-37. Markela-Lerenc, J., Ille, N., Ka iser, S., Fiedler, P., Mundt, C., & Weisbrod, M. (2004). Prefrontal-cingulate activation during ex ecutive control: Which comes first? Cognitive Brain Research, 18 278-287. Mathews, A. M., & MacLeod, C. (1985). Selective processing of threat cues in anxiety states. Behaviour Research and Therapy, 23 563-569. Mayberg, H. S. (1997). Limbic -cortical dysregulation: A pr oposed model of depression. Journal of Neuropsychiatry and Clinical Neuroscience, 9 471-481. McAllister, T. W. (2005). Mild brain injury and the postconcussion syndrome. In: J. M. Silver, T. W. McAllister, & Yudofsky, S. C. (Eds.), Textbook of traumatic brain injury (pp. 279-308). Arlington, VA: American Psychiatric Publishing, Inc. McKenna, F. P., & Sharma, D. (1995). Intrusive cognitions: An investigation of the emotional Stroop task. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21 1595-1607. Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex function. Annual Review of Neuroscience, 24 167-202. Miller, G. (2008). Growing pains for fMRI. Science, 320 1412-1414. Mohanty, A., Herrington, J. D., Koven, N. S., Fisher, J. E., Wenzel, E. A., Webb, A. G., et al. (2005). Neural mechanisms of a ffective interference in schizotypy. Journal of Abnormal Psychology, 114, 16-27. Nemeroff, C. B., Bremner, J. D., F oa, E. B., Mayberg, H. S., No rth, C. S., & Stein, M. B. (2006). Posttraumatic stress disor der: A state-of-the-science review. Journal of Psychiatric Research, 40 1-21. Parker, R. S. (2002). Recommendations for the revision of DSM-IV diagnostic categories for co-morbid posttraumatic stress disorder and traumatic brain injury. NeuroRehabilitation, 17 131-143. Perez-Edgar, K., & Fox, N. A. (2003). Individual differenc es in childrens performance during an emotional Stroop task: A behavioral and electrophysiological study. Brain and Cognition, 52, 33-51. Perlstein, W. M., Cole, M. A., Demery, J. A., Seignourel, P. J., Dixit, N. K., Larson, M. J., et al. (2004). Parametric manipulation of working memory load in traumatic brain injury: Behavioral and neural correlates. Journal of the International Neuropsychological Society, 10 724-741. 62

PAGE 63

Perlstein, W. M., Dixit N. K., Carter, C. S., Noll, D. C., & Cohen, J. D. ( 2003). Prefrontal cortex dysfunction mediates deficits in working memory and prepotent responding in schizophrenia. Biological Psychiatry, 58 25-38. Perlstein, W. M., Larson, M. J., Dotson, V. M., & Kelly K. G. (2006). Temporal dissociation of components of cognitive control dysfunction in severe TBI: ERPs and the cued-Stroop task. Neuropsychologia, 44 260-274. Phelps, E. A., & LeDoux, J. E. (2005). Contributions of the amygdala to emotion processing: From animal models to human behavior. Neuron, 48 175-187. Rao, V., & Lyketsos, C. (2000). Neuropsychi atric sequelae of traumatic brain injury. Psychosomatics, 41 95-103. Ratcliff, R. (1993). Methods for dea ling with reaction time outliers. Psychological Bulletin, 114 510-532. Sadock, B. J., & Sadock, V. A. (2007). Kaplan and Sadocks Synopsis of Psychiatry (10th Ed.) Philadel phia, PA: Lippincott Williams & Wilkins. Seignourel, P. J., Robins, D. L., Larson, M. J ., Demery, J. A., Cole, M., & Perlstein, W. M. (2005). Cognitive cont rol in closed head injury: context maintenance dysfunction or prepotent re sponse inhibition deficit? Neuropsychology, 19, 578590. Shin, L. M., Whalen, P. J., Pitman, R. K., Bu sh, G., Macklin, M. L., Lasko, N. B., et al. (2001). An fMRI study of anterior cingulate function in posttraumatic stress disorder. Society of Biological Psychiatry, 50 932-942. Smith, E. E., & Jonides, J. (1999). Storage and executive processes in the frontal lobes. Science, 283 1657-1661. Sozda, C. N. (2009). Impairment of cognitive control component processes after traumatic brain injury: An fMRI study. Unpublished master's t hesis, University of Florida, Gainesville, FL. Stein, M. B., & McAllister, T. W. (2009). Exploring the convergence of posttraumatic stress disorder and mild traumatic brain injury. American Journal of Psychiatry, 166 768-776. Stroop, J. R. (1935). Studi es of interference in serial verbal reactions. Journal of Experimental Psychology, 18 643-662. Stuss, D. T., Shallice, T., Alexander, M. P. & Picton, T. W. (1995) A multidisciplinary approach to anterior attentional functions. Annals of the New York Academy of Sciences, 769 191-211. Talairach, J., & Tournoux, P. (1988). Co-planar stereotaxic atlas of the human brain New York: Thieme Medical Publishers, Inc. van den Heuvel, O. A., Veltman, D. J., Groenewegen, H. J., Witter, M. P., Merkelbach, J., Cath, D. C., et al. (2006). Disorde r-specific neuroanatomical correlates of attentional bias in obsessive-compul sive disorder, panic disorder, and hypochondriasis. Archives of General Psychiatry, 62 922-933. 63

PAGE 64

64 Van Horn, J. D., & Poldrack, R. A. (2009). Functional MRI at the crossroads. International Journal of Psychophysiology, 73 3-9. Vasterling, J. J., Verfaellie, M., & Sullivan, K. D. (2009). Mild traumat ic brain injury and posttraumatic stress disorder in returni ng veterans: Perspectives from cognitive neuroscience. Clinical Psychology Review 29 674-684. Vogt, B. A., Finch, D. M., Olson, C. R. (1992). Functional heterogeneity in cingulate cortex: The anterior executive and pos terior evaluative regions [review]. Cerebral Cortex, 2 435-443. Vythilingam, M., Blair, K. S., McCaffrey, D., Scaramozza, M., Jones, M., Nakic, M., et al. (2007). Biased emotional attention in posttraumatic stress disorder: A help as well as a hindrance? Psychological Medicine, 37 1445-1455. Warden, D. L., & Labbate, L. A. (2005). Posttr aumatic stress disorder and other anxiety disorders. In: J. M. Silver, T. W. McAllister, & Yudofsky, S. C. (Eds.), Textbook of traumatic brain injury (pp. 231-243). Arlington, VA: American Psychiatric Publishing, Inc. Watts, E. N., McKenna, E. P., Sharrock, R. & Trezise, L. (1986). Colour naming of phobia-related words. British Journal of Psychology, 77 97-108. Whalen, P. J., Bush, G., McNally, R. J., Wilhel m, S., McInerney, S. C., Jenike, M. A., et al. (1998). The emotional counting Stroop paradigm : A functional magnetic resonance imaging probe of the anterio r cingulate affective division. Biological Psychiatry, 44 1219-1228. Williams, J. M. G., Mathew s, A., & MacLeod, C. (1996). The emotional Stroop task and psychopathology. Psychological Bulletin, 120 3-24.

PAGE 65

BIOGRAPHICAL SKETCH Callie Elizabet h Tyner graduated magna cum laude with a Bachelor of Arts degree in psychology from Agnes Scott College in December 2006. In August 2008, she began graduate school at the University of Flor ida, where Callie was awarded an Alumni Fellowship by the Department of Clinical and Health Psychology. She obtained her Master of Science degree in May 2010 and is pursuing a Doctorate in Clinical Psychology, with a specialized focus on Neuropsychology. 65