|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
PILOT DATA ON THE BEHAVIOR RATING INVENTORY OF EXECUTIVE
FUNCTION (BRIEF) AND PERFORMANCE MEASURES
OF EXECUTIVE FUNCTION IN PEDIATRIC TRAUMATIC BRAIN INJURY (TBI)
MICHELLE L. BENJAMIN
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
Michelle L. Benjamin
TABLE OF CONTENTS
L IS T O F T A B L E S ..................................................................................... .. ................. iv
LIST O F FIG U RE S ............... .......................................... ...v.... .. .... .v
ABSTRACT .................................................... ................. vi
1 IN TR O D U C T IO N ........ .. ......................................... ..........................................1.
Rationale for the Current Study..................................................... 13
A im s and H ypotheses .............. .. ............. .............................................. 15
2 M E TH O D S .............. .........................................................................................18
Subjects .................................................................................. . .. ...............18
Procedures .......................... .... ...................................21
Parent-Reported Executive Function ........................................................21
Performance Measures of Executive Function ........................... ...............22
A naly ses .............. ..................................................................... ......25
3 R E S U L T S .......................................................................................... ..................... 2 9
Post-Injury BRIEF M measures ...........................................................................29
Performance M measures of Executive Function .................................................31
Correlations between BRIEF Scales and Executive Function Performance
M measures ..........................................................................................................33
BRIEF Scales- BRIEF Scales ..................................................................33
BRIEF Scales-Neuropsychological Measures........................................33
Neuropsychological Measures-Neuropsychological Measures.................33
Differences in BRIEF Ratings for Pre- and Post-Injury .................................37
4 D ISC U SSIO N ............................................................................... ...................... 40
L im stations ..................................................................................................... 42
Strengths ............................................................................................. . 44
Future D directions ...........................................................................................45
LIST O F R EFEREN CE S ................................................................................................47
BIO GRAPH ICAL SK ETCH ..........................................................................................50
LIST OF TABLES
1 TBI classification ........ ................ .......... .............. ............... 19
2 D em graphic inform ation ........................................ ........................ ................ 20
3 M ean T scores for the 5 BRIEF scales................................................ ................ 29
4 Effect sizes for BRIEF scales post-injury ........................................... ................ 30
5 Mean T scores for the 5 performance measures of executive function.................31
6 Effect sizes for performance measures of executive function...............................32
7 Correlations between BRIEF scale scores and performance measures.................35
8 Correlations between BRIEF scale scores and performance measures.................36
9 Group x time differences in pre- and post-injury T scores for the 5 BRIEF scales .37
10 Pre- and post-injury T score differences for the 5 BRIEF scales..........................39
LIST OF FIGURES
1 B R IE F scales .............................................................................................. . 9
2 M ean T scores for the 5 BRIEF scales................................................ ................ 30
3 T scores for the 5 performance measures of executive function...........................32
4 Group x time differences pre- to post-injury for the 5 BRIEF scales ...................38
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
PILOT DATA ON THE BEHAVIOR RATING INVENTORY OF EXECUTIVE
FUNCTION (BRIEF) AND PERFORMANCE MEASURES
OF EXECUTIVE FUNCTION IN PEDIATRIC TRAUMATIC BRAIN INJURY (TBI)
Michelle L. Benjamin
Chair: Eileen B. Fennell
Major Department: Clinical and Health Psychology
Executive dysfunction has been reported in several studies of pediatric traumatic
brain injury (TBI). Research using the Behavior Rating Inventory of Executive Function
(BRIEF) in pediatric TBI has focused on cases of severe TBI evaluated several years
post-injury, and little is known about the relationship between the BRIEF and traditional
tests of executive function. Furthermore, previous studies did not explore individual
BRIEF scales and instead used global composite scores from the BRIEF. The current
study presents pilot data examining ratings of individual scales of the BRIEF and
performance on traditional tests of executive functioning for an acute pediatric TBI
sample with a range of severity. These data are compared to an orthopedic control group
within the first year of recovery.
Traumatic brain injury (TBI) occurs when the brain is injured as the result of an
external mechanical force (Loring, 1999). Injuries can be sustained in a variety of ways,
such as a direct blow, acceleration-deceleration movements of the brain within the skull
cavity, or from the penetration of the brain via an external object. Primary causes of TBI
are motor vehicle accidents, falls, and interpersonal violence. Measurements of severity
of injury include the Glasgow Coma Scale's (GSC) rating of eye, motor and verbal
responsiveness (Teasdale and Jennett, 1974) and other behavioral, physiological and
radiological evidence in addition to GSC ratings, such as abnormal MRI or CT findings,
loss of consciousness duration, and post-traumatic amnesia.
Yeates (2000) summarizes a number of neuropathological and pathophysiological
changes that occur following TBI. Primary injuries, i.e. injuries that result directly from
the trauma itself, consist of contusions, skull fractures, and mechanical injuries to blood
vessels and nerve fibers. Secondary injuries include brain swelling, increased intracranial
pressure, epidural and subdural hematomas, and seizures. Neurochemical changes
consist of the excessive production of free radicals, disruption of cellular calcium
homeostatis, and the excessive release of excitatory neurotransmitters.
The National Center for Injury Prevention and Control, Center for Disease Control
and Prevention (CDC) reports that approximately 1.5 million people sustain a traumatic
brain injury (TBI) each year in the United States. The CDC reports that this estimate is 8
times greater than the number of individuals diagnosed with breast cancer and 34 times
greater than the number of new cases of HIV/AIDS. Furthermore, the CDC estimates
that the number of people who experience long-term disability associated with a TBI
ranges from 80,000 to 90,000 (CDC, 2003).
TBI leads to an estimated 3,000 deaths, 29,000 hospitalizations and 400,000
emergency room visits each year for children under the age of 14 (Langlois and Gotsch,
2001). Kraus (1995) estimated that the average annual incidence of closed head injury is
180 per 100,000 children in children less than 15 years of age. With regards to TBI
severity, Yeates (2000) highlighted data provided by the National Pediatric Trauma
Registry and the United States National Coma Data Bank. Information from these
sources suggests that 76% to 85% of all TBIs are mild in nature, with the remaining cases
of TBI being evenly distributed between moderate and severe TBI.
In children and adolescents, traumatic brain injuries can produce dysfunction across
a number of domains, including alertness and orientation, intellectual functioning,
language skills, nonverbal skills, attention, memory, executive functions, motor skills,
academic achievement, behavioral and emotional adjustment (Yeates, 2000). Executive
function has become an area of study within the TBI literature in recent years. Executive
function research has not been as heavily published in the TBI literature as some
cognitive domains, such as memory and attention. Yeates (2000) explains that this may
be in part due to the complex and multifaceted nature of the construct. However, despite
the challenges of studying executive function, several research groups have pursued the
study of executive functions in children with TBI.
What exactly are "executive functions"? The construct of executive function
encompasses a heterogeneous sample of behaviors. Baron (2004) summarizes a range of
definitions for executive function that have been put forth by various researchers. One
definition explains executive function as an ability that helps one to maintain an
appropriate mental set in order to achieve a future goal. Another explanation describes
executive functions as the mechanisms used in order to optimize performance in
situations requiring the simultaneous operation of several different cognitive processes.
Other definitions of executive function include the following abilities: flexible thought
and action, planning and sequential processing of complex behaviors, simultaneously
attendance to multiple sources of information, resistance to distracting and interfering
stimuli from the environment, the ability to sustain certain behaviors over prolonged
periods, and the inhibition of inappropriate behaviors.
Baron (2004) provides an encompassing list of sample subdomains implicated in
executive function. These include set shifting, problem solving, abstract reasoning,
planning, organization, goal setting, working memory, inhibition, mental flexibility,
initiation, attentional control, and behavioral regulation. Typical measures of executive
function used in neuropsychological research include the Wisconsin Card Sorting Task,
Trail Making Test, various motor sequencing tasks (e.g., Go-No Go), Stroop Color and
Word Test, Booklet Category Test, n-back tasks, and verbal fluency.
Baron (2004) highlights research by Stuss and Benson (1986), who suggest that
executive functions are higher cognitive functions that serve to integrate other more basic
cognitive functions, such as memory, attention, and perception. Stuss and Benson list
among these higher functions the ability to set goals, anticipate, plan, monitor results, and
incorporate feedback. They also outline that the prefrontal cortex and interconnected
regions serves as the neuroanatomical substrates for executive function.
Given that the domains encompassed by the construct of executive function are
implicated in a variety of day-to-day tasks, it is clear that executive dysfunction could
cause a number of problems in everyday life. A few examples of executive function in
daily life include the following: being able to hold a goal in mind and perform a series of
actions to reach that goal, having the ability to inhibit inappropriate behavior so that one
does not offend another person during personal interactions, and adequately being able to
switch one's attention among several tasks as necessary. With diminished executive
function, one would struggle to adequately complete academic and occupational tasks as
well as appropriately maintain interpersonal relationships.
Researchers have evaluated the development of executive function skills in
children. Two examples of such research include Welsh and colleagues (1991) and Levin
and colleagues (1991). Welsh, Pennington, and Grossier (1991) examined children's
performance on a variety of executive function measures. Participants were 100 children
ages 3 to 12 years of age. Test measures consisted of verbal fluency, motor sequencing,
visual search, the Wisconsin Card Sorting Task, the Tower of Hanoi, and Matching
Familiar Figures Test. The authors found that adult-level performance was achieved at
three different ages within development. Children were able to perform the visual search
and Tower of Hanoi 3-disc task at adult levels by age 6. By age 10, adult-level
performance was displayed for the Matching Familiar Figures Test and the Wisconsin
Card Sorting Test. Adult-levels performance was reached for the TOH 4-disc, verbal
fluency and motor sequencing by adolescence.
Levin et al. (1991) evaluated 52 normal children ages 7 to 15 years using a wide
range of executive function measures. The testing battery consisted of the Wisconsin
Card Sorting Test, California Verbal Learning-Children's Version, Word Fluency,
Animal Naming, Design Fluency, the Twenty Questions task, the Go-No Go task, the
Tower of London, and Delayed Alternation. Children were divided into three groups by
ages, 7- to 8-year-olds, 9- to 12-year-olds, and 13- to 15-year olds, in order to determine
developmental change across the measures. With the exception of one test (Delayed
Alternation), developmental changes were found on all of the measures. Large gains
were found for various Wisconsin Card Sorting Test measures between the 7- to 8- and 9-
to 12-year-old groups. At this developmental shift, concept formation (measures by the
number of categories obtained) and problem-solving efficiency increased. In contrast,
the percentage of perseverative errors decreased with age. A developmental difference
was also found between the 7- to 8-year-olds and the 9- to 12-year-olds for false positive
errors on the Go-No Go Task. Further improvements in performance were demonstrated
up through ages 13 to 15 years for the California Verbal Learning Test, Twenty
Questions, and the Tower of London.
Research on executive function has extended into clinical populations, with
investigators utilizing a number of assessment measures with which to characterize
various neurological and psychiatric populations. Several research groups have
examined executive function in pediatric TBI. Studies in TBI have used
neuropsychological performance-based measures, behavioral reports completed by an
adult rater (e.g., parent), or a combination of these two types of assessment methods.
With regards to performance-based measures of executive function, Levin et al.
(1994) evaluated the Tower of London task performance in children with traumatic brain
injury. The Tower of London involves planning skills since the task requires the
participant to rearrange beads on three vertical rods to match a model. Task complexity
is determined by the minimum number of moves necessary to solve the problem. The
examiner reminds the child of the rules if he or she breaks a rule, such as picking up more
than one bead at a time. The authors found that children who had suffered a severe TBI
tended to break the rules despite reminders from the examiner. This tendency for rule
breaking was particularly notable in children between the ages of 6 and 10 at the time of
the assessment. Levin and colleagues also found that the initial planning time on the
Tower of London decreased with age and was prolonged in children who had sustained a
severe TBI. Such results suggest that task efficiency improved with age and was
inversely related to greater injury severity.
Levin et al. (1997) assessed performance in a pediatric TBI sample ages 5 to 18 for
three performance measures of executive function. The Twenty Questions Test, the
Tower of London, and the Wisconsin Card Sorting Test (WCST) were given to 151
children who sustained a head injury of varying severity approximately 3 years earlier
and 89 control subjects. The authors also examined a subset of this population involved
in a longitudinal study involving assessments at 3 and 36 months post-injury. In the
cross-sectional study, Levin et al. (1997) found that the severity of head injury adversely
affected performance on all three executive function tests. The severe TBI group showed
more inefficient strategies on the Twenty Questions Test, asking more questions than
either the mild TBI or control groups and asking questions that were less efficient in
allowing them to generate a correct answer. For the WCST, the severe TBI patients had
the lowest percent of conceptual responses and attained the fewest categories relative to
controls. Compared to both the mild TBI and control groups, the severe TBI group
displayed the highest percent of perseverative errors. On the Tower of London, the
severe TBI group solved a lower percentage of problems within three trials and had a
greater number of broken rules than the mild TBI and control groups. Within the severe
TBI group, young children who sustained a severe head injury had greater impairment for
solving the Tower of London relative to older severe TBI children. The young severe
TBI children demonstrated an increased tendency to break rules, and they also had more
difficulty solving Tower of London problems that were more complex. In the
longitudinal study, all three executive function measures showed improved performance
over three years, although there was a ceiling effect for the Tower of London. This
improvement over time was larger for children who had sustained a severe head injury.
Slomine et al. (2002) evaluated executive function performance in 68 children ages
7 to 15 with moderate to severe TBI 1 year post-injury as part of a structural
neuroimaging study. The authors used the WCST, the Tower of Hanoi, and the letter
fluency test. Children who had sustained a TBI at a younger age displayed more
perseverative errors on the WCST and worse performance on the letter fluency test.
Slomine and colleagues (2002) concluded that the risk for impairment on measures of
executive function is increased for children injured at a younger age.
Levin et al. (2002) assessed working memory differences in 44 normal controls, 54
mild pediatric TBI, and 26 severe TBI cases several years post-injury utilizing two
different n-back working memory paradigms. The first paradigm was a letter
identification (semantic) task in which subjects had to identify whether a specific target
letter, e.g., "X," had been presented a certain number of trials back from the current letter
presentation. The second paradigm was a phonological working memory paradigm for
which subjects had to identify whether a particular letter that rhymes with a particular
target, e.g., "C" for the letter "Z," had been presented a certain number of trials back
from the current rhyme. Both paradigms had varying memory load (ranging from 0 to 3
letters or rhymes back from the current presentation item). The correct detection of
targets as well as false alarms were measured for each task. Levin et al. (2002) found
that memory load and age significantly affect the detection of targets and false alarms in
both tasks. The identification of targets increased with age at testing across all memory
load conditions. Performance worsened for all subjects as memory load increased, with
the number of false alarms increasing and the number of targets detected decreasing with
increased load. TBI severity interacted with memory load for false alarms on the rhyme
task. The severe TBI group had more false alarms than either the mild TBI patients or
normal children on the 0-back condition. Mild TBI patients displayed more false alarms
than controls on the 0-back condition. For the 2-back condition, the severe TBI group
displayed more false alarms than the mild TBI group. The authors also found that the
rhyme condition was more difficult than the letter identification task. The authors
concluded that TBI often results in impaired working memory and diminished inhibition
Behavioral report of executive function has begun to be used in more recent years
as a measure of executive in the everyday environment. Most studies involving
behavioral report of executive function in pediatric TBI have utilized the Behavior Rating
Inventory of Executive Function (BRIEF; Gioia, Isquith, Guy, and Kenworthy, 2000).
The BRIEF is a relatively new instrument designed to assess executive function
behaviors in the home and school environment for children ages 5 to 18. Parent, teacher,
and self-report scales are available. The authors designed the instrument to be used with
children having a wide range neurological, psychiatric, developmental, and medical
conditions in order to assess executive dysfunction in a more ecologically valid way. The
Parent Form of the BRIEF contains 86 items that have been divided into eight
theoretically and empirically derived clinical scales that are purported to measure
different aspects of executive functioning: Inhibit, Shift, Emotional Control, Initiate,
Working Memory, Plan/Organize, Organization of Materials, and Monitor. Exploratory
factor analysis for the Parent Form produced two Composite Index scores. The Initiate,
Working Memory, Plan/Organize, Organization of Materials, and Monitor scales were
determined to make up the first Composite Index (labeled Metacognitive Index by the
authors), and the second Composite Index (labeled Behavioral Regulation Index) is
comprised of the Inhibit, Shift, and Emotional Control scales. A Global Composite Index
is also provided by the BRIEF and consists of the total score across all of the clinical
scales. Figure 1 outlines the scales that comprise the BRIEF.
Clinical Scales Composite Indices
Shift Behavioral Regulation
Emotional Control Index
Plan/Organize Metacognitive Index
Organization of Materials
Figure 1. BRIEF scales
The authors of the BRIEF provided preliminary information in the BRIEF manual
for the pediatric TBI population. They compared parent ratings of the eight BRIEF
scales across four groups of children consisting of 33 children with mild/moderate TBI,
34 children with severe TBI, 35 orthopedic controls, and 35 normal controls (Gioia et al.,
2000). The severe TBI group had significantly worse scores on the Inhibit, Shift,
Emotional Control, Initiate, Working Memory and Plan/Organize scales as compared to
healthy controls, suggesting global executive dysfunction. The authors reported that
children with severe TBI also had significantly worse scores on the Working Memory
scale in comparison to children in the mild/moderate TBI group and orthopedic controls.
Gioia, Isquith, Kenworthy, and Barton (2002) evaluated the individual BRIEF
scales for groups of moderate and severe TBI children and normal controls within a
larger study examining profiles of everyday executive function across a variety of
acquired and developmental disorders. The authors used BRIEF data provided by
Mangeot, Armstrong, Colvin, Yeates, and Taylor (2002), and the sample included 33
moderate TBI, 34 severe TBI, and 208 children from the BRIEF normative sample.
Gioia et al. (2002) found that the severe TBI children were rated as having more
executive dysfunction than controls for the Inhibit, Shift, Initiate, Working Memory and
Plan/Organize scales. Severe and moderate TBI were not significantly different from one
another on any of the scales, and the moderate TBI group was not significantly different
than the control group for any of the scales. It should be noted that for the Severe TBI
sample, the mean T-scores for Working Memory and Plan/Organize were in a borderline
clinical range (63.5 and 62.0, respectively), with fairly large standard deviations. This
finding suggests that a subset of the children surpassed the clinical cutoff designated for
the BRIEF scales (T-score > 65).
Studies involving executive function in pediatric TBI have also examined a
combination of performance and behavioral report measures in order to assess the
relationship between behavioral ratings and neuropsychological performance measures of
executive function. Proctor, Wilson, Sanchez, and Wesley (2000) studied the correlation
between executive function and working memory for eight adolescents with closed head
injury and eight controls. Executive functioning was measured using a respondent report
known as the Profile of Executive Functioning (Pro-Ex), and working memory was
assessed with a recognition task. When all subject data was grouped together, a positive
linear correlation was found for the Pro-Ex and the recognition task. Severity of injury
influenced test performance for both the Pro-Ex and the recognition measure, with a
significant group effect found for the recognition measure. The authors concluded that a
relationship can exist between a measure of daily executive functioning and working
memory performance, and they emphasize the clinical importance of assessing executive
functions within a TBI population.
Mangeot et al. (2002) examined executive functioning in moderate and severe TBI
using the BRIEF as a measure of everyday executive functioning within the context of a
study examining family functioning and adaptive behavior following TBI. The research
sample consisted of 33 severe TBI, 31 moderate TBI, and 34 orthopedic controls who
were evaluated approximately five years post-injury. The TBI sample in this study was
the same one used by Gioia et al. (2002) to examine the various clinical scales of the
BRIEF. Mangeot and colleagues (2002) compared the BRIEF's two Composite Index
Scores (the Behavioral Regulation Index and Metacognitive Index) and its Global
Executive Composite to several performance-based measures of executive functioning,
including Consonant Trigrams, the Rey-Osterriech Complex Figure, Word Fluency,
Contingency Naming, and the Underlining test. They found that while the 3 BRIEF
Index scores were correlated with the Consonant Trigrams test, none of the three BRIEF
Index scores correlated with any of the other performance measures of executive
function. The authors also found that the largest deficits in parent-reported executive
functions existed in the severe TBI group. The authors suggested that the BRIEF and
performance measures of executive function possibly measure different aspects of the
executive function construct. The authors implied that the BRIEF scales may have more
ecological validity and therefore may be measuring constructs within the domain of
executive function differently than performance measures.
Vriezen and Pigott (2002) examined executive functioning in 48 moderate to severe
TBI cases that were, on average, two-and-a-half years post injury. The used the BRIEF
as a measure of everyday executive functioning and compared the BRIEF's two
Composite Index Scores (the Behavioral Regulation Index and Metacognitive Index) and
its Global Executive Composite to several performance-based measures of executive
functioning, including the Wisconsin Card Sort Test, the Trail Making Test and verbal
fluency. They found that none of the three BRIEF scores correlated with any of the three
objective measures of executive function. Vriezen and Pigott noted that while the means
on the BRIEF and performance measures of executive functioning fell within the average
range for the TBI sample, there was wide variability in the types of scores that were
obtained. They also pointed out that 29 35% of the various BRIEF indices were within
a clinical range with ratings greater than 1.5 standard deviations from the mean.
Approximately 9 21% of the objective measures yielded scores that were greater than
1.5 standard deviations from the mean. Vriezen and Pigott suggested that while there
was a lack of statistical significance with their study in terms of the correlational
relationships between the BRIEF Indices and performance-based executive function
measures, their study demonstrates that a subset of the TBI population has difficulties
within the clinically significant range with regards to of executive functioning.
Rationale for the Current Study
Thus far, there are no studies of post-injury BRIEF ratings at an acute timepoint
within the pediatric TBI population. Vriezen and Pigott (2002) examined a TBI sample
that was two-and-a-half years post-injury. The TBI cohort studied by Gioia et al. (2002)
and Mangeot et al. (2002) was five years from injury.
Second, while the BRIEF manual publishes preliminary results with regards to the
performance of mild/moderate TBI cases in comparison to severe TBI years after injury,
to my knowledge no peer-reviewed studies have reported on BRIEF ratings in a mild TBI
sample for a timepoint closer to their actual injury. Furthermore, methodological
information supplied for the raw data published in the BRIEF manual suggests that the
mild/moderate group used may have been more severe than a typical mild TBI
classification. Abnormal radiological (e.g., CT scan) or neurological findings or a loss of
consciousness greater than fifteen minutes were allowed in the mild/moderate group, and
this type of criteria is not typical for the mild TBI population.
Third, although some studies have begun to explore the relationship between the
BRIEF scales and particular performance measures of executive dysfunction, such
studies have focused primarily on the two Composite Index scores from the BRIEF and
the Global Executive Composite score rather than the individual BRIEF clinical scales.
Mangeot et al. (2002) and Vriezen and Pigott (2002) found few significant relationships
between the BRIEF Index scores and performance measures of executive function for
their pediatric TBI samples. The literature on the construct of executive function
suggests that there are a wide range of specific behaviors and approaches to problem-
solving that need to occur to successfully complete various tasks of executive function.
Therefore, it may be more appropriate to compare particular scales of interest within the
BRIEF Indices. Gioia et al. (2002) found impairments across several of the BRIEF
individual clinical scales for their severe TBI sample. Using the clinical scales may also
increase the specificity and help us comprehend more thoroughly the relationship
between parental report of suboptimal executive functioning in the home and community
environment and those performance behaviors seen within the confines of
Finally, no literature on the BRIEF and TBI has attempted to look at parent reports
of pre-injury functioning in comparison to current functioning in pediatric TBI or the rate
of behavioral change that occurs acutely following TBI. This information is particularly
relevant from a clinical standpoint, since clinical referral is often based on reported
changes in behavior resulting from the injury. Yeates (2000) emphasized the need to
control for premorbid status in studies of behavioral adjustment with the pediatric TBI
population, particularly for children with mild TBI. Asarnow and colleagues (Asarnow et
al., 1995; Light et al., 1998) showed that children with mild head injuries display higher
rates of pre-injury behavior problems on the Child Behavior Checklist (Achenbach, 1991)
than children with no injury. However, when compared to children with injuries that did
not involve the head, the mild head injury children did not show differences in terms of
pre-injury behavior. While head injuries increase the risk for behavioral problems,
behavioral disturbances may also likely increase the risk of head injury. There are certain
methodological issues with trying to obtain pre-injury ratings at a timepoint following
injury, such as potential response bias and differences in report of pre-injury symptoms
based on time since injury. Despite such limitations, it was believed that the clinical
relevance of evaluating changes in behavioral ratings was important to explore within the
Aims and Hypotheses
The current study was designed to examine several dimensions of executive
functioning within mild and moderate/severe pediatric TBI samples as compared to an
orthopedic control population. The first aim was to examine children's post-injury
behavior in terms of everyday executive functioning. Post-injury BRIEF data as
completed by the child's parent was used to gauge this level of functioning. Differences
between the mild and moderate/severe TBI population and orthopedic controls were
evaluated for five specific BRIEF scales (Inhibit, Shift, Initiate, Working Memory and
Plan/Organize). Hypotheses regarding this first aim were that group differences would
be present for the five BRIEF scales. Hypothesis 1 predicted that the moderate/severe
TBI group would be rated as having more problems, i.e. higher ratings of frequency for
problems, than either the mild TBI group or the orthopedic controls across all BRIEF
scales. Hypothesis 2 predicted that the mild TBI group would be rated as exhibiting more
problems than the orthopedic controls.
The second aim was to examine children's post-injury behavior in terms of
performance measures of executive functioning. The Test of Everyday Attention for
Children (TEA-Ch; Manly et al., 1999), Trail Making Test A and B (Reitan, 1979),
Stroop Color and Word Test (Golden, 1978), and the computerized Wisconsin Card Sort
Test-64 (WCST; Heaton et al., 1993) were used to assess executive functioning
performance. Differences between the mild and moderate/severe TBI population and
orthopedic controls were evaluated for five specific measures from these instruments
(TEA-Ch Creature Counting and Opposite Worlds subtests, Trails B, Stroop Interference,
and WCST Perseverations). Hypotheses regarding this second aim were that group
differences would be present for the five performance measures. Hypothesis 3 predicted
that the moderate/severe TBI group would demonstrate greater executive dysfunction
than either the mild TBI group or the orthopedic controls on all five measures.
Hypothesis 4 predicted that the mild TBI group would exhibit more executive
dysfunction than the orthopedic control sample for these measures.
The third aim for the current study was to examine the relationship between
children's post-injury behavior as measured by the five BRIEF scales and performance
measures of executive functioning. Inconsistent results have been found in the previous
literature in regards to the relationship between performance measures of executive
function and behavioral rating scales. Proctor et al. (2000) had found significant
correlations between a working memory test and the Pro-Ex. In contrast, Mangeot et al.
(2002) and Vriezen and Pigott (2002) found few correlations between performance
measures and the BRIEF, finding that only the Consonant Trigrams test had a
relationship with the BRIEF for this clinical population. However, both of the BRIEF
studies used the Index scores provided by the BRIEF rather than the individual scales.
Therefore, the current study set out to examine the relationship between select BRIEF
scales and performance measures of executive function. Since higher T-scores for the
performance measures indicate better performance and higher T-scores on the BRIEF
suggest more impairment, Hypothesis 5 consisted of the following: 1) correlations
between paired BRIEF scales should positive, 2) correlations between paired
performance measures of executive function should be positive, and 3) correlations
between a BRIEF scale paired with a performance measure should be negative.
The final aim for the current study was to examine changes in children's behavior
in terms of everyday executive functioning from prior to the child's injury to the current
timepoint. Retrospective "pre-injury" and current "post-injury" BRIEF ratings were
completed by the child's parent in order to gauge the change in everyday executive
functioning. Differences in ratings for pre- and post-injury behavior were examined for
the mild and moderate/severe TBI population and orthopedic controls using the five
specific BRIEF scales. Hypotheses regarding this last aim were that group differences
would be present for the five BRIEF scales. Hypothesis 6 predicted that the
moderate/severe TBI group would be rated as having greater changes in executive
function than either the mild TBI group or the orthopedic controls. Hypothesis 7
predicted that the mild TBI group would be rated as exhibiting more changes in executive
function than the orthopedic control group.
This study was approved by the University of Florida Institutional Review Board
(IRB Project Number 649-2001). Informed consent was obtained from all participants'
parents or legal guardians, and informed assent was obtained from all pediatric
participants prior to inclusion in the study.
Thirty-five participants were obtained via referral from the Pediatric Intensive Care
Unit, the Emergency Department, and the Pediatric Orthopedics Clinic at Shands
Hospital at the University of Florida and by self-referral from the community. Children
between the ages of 6 to 17 years of age were invited to participate in the study. All
participants were medically stable at the time of the evaluation.
TBI participants were classified by injury severity based on several criteria,
including length of loss of consciousness immediately following the injury, post-
traumatic amnesia duration, whether they had CT/MRI abnormalities resulting from their
injuries, and Glasgow Coma Scale (GSC) scores from their hospital admission. The GCS
(Teasdale & Jennett, 1974) is a measure frequently used in the TBI literature to classify
injury severity. Individuals who have sustained a head injury are rated for the following:
verbal responsiveness, motor responsiveness, and eye opening. Scores range from 3 to
15, with lower scores indicating more severe levels of injury. Table 1 outlines how
participants in the current study were classified in terms of their head injury. Only 3
children assessed for this study were classified as having a TBI "moderate" in severity.
Given the small sample size of the group and the fact that the TBI literature often
includes moderate TBI cases with severe TBI cases when examining deficits as compared
to controls, the three moderate TBI cases from the general TBI sample were included
with the Severe TBI group for the purposes of analyses. Based on the outlined injury
criteria, 9 subjects were classified as Mild TBI, 18 as Moderate/Severe TBI, and 8
subjects were in the Orthopedic Control group.
Table 1. TBI classification
Mild Moderate Severe
Glasgow Coma Scale (admission) 13 15 9 12 < 8
Loss of Consciousness < 1 hour 1 24 hours > 24 hours
Post Traumatic Amnesia < 24 hours 1 7 days > 7 days
CT/MRI scan abnormalities No Yes Yes
Table 2 presents the demographic characteristics of the sample. Sixty-six percent
of the total sample was male, and the gender breakdown was similar to that typically seen
in the TBI literature. The Moderate/Severe TBI group had a higher proportion of female
subjects than that seen in the other two subject groups. Approximately 77% of the
participants were Caucasian, 17% were African American, 3% were Hispanic and 3% did
not provide additional information in order to determine specific race information (e.g.,
"other"). Eighty percent of the children were right-handed, although this figure may be
higher since handedness information was not provided for two subjects. Four of the
participants had a history of Attention Deficit Hyperactivity Disorder (ADHD) prior to
their injuries. One case of premorbid ADHD was in the Mild TBI group, two were in the
moderate/Severe group, and one was in the Orthopedic Control group. Mechanism of
injury included falls, motor vehicle accidents, sports injuries, and assault. The three
groups did not differ significantly in terms of time since injury, regardless of whether
they had a TBI or an orthopedic injury, F(2, 32) = 2.32, p =. 11. The TBI groups and the
Orthopedic Control group also did not differ in terms of age at assessment, F(2, 32) =
0.64,p = .54. The three groups did differ in terms of Full Scale IQ, F(2, 26) = 3.69,p
=.04. Tukey's honestly significance difference (HSD) tests revealed that the Full Scale
IQ for the Moderate/Severe TBI group was significantly lower than the Orthopedic
Control group, p = .03. Yeates (2000) describes that intelligence scores reflecting
nonverbal skills (e.g., Performance IQ scores) are particularly vulnerable following head
injury, particularly because they often require speeded motor output and fluid problem-
solving. In contrast, Verbal IQ assesses previously acquired knowledge and requires few
speeded responses. Therefore, Verbal IQ was analyzed in order to assess whether there
were any premorbid IQ differences between the three groups. The three groups did not
differ in terms of Verbal IQ, F(2, 26) = 2.40, p =. 11.
Table 2. Demographic information
Mild TBI Mod/Severe TBI Orthopedic Controls
# Subjects 9 18 8
Sex (M:F) 8:1 9:9 6:2
Age at assessment 11.94 (3.74) 13.60 (3.83) 13.64 (3.99)
Time since injury 33.71 (23.17) 35.20 (19.84) 16.77 (19.95)
Full Scale IQ 101.50 (6.63) 94.87 (11.75) 108.50 (13.76)
Verbal IQ 101.50 (9.79) 96.00 (16.76) 111.38 (17.97)
Parent-Reported Executive Function
Behavior Rating Inventory of Executive Function (BRIEF). Parent-reported
executive function was measured using the Behavior Rating Inventory of Executive
Function (BRIEF; Gioia et al., 2000). The BRIEF is an 86-item questionnaire completed
by parents or teachers of school age children to assess executive functions in the home
and school environments. The respondent rates each item's frequency of occurrence as
either occurring "Never," "Sometimes," or "Often," and a number value from 1 to 3,
respectively, is assigned to the frequency endorsed. Each of the 86 individual BRIEF
items are assigned (via factor analysis) to one of eight empirically derived scales that
measure the following aspects of executive functioning: Inhibit, Shift, Emotional
Control, Initiate, Working Memory, Plan/Organize, Organization of Materials, and
Monitor. Ratings within each scale are summed and the total raw score is transformed
into an age- and gender-corrected T-score relative to the published normative data.
Higher T-scores indicate more executive dysfunction, and a T-score > 65 is considered
clinically significant. The BRIEF also yields a Global Executive Composite score as well
as two Composite Index scores (the Behavioral Regulation Index and the Metacognition
Index). The Behavioral Regulation Index consists of the Inhibit, Shift and Emotional
Control scales, and the Metacognitive Index consists of the Initiate, Working Memory,
Plan/Organize, Organization of Materials, and Monitor scales. The internal consistency
(i.e., the degree to which the items in a single construct are measuring the same construct)
ranged from .82 to .98 for the parent form when used with a heterogeneous clinical
sample and from .80 to .97 when used to evaluate the normative sample. Test-retest
reliability for the parent form ranged from .72 to .84 for a clinical sample and from .76 to
.88 for the normative sample.
For the current study, parents completed two BRIEFs at the time of the assessment.
One BRIEF was completed to serve as an estimate of pre-injury functioning (i.e.,
retrospective rating) and one was completed to assess current behavior. The T-scores for
five clinical scales (Inhibit, Shift, Initiate, Working Memory, and Plan/Organize) were
used as the measures of parent-reported executive functioning for the current study.
Performance Measures of Executive Function
Several traditional performance-based measures of executive functioning were
given to the children participating in this study. The measures are each described in more
detail below and consist of the following: Creature Counting and Opposite Worlds from
the Test of Everyday Attention for Children (TEA-Ch; Manly et al. 1999), Trail Making
Test A and B (Reitan, 1979), Stroop Color and Word Test (Golden, 1978), and the
computerized Wisconsin Card Sort Test-64 (WCST; Heaton et al., 1993). These tests
have been widely used to assess executive dysfunction in a variety of patient populations,
including TBI (e.g., Levin et al., 1997, Vriezen and Pigott, 2002).
TEA-Ch Creature Counting. The Creature Counting subtest assesses the ability
to switch back and forth between strategies (Manly et al., 1999). For the Creature
Counting task, children were asked to complete several trials during which they count
creatures along a path. At various points within the path, arrows are present that prompt
the child to change the direction in which they are counting (e.g., "1, 2, 3" versus "3, 2,
1"). Time to complete the counting of the creatures and counting accuracy are scored for
this test. The average time for accurately completed trials provides a timing score for
which an age- and gender-corrected scaled score was obtained using normative data from
the TEA-Ch manual. The timing score was the measure of interest for the current study.
TEA-Ch Opposite Worlds. The Opposite Worlds subtest evaluates the ability to
suppress an automatic or prepotent verbal response (Manly et al. 1999). During the
precursor Same World trial of the test, children were to name along a path consisting of
tiles numbered '1' and '2'. They were to say '1' when they saw the number '1' and '2'
when they saw the number '2'. In contrast, during the Opposite World trials of the task,
as children named along the path, they were to say '1' when they saw the number '2' and
'2' when they saw the number '1'. Two trials of Same World and two trials of Opposite
World were completed. The measure of interest for this study was the age- and gender-
corrected scaled score obtained from the TEA-Ch manual for the total time that it took
children to complete the Opposite World portion of the task.
Trail Making Test. The Trail Making Test evaluates attention, visual scanning,
information processing, and switching/executive skills (Reitan, 1979). Trail A measures
the speed with which individuals are able to draw a line connecting consecutive numbers.
Trail B requires individuals to alternate between connecting consecutive numbers and
letters. Children in the current study were given the non-adult version of the Trail
Making Test. This version is similar to the adult version with the exception that there are
fewer items to complete. While the time to complete each Trail and the number of errors
committed as subjects completed the tasks were collected, only the time to complete
Trails B was used for the current study. Data acquired was converted to standardized z-
scores using normative data provided by several studies (Reitan, 1971; Klonoff and Low,
1974; Knights, 1966).
Stroop Color and Word Test. The Stroop measures the ability to inhibit a
particular response set and the ability to maintain a particular response set while
inhibiting another (Golden, 1978). The first trial of the task required the subject to read
vertical columns of color words (e.g., "RED") typed in black ink as quickly as possible.
The second trial required the subject to scan vertically columns of X's typed in a certain
color of ink (e.g., "XXXXX" in red ink) and state in what color the X's were typed. For
the third trial, subjects were asked to read vertical columns of color words, but instead of
saying the typed word, subjects were to tell the examiner in what color of ink the word
was typed. Words for this trial were typed in a color of ink incongruent to that of the
actual word (e.g., the word "RED" typed in blue ink). Each of the trials was allotted 45
seconds for completion. If a subject made an error on any of the three trials, they were
prompted to correct themselves before continuing to the next item. The three trials
provided raw scores for word reading, color naming, and color-word reading,
respectively. An interference score was calculated in order to compare the subject's
actual rate during the color-word reading trial to a predicted rate of color-word reading
that would be expected given the subject's performance during the word reading and
color-naming trials. Raw scores for the trials were converted to T-scores using the
normative data provided by Golden (1978). The interference T-score was the primary
measure of interest for the current study.
Wisconsin Card Sorting Test (WCST). The WCST assesses the ability to
apply, maintain and shift appropriate problem-solving strategies across changing
conditions in order to reach a future goal (Heaton et al, 1993). The computerized version
of the WCST-64 consists of four stimulus cards and 64 response cards that display
figures of varying colors (red, green, yellow or blue), forms (triangles, stars, crosses or
circles) and numbers (one, two, three or four figures). Subjects were asked to match each
response card to one of the four stimulus cards in the way he or she thought that the
response card should match. The subject was told whether the match was correct or
incorrect but was not told the sorting principle. Once the subject performed ten
consecutive correct matches, the sorting principle changed without warning. The subject
was then supposed to use feedback from the examiner to develop a new sorting strategy.
The WCST moves through a number of set shifts among the three sorting principles
(Color, Form and Number) until all 64 response cards have been used or until six sorting
categories (two for each of the above) have been completed. While several scoring
dimensions are produced by the test (e.g., trials to complete first category, learning to
learn, categories completed), the T-score for perseverative responses during the task was
the primary measure used for the current project.
All analyses were done using the Standard Version of SPSS 11.0.1 for Windows
(SPSS, Inc. 2001). Since the normative data provided for all of the measures used in this
study with the exception of the Stroop Color and Word test allows for both gender and
age-correction, gender was not used a covariate for any of the statistical analyses. IQ was
not used as a covariate given the previously stated rationale for evaluating Verbal IQ in
lieu of Full Scale IQ. Analyses for this study were the following and are designated by
the hypothesis for which they are being performed:
Hypotheses 1 and 2. In order to examine group differences at post-injury for the
parent report measure of executive functions, one-way ANOVAs were done for the five
BRIEF scales of interest. Group (Mild TBI, Moderate/Severe TBI, Orthopedic Control)
was the independent variable and the post-injury ratings for the five BRIEF scales
(Inhibit, Shift, Initiate, Working Memory, Plan/Organize) were the dependent variables
for the one-way ANOVAs. All scores used were age- and gender-corrected T-scores
based on the normative data provided by the BRIEF manual (Gioia et al, 2000), with
higher T-scores on the five BRIEF scales suggestive of greater impairment. Planned
post-hoc analyses for the ANOVAs that showed a significant effect for Group consisted
of Tukey's (HSD) tests.
Hypotheses 3 and 4. One-way ANOVAs were done to examine group differences
for the performance measures of executive function. Group remained the independent
variable for the one-way ANOVAs and the five dependent variables for this set of
analyses were the TEA-Ch Creature Counting (Timing scaled score), TEA-Ch Opposite
Worlds (Time scaled score), Trails B (standard score for time to complete), Stroop
Interference (T-score), and WCST Perseverations (T-score). Prior to data analysis, all
performance measures' scores were converted to T-scores in order to examine the data in
a common metric. In contrast to the BRIEF scales, higher T-scores on the five
performance measures of executive function indicated better performance. Planned post-
hoc analyses consisted of Tukey's HSD tests for the ANOVAs that showed a significant
effect for Group.
Hypothesis 5. Evaluation of the relationship between the different types of
executive function measures was done by performing one-tailed Pearson correlations on
the BRIEF scales and the performance measures. One-tailed correlations were chosen
since it was hypothesized a priori that the relationships between measures should go in a
particular direction. It was hypothesized that correlations between two BRIEF scales
should positive, correlations between two performance measures of executive function
should be positive, and correlations between a BRIEF scale paired with a performance
measures should be negative since while higher scores on the performance measures
indicate better performance, higher scores on the BRIEF suggest more impairment. Two
sets of the Pearson one-tailed correlations were performed. One set of correlations was
performed in order to examine the relationships between tests for the two types of
measures within the entire study sample. The second set of correlations was done
exclusively with the TBI participants in order to determine if there were differences in
how the test variables were related to one another as the result of being part of a clinical
population with suspected neuropsychological deficits. Due to the number of variables
examined in relation to the size of the study sample, a conservative level of significance
(p < .01) was selected to determine significant correlations.
Hypotheses 6 and 7. In order to assess changes in BRIEF rating from before the
injury to the current post-injury time point, repeated-measures ANOVAs were performed
for each set of pre- and post- injury BRIEF scale (e.g., pre-Inhibit and post-Inhibit score)
for all subjects who had pre-and post-injury ratings. Group served as the independent
variable and Time (pre-injury versus post-injury ratings for the five BRIEF scales) was
the dependent variable for this set of analyses. Difference scores were calculated for the
five BRIEF scales by subtracting the pre-injury rating T- score from the post-injury rating
T-score. One-way ANOVAs were then performed using the difference scores for the
BRIEF scales that had shown a Group x Time Interaction. Group served as the
independent variable and the BRIEF scales' difference scores served as the dependent
variables. Planned post-hoc analyses consisted of Tukey's HSD tests for the ANOVAs
that showed a significant effect for Group.
Post-Injury BRIEF Measures (Hypotheses 1 and 2)
Tables 3 and 4 and Figure 2 display the one-way ANOVA results examining post-
injury group differences for the five BRIEF scales of interest. The Working Memory
BRIEF scale demonstrated significant group differences, F(2, 26) = 3.49, p < .05.
Planned post-hoc Tukey tests did not reveal specific group differences (Mild vs.
Moderate/Severe TBI (p = .13), Mild TBI vs. Orthopedic controls (p = .94), and
Moderate/Severe vs. Orthopedic controls (p = .07). The other four BRIEF scales, Inhibit,
Shift, Initiate, and Plan/Organize, did not show significant group differences: Inhibit F(2,
26) = .53, p = .60; Shift F(2, 26) = .56, p = .58; Initiate F(2, 26) = .65, p = .53;
Plan/Organize F(2, 26) = .40, p = .67.
Table 3. Mean T scores for the 5 BRIEF scales
Mild TBI Mod/Severe Orthopedic Significance
TBI Control level
Inhibit 57.50 (16.72) 62.14 (15.84) 55.29 (12.72) .60
Shift 62.13 (16.85) 55.93 (13.29) 55.00 (15.74) .58
Initiate 61.75 (11.55) 58.71 (9.58) 55.43 (11.97) .53
Working Memory 59.13 (12.85) 71.36 (14.43) 56.71 (13.15) .05*
Plan/Organize 59.88 (12.86) 60.71 (11.01) 55.71 (13.93) .67
Note: Standard deviation data is provided in ( )
* =p <.05
Table 4. Effect sizes for BRIEF scales post-injury
Inhibit Shift Initiate Working Plan/Organize
Note: The dotted line
* =p <.05
displays mean group performance 1.5 SD above the mean.
Figure 2. Mean T scores for the 5 BRIEF scales
Mild vs. Mild TBI vs. Mod/Severe TBI
Mod/Severe TBI Orthopedic vs. Orthopedic
Inhibit .28 .15 .48
Shift .41 .44 .06
Initiate .29 .54 .31
Working Memory 1.05 .18 1.07
Plan/Organize .07 .31 .40
O Mild TBI
O Severe TBI
Performance Measures of Executive Function (Hypotheses 3 and 4)
Table 5 and Figure 3 display the one-way ANOVA results examining group
difference on the five performance measures of executive function, and Table 6 shows
effect sizes for each group and performance measure. In sum, the TEA-Ch Creature
Counting test demonstrated a significant group difference, F(2, 30) = 4.44, p = .02.
Planned post-hoc Tukey tests revealed that the group differences were significant
between the Moderate/Severe TBI and Orthopedic Control groups, p = .02. The TEA-Ch
Opposite Worlds test demonstrated a significant group difference, F(2, 32) = 7.00, p <
.01. Group differences were significant between the Moderate/Severe TBI and
Orthopedic Control groups, p = .01, as well as between the Moderate/Severe TBI and
Mild TBI groups, p = .01. There were no significant group differences for Trails B,
Stroop Interference and the WCST Perseveration scores.
Table 5. Mean T scores for the 5 performance measures of executive function
Mild TBI Mod/Severe Orthopedic Significance
TBI Control level
Creature 38.52 (9.15) 34.12 (9.47) 45.71 (5.35) .02*
Opposite Worlds 48.15 (4.12) 37.59 (10.40) 48.33 (6.42) .01**
Trails B 53.71 (6.59) 48.03 (6.84) 51.19 (9.17) .22
Stroop 47.00 (8.90) 53.43 (6.38) 54.38 (6.97) .09
WCST 53.75 (17.29) 58.50 (14.92) 63.63 (19.27) .51
Note: Standard deviation data is provided in ( )
* =p <.05
** =p <.01
aO Mild TBI
Ml55 vSevere TBI
3 5 I I I a g N ol N IO N I l
Creature Counting Opposite Worlds Trails B Stroop WCST
Note: The dotted line displays mean group performance 1.5 SD below the mean.
* =p <.05
** =p <.01
Figure 3. T scores for the 5 performance measures of executive function
Table 6. Effect sizes for performance measures of executive function
Mild vs. Mild TBI vs. Mod/Severe TBI
Mod/Severe TBI Orthopedic vs. Orthopedic
Creature Counting .47 1.00 1.60
Opposite Worlds 1.45 .04 1.27
Trails B .85 .32 .40
Stroop Interference .83 .93 .15
WCST .29 .54 .30
Correlations between BRIEF Scales and Executive Function Performance Measures
Tables 7 and 8 provide the results of the correlational analyses between the various
executive function measures.
BRIEF Scales-BRIEF Scales
All of the BRIEF scales were correlated with one another when all of the study
participants were included in the correlation matrix, p < .01. Similarly, all of the BRIEF
scales with the exception of Working Memory Shift were correlated with one another, p
< .01, when only the TBI study participants were included in the correlation matrix.
BRIEF Scales-Neuropsychological Measures
None of the BRIEF scales were correlated with the performance measures of
executive function at the significance level set for the study. However, if a less
conservative yet acceptable significance level was used (p < .05), the correlation between
the TEA-Ch Opposite Worlds and BRIEF Inhibit scores would be considered significant
when evaluating all study participants. However, the WCST Perseverations score would
also be considered significantly correlated with the BRIEF Working Memory and BRIEF
Plan/Organize scores when examining only the TBI study participants, with the direction
of the relationships suggesting that higher (i.e., worse) BRIEF scores are actually
positively correlated with better scores (i.e., less perseverations) on the WCST.
Neuropsychological Measures-Neuropsychological Measures
Only the TEA-Ch Creature Counting and the TEA-Ch Opposite Worlds tests were
significantly correlated. This was the case when all subjects were used as well as when
only the TBI subjects were examined (p < .001, p < .01, respectively). The TEA-Ch
results were unsurprising given that the two tests comprise the Attentional
Control/Switching domain for the TEA-Ch. None of the other five performance
measures were correlated with one another at the conservative significance level set for
this study. However, it should be noted that Trails B and Creature Counting
demonstrated a correlation when evaluating all study participants as well as just the TBI
subjects using a less stringent significance criteria ofp < .05. Similarly, the WCST
Perseverations and the TEA-Ch Creature Counting scores were correlated when
evaluating the TBI subjects alone.
Table 7. Correlations between BRIEF scale scores and performance measures of executive function for all study participants
BRIEF BRIEF BRIEF BRIEF BRIEF TEA-Ch TEA-Ch Trails B Stroop
Inhibit Shift Initiate Working Plan/ Creature Opposite Time Interference
Memory Organize Counting Worlds
BRIEF .589** .716**
Initiate <.001 <.001
BRIEF .593** .484** .583**
Working <.001 <.01 <.001
BRIEF .602** .651** .833** .739**
Plan/Organize <.001 <.001 <.001 <.001
TEA-Ch -.062 .047 -.135 .023 -.012
Creature .379 .408 .205 .454 .477
TEA-Ch -.373* -.143 -.256 -.283 -.266 .600**
Opposite .023 .229 .090 .069 .081 .000
Trails B -.087 .027 -.044 -.195 -.158 .415* .227
Time .327 .445 .410 .155 .207 .013 .110
Stroop .063 .140 .016 .178 .228 .152 .057 .000
Interference .375 .239 .468 .183 .121 .215 .382 .499
WCST .001 -.065 .102 .211 .257 .311 .098 .003 .283
Perseverations .498 .376 .310 .150 .103 .057 .310 .494 .072
Note: =p<.05 ** =p <.01
Table 8. Correlations between BRIEF scale scores and performance measures of executive function for TBI participants
BRIEF BRIEF BRIEF BRIEF BRIEF TEA-Ch TEA-Ch Trails B Stroop
Inhibit Shift Initiate Working Plan/ Creature Opposite Time Interference
Memory Organize Counting Worlds
BRIEF .527** .640**
Initiate <.01 <.01
BRIEF .522** .417* .501**
Working <.01 <.05 <.01
BRIEF .525** .620** .821** .700**
Plan/Organize <.01 <.01 <.001 <.001
TEA-Ch .031 .239 .059 .234 .212
Creature .447 .148 .400 .153 .178
TEA-Ch -.294 -.045 -.114 -.161 -.136 .518**
Opposite .092 .422 .307 .237 .273 <.01
Trails B -.087 .107 .031 -.198 -.072 .397* .332
Time .349 .317 .446 .188 .376 <.05 .067
Stroop .163 .182 .074 .345 .291 .129 -.085 .169
Interference .239 .215 .375 .063 .100 .284 .354 .226
WCST .078 .009 .199 .461* .384* .407* .112 -.006 .215
Perseverations .376 .485 .206 .023 .05 <.05 .319 .489 .181
Note: =p<.05 ** =p<.01
Differences in BRIEF Ratings for Pre- and Post-Injury (Hypothesis 6 and 7)
Table 9 and Figure 4 display data from the repeated-measures ANOVAs examining
differences in BRIEF scale scores from the pre-injury and post-injury BRIEF ratings. For
the Inhibit scale, there was a main effect for Time (i.e. pre- to post-BRIEF scores), F(1,
22) = 6.07, p < .05. There was no Group x Time interaction, F(2, 22) = 1.47, p = .25.
The Shift scale also showed a main effect for Time, F(1, 22) = 5.46, p < .05, but there
was no Group x Time interaction, F(2, 22) = 2.42, p =. 11. The Initiate scale showed a
main effect for Time, F(1, 22) = 14.79, p < .01, with no Group x Time interaction, F(2,
22) = 2.02, p =. 16. The Working Memory scale showed a main effect for Time, F(1, 22)
= 13.25, p < .01, and there was a Group x Time interaction for the Working Memory
scale, F(2, 22) = 3.66, p < .05. Lastly, the Plan/Organize scale showed a main effect for
Time, F(1, 22) = 8.02, p = .01, and the Group x Time interaction showed marginal
significance, F(2, 22) = 3.27, p = .057.
Table 9. Group x time differences in pre- and post-injury T scores for the 5 BRIEF scales
Mild TBI Mod/Severe TBI Orthopedic Significance
Pre: 52.25 (12.01) 48.82 (12.35) 53.00 (13.51) .25
Post: 57.50 (16.72) 58.55 (15.50) 53.83 (13.29)
Pre: 51.38 (11.11) 48.82 (10.38) 50.67 (10.67) .11
Post: 62.13 (16.85) 52.00 (11.06) 50.67 (11.81)
Pre: 49.88 (12.19) 48.64 (5.37) 51.00 (12.13) .16
Post: 61.75 (11.55) 58.45 (10.53) 52.67 (10.39)
Pre: 49.63 (12.92) 51.36 (11.34) 54.00 (13.97) .04*
Post: 59.13 (12.85) 70.55 (16.16) 55.33 (13.84)
Pre: 53.25 (16.25) 48.00 (8.76) 56.00 (18.04) .06
Post: 59.88 (12.86) 59.09 (11.87) 54.83 (15.04)
Note: Mean (SD) data from repeated-measures ANOVAs
* =p <.05
-- Mild IBI
-- Severe TBI
Note: The dotted line displays executive dysfunction reported
* =p <.05
** =p <.01
1.5 SD above the mean.
Figure 4. Group x time differences pre- to post-injury for the 5 BRIEF scales
Table 10 displays data regarding differences in BRIEF scale scores between the
pre-injury and post-injury BRIEF ratings. In order to perform post-hoc analyses,
difference scores for the BRIEF (post-injury BRIEF score minus pre-injury BRIEF score)
had been calculated for the five BRIEF scales of interest. As mentioned before, the
Working Memory scale from the BRIEF showed a significant difference in terms of pre-
to post-injury ratings for the three injury groups, F(2, 22) = 3.66, p < .05. Planned post-
hoc Tukey's HSD tests that the Working Memory scale differences were significant
between the Moderate/Severe TBI and Orthopedic Control groups, p < .05, but not
between the Mild TBI and either the Moderate/Severe TBI or Orthopedic Control groups.
Group differences for the Plan/Organize BRIEF scale ratings from pre- to post-injury
showed marginally significant differences, F(2, 22) = 3.27, p = .057. Planned post-hoc
Tukey's HSD tests revealed that the Plan/Organize scale differences were significant
between the Moderate/Severe TBI and Orthopedic Control groups, p < .05, but not
between the Mild TBI and either the Moderate/Severe TBI or Orthopedic Control groups.
Since there were no significant group differences for the BRIEF Inhibit, Shift and Initiate
scales, no post-hoc analyses were done for these three scales.
Table 10. Pre- and post-injury T score differences for the 5 BRIEF scales
Mild TBI Mod/Severe Orthopedic Significance
TBI Control level
Inhibit 5.25 (8.33) 9.73 (13.68) 0.83 (1.33) .25
Shift 10.75 (13.02) 3.18 (7.12) 0.00 (8.37) .11
Initiate 11.88 (11.46) 9.82 (9.51) 1.67 (7.74) .16
Working 9.50 (15.54) 19.18 (14.82) 1.33 (2.16) .04*
Plan/Organize 6.63 (10.68) 11.09 (10.44) -1.17 (3.92) .06
Note: Mean (SD) data
The current study attempted to examine both everyday executive functioning and
neuropsychological performance measures of executive functioning in an acute pediatric
TBI population that ranged in severity. Findings from the study suggest that in terms of
BRIEF post-injury ratings, Moderate/Severe TBI children are reported to have greater
dysfunction than either Mild TBI or Orthopedic Control children in terms of Working
Memory functioning in their day-to-day environment. The Moderate/Severe TBI group
also had significantly greater differences in ratings of behavior from pre-injury to post-
injury on the BRIEF Working Memory and Plan/Organize scales than either the Mild
TBI or Orthopedic Control groups.
For the performance measures of executive function, the two TEA-Ch subtests
were the only neuropsychological tests that displayed significant differences between the
three subject groups. The Moderate/Severe TBI group performed significantly worse
than the Orthopedic Control group for the TEA-Ch Creature Counting test, and the
Moderate/Severe TBI group performed significantly worse than both the Mild TBI and
Orthopedic Control groups on the TEA-Ch Opposite Worlds test.
Few findings resulted from the correlations between the BRIEF scales and
performance measures of executive function. The various BRIEF scales were correlated
with one another. For the comparisons between neuropsychological measures, Trails B
and Creature Counting demonstrated a correlation when evaluating all study participants
as well as just the TBI subjects using a less stringent significance criteria (p < .05), and
the WCST Perseverations and TEA-Ch Creature Counting scores were correlated when
evaluating the TBI subjects alone using this same significance criteria. When comparing
BRIEF scales to the neuropsychological measures, the TEA-Ch Opposite Worlds and
BRIEF Inhibit scores were significant when evaluating all study participants using p <
.05. Unexpectedly, the WCST Perseverations score was also significantly correlated with
the BRIEF Working Memory and BRIEF Plan/Organize scores when examining only the
TBI study participants using < .05. The direction of the relationships for the WCST
and the BRIEF Working Memory and Plan/Organize scales suggests that worse BRIEF
scores are positively correlated with better scores (i.e. less perseverations) on the WCST.
Overall, the lack of correlational findings for the BRIEF and performance measures
of executive function is similar to what has been found by Vriezen and Pigott (2002) and
Mangeot et al. (2002). Vriezen and Pigott (2002) had used the BRIEF's Composite Index
scores in comparison with performance measures of executive function, including the
WCST and Trails B, and found no relationship between the Composite Index scores and
the performance measures. Mangeot et al. (2002) had also utilized BRIEF Composite
Index scores and found only the Consonant Trigrams test to be correlated with the BRIEF
Indices. The results from the correlational analyses suggest that either the two types of
measures for executive function may be capturing different aspects of the same cognitive
domain, labeled as executive function, or that the two kinds of measures are evaluating
altogether separate constructs. Either way, the findings imply that the use of measures of
everyday executive function in addition to neuropsychological measures may provide a
more thorough examination of the executive function domain across a range of situations
and provide insight as to how problems displayed in a testing situation translate into
difficulties on tasks in the child's day-to-day environment that requires such skills.
One of the limitations for this pilot study was the size of the sample. The 35-
subject sample size was smaller than some of the previous studies of the BRIEF in a
chronic pediatric TBI population. One of the other risks of having a small sample is that
a lack of statistical findings is reported as being evidence for no group differences, when
in fact the lack of findings may be more appropriately attributed to the lack of power
resulting from an inadequate sample size. In order to address this issue, effect sizes were
calculated for the statistical results in order to thoroughly assess the findings within the
sample. Some of the effect sizes would suggest that the small sample size prevented the
detection of significant group differences. For example, the effect size for Creature
Counting in the comparison between the Mild TBI and Orthopedic Control groups is
large (1.0), but the difference between the groups on this measure did not reach a
significance level ofp < .05. Sample size issues are also pertinent when numerous
analyses are being performed. Because this study was a pilot study, Tukey's HSD post-
hoc analyses were selected instead of a more conservative analysis, such as the
Bonferroni test. However, in order to examine whether Bonferroni results would have
changed the findings, additional post-hoc analyses using Bonferroni tests were
performed. There were no differences in the results using the Bonferroni tests.
Another of the study limitations involved the heterogeneity of the sample. The
children in the TBI samples had various means of injury. As such, hypothesizing that the
TBI samples should exhibit executive function deficits, which are associated primarily
with injuries to the frontal lobes, makes interpretation of the data somewhat complex.
Neuroradiological data was not available for all participants and was not evaluated for the
current study. However, this heterogeneity issue in not specific to the current study. TBI
research typically does not limit subject recruitment to exclusively acceleration-
deceleration injuries, which are more likely to involve the frontal lobes.
Wide variability in ratings for the BRIEF existed was displayed across all of the
groups as demonstrated by the standard deviations in this study. Another study of the
BRIEF in pediatric TBI had shown this as well. Standard deviation data presented by
Mangeot et al. (2002) ranged from T-scores of 8.7 to 18.4 across BRIEF clinical scales
and the three groups. This is in contrast to the standard deviation data for the normative
sample presented in the BRIEF manual (Gioia et al., 2000), which displays standard
deviations ranging from 2.6 to 5.6 for the individual clinical scales. This finding supports
the notions of study sample heterogeneity and sample size limitations.
Four of the subjects in this pilot study had a pre-injury history of ADHD. The
inclusion of participants with premorbid ADHD could have affected some of the results
since ADHD has been implicated in specific cognitive deficits, such as inhibition
problems. However, given the small number of ADHD cases in this sample and that they
were divided evenly across the three study groups, it was decided to include these
children for this pilot study in order to increase sample size. Furthermore, including
cases of premorbid ADHD lends a certain level of generalizability to the results of this
study, since a subset of pediatric TBI cases have pre-existing ADHD in the general
pediatric TBI population.
The potential for response bias on the BRIEF parent report was another limitation.
There is no way to determine response bias for the BRIEF and whether response bias
may have affected the study's results. While some scales of behavioral report have a
measure to look at response bias, the BRIEF does not. There are several ways response
bias may affect the data. One, it may be possible that following their child's injury, the
parent was distressed by what had occurred and did not want to rate their child's behavior
as poorly as it may have been. On the other hand, following an accident, a parent may
have been hypervigilant and subsequently overreported behavior following the injury.
This limitation is one for the measure rather than for the study but is an important one to
Despite these limitations, the current study has several strengths. First, it is the first
study to examine executive functioning both in terms of observational ratings and
behavioral performance in an acute pediatric TBI population. While other studies have
looked at the BRIEF (Mangeot et al, 2002; Vriezen and Pigott, 2002), those studies'
samples were several years post-injury. None of the published studies have looked at
behavioral rating within the first year of TBI recovery. Second, this study broadened the
range of the TBI population evaluated with the BRIEF by examining mild pediatric TBI
cases in addition to the more severe TBI population. Third, this study attempted to parse
apart aspects of executive function by studying individual BRIEF scale scores rather than
Composite Indices from the BRIEF. Gioia et al. (2002) is the only study to date to have
examined the individual clinical scales. Finally, this study expands the literature by
looking at change in BRIEF ratings over time using pre-injury and post-injury BRIEF
clinical scale ratings. No studies to date have been published regarding executive
function behavioral change following TBI as measured by the BRIEF parent ratings.
One area within the development of the BRIEF that should be addressed is the
construction of a response bias measure. While parents are able to report the frequency
of certain behaviors, there is no measure on the BRIEF to assess the consistency or
validity of responses. A response bias measure would be useful in the interpretation of
the BRIEF both for clinical and research purposes for those reasons.
In terms of future directions for studying executive function and cognition
constructs associated with it, the utilization of functional imaging techniques would be
future direction for the study of executive function in pediatric TBI. In the TBI literature,
it has been difficult to come up with consistent behavioral findings for executive
function, especially for mild TBI. Some TBI patients display severe impairment on tasks
involving executive function, while others do not. However, it may be the case that in
those patients where performance output does not appear to be impaired, the underlying
brain systems involved in executive function tasks may still be compromised to a certain
degree. During two functional magnetic resonance imaging (FMRI) study of working
memory in adult mild TBI, McAllister et al. (1999, 2001) found that the pattern of frontal
activation in relation to memory load was altered in an adult mild TBI sample as
compared to uninjured controls despite similar outward task performance for both groups.
Both studies used an auditory n-back working memory task that had varying load, and in
both studies, the mild TBI group performed similar to controls in regards to behavioral
accuracy for the tasks. In the 1999 study, McAllister et al. reported that controls showed
bifrontal and biparietal activation in a low-processing load task, with a small increase in
activation associated with a medium-load (2-back) task. In contrast, mild TBI patients
showed a greater increase in activation during the medium-load task, notably in the right
parietal and dorsolateral frontal regions. McAllister et al. (2001) showed that in a high-
low condition (3-back), controls continued to increase activation within regions of
working memory circuitry. In contrast, the mild TBI group had a greater increase of
activation during the moderate-load condition, with little additional activation for the
highest processing load condition. FMRI studies of executive function concepts such as
working memory have not been done to date using a pediatric TBI sample. However,
studies in the adult TBI literature suggest that processing differences could exist between
pediatric TBI patients and age-matched healthy counterparts.
Last, the study of age-based subgroups would be an additional research direction in
order to examine the relationship between age and executive function in pediatric TBI.
Welsh, Pennington and Grossier (1991) and Levin et al. (1991) suggested that various
aspects of executive function become efficient at different points during development.
The pediatric TBI studies of executive function performance by Slomine et al. (2002) and
Levin and colleagues (1994, 1997) described greater impairment in the younger TBI
samples than in the older groups of children. Similarly, there may be age differences
within this current study sample. The children who were injured at a younger age may
have received worse BRIEF ratings and may have performed worse on the performance
measures. While this was not examined for the current study due to the small sample
size, it would be an interesting direction of study. Similar to the idea of age-based
subgroup analyses, the use of a longitudinal study design would provide information
about the relationship between age and executive function as well as the trajectory of
recovery over time.
LIST OF REFERENCES
Achenbach, T., and Edelbrock, C. (1983). Manual for the Child Behavior Checklist and
the Revised Behavior Profile. Burlington, VT: University of Vermont, Department
Asarnow, R. F., Satz, P., Light, R., Zaucha, K., Lewis, R., and McCleary, C. (1995).
The UCLA study of mild closed head injury in children and adolescents. In S. H.
Broman and M. E. Michel (Eds.), Traumatic head injury in children (pp. 117-146).
New York: Oxford University Press.
Baron, I. S. (2004). Neuropsychological evaluation of the child. New York: Oxford
Gioia, G. A., Isquith, P., K., Guy, S. C., and Kenworthy, L. (2000). Behavior Rating
Inventory of Executive Function. Odessa, FL: Psychological Assessment
Gioia G. A., Isquith, P. K., Kenworthy, L., and Barton, R. M. (2002). Profiles of
everyday executive function in acquired and developmental disorders. Child
Neuropsychology, 8, 121-37.
Golden, C. J. (1978). Stroop Color and Word Test. Wood Dale, IL: Stoelting Company.
Heaton, R. K., Chelune, G. J., Talley, J. L., Kay, G. G., and Curtiss, G. (1993).
Wisconsin Card Sorting Test manual: Revised and expanded. Odessa, FL:
Psychological Assessment Resources, Inc.
Klonoff, H., and Low, M. (1974). Disordered brain function in young children and early
adolescents: Neuropsychological and electroencephalographic correlates. In R. M.
Reitan and L. A. Davison (Eds.), Clinical neuropsychology: current status and
applications. New York: John Wiley.
Knights, R. M. (1966). Normative data on tests evaluating brain damage in children 5-
14 years of age. Research Bulletin No. 20. London, Ontario: Department of
Psychology, University of Western Ontario.
Kraus, J. F. (1995). Epidemiological features of brain injury in children: Occurrence,
children at risk, causes and manner of injury, severity and outcomes. In S. H.
Broman and M. E. Michel (Eds.), Traumatic head injury in children (pp. 22-39).
New York: Oxford University Press.
Langlois, J. and Gotsch, K. (2001). Traumatic brain injury in the United States:
Assessing outcome in children. National Center for Injury Prevention and Control,
Center for Disease Control and Prevention. Atlanta, GA.
Levin, H. S., Culhane, K. A., Hartmann, J., Evankovich, K., Mattson, A. J., Harward, H.,
Ringholz, G., Ewing-Cobbs, L., and Fletcher, J. M. (1991). Developmental
changes in performance on tests of purported frontal lobe functioning.
Developmental Neuropsychology, 7, 377-395.
Levin, H., S., Hanten, G., Chang, C. C., Zhang, L., Schachar, R., Ewing-Cobbs, L., and
Max, J. E. (2002). Working memory after traumatic brain injury in children.
Annals of Neurology, 52, 82-88.
Levin, H. S., Mendelsohn, D., Lilly, M. A., Fletcher, J. M., Culhane, K. A., Chapman, S.
B., Harward, K., Kusnerik, L., Bruce, D., and Eisenberg, H. M. (1994). Tower of
London performance in relation to magnetic resonance imaging following closed
head injury in children. Neuropsychology, 8, 171-179.
Levin, H. S., Song, J., Scheibel, R. S., Fletcher, J. M., Harward, H., Lilly, M., and
Goldstein, F. (1997). Concept formation and problem-solving following closed
head injury in children. Journal of the International Neuropsychological Society,
Light, R., Asarnow, R., Satz, P., Zaucha, K., McCleary, C., and Lewis, R. (1998). Mild
closed-head injury in children and adolescents: Behavior problems and academic
outcomes. Journal of Consulting and Clinical Psychology, 66, 1023-1029.
Loring, D. W. (1999). INS dictionary of neuropsychology. New York: Oxford
Mangeot, S.D., Armstrong, K., Colvin, A.N., Yeates, K.O., Taylor, H.G. (2002). Long-
term executive deficits in children with traumatic brain injuries: Assessing using
the Behavior Rating Inventory of Executive Function (BRIEF). Child
Neuropsychology, 8, 271-284.
Manly, T., Robertson, I., Anderson, V., and Nimmo-Smith, I. (1999). TEA-Ch: The Test
of Everyday Attention for Children. Bury St. Edmunds, England: Thames Valley
Test Company Limited.
McAllister, T.W., Saykin, A.J., Flashman, L.A., Sparling, M.B., Johnson, S.C., Guerin,
S.J., Mamourian, A.C., Weaver, J.B., and Yanofsky, N. (1999). Brain activation
during working memory 1 month after mild traumatic brain injury: A functional
MRI study. Neurology, 53, 1300-1308.
McAllister, T.W., Sparling, M.B., Flashman, L.A., Guerin, S.J., Mamourian, A.C., and
Saykin, A.J. (2001). Differential working memory load effects after mild traumatic
brain injury. Neurolmage, 14, 1004-1012.
National Center for Injury Prevention and Control, Center for Disease Control and
Prevention (CDC). (2003, September 2). Traumatic brain injury (TBI): Incidence
and distribution. Retrieved May 18, 2004 from http://www.cdc.gov/
Proctor, A., Wilson, B., Sanchez, C., and Wesley, E. (2000). Executive function and
verbal working memory in adolescents with closed head injury (CHI). Brain
Injury, 14, 633-647.
Reitain, R. (1971). Trail Making Test results for normal and brain-damaged children.
Perceptual and Motor ,kd//1, 33, 575-581.
Reitan, R. M. (1979). Manual for administration of neuropsychological test batteries for
adults and children. Tucson, AZ: Neuropsychology Laboratory.
Slomine, B. S., Gerring, J. P., Grados, M. A., Vasa, R., Brady, K. D., Christensen, J. R.,
and Denckla, M. B. (2002). Performance on measures of "executive function"
following pediatric traumatic brain injury. Brain Injury, 16, 759-772.
SPSS 11.0.1 for Windows. (2001). Chicago, IL: SPSS, Inc.
Stuss, D. T. and Benson, D. F. (1986). The frontal lobes. New York: Raven Press.
Teasdale, G. and Jennett, B. (1974). Assessment of coma and impaired
consciousnessness: A practical scale. Lancet, 2, 81-84.
Vriezen, E. R., and Pigott, S. E. (2002). The relationship between parental report on the
BRIEF and performance-based measures of executive function in children with
moderate to severe traumatic brain injury. Child Neuropsychology, 8, 296-303.
Welsh, M. C., Pennington, B. F., and Groisser, D. B. (1991). A normative-
developmental study of executive function: a window on prefrontal function in
children. Developmental Neuropsychology, 7, 131-149.
Yeates, K. 0. (2000). Closed-head injury. In K. 0. Yeates, M. D. Ris, and H. G. Taylor
(Eds.), Pediatric neuropsychology: Research, theory, and practice. (pp. 92-116).
New York: Guildford Press.
Michelle Benjamin graduated from DePaul University with a bachelor's degree in
psychology. She spent two years working in Madison, Wisconsin, at the Wisconsin Early
Autism Project as a research assistant. There she studied the outcome effects of a
behavioral modification treatment program for children diagnosed with autism and other
pervasive developmental disorders. Ms. Benjamin then went on to work as a research
assistant for the Department of Psychiatry at the University of Iowa. At the University of
Iowa, she was involved in research evaluating decisional capacity in schizophrenia and
prisoners, cognitive and psychiatric change in Huntington's disease, and the cognitive
effects of atherosclerotic vascular disease. Ms. Benjamin is currently working towards a
doctorate in clinical and health psychology with a specialization in clinical
neuropsychology at the University of Florida.