- "\ '\

T355“ LIBRARY

Michigan State
University

This is to certify that the
dissertation entitled

Mild Traumatic Brain Injury, Executive Functions, and
Post-Concussive Symptoms

presented by

Mark Lawrence Ettenhofer

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in PsychologL

 

 

/L M

 

Major Professor’s Signature

all (7/05
Date

MSU is an Affirmative Action/Equal Opportunity Institution

-.-.-—.-.-

n.-u-.-.-.—.-.—.—.—.---—-¢---u-—t-.—

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE
W 11 2 $0122

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DateDue.indd—p.1

 

MILD TRAUMATIC BRAIN INJURY, EXECUTIVE FUNCTIONS,
AND POST-CONCUSSIVE SYMPTOMS

By

Mark Lawrence Ettenhofer

A DISSERTATION

Submitted to
Michigan State University
in partial fulfilhnent of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Psychology

2006

ABSTRACT
MIILD TRAUMATIC BRAIN INJURY, EXECUTIVE FUNCTIONS, AND POST-
CONCUSSIVE SYMPTOMS

By

Mark Lawrence Ettenhofer

Relatively few studies have been conducted examining the relationships between
mild traumatic brain injury (mild TBI), neurocognitive deficits, and self-reported post-
concussive symptoms in the post-acute phase of injury. Additionally, interpretation of
the existing literature has been complicated by the small, treatment-derived samples of
participants, high rates of injury-related litigation, and inappropriate control groups
typical of studies in this area. Consequently, the significance of mild TBI to cognitive
functioning and post-concussive symptoms in the long term remains unclear.

In this study, theoretical models of long-term impairment following mild TBI
were evaluated, with a particular focus on the role of executive fimctions. Several
methodological limitations of previous studies were addressed through the use of a
relatively large sample of participants recruited in a non-treatrnent university setting (11 =
126), an orthopedic injury control group, and powerful latent and structural modeling
techniques of data analysis. Time since injury ranged fiom three months to six years.

Contrary to hypotheses, mild head injury was not associated with impairments in
any cognitive domains examined, which included executive fimctions, processing speed,
and verbal memory. Further, no relationship was found between the incidence of mild

TBI and severity of post-concussive symptoms (as measured by the PCSC; Gouvier et a1.,

1992). Instead, post-concussive symptoms were strongly related (r = .50, p < .05) to
severity of general psychiatric symptoms (as measured by the BSI; Derogatis & Spencer,
1982). These findings suggest that neurological injury may be of little clinical
significance to long-term cognitive and symptom outcome following mild TBI, relative to
other, non-neurological factors. Additional research is necessary to identify and
characterize psychological and emotional factors that may be important to long-term

recovery from injuries involving a mild TBI.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ vi
LIST OF FIGURES .......................................................................................................... vii
INTRODUCTION ............................................................................................................ 1
Mild Traumatic Brain Injury: An Overview ......................................................... 2
Executive Functions .............................................................................................. 4
Neurocognitive Sequelae of Mild TBI .................................................................. 9
Post-Concussive and Psychiatric Symptoms ........................................................ 14
Overview and Predictions ..................................................................................... 17
METHODS ....................................................................................................................... 20
Participants ............................................................................................................ 20
Procedure .............................................................................................................. 22
Measures ............................................................................................................... 23
Statistical Analyses ............................................................................................... 31
RESULTS ......................................................................................................................... 33
Sample Description ............................................................................................... 33
Descriptive Statistics ............................................................................................. 34
Confirmatory Factor Analysis of Cognitive Variables ......................................... 35
Head Injury Status Predicting Cognitive Functioning .......................................... 37
Post-Concussive Symptoms (PCS) and Cognitive Functioning ........................... 39
Relationships between Head Injury, Symptom Variables, and EF ....................... 4O
Follow-up Data Analyses ...................................................................................... 40

iv

TABLE OF CONTENTS (CONTINUED)

DISCUSSION ................................................................................................................... 42

LIST OF REFERENCES .................................................................................................. 60

LIST OF TABLES

Table 1. Demographic Characteristics of Screening Sample ............................................ 47
Table 2. Constructs and Indicators .................................................................................... 48
Table 3. Executive Function Indicators and Theoretical Components ............................. 49
Table 4. Demographic Characteristics by Participant Group ........................................... 50
Table 5. Injury Characteristics by Participant Group ....................................................... 51
Table 6. Descriptive Statistics and T-tests of Cognitive Variables .................................. 52
Table 7. Descriptive Statistics and T-tests of Self-reported Symptom Variables ............. 53
Table 8. Confirmatory Factor Analysis of Cognitive Variables ....................................... 56
Table 9. Models of Head Injury Status Predicting Cognitive Functioning ....................... 57
Table 10. Models of Post-Concussive Symptoms and Emotional/Psychiatric

Functioning ....................................................................................................................... 58
Table 11. Correlations of Selected Symptom Variables ................................................... 59

LIST OF FIGURES

Figure 1. Hypothesized Variable Relationships ................................................................ 46
Figure 2. Freely-estimated model of EF factor structure (Model EFz) ............................. 54
Figure 3. Freely-estimated model of PS and VM factor structure (Model C1) ................. 55

vii

INTRODUCTION

Although a significant amount of research has been conducted in the acute phase
of mild traumatic brain injury (TBI), relatively few studies have been conducted that
might illuminate the relationships between mild TBI, neurocognitive deficits, and post-
concussive symptoms in the long term. Additionally, most studies in this area have had
significant methodological limitations and have been primarily descriptive in nature. It
has therefore been difficult to draw firm conclusions regarding possible neurocognitive
mechanisms underlying the impairment that some individuals experience following mild
TBI. This information is critical to the development and implementation of effective
treatments.

The overarching aim of the current project was to evaluate theoretical models of
long-term impairment following mild TBI, with a particular focus on the role of
executive functions. This study addresses methodological limitations of previous studies
by recruiting a relatively large and ecologically valid mild TBI sample as well as an
orthopedic injury control group, including several theoretically and psychometrically
strong measures of the executive construct, and testing primary hypotheses using
powerful latent and structural modeling techniques of data analysis.

As used in the text, “post-concussive symptoms” or “PCS” refer to those self-
reported symptoms composing the traditional “post-concussive syndrome”, rather than all
possible symptoms that might result from mild TBI. Both “post-concussive symptoms”
(PCS) and “psychiatric symptoms” are those which the individual may experience in the

long-term, rather than those which may be experienced acutely following the injury.

Mild Traumatic Brain Injury: An Overview

Traumatic brain injury (TBI) has long been recognized as a condition with wide-
reaching effects on an individual’s personal, social, and occupational functioning. By
current estimates, approximately 200 traumatic brain injuries occur annually for every
100,000 individuals (Ryan, O’Jile, Gouvier, Parks-Levy & Betz, 1996). Of these
injuries, upwards of 75% could be classified as ‘mild’ by standard medical criteria
(Hayden, 1997). However, because the majority of mild brain injuries are never reported
to any public health agency, it is likely that the true incidence rate is much higher. Many
individuals may not believe that their injuries warrant medical attention, or they may
choose not to seek medical attention for financial reasons (Kraus & McArthur, 1996).
Unfortunately, the vast majority of studies of mild TBI have recruited participants who
have self-selected by obtaining services at a treatment facility, thereby excluding a very
large segment of the mild TBI population and biasing the sample in significant ways
(Lees-Haley, Green, Rohling, Fox & Allen, 2003; Bernstein & de Ruiter, 2000;
Newcombe, Rabbitt, & Briggs, 1994). A survey of available North American
epidemiological literature conducted by the World Health Organization (WHO)
Collaborating Centre Task Force on Mild TBI found that males are at substantially
greater risk for experiencing a mild TBI (with a relative risk of up to 2.5); teenagers and
young adults were also found to be at increased risk for mild TBI (Cassidy, Carroll,
Peloso, Borg, von Holst et al., 2004).

Some individuals who sustain a mild TBI experience chronic “post-concussive
symptoms” (PCS) such as headaches, irritability, insomnia, distractibility, depression,

and dizziness (Segalowitz & Lawson, 1995; Alves, Macciocchi, & Barth, 1993; Vamey

& Menefee, 1993). Additionally, research indicates that individuals who experience a
mild TBI may be at an increased risk for other negative health-related outcomes. For
example, epidemiological studies suggest that head injuries of any severity increase the
likelihood that an individual will subsequently develop Alzheimer’s disease or another
form of dementia (Plassman, Havlik, Steffens, Helms, Newman et al., 2000).

Evidence from experimental animal models and post-mortem studies in humans
indicates that mild TBI can result in various forms of structural damage to the brain,
including focal cortical lesions and, more commonly, diffuse axonal injury (De Kruijk,
Twijnstra, & Leffers, 2001; Rizzo & Tranel, 1996; Zohar, Schrieber, Getslev, Schwartz,
Mullins & Pick, 2003). Unfortunately, currently available structural imaging techniques
are generally poor at detecting this damage in living humans, during both the acute and
chronic phases of mild TBI (Bigler, 1999; Bigler & Snyder, 1995). The insensitivity of
structural imaging may be due to the type of damage that is commonly associated with
mild TBI, which is believed to be primarily microstructural or neurochemical in nature
(Bigler, 1999; Rizzo & Tranel, 1996). Functional imaging has demonstrated greater
sensitivity to the effects of mild TBI, but because differing patterns of activation do not
necessarily imply damage or dysfirnction, these techniques are not commonly used in
clinical settings, and great caution must be used in the interpretation of these results
(McAllister, Sparling, Flashman, & Saykin, 2001). As such, behavioral measures
(particularly those that assess executive functions, described below) within the context of
a neuropsychological assessment would appear to be the most sensitive indicators of

impairment following mild TBI (Cicerone & Azulay, 2002).

Executive Functions

Executive functions (EFs) are a class of higher-order cognitive abilities that have
traditionally been associated with the structural and fimctional integrity of the frontal
lobes (see Lezak, 1995; Stuss & Benson, 1984). The term “executive” commonly
connotes a supervisory or coordinating role over non-executive forms of cognition in the
pursuit of one or more goals. Hart, Schwartz & Mayer (1999) define EFs as those
cognitive functions which enable the purposeful organization of behavior in a manner
that is flexible and responsive to changing demands over time. Although areas of general
agreement are small regarding the cognitive functions that are most central to effective
executive control and the neuroanatomical substrate(s) of these functions, strong
associations between putative measures of EF and firnctional outcomes in a wide range of
psychiatric and medical disorders (Royall, Lauterbach, Cummings, Reeve, Rummans,
Kaufer et al., 2002) have spurred a great deal of research (and debate) in this area. While
the validity of the EF construct and the reliability of the tests commonly used to measure
it have been criticized somewhat extensively (see Lowe and Rabbitt, 1998; Phillips,
1997; Rabbitt, 1997), measures of EF continue to be used in a wide range of settings, and
researchers continue to expand the literature on BF methodology, theory, and application.

Neuroanatomically, EFs are believed to rely upon a system of reciprocal cortico-
subcortical circuits connecting the prefi'ontal cortex, anterior cingulate cortex, basal
ganglia, and cerebellum (Heyder, Suchan, & Daum, 2004). Although the prefrontal
cortex appears to be the locus for the greatest quantity and complexity of EF processing,
it is important to note that EF deficits (such as disinhibition, poor planning, behavioral

and cognitive inflexibility, and difficulty initiating, maintaining, and monitoring

behaviors) can also result from damage to fi'ontal or non-frontal components of this
network.

Concepts of EF are ever-evolving, and theories regarding the cognitive structure
of information processing underlying EF abound. Regarding what appear to be the most
significant theoretical advances within recent years, a number of leading researchers have
posited that EFs are accomplished in large part through the activity of one or more multi-
purpose forms of working memory. Engle and Kane (2004), for example, stress the
importance of working memory processes in allowing task demands/ goals to be
maintained in a sufficiently active state for task-interfering stimuli to be inhibited. This
conception describes working memory processes as being much less related to effective
“storage” of any information (a function believe to be accomplished by separate, non-
frontal neural networks), than to the inhibition of conflicting information (a function
primarily dependent upon the prefrontal cortex and the anterior cingulate cortex) (Engle
& Kane, 2004). As such, “working memory” can be reframed as an executive form of
attentional control — the selective, continued activation of goal-relevant information to the
exclusion of irrelevant information.

Consistent with the common conceptual attribution that EFs coordinate and
supervise lower-order cognitive processes, a number of researchers (Duncan & Owen,
2000; Engle and Kane, 2004; Goldman-Rakic, 1998) have suggested that the areas of the
prefrontal cortex and anterior cingulate cortex serving working memory serve to code,
process, and integrate many different types of information across modalities. However,
whereas Goldman-Rakic (1998) cites studies utilizing single-cell recording in animals

and functional imaging in humans to support sub-regional specialization of prefrontal

cortex by form of informational content (i.e., spatial vs. linguistic), cognitive experiments
by Engle and Kane (2004) have found that working memory capacity (measured within a
dual-task paradigm) appears to be predictive of a wide range of cognitive performance,
irrespective of the form or modality of the information being processed in the working
memory task. This discrepancy has not yet been resolved.

Although these theories of EF centered on working memory have empirical
support and include a large proportion of functions within the EF spectrum, they are
somewhat less compelling in terms of their ability to capture categories of function such
as abstract thinking, cognitive flexibility, and behavioral initiation. Aside from the fact
that these fimctions have been conceptually grouped beneath the traditional umbrella of
EF, studies of various tasks assessing these functions have found relationships to activity
or structural integrity of the prefrontal cortex (particularly dorsolateral areas) (Lezak,
1995; Stuss & Benson, 1984).

In a significant attempt to provide a broader theoretical context for EF that
appears more inclusive of those functions not obviously related to working memory,
Barkley (2001) posited from an evolutionary perspective that executive functions have
developed progressively as a way to allow what were previously overt behaviors to be
directed inward and represented mentally in the service of one or more goals. In this
manner, Barkley (2001) suggested that the most basic executive function, inhibition (the
delay or interruption of an overt behavior) allows for more complex executive functions
such as working memory (self-directed mental representations of experience), goal
pursuit and motivation (self-directed affect), and flexibility, fluency, and generativity

(self-directed, exploratory analysis and synthesis of prior experience).

A number of studies have been conducted examining the relationships between
various putative measures of EF in order to evaluate existing theories of EF and elaborate
on the composition and organization of the EF construct. However, when quantitative
studies on the issue of EF structure (most commonly, factor analyses evaluating unity vs.
diversity of functions) are examined as a whole (as in Royall et al., 2002), results appear
largely inconclusive. Whereas some investigations have yielded support for the
modularity of executive functions (Burgess, Alderman, Evans, Emslie & Wilson, 1998;
Duncan, Johnson, Swales & Freer, 1997), other studies have produced ambiguous results,
with separable but related factors (Boone, Ponton, Gorsuch, Gonzalez & Miller, 1998;
Miyake, Friedman, Emerson, Witzki & Howerter, 2000), or results that support a unitary
model of executive function (Della Sala, Gray, Spinnler & Trivelli, 1998; Ettenhofer,
Harnbrick & Abeles, in press).

It appears likely that this ambiguity stems, at least in part, from divergent methods
used to measure EF, as the particular measures and indices used can have a strong impact
on a the apparent factor structure of EF. Potential sources of error could include the
exclusion of measures tapping one or more important subcomponents of EF, as well as
the inclusion of non-executive indices fi'om putatively EF measures (e.g., the word and
color trials from the Stroop task). Additionally, it is possible that the discrepancies
between these various results are related to differences in the populations studied. For
example, if current findings of functional imaging studies of EF are correct in suggesting
anatomical modularity of executive functions to subdivisions of frontal and sub-cortical
brain areas within reciprocally-connected networks (Heyder, Suchan & Daum, 2004),

then populations with injuries/illnesses causing diffuse damage to the executive network

might be expected to exhibit greater communality of BF than populations with
injuries/illnesses causing more restricted damage to fewer subcomponents of that
network. Further, the structure of the EF construct might appear different still in
participant groups whose EFs are operating in a more “normal” fashion (such as control
groups or participants in non-clinical studies).

In consideration of the issues described above, the present study included a wide
variety of EF measures representing the component EFs of working memory, inhibition,
cognitive flexibility (i.e., set shifting), abstract reasoning, and behavioral initiation (e. g.,
verbal fluency/generativity). Interpretation of the obtained results and the evaluation of
primary hypotheses (see Overview and Predictions) are guided by an examination of the
structure of EF in the current sample and its coherence in terms of the relevant
neurocognitive theories presented.

Rationale for Latent Modeling of EF. Latent or structural modeling techniques
were used to test all theoretical models. Because executive firnctions serve to coordinate
and monitor other “lower level” cognitive functions in the service of one or more goals,
all measures of executive functioning necessarily contain one or more non-executive
components (Lezak, 1995; Stuss & Levine, 2002). In this study, in order to permit a
“purer” examination of executive function, multiple measures of EF were used to
construct a latent EF factor. Whereas any single index of executive function would
necessarily contain an indeterminable combination of “executive firnction” variance,
“non-executive” variance, and error variance, a latent EF variable pools shared variance
from several separate indices of executive function, maximizing construct-relevant

variance and excluding variance that is unique to any single measure.

Because the category of functions described as “executive” is rather under-
specified, tasks which assess executive function vary considerably. Although the
relatively divergent nature of the EF measures utilized in this study would serve as an
obstacle to more traditional methods of data analysis, latent modeling techniques are able
to capitalize on this divergence by extracting elements that are common across measures.
In this case, it would be expected that these common elements would be very highly
related to the construct in question (executive firnction), since there appears to be little
commonality between the non-executive components of these measures as a whole.

One powerful illustration of these principles comes from a prior study conducted
by the present author and colleagues in which a single latent executive function variable
(EF) was derived from four measures of executive function (Ettenhofer, Harnbrick &
Abeles, 2005). Although the average observed test-retest correlation of these individual
variables was .67, the test-retest correlation of the latent EF variable constructed from
these same variables was much higher, at .96. The present study employed similar latent
modeling techniques, in which a latent EF factor model was derived from seven measures
of executive firnctioning. This method of analysis is designed to increase the reliability
and construct validity of EF, and allow conclusions to be drawn with greater confidence.
Neurocognitive Sequelae of Mild TBI

Although the clinical significance of mild TBI has been downplayed by a number
of researchers (Dikmen, Macharner & Temkin, 2001; Lees-Haley et al., 2003; Satz,
Alfano, Light, Morgenstem, Zaucha, Asamow et al., 1999), a significant number of
studies suggest that even mild head injuries may negatively impact an individual’s

attention (Binder, Rohling, & Larrabee, 1997; Brewer, Metzger & Therrien, 2002)

memory (Fisher, Ledbetter, Cohen, Marmor, & Tulsky, 2000; Mangels, Craik, Levine,
Schwartz & Stuss, 2002; Thornton, 2003), visuospatial abilities (Cremona-Meteyard &
Geffen, 1994), processing speed (Hinton-Bayre, Geffen, & McFarland, 1997; Pinkston,
Gouvier, & Santa Maria, 2000), and executive abilities (Brooks, Fos, Greve &
Hammond, 1999; Raskin & Rearick, 1996; Stablum, Mogentale & Umilta, 1996;
Wallesch, Curio, Kutz, Jost, Bartels et al., 2001). Nevertheless, it has long been assumed
that these cognitive deficits resolve within a few weeks post-injury, and consequently, the
majority of studies examining the cognitive deficits following mild TBI have focused on
individuals in the acute stage of recovery. However, some studies that have been
conducted with post-acute mild TBI populations have found that deficits may indeed
persist beyond the acute phase, with approximately 14% experiencing long-term
disability in terms of employment (Larrabee, 1999). Interestingly, impairments in
executive processes such as working memory, behavioral inhibition, and abstract
problem-solving appear to be the most commonly-reported types of cognitive deficits
found in long-term follow-ups after mild TBI (McDonald, F lashman & Saykin, 2002;
Vamey & Menefee, 1993). It is possible that the relative vulnerability of the frontal
lobes to damage in mild TBI (Capruso & Levin, 1996; McDonald, Flashman & Saykin,
2002) accounts this finding, as executive functions are believed to rely heavily upon
fiontal areas (Lezak, 1995; Stuss & Benson, 1984).

Two widely-cited studies have employed meta-analytic techniques in examining
cognitive functioning in the long-term following mild TBI. From their analysis, Binder,
Rohling, & Larrabee (1997) reported a small but significant mean effect size of mild TBI

on overall long-term cognitive performance ((1 = .12). Only one cognitive sub-domain

10

(Attention/Concentration) reached statistical significance (d = .20) in this meta-analysis.
It should be noted that this sub-domain was defined broadly to also include a number of
tasks that could be considered measures of executive functions, such as Trail Making
Test Part B, and the Stroop Color-Word Test (see Stuss et al., 2001; Arbuthnott & Frank,
2000). Using alternative methodology, Frencham, Fox, and Mayberry (2005) found no
significant cognitive effects of mild TBI in studies three months or later post-injury.
However, the strength of this finding is limited somewhat by the small number of original
post-acute studies included in the analysis (four), each with relatively small sample sizes
(N: 21 to N= 48).

Additional data, derived from methodologically strong studies, would greatly
enhance our ability to draw firm conclusions about cognitive functioning in the post-
acute phase of mild TBI. Additionally, if long-term impairment is found to be present, a
firller examination of the precise relationships between mild TBI and specific forms of
cognitive impairment would be an important secondary area of investigation. For
example, any of a number of theoretical patterns of neuronal damage could be
responsible for cognitive impairment, if present. One possibility is that mild TBI could
cause direct, but non-selective damage to a variety of areas throughout the brain,
affecting general information processing as well as the neurocognitive networks
underlying executive, memory, and visuospatial functions. While relatively
straightforward, this possibility fails to account for the theoretical interrelationships
between neurocognitive networks, and appears to conflict with evidence demonstrating

relative localization of both focal lesions as well as more diffuse axonal injury to the

11

frontal (and to a lesser degree, temporal) lobes (Capruso & Levin, 1996; McDonald,
Flashman & Saykin, 2002).

Another, conceptually-related possibility is that mild TBI could result in diffuse
forms of damage that, rather than directly affecting a wide range of neurocognitive
networks, cause generalized cognitive impairment indirectly by slowing the overall rate
of information processing (W allesch et al., 2001). Theoretically, a reduced rate of
information processing could impair component cognitive abilities by slowing task
performance and reducing the availability of general cognitive resources (Salthouse,
1991). Providing some support for this possibility, Hart, Schwartz and Mayer (1999)
found that those who had sustained a mild TBI made more “errors of action” when
cognitive resources were limited by task demands. These errors were significantly
correlated with measures of processing speed.

Although a model focused on reduced processing speed appears to offer a more
coherent explanation of impairment following mild TBI than a direct-effects model, it is
unclear whether deficits in processing speed can account for the full range of cognitive
deficits that may follow mild TBI. For example, in a sample of individuals who had
experienced a mild TBI, Chan (2002) found significant impairment in EF even after
controlling for the presence of processing speed deficits. Additionally, in the study
conducted by Hart, Schwartz and Mayer (1999) “errors of action” were also significantly
correlated with measures of EF, which leaves room for another distinct mechanism of
impairment.

If mild TBI can result in forms of cognitive dysfirnction that are not explained by

reductions in processing speed, a third possibility is that cognitive impairments following

12

mild TBI may result from the disruption of neurocognitive networks underlying the
executive control of component cognitive processes. Because EFs have been theorized to
be responsible for planning, execution, and on-line monitoring of component cognitive
functions, and the networks underlying EF are reciprocally connected with many areas of
the brain (Duncan & Owen, 2000; Goldman-Rakic, 1998; Heyder, Suchan, & Daum,
2004), it appears reasonable that EF impairment could negatively affect performance on a
wide range of cognitive tasks. In such a scenario, reductions in apparent processing
speed might reflect inefficient task performance as a result of poor planning,
coordination, or monitoring of behaviors, rather than slowed speed of information
processing as such. Limited support for this executive-mediated theory of cognitive
dysfunction comes from a study in which individuals with mild TBI exhibited memory
deficits relative to controls, but only under task conditions in which there were also
demands upon the executive system (Mangels et al., 2002). Additionally, the possibility
that mild TBI might impair executive control by disrupting communication between
frontal and non-fiontal areas appears to receive some support from an EEG study which
found altered frontal-parietal coherence during an executive task among individuals who
sustained a mild TBI (Cudmore, Segalowitz, & Dywan, 2000).

To summarize, there are a number of possible mechanisms by which mild TBI
could affect long-term neurocognitive function. Those that were considered for
examination in the present study included a) the direct effects of injury upon a wide
variety of neurocognitive systems, b) diffuse injury that affects general cognitive

functioning by reducing processing speed, and c) injury-related disruption of

13

frontal/executive networks that affects general cognitive functioning by preventing
effective executive control.

It is also important to note that factors other than brain injury can affect the results
of neuropsychological testing among some individuals who have experienced a mild TBI.
Perhaps most importantly, many individuals who are (or who expect to be) involved in
litigation related to a mild TBI may be unmotivated to perform well, or even motivated to
perform poorly, because of factors surrounding compensation (Binder & Rohling, 1996;
Larrabee, 1999). Therefore, the results of studies of mild TBI that draw participants from
clinical treatment populations may be biased, especially if they are selected on the basis
of symptoms complaints (Larrabee, 1999). As such, alternate sampling techniques were
employed in the present study in order to minimize these biases.

Post-Concussile and Psychfirtric Symptoms

Although the current study focused primarily upon cognitive abilities as measured
through neuropsychological testing, it is also important to consider the importance of
subjective symptom reports to the overall picture of mild TBI and the varying trajectories
of recovery experienced by different individuals. During the acute phase of recovery
from mild TBI, it is common for patients to report symptoms such as dizziness, headache,
fatigue, nausea, irritability, confusion, and difficulty concentrating. In the majority of
cases, these symptoms appear to remit within the first few weeks or months of recovery;
however, a substantial proportion of individuals who have sustained a mild TBI continue
to report these or similar symptoms for a year or more following injury. A DSM-IV
diagnosis of Post-Concussive Disorder applies when an individual experiences these

symptoms to an impairing or distressing degree for more than 3 months following a mild

l4

TBI. In terms of the typical trajectory of recovery fi'om these symptoms, epidemiological
studies of mild TBI have found the prevalence of clinically-significant post-concussive
symptoms to be approximately 51% at six weeks post-injury (Rutherford, Merrett &
McDonald, 1977), and 34% at three years post-injury (Fee & Rutherford, 1987).

Surprisingly, many studies have been unable to find solid relationships between
post-concussive symptoms (PCS) and the actual occurrence of mild TBI. For example,
despite a large sample size, Paniak, Reynolds, Phillips, Teller-Lobe, Melnyk et al. (2002)
were only partially successful at discriminating mild TBI patients from controls using
logistic regression based upon PCS. Likewise, neither Sawchyn, Brulot, & Strauss
(2000) nor Gouvier, Cubic, Jones, Brantley & Cutlip (1992) were able to find a
significant relationship between PCS and head injury status, but PCS were correlated
significantly with levels of daily stress (Gouvier et al., 1992) and litigation activity
(Bernstein & de Ruiter, 2000; Larrabee, 1999; Lees-Haley et al., 2003; Newcombe,
Rabbitt, & Briggs, 1994). Thus, evidence strongly suggests that factors other than mild
TBI impact the degree to which an individual reports these “post-concussive” symptoms.

Another such factor appears to be emotional functioning; individuals who
experience high levels of psychiatric symptoms in addition to mild TBI appear to exhibit
a greater number of PCS at long-term follow-up (Mooney & Speed, 2001; Sawchyn,
Brulot, & Strauss, 2000). Further, individuals who have experienced a mild TBI appear
to be more susceptible to the effects of experimentally-induced stress; which increases
the frequency and severity of their pcst-concussive symptoms (Hanna-Pladdy, Berry,

Bennett, Phillips & Gouvier, 2001).

15

Clearly, a theoretical model is needed to explain the somewhat enigmatic
relationships between mild TBI, PCS, and psychiatric symptoms. Although previous
research suggests that mild TBI might not directly increase levels of PCS, evidence
appears to indicate that emotional/psychiatric factors play an important role. Psychiatric
symptoms such as depression may be more common among individuals who have
sustained a mild TBI (Mathias & Coats, 1999), and may also increase the severity of
post-concussive complaints (Sawchyn, Brulot, & Strauss, 2000). Thus, models in which
psychiatric symptoms mediate the relationship between mild TBI and PCS are worthy of
examination.

In addition to the possible role of psychiatric symptoms in the expression of post-
concussive symptoms following mild TBI, hypotheses have also been raised regarding
the role of cognitive deficits in mediating the effects of mild TBI on levels of post-
concussive symptoms. For example, in a relatively small sample of individuals with mild
TBI, Pinkston, Gouvier, and Santa Maria (2000) found a significant relationship between
processing speed and severity of post-concussive symptoms, suggesting that the
relationship between mild TBI and PCS may be mediated by neurocognitive factors.
Replication and extension of these preliminary results into a wider context might provide
additional information about the relationship between mild TBI, neurocognitive firnction,
and PCS. For example, because attentional abilities are conceptually related to EF (see
Lezak, 1995), it could be argued that post-concussive complaints of inattention and
distractibility may be accounted for, at least in part, by executive dysfunction. In
addition, for both processing-speed- and EF-mediated models of PCS, impairments in

cognitive functioning could be hypothesized to impact psychiatric symptoms by

16

decreasing stress tolerance levels and/or the effectiveness of psychological coping
mechanisms.

To summarize, it appears likely that the presence of emotional/psychiatric
symptoms increases the expression of self-reported post-concussive symptoms.
However, a number of possibilities have been raised regarding mechanisms by which
mild TBI may affect long-term PCS. The include a) the direct effects of injury in causing
PCS, b) injury-related increases in psychiatric symptoms, which in turn result in post-
concussive symptom complaints (PCS), c) diffuse injury that increases symptom
complaints by reducing processing speed, and d) injury-related disruption of
frontal/executive networks that increases symptom complaints by preventing effective
executive control.

OvervievL and Predictions

Mild TBI is a highly prevalent public health concern, with inestimable costs to the
individuals who experience long-term symptoms. However, directed studies of the
effects of mild TBI are limited, and available evidence is insufficient to draw firm
conclusions regarding the potential long-term effects. Additionally, it is unclear what
relationship, if any, exists between mild TBI, psychiatric symptoms, and reported post-
concussive symptoms (PCS). Clarification of the causal factors underlying the cognitive
impairments that some individuals experience following mild TBI, and their relationship
to psychiatric symptoms and PCS, is essential to the development and implementation of
methods of remediation. In the present research, a number of theoretical models were
evaluated in order to investigate the neurocognitive mechanisms that may result in

cognitive impairments following mild TBI. These models also include an examination of

17

the theoretical relationships between mild TBI, psychiatric symptoms, and self-reported
symptom complaints (PCS). Because it was believed that injury-related deficits in
executive function (EF) were likely to be primarily responsible for the dysfunctions that
may follow mild TBI (both cognitive and subjective/symptomatic), a central
(mediational) role was hypothesized for EF. Primary hypotheses are depicted (in
aggregate form) in Figure 1.

In addition to providing a theoretical framework for the examination of
fundamental hypotheses of impairment following mild TBI, the present research
addresses a number of limitations of prior studies. First, mild TBI is often associated
with such factors as injury to other bodily systems and emotional distress; therefore it has
been suggested that studies of mild TBI need to include an appropriate control group to
rule out the possible effects of these non-neurological factors (Satz et al., 1999; Dikmen
& Levin, 1993). This study included a control group composed of individuals with mild
orthopedic injuries, thereby controlling for generic effects of injury. Second, the vast
majority of previous studies have obtained participants from treatment centers, a practice
which excludes a very large segment of the injured population and introduces potentially
confounding psychogenic factors such as malingering (Lees-Haley et al., 2003; Bernstein
& de Ruiter, 2000; Newcombe, Rabbitt, & Briggs, 1994). Ecological validity in this
study was greatly enhanced by the inclusion of participants without regard to whether
they sought medical attention for their injuries. Third, many studies of mild TBI have
been statistically underpowered, limiting their ability to detect more subtle effects;
sample size in this study (N = 126) provides sufficient statistical power (.80 or higher) to

detect meaningful relationships for all analyses of interest. Finally, this study utilized a

18

battery of widely-used neuropsychological tests that were selected based upon their
theoretical relevance to mild TBI and their strong psychometric properties. These tests
provide a high level of sensitivity in detecting the potential effects of head injury, and
facilitate the application of results to clinical assessment and treatment planning.

Specific Aim 1. To evaluate theoretical models predicting performance on
measures of cognitive firnctioning based upon long-term mild TBI status. Three general
classes of models were considered, and tested as appropriate: direct effects models,
processing speed (PS)-mediated models, and executive functions (EF)-mediated models.
It was hypothesized that head injury status would be a significant predictor of all general
areas of cognition measured, but that an EF-mediated model would best represent the
data collected.

Specific Aim 2. To evaluate theoretical models predicting patient-reported “post-
concussive symptom” complaints (PCS) from long-term mild TBI status. Three general
classes of models were considered, and tested as appropriate: direct effects models,
psychiatrically-mediated models, processing speed (PS)-mediated models, and executive
function (EF)-mediated models. It was hypothesized that mild TBI would be a
significant predictor of PCS, but that an EF-mediated model would best represent the data

collected.

19

METHODS
Participants

Part 1: Screening. All individuals enrolled in courses at Michigan State
University that utilized the Human Participants Subject Pool (HPSP) were eligible for
participation in the initial screening phase of the study. The HPSP required all students
who wished to participate in research to register at a central web site maintained by the
university. Through this web site, students were able to obtain course credit for the
completion of web-based measures. The screening questionnaire described below was
one such measure, identified as part of a “study of injury” for which all students qualified
to participate in the subject pool were eligible. Informed consent for participation was
obtained with the use of electronic signatures, arr IRB approved practice that is self-
contained within the HPSP web page. In addition to collecting information related to
eligibility for Part 2 of the study, name, phone number, and e-mail address were obtained
as contact information for participant recruitment. A total of 3,347 students participated
in the HPSP screenings for this study. Demographic characteristics of the screening
sample are shown in Table 1.

Part 2: Neuropsychological Testing. Of those individuals who participated in
screening, 332 (9.92%) were invited to participate in Part 2 of the study
(neuropsychological testing). Of these individuals, 141 (42.47%) volunteered to
participate. Fifteen participants were excluded fi'om analyses because of information
obtained during interview which violated group eligibility criteria and was inconsistent
with information obtained during screening. This included 4 individuals whose mild

TBIs had occurred greater than six years previously, 3 individuals who had experienced

20

two or more orthopedic injuries within the previous five years, 2 individuals who had
experienced a head injury less than three months prior to testing, 2 individuals who had
experienced three or more lifetime head injuries, 2 individuals who had no injuries of
sufficient severity to qualify as a lifetime mild TBI or an orthopedic injury within the last
five years, 1 individual who had been hospitalized for greater than 24 hours as the result
of an injury, and 1 individual who had experienced PTA for greater than 24 hours as the
result of an injury.

This study follows the Kay, Harrington, Adams, Anderson, Berrol et a1. (1993)
definition of mild TBI: an episode in which an individual reports sustaining blunt head
trauma that is accompanied by a loss of consciousness (LOC) that does not exceed 30
minutes in duration, and/or a period of post-traumatic amnesia (PTA) of less than 24
hours. Individuals who report ever experiencing LOC for greater than 30 minutes, PTA
for greater than 24 hours, a penetrating head injury, or more than two lifetime head
injuries were excluded from both the mild TBI (MTBI) and orthopedic injury (01)
groups. The 01 group is composed of those individuals who do not report having ever
sustained a head injury (of any severity) but do report having sustained a broken bone.
Individuals who were hospitalized for greater than 24 hours for treatment were excluded
from both MTBI and 01 groups in order to control for injury severity. Other exclusions
for both the MTBI and 01 groups include the following: evidence of current substance
dependence that may prevent sober testing, history of psychotic illness, history of autism,
sensory-motor handicap, neurological illness, or native language not English. Individuals
who were prescribed stimulant, anti-psychotic, anti-depressant, or anti-convulsant

medications were also excluded.

21

One aspect of backup recruitment was also utilized, in which participants who had
already completed the screening described above were offered and provided with
monetary compensation for participation in the neuropsychological assessment portion of
the study. This strategy enabled the research team to recruit individuals who had
participated in the pre-screening in prior semesters, as well as individuals who
participated in the pre-screening in the current semester but who may not need and/or
want course credit for other reasons. Monetary compensation was offered at fair and
appropriate levels (i.e., $10 to $15 per hour of participation) that minimized any possible
coercive influence. Written informed consent was obtained from all individuals at the
beginning of the testing session.

The final sample consisted of 63 participants who had sustained a mild TBI within
the past 6 years (MTBI Group), and 63 participants who had sustained a mild orthopedic
injury within the past 6 years (01 Group). A priori power analyses indicated that the
power to detect medium effects sizes with this N would be .80 for independent-sample t-
tests and ANOVAs, .86 for Chi-squares, .94 for correlations, and .93 to .99 for
regressions analyses with 1 to 5 predictors. Power to test variable interrelationships in
SEM models tends to be comparable to that in more traditional analytic techniques, but
could not be estimated directly. However, the samples of the present size are considered
“medium” by Kline (2004).

Procedure

Following the online screening, participants who met criteria for the MTBI or 01

groups were contacted and asked if they would like to participate in a study of

“individual differences in cognition” for additional course credit. Those who volunteered

22

met with an examiner, who was blind to the participants’ personal history of injury, for a
one-on-one neuropsychological testing session. The participant then completed a fixed
battery of tests and questionnaires, as described below. Completion of this battery took
two hours or less. Following testing, the examiner engaged the participant in a brief
structured interview, described below. The participant was then be debriefed in full,
awarded course credit or paid as necessary, and released.
Measures

Constructs of interest and their indicators are summarized in Table 2. Indicators
of executive function and their theoretical components are summarized in Table 3.

Screening Questionnaire. This online questionnaire was been designed to obtain
the following information: name, phone number, e-mail address, gender, age,
racial/ethnic group membership, native language, level of education, handedness, current
sensory/motor capabilities, fiequency and type of alcohol and substance use, current
medications prescribed, history of neurological illness, psychotic illness, and
developmental disorder, and a detailed history of injuries. For head injuries, detailed
information was obtained to allow classification of injury severity. Additionally, all
items from the Post-Concussion Syndrome Checklist (PCSC; Gouvier et al., 1992)
were included in this questionnaire in order to obtain data regarding participants’ current
level of post-concussive symptomatology. Total PCSC score was used as a measure of
post-concussive symptoms, as it has demonstrated validity and specificity for this
purpose (Gouvier et al., 1992).

Wisconsin Card Sorting Test (W CST). The WCST is a measure of abstract

reasoning, concept formation, and set-shitting. This test requires the participant to sort a

23

number of cards based upon shifting criteria of color, shape, and number (Heaton et al.,
1993). Along with studies employing functional brain imaging which have shown that
this task produces activation in prefrontal areas (Rezai, Andreasen, Alli ger, Cohen,
Swayze et al., 1993; Konishi, Kawazu, Uchida, Kikyo, Asakura et al., 1999; Berrnan,
Ostrem, Randolph, Gold, Goldberg etal., 1995; Rogers, Andrews, Grasby, Brooks, &
Robbins, 2000; Wang, Kakigi, & Hoshiyarna, 2001), studies have also found that greater
numbers of perseverative errors on this test are associated with prefrontal lesions
(Anderson, Darnasio, Jones, & Tranel, 1991) and frontal metabolic deficits (Lombardi,
Anderson, Sirocco, Rio, Gross et al., 1999). This study employed a computerized 128-
card version of the WCST that takes 10 to 15 rrrinutes to administer. Total number of
perseverative errors on the WCST was used as a measure of executive firnction.
Color-Word Interference Test (CWIT). The CWIT is a subtest of the Delis-
Kaplan Executive Functioning System (D-KEFS; Delis, Kaplan, & Kramer, 2001a; Delis,
Kaplan, & Kramer, 2001b) that has been adapted from the Stroop Color-Word Test
(Stroop, 1935). This widely-used test contains four trials which are designed to measure
the ability to focus attention, inhibit prepotent responses, and shift mental set.
Participants are required to read as quickly as possible through a list of “xxxx” in blue,
red, or green ink (Condition 1: Color Naming), a list of color words (“blue”, “red”, or
“green”; Condition 2: Word Reading), a list of color words in which the color of the text
is incongr'uent with the word itself (Condition 3: Inhibition), and a list whichrequires the
participant to alternate between list demands of Conditions 1, 2, and 3 (Condition 4:
Inhibition/ Switching). Total time to administer this task is approximately 6 minutes.

Studies have shown that individuals with frontal lobe lesions have greater difficulty with

24

the Inhibition trial than both matched controls (Vendrell, J unque, Pujol, Jurado, Molet et
al., 1995) and individuals with non-fiontal brain lesions (Stuss, Floden, Alexander,
Levine, & Katz, 2001). Total time on Condition 3 (Inhibition) was used as an index of
executive functioning for this study.

Trail Making Test (TMT). The TMT is a subtest of the D-KEF S (Delis, Kaplan,
& Kramer, 2001a; Delis, Kaplan, & Kramer, 2001b) that has been adapted from Reitan’s
(1958) original test of the same name. This widely-used, timed paper-and-pencil test
consists of five conditions. In Condition 1 (Visual Scanning), the participant is presented
with a page of circles containing various letters and numbers, and they are required to
identify and “cross ou ” all of the circles containing the number 3 only. In Condition 2
(Number Sequencing), the participant is presented with a similar page of circled letters
and numbers and is required to draw a line connecting the numbered circles in sequential
order. Similarly, in Condition 3 (Letter Sequencing) the participant is required to connect
only letters in sequential order. Condition 4 (Number-Letter Switching) requires the
participant to draw a line connecting numbered and “lettered” circles in alternating
sequential-alphabetic order. Finally, Condition 5 (Motor Speed) requires participants to
draw a line connecting empty circles in a pre-specified order (indicated by a dotted line
on the page). Total time to administer this task is approximately 7 minutes. Score on each
part of the TMT is determined by the time required to complete each trial. Whereas
performance on Conditions 1, 2, 3, and 5 of the TMT depend largely upon the
participant’s psychomotor speed and visual search abilities, Condition 4, analogously to
Trial B of Reitan’s (1958) TMT, places additional demands upon the participant’s

working memory and cognitive flexibility (Crowe, 1998; Arbuthnott & Frank, 2000).

25

Total time on trial 4 (Letter-Number Switching) was used as a measure of executive
function in this study.

Verbal Fluency Test (V FT). The VFT is a subtest of the D-KEFS (Delis,
Kaplan, & Kramer, 2001a; Delis, Kaplan, & Kramer, 2001b) measuring verbal
generativity/initiation (fluency) that has been adapted fi'om a number of similar tests
(Benton, Harnsher, & Sivan, 1983). This test consists of three conditions. In Condition 1
(Letter Fluency), the participant is asked to say as many words as possible within one
minute that begin with a particular letter of the alphabet, excluding proper nouns as well
as different forms of the same word. Condition 1 has three trials, one each for the letters
“F”, “A”, and “S”. In Condition 2 (Category Fluency), the participant is asked to say as
many words as possible that fall within a certain category (animals or boy’s names)
within one minute each. Total scores for Conditions 1 and 2, respectively, are the total
number of correct responses for all trials. Condition 3 (Category Switching) is similar to
condition 2, but the participant is required to alternate between two categories (fi'uits and
furniture) during a single one-minute trial. Total score (correct switches) for Condition 3
is the number of responses that fall within the specified categories and alternation pattern.

The total time to administer this task is approximately 7 minutes. Performance on
measures of verbal fluency has been found to be sensitive to frontal lobe dysfunction
(Raskin & Rearick, 1996), and is associated with prefrontal brain activation in studies of
functional imaging (Frith, F riston, Liddle, & Frackowiak, 1991; Gourovitch, Kirkby,
Goldberg, Weinberger, Gold et al., 2000). This study utilized total score for Condition 1
(Letter Fluency), total score for Condition 2 (Category Fluency), and total number of

correct switches for Condition 3 (Category Switching), as indices of executive function.

26

Operation Span (OS). OS is a measure of working memory capacity adapted
from Turner & Engle’s (1989) study that is administered with the aid of a computer. The
task consists of 12 trials total. During each trial, the participant is presented with a series
of arithmetic problems followed by a word (e.g., “Is (6 X 2) — 5 = 7? DOG”). The
participant is instructed to respond “Yes” or “No” to the arithmetic problem, say the word
aloud, and try to remember the word for later recall. Immediately after the participant
says the word, the examiner presses a key that presents a blank screen for 500 ms.
Subsequently, either another arithmetic problem and word, or a recall cue appears,
depending upon the total number of items and the number of items already presented in
the selected trial. When the recall cue appears, the participant is instructed to repeat all
words from that trial. The number of items in each trial increases sequentially from two
to five over each of the 12 trials. This study used total number of word errors as an index
of executive function. Total time to administer this test was approximately 5 nrinutes.

Rotation Span (RS). RS is a measure of working memory capacity adapted
from Shah and Miyake’s (1996) study that is administered with the aid of a computer.
The task consists of 12 trials total. During each trial, the participant is first presented (for
1000 ms) with a long or a short arrow in 1 of 8 rotational orientations. Subsequently, a
normal or mirror-reversed letter (G, F, or R) is presented, in l of 8 rotational orientations.
The participant is instructed to mentally rotate the letter so that they can verbally indicate
whether it is normal (“Yes”) or mirror-reversed (“No”). Immediately after the participant
responds, the examiner presses a key that presents a blank screen for 500 ms.
Subsequently, either another arrow or a recall cue appears, depending upon the total

number of arrow-letter pairs and the number of arrow-letter pairs already presented in the

27

selected trial. When the recall cue appears, the participant is instructed to indicate all
arrows (long or short) in their correct rotational orientations on an answer sheet. The
number of arrow-letter pairs in each trial increases sequentially from two to five over
each of the 12 trials. This study used total number of arrow errors as an index of
executive function. Total time to administer this test was approximately 5 minutes.

California Verbal Learning Test — Second Edition (CVLT-II). The CVLT-II
is a l6-item word-learning task that requires the participant to learn four words from each
of four semantic categories over a series of four presentations, as well as after a short and
a long delay (Delis, Kramer, Kaplan, & Ober, 2000). This study used the total number of
correct Immediate Recall responses (trials 1-5), the total number of correct Short Delayed
Recall responses, and the total number of correct Long Delay Recall responses as indices
of memory functioning. Total time to administer this task was approximately 20 minutes,
plus a 20-minute delay, during which other tasks were administered.

Letter Comparison (LC), Number Comparison (NC), and Pattern
Comparison (PC). For these paper-and-pencil tests of processing speed, participants are
required to examine a series of pairs of letter groups (LC), numbers groups (NC), or line
segments (PC), and decide as rapidly as possible if each pair of stimuli is the same or
different (Salthouse & Babcock, 1991). For each stimulus pair, the participant writes S
on a line in between the pair if they are the same, or D if they are different. After being
provided with 3 example pairs, the participant is given 30 seconds to complete as many
items as possible for each of two total pages of the task. Total time to administer all three
of these tasks is approximately 5 minutes. For each task, the participant received an

overall score of total number correct minus total number incorrect; items that were not

28

attempted did not affect overall task score. This study used total score for LC, NC, and
PC, respectively, as indices of processing speed (PS).

Intelligence. A two-subtest short form equivalent of the WAlS-III (W echsler,
1997) was used to estimate Full-Scale IQ (FSIQ) for the purposes of demographic
description of the sample and various between-group comparisons. Reports indicate that
this short form, which utilizes only the Vocabulary and Block Desigr subtests of the
WAIS-III, has good reliability (.94 to .95) and validity (.89 to .90) within an 18-34 year
old sample (Jeyakumar, Warriner, Raval, & Ahmad, 2004). Data fi'om Sattler (2001)
were used to compute an estimate of F SIQ from the WAIS-III scaled scores fiom these
tasks.

Vocabulary. Vocabulary is a subtest of the WAIS-III in which the participant is
required to orally define a series of word that are presented both orally and in writing by
the examiner (Wechsler, 1997). Each word is scored according to various criteria as
inadequate (0 points), marginally adequate (1 point), or correct (2 points). Total score
from this measure was used in combination with Block Design for the purpose of
estimating FS IQ.

Block Design. Block Design is a subtest of the WAIS-III in which the
participant is asked to use small blocks to reproduce a series of two-dimensional designs
presented by the examiner (W echsler, 1997). Each block has two white sides, two red
sides, and two sides that are half red and half white. Up to 14 trials are administered
(more commonly, 7 to 10) in order of increasing difficulty. Average time to administer

this task is approximately 12 minutes. This study used total score as an index of

29

visuospatial functioning. Total score from this measure was used in combination with
Vocabulary for the purpose of estimating Full-Scale IQ.

Brief Symptom Inventory (BSI). The BSI is a 53-item questionnaire measure
that assesses nine psychiatric symptom dimensions, including somatization, obsessive-
compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety,
paranoid ideation, and psychoticism. Participants are asked to rate the amount of distress
associated with each symptom on a 5-point rating scale (Derogatis & Spencer, 1982). In
this study, total symptom severity score was used as an overall index of psychiatric
symptoms for primary path analyses. Individual dimension scores were used for further
exploratory analyses. Normative data for a university was used to determine clinical
cutoffs levels of psychiatric symptoms (Cochran & Hale, 1985). Individuals with
moderate to severe levels of overall symptoms were provided with an appropriate referral
for assessment/treatment.

Structured Interview. The structured interview contained the same questions
related to injury history that the participant provided in the screening questionnaire.
Additionally, information was collected regarding the use of medications and other
substances to ensure that screening criteria for medications, alcohol, or drugs were not
violated. Because of the presumed greater validity of an in-person interview with a
trained staff member than the initial on-line screening questionnaire, the structured
interview was used as the primary source of information when verifying group inclusion
criteria. However, some participants were re-contacted to voluntarily provide additional

information for more information to resolve injury history discrepancies.

30

fluistfl Analyses

Data Preparation. As recommended in recent methodological texts, extreme
outliers (z > 4.0 and more than .5 SD fiom next score) were truncated to within .5 SD of
the next nearest score to prevent the undue influence of single scores on linear models
and reduce Type I and Type 11 error (see Wilcox, Keselman, & Kowalchuk, 1998). The
expectation maximization (EM) method was used to irnpute missing data (0.53% of all
data points).

Analysis. The AMOS 5.0 (2003) statistical package, using the Maximum
Likelihood method, was used for all latent variable and structural analyses. The
GPOWER (1996) statistical program was used for all power analyses. The SPSS 11.01
(2001) statistical package was used for all other analyses. For all SEM analyses, the
following fit indices have been reported and interpreted as outlined by Kline (2004): (1)
Pearson chi-square for absolute comparisons of fit between nested models; (2) Goodness
of Fit Index (GFI; Joreskog & Sorbom, 1981) for which values .90 or greater are
considered good fit; (3) Comparative Fit Index (CF I; Bentler, 1990) for which values .90
or greater are considered good fit; and (4) Root Mean Square Error of Approximation
(RMSEA; Steiger, 1990) for which values .08 or less are considered acceptable and .05
or less are considered good fit. Although the use of structural equation modeling with
manifest variable models is statistically equivalent to a regression-based path analytic
approach (Kline, 2004), structural equation modeling also provides a number features and
additional statistics that are useful in evaluating and comparing models. Hypotheses for
specific variable relationships have been evaluated based upon the statistical significance

of relevant path/covariance coefficients inherent to each model. Hypotheses regarding

31

the relative fit of a given model haven been examined through a comparison of relevant
indices of overall model fit (e. g., xi, GFI, CFI, RMSEA).

As a secondary check on the accuracy of structural SEM models, standardized
composite variables were computed for all cognitive constructs of interest (executive
functions, processing speed, and verbal memory) with the same indicators utilized in the
SEM models, and all hypotheses were evaluated using traditional statistical techniques
(e. g., t-tests & correlations). Results from these analyses were consistent in all cases with
those reported below. Additionally, considering the possible effects of gender and the
overrepresentation of females in the overall sample, all structural SEM models were re-
evaluated with the inclusion of gender as a manifest variable in order to control for
gender effects; all relationships in these models were consistent with those reported

below.

32

RESULTS

Sample Description

Demographic information for the participant groups is provided in Table 4. Mean
years of education (12.86 years for both MTBI and 01 groups) reflects the recruitment
methods, which obtained primarily Freshmen or Sophomores at the University. Mean
Estimated F SIQ (113.32 and 111.90 for MTBI and 01 groups, respectively) was elevated
relative to the general population but also consistent with the university sample. Mean
age, education, and FSIQ were not significantly different between the groups. The
majority of participants were female (63.5% and 69.8% for the MTBI and 01 groups,
respectively), a proportion which was not significantly different between the groups.
Mean alcoholic drinks per month (30.44 and 22.72 for the MTBI and 01 groups,
respectively) and mean days per month in which other drugs were used (1.11 and .62 for
the MTBI and 01 groups, respectively) were somewhat higher in the MTBI than the 01
group (marginally significant at p = .06 for both comparisons). Consistent with the
screening sample, both participant groups were primarily Caucasian (85.7% and 95.2%
for the MTBI and 01 groups, respectively). The overall chi-square for the between-
groups comparison of ethnicity/race was non-significant (38(4) = 7.52, p > .05). Results
were also non-significant when examining race/ethnicity as a dichotomous variable
(Caucasian and Non-Caucasian; 78(1) = 3.32, p > .05).

Regarding characteristics of the injuries sustained between the two participant
groups, a significantly smaller proportion of participants received treatment for their
injury in the MTBI group than in the 01 group (73.0% and 92.1%, respectively, p < .01).

However, neither group was significantly more likely to be hospitalized for their injury

33

(39.7% and 52.4% for the MTBI and 01 groups, respectively). Participants were
significantly more likely to be involved in litigation related to their injury in the MTBI
group than in the 01 group (4.8% and 0%, respectively, p < .01 ). However, it should be
noted that the number of participants involved in litigation was very small overall relative
to other available studies of injury. An examination of the cause of injury by participant
group (Table 5) demonstrates that cause of injury in the MTBI group is somewhat more
varied than that in the 01 group; the chi-square statistic was significant at the p < .05
level for the overall comparison. However, participation in sports activities was the
primary source of injury for individuals in both the MTBI and 01 groups (50.8% and
74.6%, respectively), with accidental falls (22.2% and 11.1%, respectively) being the
second most common cause. The cause of injury in this sample is somewhat discrepant
from other studies of the US population as a whole, which cite motor vehicle accidents
and accidental falls as the primary causes of mild TBI (Bazarian, McClung, Shah,
Cheng, Flesher et al., 2005; Cassidy et al., 2004). This contrast is likely due to
differences between the university sample used in this study and the hospital/treatment
samples typical of other studies.
Descriptive Statistics

Independent samples t-tests were conducted to examine potential differences
between the MTBI and 01 groups on all relevant cognitive and symptom variables (see
Tables 6 and 7). As shown, the MTBI group performed significantly better than the 01
group on two of the three processing speed tasks: LC [t( 124) = 2.34, p < .05] and PC
[t(124) = 2.34, p < .05]. A lack of significant differences between groups on TMT Trial 5

(Motor Speed, t(124) = -.26, p > .05), and significant correlations between head injury

34

and these variables when TMT Trial 5 was covaried, suggested that slowed motor
performance could not account for these differences in processing speed. Significant
differences were not found for any other cognitive variables, or for any of the symptom
variables examined.
Confirmatog Factor Anplysis of Cognitive Vaerles

In the first stage of model analysis, a series of CFAs was conducted to determine
the appropriate factor structure of the EF construct. An initial CFA was conducted in
which all EF variables (WCST perseverative errors, CWIT Trial 3 total time, TMT Trial
4 total time, VFT Letter, Category, and Switching scores, OS total correct trials, and RS
total correct trials) loaded onto a single latent EF factor. In this model (denoted EFI), the
unstandardized factor loading of VFT Letter score was set to l to scale the latent
EF variable; all other factor loadings were freely estimated. Additionally, in order to
properly account for method covariance, the error terms for VFT Letter, Category, and

Switching were permitted to correlate with one another, as were error terms for OS and
RS. Model EFl demonstrated good overall fit, 3806, N = 126) = 15.91, p > .05, GFI =

.97, CFI = 1.00, RMSEA = .00; however, the loading of VFT Switching on the EF factor

was non-significant. Therefore, VFT Switching was excluded from subsequent models.
Model EF2, which demonstrated good fit [x2(12, N = 126) = 14.59, p > .05, GFI = .97,

CFI = .98, RMSEA = .04], was constructed by removing VFT Switching fiom EF 1. All
factor loadings in this model were statistically significant at the p < .05 level, a result
which was unchanged by testing the model with alternate indicators’ factor weights set to

1 for scaling. Model EF2 is depicted in Figure 2.

35

In the next stage of analysis, a multiple—groups confirmatory factor analysis was
conducted to evaluate the appropriateness of using the EF2 factor model with both the
MTBI and 01 groups, respectively. This is accomplished by evaluating the measurement
model in both groups simultaneously with appropriate levels of cross- group equality
constraints. If there is no significant difference in fit (as determined by a comparison of
the chi-square statistic) of an unconstrained model to a model in which factor loadings
are equality-constrained, then the construct’s factor structure is judged to be comparable
across the groups. However, a significant reduction in fit would suggest that the relations
of the indicators to one another varied by group.

The baseline, unconstrained tvvo-group model (EF 3) demonstrated acceptable
overall fit [x2(24, N = 126) = 34.54, p > .05, GFI = .93, CFI = .91, RMSEA = .06]. The
following model (EF4), which constrained the EF factor loadings to be equal across
groups, demonstrated good overall fit [3360, N = 126) = 37.79, p > .05, GF I = .93, CFI =
.94, RMSEA = .05], which was not significantly changed from that of the baseline,

unconstrained multi-group model EF3 [Ax2(6, N = 126) = 3.25, p > .05]. It was therefore

concluded that the factor structure of the EF construct was similar between the two
groups, a finding which permitted interpretation of results from latent models of EF
including all participants. Fit statistics for EF models are presented in Table 8.

Analytic procedures were modified slightly in order to evaluate latent factors for
processing speed (PS) and verbal memory (VM), factors whose number of indicators (3
each) prevents the evaluation of each factor’s fit as an isolated model. Instead, a
measurement model (C1) was created with two latent factors: one factor for PS (whose

indicators included LC, NC, and PC totals), and one factor for VM (whose indicators

36

included Immediate Memory total, Short Delay total, and Long Delay total).
Unstandardized factor loadings for LC and Immediate Memory were set to 1 for sealing
of the latent PS and VM factors, respectively, and these latent factors were permitted to
correlate. The initial fit of model Cl was good, x2(8, N = 126) = 7.41 , p > .05, GFI = .98,
CF I = 1.00, RMSEA = .00. All factor loadings in this model were statistically significant
at the p < .05 level, a result which was unchanged by testing the model with alternate
indicators’ factor weights set to l for scaling. Model C1 is depicted in Figure 3.

In the next stage of confirmatory factor analysis for the combined PS-VM model,

a baseline, unconstrained two-group model was created. This model (C2) demonstrated
good overall fit, 7806, N = 126) = 11.60, p > .05, GFI = .97, CFI = 1.00, RMSEA = .00.
The following model (C3), which constrained all factor loadings to be equal across
groups, demonstrated good overall fit [x2(20, N = 126) = 17.29, p > .05, GFI = .96, CF I =
1.00, RMSEA = .00], which was not significantly changed from that of the baseline,
unconstrained multi-group model C2 [Ax2(4, N = 126) = 5.70, p > .05]. It was therefore

concluded that the factor structure of the PS and VM constructs was similar between the
two groups, which permitted interpretation of results from latent models of these
constructs including all participants. Fit statistics for PS and VM models are presented in

Table 8.

Head Injm Status Predicting Cogpitive Functioning

 

Executive Functions (EF). Next, structural equation modeling was used to
evaluate primary hypotheses about the relationship between head injury status and the

cognitive constructs of interest. EF was examined first, as this construct was

37

hypothesized to play a mediational role between head injury status and other cognitive
functions. The path model (denoted H1) was constructed such that the dichotomous

variable representing head injury status directly predicted the latent EF variable
(established in model EF2). Overall fit for model H1 was good, 780 8, N = 126) = 18.22,

p > .05, GFI = .97, CFI = 1.00, RMSEA = .01. Head injury status was not a significant
predictor of EF (B = .05, p > .05), indicating that primary hypotheses about the
relationship between head injury status and EF could not be supported. Further, this
finding eliminated EF fi'om consideration as a possible mediator between head injury
status and other variables of interest.

Processing Speed (PS). The relationship between head injury status and PS was
evaluated next. The path model (denoted H2) was constructed such that the dichotomous
variable representing head injury status directly predicted the latent PS variable
(established in model Cl). Overall fit for model H2 was good, 780, N = 126) = 1.41 , p >
.05, GF I = .99, CFI = 1.00, RMSEA = .00. Head injury status significantly predicted PS
(B = -.28, p < .05), indicating that OI group membership was related to poorer PS
performance relative to MTBI group membership, a relationship opposite of that
hypothesized. This finding was not affected by the inclusion of gender as a control in the
model.

Verbal Memory (VM). A third path model of this type (denoted H3) was
constructed with the dichotomous variable representing head injury status directly

predicted the latent VM variable (established in model C1). Overall fit for model H3 was
good, {(2, N = 126) = .10, p > .05, GFI = 1.00, CFI = 1.00, RMSEA = .00. Head injury

status was not significantly predictive of VM (B = -.01, p > .05), indicating that direct-

38

effects hypotheses about the relationship between head injury status and VM could not be
supported. Fit statistics for models in which head injury status predicts cognitive
functioning are presented in Table 9.
Post-Concussive Symptoms LPCS) and Cognitive Functioning

Executive Functions (EF). Next, structural equation modeling was used to
examine the relationships between PCS and the cognitive constructs of interest. This
model (denoted PCSI) was constructed such that the manifest variable PCSC was
permitted to correlate with the latent EF variable (established in model EF 2). Overall fit
for model PCS] was good, 7808, N = 126) = 18.52, p > .05, GFI = .97, CFI = 1.00,
RMSEA = .02. PCSC was not significantly related to EF (r = .17, p > .05).

Processing Speed (PS). The relationship between head injury status and PS was
evaluated next. This model (denoted PCSZ) was constructed such that the manifest

variable PCSC was permitted to correlate with the latent PS variable (established in
model c1). Overall fit for model Pcsz was good, {(2, N= 126) = 1.50, p > .05, GFI =
.99, CFI = 1.00, RMSEA = .02. PCSC was not significantly related to PS (r = .02, p >
.05).

Verbal Memory (VM). The third model of this type (denoted PCS3) was constructed
such that the manifest variable PCSC was permitted to correlate with the latent VM
variable (established in model Cl). Overall fit for model PCS3 was good, {(2, N = 126)
= .49, p > .05, GFI = 1.00, CFI = 1.00, RMSEA = .00. PCSC was significantly correlated
with VM, r = .20, p < .05, indicating better memory performance for individuals with

more severe symptoms reported on the PCSC. This finding was not affected by the

39

inclusion of gender as a control in the model. Fit statistics for models examining PCS
and cognitive functioning are presented in Table 10.
mtionships between Head Injugy, Supptom Variablesyand EF

Hypotheses regarding the relationships between head injury status, EF,
psychiatric symptoms, and PCS were evaluated by constructing a model (denoted SYMI)
which included the latent EF variable (established in model EF 2) as well as manifest
variables BSI (representing psychiatric symptoms), PCSC, and head injury status. All
latent and manifest variables were permitted to correlate. Overall fit for model SYMI
was good, {(30, N= 126) = 39.74, p > .05, GFI = .94, CFI = .94, RMSEA = .05. No
variables in this model were significantly related to EF or head injury status. BSI was
significantly correlated with PCSC (r = .50, p < .01), indicating that increased severity of
symptoms on the PCSC was related to increased severity of symptoms on the BSI. The
bivariate correlation of PCSC and BSI was also r = .50, p < .01, when examined using
traditional statistical techniques. Consistent with the results of partial correlations, this
finding was not affected by the inclusion of gender as a control in the SEM model.
Follow-up Data Analyses

Pearson correlations between PCSC and BSI subscales (shown in Table 11) were
examined in order to more fully explore the relationship between these two measures.
Notable relationships included the following: BSI Somatization and PCSC Dizziness (r =
.33, p < .01); BSI Somatization and PCSC Fatigue (r = .32, p < .01); BSI Hostility and
PCSC Irritability (r = .46, p < .01); BSI Obsessive-Compulsive and PCSC Memory (r =
.26, p < .01); BSI Obsessive-Compulsive and PCSC Difficulty Concentrating (r = .37, p

< .01); and PCSC Anxiety and the BSI Total (r = .37, p < .01). Examination of BSI and

40

PCSC items revealed significant content overlap in each of these areas. As a result, a BSI
adjusted total was calculated, excluding items #2 (“dizziness”), #5 (“trouble
remembering”), #6 (“feeling . .irritated”), #36 (“trouble concentrating”), and #37
(“feeling weak”). Additionally, a PCSC adjusted total was calculated excluding the
Anxiety subscale. The relationship between the BSI and PCSC totals adjusted in this
manner was similar, r = .46, p < .01, suggesting that the relationship between these two
variables was not due solely to similarities in item content.

In order to explore further the relationship demonstrated between PCSC total
score and the latent Verbal Memory variable in model PCS3, correlations were also
examined between PCSC subscales and a Verbal Memory composite variable computed
as the average of standardized Immediate Memory, Short Delay, and Long Delay
variables. Significant correlations were found between the Verbal Memory composite
variable and Total score of the PCSC (r = .19, p <. 05), as well as the Memory (r = .19, p

< .05) and Visual Disturbances (r = .18, p < .05) PCSC subscales.

41

DISCUSSION

This study was conducted to investigate the effects of mild TBI on cognitive
functioning and post-concussive syrnptomatology in the post-acute phase of recovery,
with particular focus on the possible role of executive functions. Results failed to
provide support that mild TBI leads to impairment in any cognitive domains examined
(including executive functions, processing speed, and verbal memory). Although this
finding is not unique in and of itself (see Bernstein & de Ruiter, 2000; Frencham, Fox, &
Mayberry, 2005), methodological strengths of the study permit a greater degree of
interpretation of these results than has been possible in previous studies with similar
findings, and offer insights into some non-neurological factors that may be more relevant
to an individual’s long-term firnctioning after a mild TBI than the head injury itself.

Paradoxically, the orthopedic injury control group in this study performed worse
on measures of processing speed than the mild TBI group (although these differences
were rather small in magnitude). Differences between the groups in motor performance
could not account for this finding, which appears to rule out the possibility that certain
forms of injury (e.g., previously broken arms, wrists, or fingers) could have slowed
manual task performance. Similarly, gender did not appear to explain this relationship.
Rather, differences in injury characteristics that were found between these two groups
suggest that the orthopedic injuries sustained by individuals in this sample may have had
more impact than the mild TBIs in some aspects. For example, individualswith
orthopedic injuries received medical treatment at a significantly higher rate than
individuals with mild TBIs, which could be related to more severe injuries in general, or

also simply to the greater perceived importance of the need for treatment with a broken

42

bone. In any case, the relatively greater attention given to a broken bone than to a mild
TBI might be expected to be related to increased psychological impacts (such as distress)
related to the injury. Further, individuals with broken bones may experience physical
disability that altogether prevents them fi'om engaging in certain activities for a period of
time, potentially increasing the psychological impact relative to the short-term cognitive
impairment that may follow a mild TBI.

Additionally, as reported previously by others (Sawchyn, Brulot, & Strauss, 2000;
Suhr & Gunstad, 2002), no relationship was found between head injury status and
purportedly “post-concussive” symptomatology within the time frame of the current
study (3 months to 5 years post-injury). Instead, these self-reported symptoms appeared
better explained by factors related to general emotional/psychiatric distress. Though
substantial item content overlap existed between the measures of post-concussive
symptoms and emotional/psychiatric symptoms utilized in this study, elimination of
overlapping items and subscales did little to mitigate the relationship between these two
variables. Similar results have also been found previously for other measures of these
symptom domains, such as the Post-Concussional Index (PCI), the Beck Depression
Inventory-2 (BDI—II), and the Beck Anxiety Inventory (BAI) (Suhr & Gunstad, 2002;
Trahan, Ross, & Trahan, 2001). As such, the validity of scales such as the PCSC as
measures of “post-concussive” symptoms is suspect, and clinicians and researchers
would be well-advised to collect comprehensive information on the emotional
functioning of individuals who report high levels of these types of symptoms.

Post-concussive symptoms were not significantly related to executive function or

processing speed, findings which failed to provide support for hypotheses linking these

43

aspects of neuropsychological functioning to self-reported problems with attention and
concentration. PCSC was significantly correlated with VM, indicating slightly better
memory performance for individuals with more severe symptoms reported on the PCSC.
Analysis of PCSC subscales revealed that this finding could be attributed in part to items
on the PCSC relating to reports of memory problems and visual disturbances.
Complaints of memory problems on the PCSC were therefore somewhat more common
in individuals with better memory abilities; these findings are difficult to fully explain.

Though the results of this study failed to support primary hypotheses, strengths of
the study’s design give these findings substantially greater weight than many previously
conducted studies. For example, few studies of mild TBI have had samples of the size
available to this study (n = 126), drawn outside of a treatment setting, with a very low
rate of involvement in litigation (2.4% overall). Additionally, no known previous studies
of mild TBI have examined executive functions as comprehensively, using a large
number of individual measures combined with latent modeling techniques of analysis to
aid in data interpretation. Further, the inclusion of an injury control group is believed to
have substantially mitigated the effects of a number of confounding factors related to
generic effects of injury that are important to neuropsychological performance (see
Dikmen & Levin, 1993; Satz et al., 1999). Additional factors that may affect cognition or
symptom presentation related to injury type and treatment have also been highlighted for
further investigation.

Many of these same methodological issues are also important when considering
the study’s limitations. Even though young adults are at a greatly increased risk for mild

T BI relative to older adults (Cassidy et al., 2004; Bazarian et al., 2005), the restricted age

44

range of the present sample may limit generalizability of the present results to
populations of older adults. Females were also over-represented in the current sample;
although gender was not found to moderate any relationships of interest, this could have
biased the results in subtler ways. Additionally, the inclusion of a second, non-injury
control group might have allowed for more comprehensive evaluation of injury-related
factors affecting neuropsychological performance. Finally, although the
neuropsychological tasks used in this study are believed to have adequately assessed the
cognitive domains of executive functioning, verbal memory, and processing speed, a
number of cognitive domains (e.g., visuospatial ability, language) were not examined,
and it is possible that an alternate battery of neuropsychological tasks might provide
somewhat differing results.

In conclusion, clinically significant deficits in executive functions, processing
speed, and verbal memory do not appear to be associated with mild TBI in the post-acute
phase of recovery. Additionally, long-term post-concussive symptoms appear to be
related to general emotional distress, rather than to the actual occurrence of a mild TBI.
Valuable aspects of future studies in this area could include the examination of alternate
cognitive domains, additional non-injury control groups, and more comprehensive
evaluations of non-neurological factors related to neuropsychological performance

following injury.

45

 

 

 

2:03:50

2:025:00 . ,

nanom—

mucosa—chm
uranium»;

 

 

 

 

 

 

 

Eon—OE

 

 

 

coon»

553002.—

 

 

/ \

 

.3325“—
02500:“

 

 

Enigma—om 033:; ecfimofiogm mo eowommom .l_ 31me

 

 

_fl.—. i=2

 

 

46

Table 1. Demographic characteristics of screening sample

 

 

(Total N = 3,347) Valid N 8:22:31?
Mean Age in Years 3289 19.49
(SD) (1.78)
Mean Years Education 3325 13.02
(SD) (1.02)
% Female 3337 70.6%
% Hispanic or Latino/a 3337 1.9%
% Native American 3337 0.3%
% Asian 3337 4.9%
% African American / Black 3337 7.6%
% Native Hawaiian / Pacific Islander 3337 0.3%
% Caucasian / White 3337 81.4%
% “Other” Ethnicity/ Race 3337 3.3%

 

47

Igble 2. Congtructs and indicators‘.

 

 

 

 

 

 

 

Construct Measures (Indicators)

Executive Function WCST (Perseverative Errors); CWIT (Trial 3 Total Time); TMT (Trial 4
Total Time); VF I (Letter, Category, and Switching scores); OS (Total
word errors); RS (Total arrow errors)

Processing Speed LC (Total Score); NC (Total Score); PC (Total Score)

Verbal Memory CVLT-II (Immediate Recall total, Short Delayed Recall total, Long Delay
Recall total)

Post-Concussive PCSC (Total score)

Symptoms

Psychiatric BSI (Total score)

Symptoms

Intelligence Estimated F SIQ (W AIS-III Vocabulary and Block Design)

 

 

 

 

' WCST = Wisconsin Card Sorting Test, CWIT = D-KEFS Color-Word Interference Test, TMT = D-KEFS
Trail Making Test, VFT = D-KEFS Verbal Fluency Test, OS = Operation Span, RS = Rotation Span;
SDMT = Symbol Digit Modalities Test, LC = letter Comparison, NC = Number Comparison; PC = Pattern
Comparison, CVLT-II = California Verbal Learning Test 2, PCSC = Post-Concussion Syndrome Checklist,
BSI = Brief Symptom Inventory.

48

mle 3. Executive Function Indicators and Theoretigl Components 2.

 

 

 

 

 

 

Measure Indicator(s) Theoretical EF Component(s)
WCST Perseverative Errors Abstract Reasoning, Cognitive
Flexibility
CWIT Trial 3 (Inhibition) Total Time Inhibition, Cognitive Flexibility
TMT Trial 4 (Number-Letter Switching) Cognitive Flexibility
Total Time
VF I Letter, Category, and Switching Behavioral Initiation / Fluency,
scores Cognitive Flexibility
OS/RS Total Word/Arrow Errors Working Memory Capacity

 

 

 

 

 

2 WCST = Wisconsin Card Sorting Test, CWIT = D-KEF S Color-Word Interference Test, TMT = D-KEF S
Trail Making Test, VF'I = D-KEF S Verbal Fluency Test, OS = Operation Span, R8 = Rotation Span

49

'_1‘_able 4. Demogaphic Characteristics by Pficipant Group

 

 

 

 

MTBI 01 1
Group Group p
N 63 63 --
Mean AgeinYears 19.19 19.10
(SD) (1.06) (1.17) ns
Mean Years Education 12.86 12.86
(SD) (.91) (.91) ns
Mean Estimated FSIQ 113.32 111.90
(SD) (12.49) (9.21) ns
Mean Alcoholic Drinks Per Month 30.44 22.72
(SD) (30.31) (25.97) .06
Mean Days Per Month Drugs Used 1.11 .63
(SD) (3.47) (1.55) .06
% Female 63.5% 69.8% ns
Ethnicig / Racez MTBI 01
Group Group
% Hispanic or Latino/a 3.2% 0.0%
% Native American 0.0% 0.0%
% Asian 4.8% 0.0%
% African American / Black 3.2% 4.8%
% Native Hawaiian / Pacific
Islander 0.0% 0.0%
% Caucasian / White 85.7% 95.2%
% “Other” Ethnicity / Race 3.2% 0.0%

 

’ Results of independent-samples T-test.
2 Chi-square statistic was non-significant for the overall comparison.

50

Table 5. Injm Characterigics by Participant Group

 

 

 

MTBI OI

Group Group p
% Received Treatment for Injury 73.0% 92.1% <.01
% Hospitalized for Injury 39.7% 52.4% ns
% Involved in Litigation 4.8% 0% <0]

Cause of Injugrl gigs; (,ng

% Motor Vehicle Accidents 14.3% 3.2%
% Sports Participation 50.8% 74.6%
% Accidental Falls 22.2% 11.1%
% Interpersonal Violence 1.6% 0.00%
%Other Cause ofInjury 11.1% 11.1%

 

I Chi-square statistic was significant at the p < .05 level for the overall comparison.

51

Table 6. Descriptive Statistics ml T-tests of Cognitive Variablesl

 

 

vmable df w 153.13%) Difhéfzrlice' b d
6.57 6.25
WCST Perseverative Errors 124 (341) (303) _32 -55 .10
TMT Trial 4 Time Total 124 8:2?) 823;) -31 -.13 —.02
TMT Trial 5 Time Total 124 (251.31) (2531):) -.26 -.26 -.05
VFT Trial 1 Correct 124 (3878768) (1:33?) -1.06 -.60 -.ll
VFT Trial 2 Correct 124 3:88) 3528617) 1.41 1.14 .20
VFT Trial 3 Switches Total 124 (123317) (12259:) .18 .37 .07
OS Word Errors Total 124 (2:90;) (15941.?) .87 .79 .14
RS Arrow Errors Total 124 (12:29) (11:13.68) .11 .09 .02
CWIT Trial 3 Time Total 124 38532:) 285.4025) .17 .1 1 .02
CVLT Immediate Recall Total 124 (584-9581) (5:90;) 26 .17 .03
CVLT SD Free Total 124 (12137:) (121.36) .04 .10 .02
CVLT LD Free Total 124 (121.378 (121.511) -.02 -.04 -.01
LC Total 124 (253,75 (231.3391) 1.92 234* .42
NC Total 124 gig?) (129$) .60 .87 .16
PC Total 124 $72278) (3:35;) 2.83 240* .43

 

I WCST = Wisconsin Card Sorting Test, CWIT = D-KEF S Color-Word Interference Test, TMT = D-KEFS
Trail Making Test, VF'I = D—KEFS Verbal Fluency Test, OS = Operation Span, RS = Rotation Span;, LC
= Letter Comparison, NC = Number Comparison; PC = Pattern Comparison, CVLT-II = California Verbal
Learning Test-2

’ MTBI Group Mean — 01 Group Mean

b t value of independent-samples t-test

* Statistically significant at the (p < .05) level (2-tailed)

52

Table 7. Desm'ptive Statisticsemd T-tests of Self—Reported Symptom Variables

 

 

 

MTBI Q1
Group Group
Variable df . Mean , b d
Mean Mean Difference
(SD) (SD)
BSI Somatization 124 .40 .32 .08 1.05 .19
(.48) (.38)
BSI Obsessive-Compulsive 124 1.08 1.16 -.08 -.60 -.11
(.77) (.66)
BSI Interpersonal 124 .54 .61 -.08 -.71 -.13
Sensitivity (.61) (.59)
BSI Depression 124 .44 .43 .01 .12 .02
(.54) (.46)
BSI Anxiety 124 .43 .49 -.06 -.75 -.13
(.45) (.42)
BSI Hostility 124 .52 .44 .09 .94 .17
(.57) (.44)
BSI Phobic Anxiety 124 .15 .08 .07 1.29 .23
(.34) (.24)
BSI Paranoid Ideation 124 .48 .52 -.04 -.37 -.07
(.54) (.52)
BSI Psychoticism 124 .32 .32 .01 .08 .01
(.51) (.38)
BSI Total (GSI) 124 .49 .49 .01 .12 .02
(.41) (.32)
PCSC Total 124 62.03 60.02 2.01 .89 .16
(12.77) (12.75)

 

* Statistically significant at the (p < .05) level (2-tailed)
" MTBI Group Mean — 01 Group Mean

b t value of independent-samples t-test

53

.38 .35

TMT VFT] VFT2 OS RS WCST CWIT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

-.44 .49 .25 -.42 -.55 -.47 -.53

Figure 2. Freely-estimated model of EF factor structure (Model EF2).
Standardized values Shown. All factor loadings significant at the p < .05 level (2-tailed).

54

 

m 73°

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LC NC PC [M SD LD
\94 .39 f46 .8
(n8)

F igpre 3. Freely-estimated model of PS and VM factor structure (Model C1).
Standardized values shown. All factor loadings Significant at the p < .05 level (2-tailed).

55

Ipble 8. Confirmatory Factor An_al\§is of Cognitive Variables

 

 

Model X: df p GFI crr RMSEA
EF, 15.91 16 > .05 .97 1.00 .00
EF2 14.59 12 > .05 .97 .98 .04
EF3 34.54 24 > .05 .93 .91 .06
EF, 37.79 30 > .05 .93 .95 .05
cl 7.41 8 > .05 .98 1.00 .00
c2 11.60 16 > .05 .97 1.00 .00
c3 17.29 20 > .05 .96 1.00 .00

 

Note. p = 126 for all analyses; GFI = Goodness-of-fit index; CFI = Comparative fit index; RMSEA =

Root Mean Square Error of Approximation.

56

Table 9. Models of Head Injrmy Sptatus Predicting Cognitive Functioning

 

 

Model x2 df p GFI CFI RMSEA
H, 18.22 18 > .05 .97 1.00 .01
H2 1.41 2 > .05 .99 1.00 .00
H3 .10 2 > .05 1.00 1.00 .00

 

Note. n = 126 for all analyses; GF I = Goodness-of-fit index; CFI = Comparative fit index; RMSEA =

Root Mean Square Error of Approximation.

57

Table 10. Models of Post-Concussive Smptcms and Emotional/Psychiatric Functioning

 

 

Model x2 df p GFI CFI RMSEA
PCSl 18.52 18 > .05 .97 1.00 .02
PCS2 1.50 2 > .05 .99 1.00 .02
PCS3 .49 2 > .05 1.00 1.00 .00
SYM, 39.74 30 > .05 .94 .94 .05

 

Note. r_r = 126 for all analyses; GFI = Goodness-of-fit index; CFI = Comparative fit index; RMSEA =

Root Mean Square Error of Approximation.

58

625-3 35 2e. v 5 as a sesame 2.83% r.

325$ :32 so. v s as a 58%.? baccaesm ..

56%. «.8328 E» .8 09 u e N

Emotonoxmm .1. >mm EocfiE 20:85

n Mai SE38: n m0: 90252 n XZ< E23230 .1. nEm “bSEEom Know—0&3:— H mm moZm—anUczmmomn—O M 0-0 ”cementeEom u 20m 90852:
EBA—ENE norm u Em ”banana u XZ< $80505 :55ng n 82. ”882 3 eose>fiww< n 52 3859535 33.5 .1. m; moawcem u 0.5m MERE—8:00
£855 n 028 messed case: n 52 Emcee: u 5: ”82:55 u NE m2.23.8: u 94m: 4.5985 seesaw eeaaoeooaaca n omen. _

 

 

Ion. _ mien. INm. 1.94. *st. :3. *L0. 1.3. 1.3. :94. :E. .58. Umom
L stem. scum. aaoN. 2. 2. .13. :Nm. SEN. INm. :N. *3. :38. Em
Ebb. INm. :oN. o_. 5.- 2. *9. *mN. N_. INN. IN. 2. >mm Em
Ibo. :N. 2. no. 2. 2. we. no. *2. SEN. S. 8. mafia— Em
:3. 1.94. seem. a... _ m. mo. N_. «N. ..:.&N. ac. asmN. :mN. 5. CIA— Em
3:. INv. :oN. 3. RN. *2. *LHN. *2. N_. 1.94. 2. 3. m0: Em
:5. :34. :mm. *coN. *MN. 8. SEN. seem. aimN. seem. :mN. LN. XZ< Em
INw. tom. :oN. *NN. mo. 3. LN. :oN. *NN. *2. SEN. C. me Em
1v». sswm. .4st. .IoN. ea. 2. *LdN. *2. «ON. *NN. «N. no. mm Em
1.2.. :94. “SN. IN. 2. no. LN. scum. SEN. IbN. *NN. 3. 0-0 Em
save. .83. *NN. SEN. S. *2. :Nm. :mN. LN. LN. summ. LN. EOE Em
:38. :30... XZ< 83—. ~02 m_> 0.5?— UZOU 352 My: NLQ Q<mm
Em “EDA— Omom Umom “EDA— Umom “EDA “EDA UmUm ”EDA. UmUm Umom

 

 

emu—paxw EoEEE cameo—om mo 32:22.80 .: DBeIH

59

References

Anderson, S. W., Darnasio, H., Jones, R. D., Tranel, D. (1991). Wisconsin Card Sorting
Test performance as a measure of frontal lobe damage. Journal of Clinical &
Experimental Neuropsychology, 13(6), 909-922.

Alves, W., Macciocchi, S. N., & Barth, J. T. (1993). Postconcussive symptoms after
uncomplicated mild head injury. Journal of Head Trauma Rehabilitation, 8(3)
48-59.

Arbuthnott, K., & Frank, J. (2000). Trail Making Test, part B as a measure of executive
control: Validation using a set-switching paradigm. Journal of Clinical and
Experimental Neuropsychology, 22(4), 518-528.

Benton, A. L., Harnsher, K., & Sivan, A. B. (1983). Multilingual aphasia examination
(3rd Ed.). Iowa City, IA: AJA Associates.

Berman, K. F., Ostrem, J. L., Randolph, C., Gold, J., Goldberg, T. E., Coppola, R.,
Carson, R. E., Herscovitch, P., & Weinberger, D. R. (1995). Physiological
activation of a cortical network during performance of the Wisconsin Card
Sorting Test: A positron emission tomography study. Neuropsychologia, 33(8),
1027-1 046.

Barkley, R. A. (2001). The executive functions and self-regulation: An evolutionary
neuropsychological perspective. Neuropsychology Review, 11 (1), 1-29.

Bazarian, J. J, McClung, J., Shah, M. N., Cheng, Y. T., Flesher, W., & Kraus, J. (2005).
Mild traumatic brain injury in the United States, 1998-2000. Brain Injury, 19, 85-
91.

Bentler, P. M. (1990). Significance tests and goodness of fit in the analysis of
covariance structures. Psychological Bulletin, 107, 238-246.

Bernstein, D. M., & de Ruiter, S. W. (2000). The influence of motivation on
neurocognitive performance long after mild traumatic brain injury. Brain and
Cognition, 44, 50-66.

Bigler, E. D. (1999). Neuroimaging in Mild TBI. In Vamey, N. R., & Roberts, R. J.
(Eds.), The Evaluation and Treatment of Mild Traumatic Brain Injury. Mahwah,
NJ: Lawrence Erlbaurn Associates.

Bigler, E. D. & Snyder, J. L. (1995). Neuropsychological outcome and quantitative
neuroimaging in mild head injury. Archives of Clinical Neuropsychology, 10(2),
1 59-174.

Binder, L. M., Rohling, M. L., & Larrabee, G. J. (1997). A review of mild head trauma.

60

Part I: Meta-analytic review of neuropsychological studies. Journal of Clinical
and Experimental Neuropsychology, 19, 421-43 1.

Boone, K. B., Ponton, M. 0., Gorsuch, R. L., Gonzalez, J. J ., & Miller, B. L. (1998).
Factor analysis of four measures of prefrontal lobe functioning. Archives of
Clinical Neuropsychology, 13(7), 585-595.

Brewer, T. L., Metzger, B. L., & Therrien, B. (2002). Trajectories of cognitive recovery
following a minor brain injury. Research in Nursing and Health, 25, 269-281.

Brooks, J ., Fos, L. A., Greve, K. W., & Hammond, J. S. (1999). Assessment of
executive function in patients with mild traumatic brain injury. Journal of
Trauma, 46(1), 159-163.

Burgess, P. W., Alderman, N., Evans, J ., Emslie, H., & Wilson, B. A. (1998). The
ecological validity of tests of executive function. Journal of the International
Neuropsychological Society, 4, 547-558.

Capruso, D. X., & Levin, H. S. (1996). Neurobehavioral outcome of head trauma. In
Evans, R. W. (Ed.), Neurology and T rauma_(pp. 3-17). W. B. Saunders
Company: Philadelphia.

Cassidy, J. D., Carroll, L. J., Peloso, P. M., Borg, J ., von Holst, H., Holm, L., et a1.
(2004). Journal of Rehabilitation Medicine, 43 (Suppl.), 28-60.

Chan, R. (2002). Attentional deficits in patients with persisting postconcussive
complaints: A general deficit or specific component deficit? Journal of Clinical
and Experimental Neuropsychology, 24(8), 1 081-1093.

Cicerone, K., & Azulay, J. (2002). Diagnostic utility of attention measures in
postconcussion syndrome. The Clinical Neuropsychologist, 16(3), 280-289.

Cochran, C. D., & Hale, W. D. (1985). College student norms on the Brief Symptom
Inventory. Journal of Clinical Psychology, 41, 777-779.

Cremona-Meteyard, S. L., & Geffen, G. M. (1994). Persistent visuospatial attention
deficits following mild head injury in Australian rules football players.
Neuropsychologia, 32(6), 649-662.

Crowe, S. F. (1998). The differential contribution of mental tracking, cognitive
flexibility, visual search, and motor speed to performance on parts A and B of the
Trail Making Test. Journal of Clinical Psychology, 54(5), 585-591.

Cudmore, L., Segalowitz, S., & Dywan, J. (2000). EEG coherence shows altered

frontal-parietal communication in mild TBI during a dual-task. Brain and
Cognition, 44(1), 86-90.

61

De Kruijk, J. R., Twijnstra, A., & Leffers, P. (2001). Diagnostic criteria and differential
diagnosis of mild traumatic brain injury. Brain Injury, 15(2), 99-106.

Delis, D. C., Kaplan, E., & Kramer, J. H. (20013). The Delis-Kaplan Executive Function
System: Examiner's Manual. San Antonio: The Psychological Corporation.

Delis, D. C., Kaplan, E., & Kramer, J. H. (2001b). The Delis-Kaplan Executive Function
System: Technical Manual. San Antonio: The Psychological Corporation.

Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. (2000). The California Verbal
Learning Test — 2"d edition. San Antonio, TX: The Psychological Corporation.

Della Sala, 8., Gray, C., Spinnler, H., & Trivelli, C. (1998). Frontal lobe fimctioning in
man: the riddle revisited. Archives of Clinical Neuropsychology, 13(8), 663-682.

Derogatis, L. R., & Spencer, P. M. (1982). The Brief Symptom Inventory (BSI)
Administration. Scoring & Procedures Manual-I. Johns Hopkins University
School of Medicine.

Dikmen, S. S., & Levin, H. S. (1993). Methodological issues in the study of mild head
injury. Journal of Head Trauma Rehabilitation, 8(3), 30-37.

Dikmen, S. S., Machamer, J ., & Temkin, N. (2001). Mild head injury: facts and artifacts.
Journal of Clinical and Experimental Neuropsychology, 23(6), 729-73 8.

Duncan, J ., Johnson, R., Swales, M., & Freer, C. (1997). Frontal lobe deficits after head
injury: unity and diversity of function. Cognitive Neuropsychology, 14(5), 713-
741.

Duncan, J ., & Owen, A. M. (2000). Common regions of the human frontal lobe recruited
by diverse cognitive demands. Trends in Neurosciences, 23(10), 47 5-483.

Engle, R. W., & Kane, M. J. (2004). Executive attention, working memory capacity, and
a two-factor theory of cognitive control. The Psychology of Learning and
Motivation, 44, 145-199.

Erdfelder, E., F aul, F ., & Buchner, A. (1996). GPOWER: A general power analysis
program. Behavior Research Methods, Instruments, & Computers, 28, 1-11.

Ettenhofer, M. L., Harnbrick, D. Z., & Abeles, N. (in press). Reliability and structural
stability of executive function in a normal aging population. Neuropsychology.

Fee, C. R., & Rutherford, W. H. (1987). A study of the effect of legal settlement on
postconcussion symptoms. Archives of Emergency Medicine, 4, 159.

Fisher, D., Ledbetter, M., Cohen, J ., Marmor, D., & Tulsky, D. (2000). WAIS-III and

62

WMS-III profiles of mildly to severely brain-injured patients. Applied
Neuropsychology, 7(3), 126-132.

Frencham, K. A., Fox, A. M., & Mayberry, M. T. (2005). Neuropsychological studies of
mild traumatic brain injury: A meta-analytic review of research since 1995.
Journal of Clinical and Experimental Neuropsychology, 27, 334-351.

Frith, C. D., Friston, K. J ., Liddle, P. F ., Frackowiak, R. S. (1991). A PET study of word
finding. Neuropsychologia, 29(12), 1137-1148.

Goldman-Rakic, P. S. (1998). The prefrontal landscape: implications of functional
architecture for understanding human mentation and the central executive. In
Gouvier, W. D., Cubic, B., Jones, G., Brantley, P., & Cutlip, Q. (1992).
Postconcussion symptoms and daily stress in normal and head-injured college
populations. Journal of Clinical Neuropsychology, 7, 193-211.

Gourovitch, M. L., Kirkby, B. S., Goldberg, T. E., Weinberger, D. R., Gold, J. M.,
Esposito, G., Van Horn, J. D., & Berman, K. F. (2000). A comparison of rCBF
patterns during letter and semantic fluency. Neuropsychology, 14(3), 353-360.

Hanna-Pladdy, 3., Berry, Z., Bennett, T., Phillips, H., & Gouvier, W. (2001). Stress as a
diagnostic challenge for postconcussive symptoms: Sequelae of mild traumatic

brain injury or physiological stress response. The Clinical Neuropsychologist,
15(3), 289-304

Hart, T., Schwartz, M. F., & Mayer, N. (1999). Executive function: some current
theories and their applications. In Vamey, N. R. & Roberts R. J. (Eds), The
evaluation and treatment of mild traumatic brain injury, (pp. 133-148). Lawrence
Erlbaum Associates: Mahwah, New Jersey.

Hayden, M. E. (1997). Mild traumatic brain injury: a primer for understanding its
impact on employee return to work. AAOHN Journal, 45(12), 63 5-645.

Heaton, R. K., Chelune, G. J ., Talley, J. L., Kay, G. G., & Curtiss, G. (1993). Wisconsin
Card Sorting Test Manual. USA: Publishing Assessment Resources.

Heyder, K., Suchan, B., & Daum, I. (2004). Cortico-subcortical contributions to
executive control. Acta Psychologica, 115, 271-289.

Hinton-Bayre, A., Geffen, G., & McFarland, K. (1997). Mild head injury and speed of
information processing: A prospective study of professional rugby league players.
Journal of Clinical and Experimental Neuropsychology, 19(2), 275-289

Jeyakumar, S. E., Warriner, E. M., Raval, V. V., & Ahmad, A. (2004). Balancing the

63

need for reliability and time efficiency: short forms of the Wechsler Adult
Intelligence Scale-III. Educational and Psychological Measurement, 64(1), 71-
87.

J oreskog, K. G., & Sorbom, D. (1981). Analysis of linear structural relationships by
maximum likelihood and least squares methods (Research Report No. 81 -8).
Uppsala, Sweden: University of Uppsala.

Kay, T., Harrington, D.E., Adams, R., Anderson, T., Berrol, S., Cicerone, K., et a1.
(1993). Definition of mild traumatic brain injury. Journal of Head Trauma
Rehabilitation, 8(3), 86-87.

Kline, R. B. (2004). Principles and Practice of Structural Equation Modeling. Guilford
Press: New York.

Konishi, S., Kawazu, M., Uchida, I., Kikyo, H., Asakura, I., & Miyashita, Y. (1999).
Contribution of working memory to transient activation in human inferior
prefrontal cortex during performance of the Wisconsin Card Sorting Test.
Cerebral Cortex, 9(7), 745-753.

Kraus, J. F., & McArthur, D. L. (1996). Epidemiology of brain injury. In Evans, R. W.
(Ed.), Neurology and Trauma, (pp. 3-17). W. B. Saunders Company:
Philadelphia.

Larrabee, G. J. (1999). Current controversies in mild head injury: scientific and
methodological considerations. In Varney, N. R. & Roberts R. J. (Eds), The
evaluation and treatment of mild traumatic brain injury, (pp. 327-346). Lawrence
Erlbaum Associates: Mahwah, New Jersey.

Lees-Haley, P. R., Green, P., Rohling, M. L., Fox, D. D., Allen, L. M. (2003). The
lesion(s) in traumatic brain injury: implications for clinical neuropsychology.
Archives of Clinical Neuropsychology, 18, 585-594.

Lezak, M. (1995). Neuropsychological Assessment (3rd ed.). New York: Oxford
University Press.

Lombardi, W. J ., Andreason, P. J ., Sirocco, K. Y., Rio, D. E., Gross, R. E., Umhau, J. C.,
& Hommer, D. W. (1999). Wisconsin Card Sorting Test performance following
head injury: dorsolateral fronto-striatal circuit activity predicts perseveration.
Journal of Clinical and Experimental Neuropsychology, 21(1), 2-16.

Lowe, C., & Rabbitt, P. (1998). Test/retest reliability of the CANT AB and ISPOCD
neuropsychological batteries: theoretical and practical issues. Neuropsychologia,
36(9), 915-923.

Mangels, J. A., Craik, R, Levine, B., Schwartz, M. L., & Stuss, D. T. (2002). Effects of

64

divided attention on episodic memory in chronic traumatic brain injury: a function
of severity and strategy. Neuropsychologia, 40, 2369-23 85.

Mathias, J. L., & Coats, J. L. (1999). Emotional and cognitive sequelae to mild
traumatic brain injury. Journal of Clinical and Experimental Neuropsychology,
21(2), 200-215.

McAllister, T., Sparling, M., Flashman, L., & Saykin, A. (2001). Neuroimaging findings
in mild traumatic brain injury. Journal of Clinical and Experimental
Neuropsychology, 23(6), 775-791.

McDonald, B. C., Flashman, L. A., & Saykin, A. J. (2002). Executive dysfimction
following traumatic brain injury: Neural substrates and treatment strategies.
Neurorehabilitation, 17(4), 333-344.

Miyake, A., Friedman, N., Emerson, M. J ., Witzki, A. H., Howerter, A., & Wager, T. D.
(2000). The unity and diversity of executive functions and their contributions to

complex “fiontal lobe” tasks: a latent variable analysis. Cognitive Psychology,
41, 49-100.

Mooney, G., & Speed, J. (2001). The association between mild traumatic brain injury
and psychiatric conditions. Brain Injury, 15(10), 865-877.

Newcombe, F., Rabbitt, P., Briggs, M. (1994). Minor head injury: pathophysiological or
iatrogenic sequelae? Journal of Neurology, Neurosurgery, and Psychiatry, 5 7,
709-716.

Pinkston, J ., Gouvier, W., & Santa Maria, M. (2000). Mild head injury: Differentiation
of long-term differences on testing. Brain and Cognition, 44(1), 74-78.

Paniak, C., Reynolds, 8., Phillips, K., Toller-Lobe, G., Melnyk, A., & Nagy, J. (2002).
Patient complaints within 1 month of mild traumatic brain injury: A controlled
study. Archives of Clinical Neuropsychology, 1 7(4), 319-334.

Phillips, L. H. (1997). Do “fiontal tests” measure executive function? Issues of
assessment and evidence from fluency tests. In P. Rabbit (Ed.), Methodology of
Frontal and Executive Function (pp. 191-213). Psychology Press: East Sussex,
UK.

Plassman, B. L., Havlik, R. J ., Steffens, D. C., Helms, M. J ., Newman, T. N., et a1.
(2000). Documented head injury in early adulthood and risk of Alzheimer’s
disease and other dementias. Neurology, 55(8), 1158-1166.

Powell, B. D., & Holmes, C. B. (1995). Prevalence of head injury among four-year

college students: a replication and combined results. Perceptual and Motor Skills,
81(1), 227-230.

65

Rabbitt, P. (1997). Methodology and models in the study of executive functions. In P.
Rabbitt (Ed.), Methodology of Frontal and Executive Function (pp. 1-38).
Psychology Press: East Sussex, UK.

Raskin, S. A., & Rearick, E. (1996). Verbal fluency in individuals with mild traumatic
brain injury. Neuropsychology, 10(3), 416-422.

Reitan, R. M. (1958). Validity of the Trail Making Test as an indicator of organic brain
damage. Perceptual & Motor Skills, 8, 271-276.

Rezai, K., Andreasen, N. C., Alliger, R., Cohen, G., Swayze, V., & O’Leary, D. S.
(1993). The neuropsychology of the prefrontal cortex. Archives of Neurology,
50(6), 636-642.

Rizzo, M., & Tranel, D. (1996). Overview of head injury and post-concussive
syndrome. In Rizzo, M., & Tranel, D. (Eds.), Head Injury and Postconcussive
Syndrome, (pp. 1-18). Churchill Livingstone: New York.

Rogers, R. D., Andrews, T. C., Grasby, P. M., Brooks, D. J ., & Robbins, T. W. (2000).
Contrasting cortical and subcortical activations produced by attentional-set
shifting and reversal learning in humans. Journal of Cognitive Neuroscience,
12(1), 142-162.

Royall, D. R., Lauterbach, E. C., Cummings, J. L., Reeve, A., Rummans, T. A., Kaufer,
D. I., et a1. (2002). Executive control function: a review of its promise and

challenges for clinical research. Journal of Neuropsychiatry and Clinical
Neurosciences, 14(4), 377-405.

Rutherford, W. H., Merrett, J. D., & McDonald, J. R. (1977). Sequelae of concussion
caused by minor head injuries. Lancet, 1, 1-4.

Ryan, L. M., O’Jile, J. R., Gouvier, W. D., Parks-Levy, J ., & Betz, B. (1996). Head
injury in a college population: analysis of epidemiological factors. Applied
Neuropsychology, 3, 49-54.

Salthouse, T. A. (1991). Theoretical Perspectives on Cognitive Aging. Hillsdale, New
Jersey: Lawrence Erlbaum Associates.

Salthouse, T. A., & Babcock, R. L. (1991). Decomposing adult age differences in
working memory. Developmental Psychology, 27(5), 763-776.

Sattler, J. M. (2001). Assessment of Children: Cognitive Applications. 4‘11 Edition. La
Mesa, CA: Sattler Publishing.

Satz, P., Alfano, M. 8., Light, R., Morgenstem, H., Zaucha, K., Asarnow, R. F., et al.

66

(1999). Persistent post-concussive syndrome: A proposed methodology and
literature review to determine the effects, if any, of mild head injury and other
bodily injury. Journal of Clinical and Experimental Neuropsychology, 21(5),
620-628.

Sawchyn, J ., Brulot, M., & Strauss, E. (2000). Note on the use of the postconcussion
syndrome checklist. Archives of Clinical Neuropsychology, 15(1), 1-8.

Segalowitz, S., & Lawson, S. (1995). Subtle symptoms associated with self-reported
mild head injury. Journal of Learning Disabilities, 28(5), 309-319.

Shah, P., & Miyake, A. (1996). The separability of working memory resources for
spatial thinking and language processing: an individual differences approach.
Journal of Experimental Psychology: General, 125(1), 4-27.

Stablum, F ., Mogentale, C., & Umilta, C. (1996). Executive functioning following mild
closed head injury. Cortex, 32, 261-278.

Steiger, J. H. (1990). Structural model evaluation and modification: An interval estimate
approach Multivariate Behavioral Research, 25, 173-180.

Stroop, J. R. (193 5). Studies of interference in serial verbal reactions. Journal of
Experimental Psychology, 18, 643-662.

Stuss, D. T., & Benson, D. F. (1984). Neuropsychological studies of the frontal lobes.
Psychological Bulletin, 95(1), 3-28.

Stuss, D. T., Floden, D., Alexander, M. R, Levine, B., & Katz, D. (2001). Stroop
performance in focal lesion patients: dissociation of processes and frontal lobe
lesion location. Neuropsychologia, 39, 771-786.

Stuss, D. T., & Levine, B. (2002). Adult clinical neuropsychology: Lessons from studies
of the frontal lobes. Annual Review of Psychology, 33, 401-434.

Suhr, J. A., & Gunstad, J. (2002). Postconcussive symptom report: the relative influence
of head injury and depression. Journal of Clinical and Experimental
Neuropsychology, 24, 981-993.

Trahan, D. E., Ross, C. E., & Trahan, S. L. (2001). Relationships among
postconcussional-type symptoms, depression, and anxiety in neurologically
normal young adults and victims of mild brain injury. Archives of Clinical
Neuropsychology, 16, 43 5-445.

Turner, M. L., & Engle, R. W. (1989). Is working memory capacity task dependent?
Journal of Memory and Language, 28, 127-154.

67

Thornton, K. (2003). The electrophysiological effects of a brain injury on auditory
memory fimctioning: The QEEG correlates of impaired memory. Archives of
Clinical Neuropsychology, 18(4), 363-378.

Vamey, N. R., & Menefee, L. (1993). Psychosocial and executive deficits following
closed head injury: Implications for orbital fiontal cortex. Journal of Head
Trauma Rehabilitation, 8(1), 32-44.

Vendrell, P., Junque, C., Pujol, J ., Jurado, M. A., Molet, J ., & Grafinan, J. (1995). The
role of prefrontal lesions in the Stroop task. Neuropsychologia, 33(3), 341-3 52.

Wallesch, C. W., Curio, N., Kutz, S., Jost, S., Bartels, C., & Synowitz, H. (2001).
Outcome after mild-to-moderate blunt head injury: effects of focal lesions and
diffuse axonal injury. Brain Injury, 15(5), 401-412.

Wang, L., Kakigi, R., & Hoshiyarna, M. (2001). Neural activities during Wisconsin Card
Sorting Test — MEG observation. Cognitive Brain Research, 12, 19-31.

Wechsler, D. (1997). Wechsler Adult Intelligence Scale - Third Edition. San Antonio,
Texas: The Psychological Corporation.

Wilcox, R. R., Keselman, H. J ., & Kowalchuk, R. K. (1998). Can tests for treatment
group equality be improved?: The bootstrap and trimmed means conjecture.
British Journal of Mathematical and Statistical Psychology, 51, 123-134.

Zohar, O., Schrieber, S., Getslev, A., Schwartz, J. P., Mullins, P.G., & Pick, C. G.

(2003). Closed-head minimal traumatic brain injury produces long-term cognitive
deficits in mice. Neuroscience, 118, 949-955.

68

 

 

ul’lgmiljgjtflnjjj[ij

 

-.‘Jl _‘—.—.—n _ _