Cognitive impairment and functioning in PTSD related to intimate partner violence.

Abstract

Posttraumatic stress disorder (PTSD) has been associated with neuropsychological impairments across multiple domains, but consensus regarding the cognitive profile of PTSD has not been reached. In this study of women with PTSD related to intimate partner violence (n = 55) and healthy, demographically similar comparison participants (NCs; n = 20), we attempted to control for many potential confounds in PTSD samples. All participants were assessed with a comprehensive neuropsychological battery emphasizing executive functioning, including inhibition, switching, and abstraction. NCs outperformed PTSD participants on most neuropsychological measures, but the differences were significant only on speeded tasks (with and without executive functioning components). The PTSD group's mean performance was within the average range on all neuropsychological tests. Within the PTSD group, more severe PTSD symptoms were associated with slower processing speed, and more severe dissociative symptoms were associated with poorer reasoning performance. These results suggest that women with PTSD related to intimate partner violence demonstrate slower than normal processing speed, which is associated with the severity of psychiatric symptoms. We speculate that the cognitive slowing seen in PTSD may be attributable to reduced attention due to a need to allocate resources to cope with psychological distress or unpleasant internal experiences.

Full text

Page 1

Journal of the International Neuropsychological Society (2009), 15, 879 – 887 .
Copyright © INS. Published by Cambridge University Press, 2009.
doi:10.1017/S135561770999049X
879
COGNITIVE IMPAIRMENT AND
FUNCTIONING IN PTSD
Posttraumatic Stress Disorder (PTSD) was included in the
third edition of the Diagnostic and Statistical Manual
(DSM-III) in 1980 (American Psychiatric Association,
1980 ) and began to be studied from a neuropsychological
perspective soon afterward (e.g., Everly & Horton, 1989 ).
Interest in the neuropsychological effects of stress and
trauma has risen during the past decade, particularly in light
of the current confl icts in Iraq and Afghanistan, along with
research showing that deployment alone is a risk factor
for neurocognitive dysfunction (Vasterling et al., 2006 ).
Research on the cognitive impairments of trauma and PTSD
over the past three decades has yet to yield a consensus re-
garding which cognitive domains are most affected, how-
ever. Impairments have been identifi ed in almost every
cognitive domain, including premorbid intellectual func-
tioning (Vasterling et al., 2002 ), attention/working memory
(Horner & Hamner, 2002 ; Samuelson et al., 2006 ; Stein,
Kennedy, & Twamley, 2002 ; Vasterling, Brailey, Constans,
& Sutker, 1998 ; Vasterling et al., 2002 ), processing speed
(Samuelson et al., 2006 ; Stein et al., 2002 ), learning (Horner
& Hamner, 2002 ; Samuelson et al., 2006 ; Vasterling et al.,
SYMPOSIUM
Cognitive impairment and functioning in PTSD related
to intimate partner violence
ELIZABETH W. TWAMLEY , 1 , 2 , 3 CAROLYN B. ALLARD , 1 , 3 STEVEN R. THORP , 1 , 2 , 4
SONYA B. NORMAN , 1 , 2 , 3 SHADHA HAMI CISSELL , 1 , 4 KELLY HUGHES BERARDI , 3
ERIN M. GRIMES , 1 , 4 and MURRAY B. STEIN 1 , 5 , 6
1 Department of Psychiatry , University of California , San Diego , La Jolla , California
2 Center of Excellence for Stress and Mental Health , VA San Diego Healthcare System , San Diego , California
3 Psychology Service , VA San Diego Healthcare System , San Diego , California
4 Research Service , VA San Diego Healthcare System , San Diego , California
5 Psychiatry Service , VA San Diego Healthcare System , San Diego , California
6 Department of Family and Preventive Medicine , University of California , San Diego , La Jolla , California
(Received December 23 , 2008 ; Final Revision June 16 , 2009 ; Accepted June 16 , 2009 )
Abstract
Posttraumatic stress disorder (PTSD) has been associated with neuropsychological impairments across multiple
domains, but consensus regarding the cognitive profi le of PTSD has not been reached. In this study of women with
PTSD related to intimate partner violence ( n = 55) and healthy, demographically similar comparison participants
(NCs; n = 20), we attempted to control for many potential confounds in PTSD samples. All participants were assessed
with a comprehensive neuropsychological battery emphasizing executive functioning, including inhibition, switching,
and abstraction. NCs outperformed PTSD participants on most neuropsychological measures, but the differences were
signifi cant only on speeded tasks (with and without executive functioning components). The PTSD group’s mean
performance was within the average range on all neuropsychological tests. Within the PTSD group, more severe PTSD
symptoms were associated with slower processing speed, and more severe dissociative symptoms were associated with
poorer reasoning performance. These results suggest that women with PTSD related to intimate partner violence
demonstrate slower than normal processing speed, which is associated with the severity of psychiatric symptoms.
We speculate that the cognitive slowing seen in PTSD may be attributable to reduced attention due to a need to allocate
resources to cope with psychological distress or unpleasant internal experiences. ( JINS , 2009, 15, 879–887.)
Keywords : Posttraumatic stress disorder , Trauma , Stress , Neuropsychological , Cognition , Domestic violence
Correspondence and reprint requests to: Elizabeth W. Twamley, Depart-
ment of Psychiatry, University of California, San Diego, 140 Arbor Drive
(0851), San Diego, CA 92103. E-mail: etwamley@ucsd.edu

Page 2

E.W. Twamley et al.
880
1998 , 2002 ), and executive functioning (Bremner et al.,
2004 ; Leskin & White, 2007 ). Most studies have found
no association between PTSD and impairment in memory
retention or recognition (see Golier & Yehuda, 2002 , for
review). The pattern of neuropsychological impairments
found across studies suggests disruption in frontal-
subcortical circuits (Stein et al., 2002 ; Vasterling et al., 1998 ,
2002 ), although decreased hippocampal size and function-
ing have also been associated with PTSD (see Bremner,
2006 , for review). Importantly, not all studies have found
differences in cognitive functioning between PTSD and non-
PTSD samples (Gurvits et al., 1993 ; Stein, Hanna, Vaerum,
& Koverola, 1999 ; Twamley, Hami, & Stein, 2004 ).
Certain aspects of trauma exposure and PTSD have been
associated with high likelihood of neurocognitive dysfunc-
tion and real-world functional impairments. For example,
exposure to childhood trauma is associated with worse aca-
demic performance and numerous areas of cognitive defi -
cits (Carrey, Butter, Persinger, & Bialik, 1995 ; Perez &
Widom, 1994 ; Saigh, Mroueh, & Bremner, 1997 ). Among
Axis I psychiatric disorders including PTSD, greater
cognitive defi cits are associated with greater psychiatric
symptoms and functional impairment (Geuze, Vermetten,
de Kloet, Hijman, & Westenberg, 2009 ; Green, Kern, Braff,
& Mintz, 2000 ; Kalechstein, Newton, & van Gorp, 2003 ;
Marvel & Paradiso, 2004 ; Twamley et al., 2002 ). There-
fore, cognitive impairments are an important area of focus
in PTSD research. However, a consistent picture of the
neuropsychological sequelae of PTSD has been elusive,
partly due to the numerous confounds associated with
trauma exposure. Potential factors that result in inconsis-
tencies across samples include: variation in the history of
PTSD (e.g., number, timing, spacing, and severity of trau-
matic events; age of onset, severity, and chronicity of PTSD
symptoms); variation in types of traumatic events (e.g.,
betrayal-related trauma such as rape or incest, nonbetrayal
events such as accidents or acts of nature, and other events
such as killing or being injured in combat or terrorist at-
tacks); and variation in comorbid conditions (e.g., current
or historical depression, substance use disorders, dissocia-
tion, brain injury, or other medical illnesses known to af-
fect cognition).
Our program of research on PTSD-associated neuropsy-
chological impairment has attempted to reduce the infl uence
of some of these factors by focusing on one type of trauma—
domestic or intimate partner violence (IPV). We have lim-
ited our trauma sample to women survivors of IPV. Thus,
each of the women experienced betrayal trauma, which has
been shown to result in a greater likelihood and severity
of posttraumatic symptoms (Freyd, 1996 ; Freyd, Klest, &
Allard, 2005 ). All of the women in our sample experienced
trauma during adulthood (although some also experienced
childhood trauma), and none had comorbid alcohol or sub-
stance use disorders when they participated in our research.
Furthermore, we do not include women with comorbid
medical conditions or medications that could affect brain
functioning. In previous studies of women who meet these
criteria, we have found impairments primarily on timed
tasks of attention/working memory, inhibition, and switch-
ing (Stein et al., 2002 ). To further examine these executive
functioning domains, the current study included an expanded
neuropsychological battery with fi ner-grained assessments
of executive functions in a new sample of women survivors
of IPV.
We hypothesized that compared with non-traumatized
comparison participants (NCs), PTSD participants would
perform worse on neuropsychological tests of executive
functions, particularly timed tasks. Furthermore, we expected
that within the PTSD group, worse neuropsychological per-
formance would be associated with greater childhood his-
tory of trauma, more severe current psychiatric symptoms,
and greater levels of functional disability.
METHOD
Participants
From a larger initial sample, 3 PTSD participants with low
levels of education ( ≤ 10 years) and 11 NC participants with
high levels of education ( ≥ 21 years ) were excluded from the
current study to create groups that did not differ statistically
on education. The fi nal sample of participants included
55 women with full or partial PTSD related to IPV experi-
enced within the past 5 years and 20 women who never ex-
perienced IPV and never met full or partial criteria for PTSD
(current or lifetime) related to any trauma (NCs). Partici-
pants were recruited through ads in a local events and enter-
tainment magazine, on a community volunteer Web site, and
fl yers distributed to community agencies (including agen-
cies that provide IPV-related services). Of the 55 subjects
with IPV exposure, 48 met full DSM-IV criteria for PTSD,
while seven partially met criteria for PTSD (i.e., fulfi lled
Criterion A and the impairment/distress criterion, and had
one less symptom than necessary to fulfi ll C or D criteria).
Excluding the seven subjects with partial PTSD did not
change the results in any meaningful way. Therefore, they
were included in the PTSD group. Partial PTSD has been
associated with the same level of distress and impairment as
full PTSD (Stein, Walker, Hazen, & Forde, 1997 ), although
this has not been observed uniformly across studies ( Breslau,
Lucia, & Davis, 2004 ).
Participants were excluded for factors known to affect
neurocognitive functioning because these would interfere
with understanding the relationship between cognitive func-
tioning and PTSD. These exclusion criteria were: (1) current
abuse of alcohol or other substances; (2) abuse of alcohol or
other substances for a period of 5 years or more; (3) use of
psychotropic medication within the last 4 weeks (fl uoxetine
within the last 6 weeks) or steroids in the past 4 months;
(4) history of bipolar disorder, schizophrenia, attention defi -
cit disorder, learning disability, loss of consciousness greater
than 10 min or requiring hospitalization for 24 hr or more, or
any neurological illness; and (5) English reading ability be-
low the 8th grade level.

Page 3

Cognitive impairment and functioning in PTSD
881
PTSD participants were not excluded if they had other
mood disorders, such as major depressive disorder, because
comorbid mood disorders are common among patients with
PTSD and excluding them would have yielded a non-
representative sample. NCs were excluded if they had any
DSM-IV Axis I disorder. All participants gave informed
written consent to participate in this study, which was
approved by the University of California San Diego Human
Research Protection Program and the VA San Diego
Healthcare System Research and Development Review
Committee.
Procedure
After obtaining verbal informed consent, a telephone inter-
view was used to screen potential participants for the inclu-
sion and exclusion criteria above and to gather demographic
data. Callers were asked fi rst about IPV history. If present,
the Posttraumatic Stress Disorder Checklist - Civilian (PCL-C;
Weathers, Litz, Herman, Huska, & Keane, 1993 ) was admin-
istered with regard to the IPV history. This brief, widely used
self-report instrument was used to quantify PTSD symptoms
and assess the likelihood of PTSD. If the caller met all crite-
ria, she was invited for an in-person interview (see measures
below). If the caller denied IPV history, she was asked about
other trauma history and screened for PTSD using stem
questions from the PTSD section of the Structured Clinical
Interview for DSM-IV Axis I Disorders (SCID-I; First,
Spitzer, Gibbon, & Williams, 2002 ).
To screen for other disorders, the telephone interview in-
cluded stem questions from other diagnostic categories in
the SCID-I. Screeners administered stem questions from the
psychotic and bipolar disorder sections to all callers and ad-
ditionally the major depression, panic, generalized anxiety,
and PTSD stem questions to potential NCs. When a caller
answered any of the stem questions in the positive, the
screener administered the full section of the SCID for that
disorder. Screeners were doctoral level clinicians.
All participants gave written informed consent at their ini-
tial in-person visit. IPV participants completed a set of diag-
nostic interviews in person, took home a set of questionnaires
to complete, and were administered a neuropsychological
assessment battery at a subsequent visit within 6 weeks. NCs
underwent neuropsychological testing and completed the
take-home questionnaires, but were not administered further
diagnostic interviews.
Measures
IPV participants were administered the Clinician-Adminis-
tered PTSD Scale (CAPS; Blake et al., 1995 ), a standard
semi-structured interview to assess PTSD status and severity.
Respondents were asked to describe the most traumatic IPV
event(s) they experienced in their most recent abusive rela-
tionship, to be used as the basis of assessing PTSD. The
CAPS assesses the frequency and intensity of each of the
17 items from the DSM-IV criteria B, C, and D during the past
month. The F1/I2 method of scoring was used in the current
study to determine diagnostic status (Weathers, Keane, &
Davidson, 2001 ), and a severity score was computed by
summing the frequency and intensity scores over all 17 items.
Ten percent of the CAPS outcome assessments were ran-
domly selected throughout the study for audiotape review to
establish inter-rater reliability. The intraclass correlation
coeffi cient obtained on seven CAPS independently rated by
two raters was .99 for the total score.
A take-home battery of self-report questionnaires was
completed by all participants, including the following mea-
sures of child abuse history, current psychiatric symptom
severity, and level of functioning. The Childhood Trauma
Questionnaire (CTQ; Bernstein, Fink, Handelsman, & Foote,
1994 ) measures childhood maltreatment across fi ve domains:
emotional abuse, physical abuse, sexual abuse, emotional
neglect, and physical neglect. The total score, ranging from
5 to 125, was used in the current study, with higher scores
refl ecting higher levels of childhood trauma. The Beck De-
pression Inventory (BDI-II; Beck, Steer, & Brown, 1996 )
was used to determine the presence and severity of symp-
toms of depression during the past week. Higher scores
indicate greater severity of depressive symptoms. The Dis-
sociative Experiences Scale Taxon (DES-T; Waller, Putnam,
& Carlson, 1996 ) is a modifi ed version of the Dissociative
Experiences Scale that relies on a subset of eight categorical
items that refl ect abnormal or severe dissociative experi-
ences. Scores on the DES-T range from 0 to 100, with higher
scores refl ecting higher levels of dissociative experiences.
The Sheehan Disability Scale (SDS; Sheehan, 1983 ) as-
sessed level of disablement using visual 0 to 10 rating scales
and verbal descriptive anchors in three domains: work, so-
cial life, and family/home life. Higher scores on these three
subscales refl ect greater levels of impairment.
All participants underwent neuropsychological testing.
Premorbid intellectual functioning was measured with the
American National Adult Reading Test (ANART; Grober &
Sliwinski, 1991 ). The neuropsychological battery empha-
sized executive functioning, using multiple subtests of the
Delis-Kaplan Executive Function System (D-KEFS; Delis,
Kaplan, & Kramer, 2001 ). The Wisconsin Card Sorting Test
(Kongs, Thompson, Iverson, & Heaton, 2000 ) was used as a
general measure of reasoning. Verbal learning and memory
tests were not included because in our previous work, we
found no difference between women with PTSD due to IPV
and healthy comparison subjects on verbal learning and
memory tests using story, list, or paired associate stimuli.
The battery included assessments of multiple domains, in-
cluding the following: (1) Processing Speed (scaled scores
from the D-KEFS Trail Making Test Visual Scanning,
Number Sequencing, Letter Sequencing, and Motor Speed
conditions); (2) Visuoconstruction (Rey-Osterrieth Complex
Figure Test copy t score; Rey, 1941 ; norms from Spreen &
Strauss, 1998 ); (3) Visual Memory (Rey-Osterrieth Complex
Figure Test 30-min delayed recall t score; Rey, 1941 ; norms
from Spreen & Strauss, 1998 ); (4) Speeded Fluency (scaled
scores from the D-KEFS Design Fluency Test [Combined

Page 4

E.W. Twamley et al.
882
Filled Dots and Empty Dots], Letter Fluency, and Category
Fluency); (5) Inhibition (D-KEFS Color-Word Interference
Test Inhibition vs. Color Naming scaled score); (6) Switching
(scaled scores from the D-KEFS Trail Making Test Number-
Letter Switching vs. Combined Number Sequencing and
Letter Sequencing Contrast Score; D-KEFS Design Fluency
Test Switching vs. Combined Filled Dots and Empty Dots
Contrast Score; D-KEFS Category Fluency Switching vs.
Category Fluency Contrast Score; D-KEFS Color-Word In-
terference Test Inhibition/Switching vs. Inhibition Contrast
Score); (7) Reasoning (Wisconsin Card Sorting Test-64 total
errors t score [WCST]; Kongs et al., 2000 ).
The t scores (characterized in a normal population by a
mean of 50 and standard deviation of 10) were used for the
Rey-Osterrieth Complex Figure Test and for the WCST;
scaled scores (characterized in a normal population by a
mean of 10 and standard deviation of 3) were used for all
other tests.
Analyses
All variables were normally distributed. To assess the extent
to which our groups were demographically matched, we
conducted t tests to compare groups on age and education,
and ? χ? 2 analyses to measure distribution differences in minor-
ity status. To reduce the possibility of Type I error, hypothe-
sized differences between the two groups were tested with
MANOVA (multivariate analysis of variance; when there
was more than one test per neuropsychological domain) and
t tests (when there was only one test per neuropsychological
domain or when the MANOVA was signifi cant). Associa-
tions between neuropsychological performance and symptom
and disability severity were tested with Pearson correlations.
Alpha for signifi cance was set at .05 for the MANOVAs and
.01 for all other tests, to reduce the chance of Type I error
due to multiple comparisons.
RESULTS
The groups did not differ statistically on age, education, eth-
nic minority status, or premorbid intellectual functioning as
measured by the ANART (see Table 1 ). Within the PTSD
group, the mean CAPS score was 73.3 ( SD = 21.9); the mean
PCL-C score was 62.3 [ SD = 11.7; CAPS scores above 60
and PCL-C scores above 50 are each suggestive of PTSD
(Weathers et al., 1993 , 2001 )]; the mean BDI-II score was
17.9 ( SD = 8.9), in the range of moderate depressive symp-
toms (Beck et al., 1996 ); the mean score on the DES-T was
11.0 ( SD = 14.3), indicative of high average levels of disso-
ciation (Seedat, Stein, & Forde, 2003 ). PTSD participants’
mean CTQ score was 52.4 ( SD = 20.4), suggesting that many
participants had experienced childhood abuse. On the SDS,
PTSD participants reported moderate levels of disablement
in work, social, and home/family domains [means ( SDs )
were 5.7 (2.8), 6.7 (2.7), and 5.4 (2.9), respectively].
On the neuropsychological measures, within the PTSD
group, the percentage of the sample that scored within the
impaired range (<1.5 SD below the mean) was 13% on the
Rey-Osterrieth copy, 17% on the Rey-Osterrieth delay, 17%
on the WCST, and less than 10% on all other tests. By com-
parison, rates of impairment among the NC group were 14%
on Rey-Osterrieth delay and less than 10% on all other
tests.
In partial support of our fi rst hypothesis, NCs outper-
formed participants in the PTSD group on most neuropsy-
chological measures. However, these differences were only
signifi cant in the domains of processing speed and speeded
fl uency, with a trend toward signifi cance in reasoning (see
Table 1 ). Within the processing speed domain, the PTSD
group performed worse than did the NC group on D-KEFS
Trail Making Letter Sequencing and Motor Speed. Within
the speeded fl uency domain, the PTSD group performed
worse than did NCs on D-KEFS Design Fluency and Letter
Fluency. It is important to note that on all tests, the PTSD
group’s mean performance was within the average range.
Supporting our second hypothesis (see Table 2 ), we found
that, within the PTSD group, more severe current PTSD
symptoms (measured with the CAPS interview) were associ-
ated with lower processing speed (Trail Making Visual Scan;
r = −.36; p = .009). More severe dissociative symptoms
(DES-T) were associated with worse reasoning (WCST; r =
−.54; p < .001). More severe childhood maltreatment (CTQ)
was associated with worse performance on D-KEFS Cate-
gory Fluency Switching vs. Category Fluency (r = −.42; p =
.005). On the SDS, unexpectedly, self-reported family/home
impairment was associated with better visuoconstruction
performance (Rey-Osterrieth copy; r = .51; p < .001). Cur-
rent level of depressive symptomatology (BDI-II) was not
associated with performance on any neuropsychological
measure.
DISCUSSION
Although we expected PTSD participants to perform signifi -
cantly worse than would NCs on tasks of executive function-
ing, the pattern of results showed that they differed primarily
on speeded tasks (some of which involved the executive
functioning domain of fl uency). Thus, our results are similar
to those fi nding PTSD-associated defi cits in processing
speed and executive functioning in recent U.S. veterans of
the Iraq and Afghanistan confl icts (Nelson, Yoash-Gantz,
Pickett, & Campbell, 2009 ). Notably, the differences we
found between groups do not appear to be attributable to de-
pressive symptom severity, as BDI-II scores were not associ-
ated with any performance on neuropsychological measure.
Our PTSD sample did not differ signifi cantly from the NC
sample on any of the switching tests in the D-KEFS. It is
important to note that most group differences were in the
direction of NCs outperforming PTSD participants; larger
sample sizes may have resulted in more statistically signifi -
cant differences. However, such differences between groups
may have little clinical signifi cance, considering that the
mean performance of the PTSD group on all tests was in
the average range. Importantly, as a group, they were not

Page 5

Cognitive impairment and functioning in PTSD
883
impaired on any test. Although individuals with PTSD fre-
quently present with subjective cognitive diffi culty, previous
research has found that cognitive complaints tend to be re-
lated more to psychiatric symptom severity than to actual
cognitive performance (Binder, Storzbach, Anger, Campbell,
& Rohlman, 1999 ; Garcia et al., 2004 ). Indeed, some well-
controlled research on combat veterans with PTSD found no
neuropsychological differences between patients with cur-
rent PTSD, past PTSD, psychiatric comparison subjects,
and healthy comparison subjects, despite large sample sizes
(80 per group; Crowell et al., 2002 ). We found that higher
levels of PTSD symptoms were associated with worse
speeded attention, and higher levels of dissociation were as-
sociated with poorer reasoning performance. These results
are similar to previous work fi nding associations between
dissociation severity and greater Stroop interference (Freyd,
Martorello, Alvarado, Hayes, & Christman, 1998 ).
Why would PTSD adversely affect processing speed?
Cognitive slowing may be attributable to reduced attentional
resources, which could be a consequence of brain resources
being directed toward coping with psychological distress,
unpleasant internal experiences (see DePrince & Freyd,
1999 , 2004 ), or potential threats in the environment (i.e.,
hypervigilance) rather than the task at hand. In effect, even
simple tasks could become exercises in multitasking. If
people with PTSD experience micro-dissociations when
confronted with intrusive memories, for example, reduced
attentional resources could yield slower performance on pro-
cessing speed measures. Other factors that could infl uence
speeded tasks include sleep defi cits, heightened anxiety or
Table 1. Comparison of demographic characteristics and cognitive function in the PTSD and NC groups
( p values <.05 for MANOVAs and <.01 for other comparisons are highlighted in bold)
PTSD NC Test statistic df p

Age, years
Education, years
Mean
36.7
14.0
%
51
Mean
30.2

10.6
10.3
9.5
10.6
SD
8.8
1.9


SD
9.7

1.8
2.6
2.8
1.7
Mean
35.0
14.6
%
50
Mean
31.1
Hotelling’s Trace F = 4.69
11.9 2.1
10.7 2.3
11.7 1.1
12.0 1.7
SD
8.6
1.5


SD
10.4

t = 0.74
t = 1.32
73
73
.460
.191

.944

.808
.002
.011
.572
<.001
.004

% ethnic minority status

ANART
Processing Speed MANOVA
Trail Making Visual Scanning
Trail Making Number Sequencing
Trail Making Letter Sequencing
Trail Making Motor Speed
Visuoconstruction
Rey-Osterrieth Copy
Visual Memory
Rey-Osterrieth Delay
Speeded Fluency MANOVA
Design Fluency (Filled + Empty)
Letter Fluency
Category Fluency
Inhibition
Color-Word Interference Inhibition
vs. Color Naming
Switching MANOVA
Trail Making Number-Letter Switching
vs. Combined Number Sequencing +
Letter Sequencing
Design Fluency Switching vs.
Combined Filled + Empty
Category Fluency Switching vs.
Category Fluency
Color-Word Interference Inhibition/
Switching vs. Inhibition
Reasoning
Wisconsin Card Sorting Test total
errors t score
χ 2 = 0.005 1

t = 0.25 44
4,64
68
68
65.9 *
67
t = 2.61
t = 0.57
t = 4.63
t = 2.95
49.0 13.4 52.5 11.8 t = 0.87 65 .388
45.5

9.1
10.1
11.0
9.0

2.3
3.5
3.3
44.6
Hotelling’s Trace F = 4.71
11.1 2.0
12.6 2.7
11.8 3.5
7.6 t = 0.34 65 .737
.005
.002
.009
.354
3,66
68
68
68
t = 3.20
t = 2.69
t = 0.93
11.1 2.3 11.3 1.8 t = 0.34 68 .739
Hotelling’s Trace F = 0.47
9.2 1.9
4,65
—
.758
— 9.0 2.7 —
10.9 3.0 10.7 2.1 — — —
11.1 3.1 10.6 2.9 — — —
9.3 2.9 10.2 2.3 — — —
45.1 9.0 51.3 5.9 t = 2.40 59 .020
Note. ANART = American National Adult Reading Test; MANOVA = multivariate analysis of variance; PTSD = posttraumatic stress
disorder group; NC = normal comparison group; SD = standard deviation.
* degrees of freedom adjusted for unequal variances