Journal of Psychiatric Research 79 (2016) 70e77

Contents lists available at ScienceDirect

Journal of Psychiatric Research

journal homepage: www.elsevier.com/locate/psychires

Early life trauma is associated with altered white matter integrity and
affective control
Vincent Corbo a, b, *, Melissa A. Amick a, William P. Milberg a, c, Regina E. McGlinchey a, c,
David H. Salat a, d
a
Translational Research Center for TBI and Stress Disorders/Geriatric Research Education and Clinical Centers (GRECC), VA Boston Healthcare System,
Boston, MA, USA
b
Boston University School of Medicine, Boston, MA, USA
c
Department of Psychiatry, Harvard Medical School, Cambridge, MA, USA
d
Athinoula A. Martinos Center for Biomedical Imaging, Charlestown, MA, USA

a r t i c l e i n f o a b s t r a c t

Article history: Early life trauma (ELT) has been shown to impair affective control and attention well into adulthood.
Received 26 January 2016 Neuroimaging studies have further shown that ELT was associated with decreased white matter integrity
Received in revised form in the prefrontal areas in children and adults. However, no study to date has looked at the relationship
15 April 2016
between white matter integrity and affective control in individuals with and without a history of ELT. To
Accepted 3 May 2016
examine this, we tested 240 Veterans with (ELT N ¼ 80) and without (NoELT N ¼ 160) a history of
childhood sexual abuse, physical abuse or family violence. Affective control was measured with the
Keywords:
Affective Go/No-Go (AGN) and attention was indexed with the Test of Variable Attention (TOVA). White
Trauma
Childhood
matter integrity was measured using fractional anisotropy (FA). Results showed greater number of errors
Fractional anisotropy on the AGN in ELT compared to NoELT. There was no difference on the TOVA. While there were no mean
Executive function differences in FA, there was an interaction between FA and reaction time to positive stimuli on the AGN
PTSD where the ELT group showed a positive relationship between FA and reaction time in right frontal and
Attention prefrontal areas, whereas the NoELT group showed a negative or no association between FA and reaction
time. This suggests that ELT may be associated with a distinct brain-behavior relationship that could be
related to other determinants of FA than those present in healthy adults.
Published by Elsevier Ltd.

1. Introduction that the effects of trauma are greatest during specific sensitive
periods of brain development (Andersen, 2003). This increased
The aim of the current study was to examine the association impact of stress on the brain has already been illustrated in animal
between white matter integrity and cognitive performance in a models, especially in the area of the hippocampus, amygdala and
sample of Veterans with and without a history of early life trauma. prefrontal cortex (PFC) (McEwen, 2008). Based on these animal
Emerging literature has alerted researchers and clinicians to the models, studies using Magnetic Resonance Imaging (MRI) have
long-term impact of exposure to early life trauma (ELT) on physical provided evidence of the important and lasting impact of ELT on
and psychological health and well being (Anda et al., 2006). While brain integrity potentially suggesting altered development. For
the exact definition of what constitutes ELT or childhood adversity example, ELT has been associated with alteration in the funda-
differs between studies, most authors include childhood sexual mental symmetry of the PFC of children exposed to ELT compared
abuse, physical abuse and family violence under the umbrella term to age-matched control subjects (Carrion et al., 2001). Another
ELT, since these three traumas share an interpersonal dimension important region for stress, the anterior cingulate cortex, has been
(De Bellis and Zisk, 2014; Perry et al., 1995). Importantly, it appears shown to be smaller in children exposed to trauma (Cohen et al.,
2006). Similarly, another study (Richert et al., 2006) reported
smaller dorso-lateral PFC but larger ventral and middle PFC in
children exposed to ELT. Altered gray matter volume within the
* Corresponding author. 150 South Huntington Ave, C-11-36, Boston, MA 02120,
USA.
frontal regions of children exposed to ELT have since been repli-
E-mail address: [email protected] (V. Corbo). cated, underscoring the consistency of these effects (Carrion and

http://dx.doi.org/10.1016/j.jpsychires.2016.05.001
0022-3956/Published by Elsevier Ltd.
 V. Corbo et al. / Journal of Psychiatric Research 79 (2016) 70e77 71

Wong, 2012; Hanson et al., 2010). Studies of adults reporting a anterior cingulate. We also hypothesized that individuals with ELT
history of ELT have shown altered gray matter integrity lasting into would evidence impaired attention performance, indexed by
adulthood in regions overlapping the findings of studies in chil- greater number of commission errors, on the sustained attention
dren, such as the PFC (Tomoda et al., 2009), dorsal anterior cingu- task and greater interference of negative emotional distracters on
late cortex (Thomaes et al., 2010), orbito-frontal and anterior the affective go/no-go task, indexed by greater number of com-
cingulate cortices (Dannlowski et al., 2012), as well as mid- mission errors for the negative distracter (positive target).
cingulate cortex (Corbo et al., 2014).
Brain development presents different windows for gray and 2. Material and methods
white matter. White matter development lasts longer, with mye-
lination in the frontal areas extending into early adulthood 2.1. Recruitment
(Andersen, 2003). Despite what may be a larger window for the
effect of stress on the white matter, studies investigating the impact Three hundred and seven service members were initially
of ELT on the brain have mainly focused on gray matter and our recruited from the cohort of the VA Rehabilitation Research &
knowledge of the impact of ELT on white matter development is Development-supported Traumatic Brain Injury (TBI) National
sparser. Recent studies have used diffusion tensor imaging to Center for TBI Research (NCR) at VA Boston Healthcare System: The
investigate the integrity of the underlying white matter tracts Translational Research Center for TBI and Stress Disorders
across the central nervous system. These studies have shown (TRACTS). Participants enrolled in the TRACTS NCR cohort are
reduced fractional anisotropy (FA), a marker of white matter recruited from the Boston Metropolitan area and the surroundings.
integrity, in the anterior and posterior sections of the corpus cal- Individuals were ineligible for enrollment in TRACTS if they met any
losum in maltreated children compared to healthy control subjects of the following criteria: (a) history of neurological illness (other
(Jackowski et al., 2008), in the left arcuate fasciculus and cingulum than TBI); (b) history of seizures; (c) current diagnosis of schizo-
bundle in adults victim of verbal abuse during childhood (Choi phrenia spectrum or other psychotic disorders (not related to
et al., 2009), and in the inferior longitudinal fasciculus of the oc- PTSD); (d) current active suicidal and/or homicidal ideation, intent,
cipital region (Choi et al., 2012) of adults reporting a history of or plan requiring crisis intervention; or (e) cognitive disorder due to
domestic abuse. A recent study (Teicher et al., 2014) extended these general medical condition other than TBI. The Institutional Review
findings by showing how ELT (verbal abuse) was associated with Board of Human Studies Research at the VA Boston Healthcare
reduced FA in the corpus callosum and corona radiata. Similarly, System approved all research procedures and all participants pro-
one study (Benedetti et al., 2014) showed that ELT affected the vided informed consent and were reimbursed for their time and
white matter integrity of the bilateral superior longitudinal travel expenses. From the original 307 participants, we excluded
fasciculi and the left anterior thalamic radiation. Collectively, these participants with missing ELT rating (N ¼ 13), individuals with
studies suggest an impact of ELT on the development of white missing neuropsychological data (N ¼ 17) and clinical data (N ¼ 10).
matter, though the specific localization of the impact remains Additionally, to ensure that performance accurately reflected sub-
largely unclear. ject’s abilities, the Medical Symptom Validity Test [MSVT, (Green,
Brain imaging studies of ELT are critical to understanding re- 2003)] was administered. Subjects who failed the MSVT were
ported behavioral impairments in victims of ELT, especially the excluded from the analyses (N ¼ 27). The final sample consisted of
cognitive domains of attention and affective control. Studies two hundred and forty participants, including a total of 22 women
showed that children reporting maltreatment presented a bias in and 218 men.
attention to threat-related stimuli, suggesting altered attention
control (Gamble and Rapee, 2009; Pine et al., 2005; Roy et al., 2008; 2.2. Participants
Waters et al., 2008). Beyond the attention bias, children exposed to
ELT have been shown to perform more poorly on cognitive tasks History of ELT was determined using the Traumatic Life Events
involving attention and affective control compared to healthy age- Questionnaire (see below). Based on this questionnaire, the ELT
matched control subjects (DePrince et al., 2009; Gould et al., 2012; group (N ¼ 80) was composed of individuals reporting a history of
Kaplow et al., 2008; Nikulina and Widom, 2013; Porter et al., 2005). physical abuse, sexual abuse and/or family violence before the age
Similar impairments have been shown in neutral (Navalta et al., of 18 coupled with an A2 reaction of fear/helplessness/horror as
2006) and affective (Gould et al., 2012) Go/No-Go paradigms, defined by DSM-IV-TR, (APA, 2004). The age cut-off insured that the
suggesting that ELT may impair inhibition and affective control, in traumatic event would be pre-deployment. The control group
general, as well as attention. In sum, there is mounting evidence for (NoELT, N ¼ 160) was composed of individuals who reported no
both impaired attention in individuals with a history of ELT and interpersonal trauma before the age of 18. This did not preclude
decreased integrity of brain structures thought to support attention exposure to other events not of an interpersonal nature (e.g. natural
and affective control. However no study to date has investigated disaster, motor-vehicle accidents, witness of robbery, criterion A2
how ELT may influence the association between white matter of the DSM-IV-TR), though none were diagnosed with PTSD as a
integrity, especially in tracts located in the PFC (Casey et al., 2000; result of this pre-deployment event. Effectively, this means that
Manna et al., 2010; Wood and Grafman, 2003), and performance on subjects of the ELT group all had trauma exposure before deploy-
tasks of attention/affective control, a brain region and cognitive ment, while some of the participants in the NoELT group reported a
functions that may be specifically impacted by ELT. first trauma exposure before deployment (e.g. earthquake that did
The objectives of the current study were to examine the impact not result in PTSD).
of ELT on prefrontal white matter integrity and to evaluate a
possible link between white matter structural changes and 2.3. Clinical assessment
behavioral performance on tasks of affective control and attention.
White matter integrity was evaluated across the whole brain using All clinical assessments were conducted by a doctoral-level
diffusion tensor imaging. Attention and affective control were psychologist and reviewed by at least three doctoral-level psy-
measured using a continuous performance task and affective Go/ chologists to achieve consensus
No-go task, respectively. We hypothesized that individuals with Clinician Administered PTSD Scale (CAPS): Current PTSD Diag-
ELT would show decreased FA in the corpus callosum and the PFC/ nosis and symptoms severity were assessed using the Clinician-
72 V. Corbo et al. / Journal of Psychiatric Research 79 (2016) 70e77

Administered PTSD Scale [CAPS; (Blake et al., 1995)]. The CAPS is a to stimuli of the other valence. Order of presentation was coun-
30-item structured interview that corresponds to the DSM-IV terbalanced across participants. In total, participants are asked to
criteria for PTSD. Since TRACTS was initiated before the transition complete ten blocks. The first two blocks of the task served as
to DSM-V criteria, we maintained use of the DSM-IV-TR version practice trials. The critical dataset consisted of the remaining eight
(although since released, participants also received a DSM-V diag- blocks. Each block contains nine words congruent with the target
nosis via the CAPS-V). Severity of symptoms was computed by valence (“Go” target words) and nine incongruent words (“No-Go”
summing the frequency and intensity of clusters of symptoms B distracter words). Dependent measures included, for both the
(Flashback and Intrusive Memories), C (Avoidance and Emotional positive and negative valence, the number of omission errors (i.e.,
Numbing) and D (Hyperarousal). failure to respond for a word that matched the targeted valence),
Structured clinical interview for DSM-IV axis I disorders (SCID-I/ commission errors (i.e., responding for a non-matched word for a
NP): The SCID-I/NP (http://www.scid4.org/psychometric/) was targeted valence), and mean correct latency (LAT; in ms).
administered to determine eligibility and to characterize the in-
dividual’s psychological history. The SCID-I/NP is a semi-structured
2.6. Statistical analyses
interview that covers main diagnoses of DSM-IV Axis I psychiatric
disorders both currently and as a lifetime condition.
All analyses were conducted with SPSS 16.0 for Mac OS X 10.6.
Boston Assessment of TBI-Lifetime (BAT-L): Since traumatic brain
Independent t-tests were computed to compare group means on
injuries and exposures to blast munitions are prevalent in military
possible continuous variables (except for sex, because of the sig-
samples, TBI exposure was assessed using the Boston Assessment of
nificant discrepancy in sample sizes, in which case we used the
TBI-Lifetime (Fortier et al., 2013), a measure validated by TRACTS
Mann-Whitney U non-parametric test), whereas c2 tests were used
and developed in collaboration with a number of services at VA
for discrete measures. For each performance measure on the AGN
BHS. Initial inter-rater reliability and validity have been established
and TOVA, we used a General Linear Model with ELT (±) as
for the BAT-L and is comparable to the Ohio State University TBI
between-subjects factor and age, PTSD severity, level of education
Identification Method (Corrigan and Bogner, 2007; Fortier et al.,
and number of mTBI across the lifetime were included as a covar-
2013). For the purpose of this study, number of mild TBIs across
iate for all models. All p-values were set at 0.05.
lifetime was the variable retained as a measure for brain injuries. A
single individual reported moderate/severe TBI and was not
excluded from the initial analyses. 2.7. MRI acquisition and processing

2.4. Questionnaires Imaging procedures for this protocol have been described
elsewhere (Trotter et al., 2015). All scans were performed at the
History of interpersonal early life trauma was determined using Neuroimaging Research for Veterans (NeRVe) Center at the VA
the Traumatic Life Events Questionnaire [TLEQ, (Kubany et al., 2000)]. Boston Healthcare System in Jamaica Plains, using a Siemens Tim
The TLEQ is a 23-item self-report measure of 22 types of potentially Trio 3-T scanner with 12 RF channels head coil. Image acquisition
traumatic events including natural disasters, exposure to warfare, was based on a 2 mm isotropic, 60 direction single shot echo planar
robbery involving a weapon, physical abuse and being stalked. It sequence with a twice-refocused spin echo pulse sequence (Reese
records age of onset and reaction of horror and helplessness. The et al., 2003), 10 T2 (b ¼ 0)þ60 diffusion directions; b ¼ 700,
TLEQ has good temporal stability, reliability, and validity (Kubany TR ¼ 10,000 ms, TE ¼ 103 ms; bandwidth ¼ 1395 Hz/px, slice
et al., 2000). Total alcohol consumption adjusted for weight was thickness ¼ 2.0 mm, FOV ¼ 256  256 mm; 128  128 matrix; 64
measured using the Lifetime Drinking History (Koenig et al., 2009). slices; with 0 gap]. To increase signal-to-noise ratio, the 60
Severity of combat exposure was assessed using the Combat scale diffusion-weighted directions were obtained using the electrostatic
of the Deployment Risk and Resilience Inventory (Vogt et al., 2008). shell method (Jones et al., 1999). The diffusion tensor was calcu-
General IQ was estimated using the Wechsler Test of Adult Reading lated on a voxel-by-voxel basis (Basser et al., 1994; Fischl et al.,
(WTAR). 1999; Smith et al., 2004). The FreeSurfer image analysis suite
(http://surfer.nmr.mgh.harvard.edu/) and FSL Diffusion Toolbox
2.5. Neuropsychological assessment (http://www.fmrib.ox.ac.uk.fsl/) were used to preprocess the
diffusion data. Briefly, the raw data was motion and eddy corrected
The Test of Variable Attention (TOVA): The TOVA (Greenberg and using FSL. A T2-weighted structural volume (b0) volume, collected
Waldmant, 2006) is an objective, neurocognitive measure of sus- using identical sequence parameters, was then used as an affine,
tained attention commonly known as a continuous performance and rigid body registration target for eddy current, and motion
task (CPT). It is a 20-min long computerized, non-language based, correction of the raw DTI data, respectively (Jenkinson et al., 2002).
fixed interval, visual performance test. The test is comprised of two The fitting of a tensor model to the corrected images was per-
phases presented in the following order for all participants: infre- formed using FreeSurfer. The brain extraction toolkit of FSL was then
quent target presentation and frequent target presentation. During used to extract the diffusion images (Smith et al., 2004). Data were
each phase, participants are asked to respond as quickly as possible then prepared for statistical analysis using Tract-Based Spatial
to a geometrical figure (target) while refraining from responding to Statistics [TBSS, (Smith et al., 2006)]. For the purpose of this study,
the non-target geometrical figure. Measures of commission errors we used the FMRIB58 FA skeleton template after coregistering the
(responding to non-target), omission errors (not responding to data. Voxel-wise analyses of the white matter skeleton was per-
target), reaction time (RT) and reaction time variability (RTV) were formed with the TBSS tool Randomise (Winkler et al., 2014), a
recorded for this study. nonparametric permutation testing method for statistical analysis
The affective Go/No-Go (AGN): The AGN task of the CANTAB that avoids assumptions about the distribution of DTI data. To
(www.cambridgecognition.com) consists of series of stimuli words compensate for multiple comparisons issues, threshold-free cluster
of either positive (e.g. Happy, Victory) or negative (e.g. Weakness, enhancements were used (Smith et al., 2006; Winkler et al., 2014).
Defeat) valence presented on the center of a monitor for 300 ms Analyses tested for main effect of ELT on FA as well as formal
with a 900 ms inter-stimulus-interval (ISI). Participants are asked interaction of ELT on the association between FA and measures of
to respond to the words of the target valence while not responding performance on the TOVA/AGN.
 V. Corbo et al. / Journal of Psychiatric Research 79 (2016) 70e77 73

3. Results

3.1. Clinical/demographic group differences

All clinical and demographic information is summarized in

Table 1. Analyses revealed that the ELT þ group was older
[T(239) ¼ 2.45, p ¼ 0.015], had an earlier age of first trauma
exposure [T(239) ¼ 10.14, p < 0.001], and had lower premorbid
estimated IQ [T(239) ¼ 2.44, p ¼ 0.02]. While groups did not differ
in current severity of PTSD symptoms, the ELT þ group presented
with lower levels of combat exposure [T(239) ¼ 3.02, p ¼ 0.003].

3.2. Cognitive performance

All performance measures are displayed in Table 1. Individuals

in the ELT þ group committed more commission errors for both
positive and negative target in the AGN (see Fig. 1). This result
remained significant after controlling for age, PTSD severity and
education level. However, groups did not differ on omission errors
and in reaction time for either positive or negative valence. When
measuring performance on the TOVA, both groups performed Fig. 1. Number of errors per group on the Affective Go/No-Go (both commission and
similarly in errors, reaction time and reaction time variability. omission errors). Individuals with a history of ELT showed on average more errors of
Across both groups on the TOVA, women presented a slower RT commission, independent of the valence of the stimulus presented.
compared to male subjects [Mann-Whitney U Z ¼ 2.162, p < 0.03].
There was no effect of sex on AGN performance.
140/102; ELT þ r ¼ 0.480, p < 0.001; NoELT r ¼ 0.194,
p < 0.015)(xyz-peak 59/151/108; ELT þ r ¼ 0.341, p < 0.002; NoELT
3.3. FA and behavior association
r ¼ 0.087, n.s.)(xyz-peak 52/125/114; ELT þ r ¼ 0.202, n.s.; NoELT
r ¼ 0.139, n.s.), precentral gyrus white matter (xyz-peak 52/119/
There was no significant difference in mean FA between groups
104; ELT þ r ¼ 0.477, p < 0.001; NoELT r ¼ 0.110, n.s.), external
across the whole brain. When looking at the impact of ELT on the
capsule/pars orbitalis of the inferior frontal gyrus (xyz-peak 60/135/
association between performance on the AGN and FA, a significant
86; ELT þ r ¼ 0.355, p < 0.002; NoELT r ¼ 0.164, p < 0.040),
interaction was found for the mean correct latency for positive
postcentral gyrus white matter (xyz-peak 46/108/105;
target (see Fig. 2) in clusters located in the right frontal/prefrontal
ELT þ r ¼ 0.450, p < 0.001; NoELT r ¼ 0.200, p < 0.012), anterior
white matter. Specifically, ELT þ participants presented a positive
corona radiata (xyz-peak 63/156/84; ELT þ r ¼ 0.350, p < 0.002;
association between mean correct latency and FA, while the NoELT
NoELT r ¼ 0.046, n.s.), pars triangularis of the inferior frontal gyrus
participants presented a negative or no association, in the right
(xyz-peak 56/161/71; ELT þ r ¼ 0.338, p < 0.003; NoELT r ¼ 0.181,
caudal/rostral middle frontal gyrus white matter (xyz-peak 56/157/
p < 0.024), putamen white matter (xyz-peak 72/129/64;
98; ELT þ r ¼ 0.217, n.s.; NoELT r ¼ 0.160, p < 0.045)(xyz-peak 63/
ELT þ r ¼ 0.274, p < 0.016; NoELT r ¼ 0.270, p < 0.001), lateral
orbitofrontal gyrus white matter/anterior corona radiata (xyz-peak
Table 1 68/181/65; ELT þ r ¼ 0.245, p < 0.032; NoELT r ¼ 0.155, p < 0.052),
Sociodemographic, clinical and behavioral results per group. lateral orbitofrontal cortex (xyz-peak 64/151/61; ELT þ r ¼ 0.313,
NoELT ELT T/F/c P p < 0.006; NoELT r ¼ 0.219, p < 0.006), superior/anterior corona
(N ¼ 160) (N ¼ 80) radiata (xyz-peak 64/139/92; ELT þ r ¼ 0.375, p < 0.001; NoELT
Age 30.85 (8.39) 33.73 (8.18) 2.40 0.017*
r ¼ 0.022, n.s.), and anterior limb of the external capsule/superior
Sex (women:men) 10:150 12:68 4.90 0.027* fronto-occipital fasciculus (xyz-peak 68/136/89; ELT þ r ¼ 0.363,
Education 14.07 (1.97) 13.73 (1.86) 1.30 0.20 p < 0.001; NoELT r ¼ 0.077, n.s.). All other regions of the white
CAPS Curr 42.71 (29.24) 49.37 (26.59) 4.63 0.10 matter skeleton were not associated with the mean correct latency
Depl Duration 13.78 (8.73) 12.86 (5.90) 0.92 0.36
of the AGN. Furthermore, there was no voxel-wise association be-
Combat exposure 16.87 (12.13) 12.64 (9.80) 2.76 0.006*
Lifetime mTBI 1.31 (2.00) 1.77 (2.79) 1.38 0.17 tween FA and other metrics of the AGN or of the TOVA.
Depression 6.82 (8.81) 7.73 (7.17) 0.76 0.45
IQ 103.24 (11.21) 99.41 (12.83) 2.26 0.025*
AGN Lat Positive 499.88 (70.72) 489.08 (72.46) 1.05 0.29 4. Discussion
AGN Lat Negative 502.21 (68.72) 495.26 (76.63) 0.68 0.50
AGN Comm Positive 5.16 (4.25) 6.89 (5.58) 2.33 0.021* In this study, we examined how, in a sample of Veterans, a
AGN Comm Negative 4.81 (3.94) 6.37 (5.70) 2.09 0.039* history of ELT influenced the association between white matter
AGN Omiss Positive 3.15 (4.06) 3.56 (4.06) 0.701 0.48
AGN Omiss Negative 2.61 (3.38) 3.44 (4.17) 1.47 0.145
integrity and performance on attention and affective control tasks.
TOVA Comm 11.83 (9.89) 13.12 (9.87) 0.913 0.362 Our findings indicated impaired affective control in subjects with
TOVA Omis 16.24 (37.69) 18.68 (37.14) 0.453 0.651 ELT, as revealed by greater errors of commission on the AGN.
TOVA RT 354.93 (62.52) 380.05 (111.35) 1.79 0.077 Additionally, reaction time for positive targets on the AGN was
TOVA RTV 100.73 (38.45) 111.45 (52.27) 1.55 0.123
positively associated with white matter integrity in the ELT þ group
Sociodemographic, clinical and behavioral results per group: Curr ¼ current, within while negatively associated in the NoELT group. This suggests that
the last month; Depl ¼ deployment (in months); AGN ¼ Affective Go/No-Go; stressful events occurring during periods of brain development
Lat ¼ mean correct latency; Comm ¼ errors of commission; Omis ¼ errors of
omission; TOVA ¼ Test of Variable Attention; RT ¼ reaction time; RTV ¼ reaction
(e.g. myelination) may influence how behavior becomes associated
time variability. with markers of brain integrity. This finding offers a new
* ¼ p < 0.05. perspective on the impact of ELT on the brain. Prior studies using
74 V. Corbo et al. / Journal of Psychiatric Research 79 (2016) 70e77

Fig. 2. Clusters of significant interaction between ELT/NoELT on the relationship between FA and mean correct latency for positive stimuli on the Affective Go/No-Go, controlling for
PTSD symptoms severity, age and level of education. Fourteen (clusters 1e4 illustrated here) clusters were detected within the right frontal tracts, all significant at p < 0.05 corrected
for multiple comparisons.

brain imaging techniques have generally approached the question assumption behind these studies was that stress and trauma should
of the impact of ELT on white matter by testing for mean differences reduce FA levels. Instead, what our findings suggest is that ELT may
in FA between groups (Benedetti et al., 2014; Choi et al., 2012; Choi act to shape differently the association between brain and behavior.
et al., 2009; Jackowski et al., 2008; Teicher et al., 2014). The As was noted by Beaulieu (Beaulieu, 2002), diffusion of water in the
 V. Corbo et al. / Journal of Psychiatric Research 79 (2016) 70e77 75

cortex is influenced by many factors outside myelination. One study observed. Future studies should address this specific question.
illustrated how gliosis, i.e. the process of generation of glial cell, Despite these limitations, our study does contribute to the impor-
contributed significantly to observed levels of fractional anisotropy tant field of the cognitive neuroscience of ELT. Since our findings
(Budde et al., 2011). Furthermore, stress hormones have been were exclusive to the AGN, this suggests that the affective control
shown to stimulate gliosis and even be protective of the myelin domain may suffer more importantly from the impact of ELT when
sheet, thus contributing potentially to higher FA (Jauregui-Huerta contrasted to sustained attention. As mentioned earlier, no study so
et al., 2010). This proliferation of glial cells would effectively lead far has investigated the impact of ELT on how behavior is associated
to an increase in FA concomitant to an increase in reaction time, as with white matter integrity. This approach may prove helpful in
noted in our study. In adults without ELT, the decrease in FA would revealing new and subtler effects of stress and trauma. It may also
likely be due to neural death, leading to slower processing speed help us understand further how various aspects of brain structural
and therefore longer reaction times. Thus, the opposite patterns integrity are associated with different components of cognition,
observed between FA and behavioral performances would be due from processing of the stimuli to decision-making and its behav-
to different cellular mechanisms influencing FA. ioral result. Through these findings, we may better understand the
The association of white matter integrity and cognitive perfor- long-lasting emotional and cognitive impact of ELT.
mance has been shown in healthy young adults and in aging. Initial
studies showed a moderate association between decline in cogni- Author contributions
tive function and non-specific white matter lesions (Gunning-
Dixon and Raz, 2000). However, the studies based on diffusion  Vincent Corbo: Collection of data, data analysis, and writing of
are less clear regarding the presence and directionality of the as- manuscript.
sociation between white matter integrity and reaction time. One  Melissa A Amick: Provided clinical expertise for clinical vari-
study of healthy adults ranging from 35 to 74 years old found that, ables, contributed to data analyses, manuscript editing.
controlling for the effect of age, an increase in fractional anisotropy  William P Milberg: provided expertise for neuropsychological
was related to a decrease in reaction time (Grieve et al., 2007). This assessment, data analyses and manuscript editing.
inverse association, which echoes the finding in our NoELT group,  Regina E McGlinchey: supervised clinical and neuropsycholog-
was replicated in a different sample of healthy young adults, aged ical data collection, provided expertise for trauma and PTSD,
on average 23 years old (Konrad et al., 2009). Sasson et al. 2012 provided expertise for neuropsychological assessment, provided
extended these findings by showing the inverse relationship be- expertise for statistical analyses, manuscript editing.
tween FA and a composite measure of reaction time drawn from  David H Salat: supervised MRI data collection, provided exper-
multiple cognitive tests. Children and adolescents who survived tise in MRI data analyses, provided expertise in statistical
brain tumors also presented this inverse relationship (Rueckriegel modeling, manuscript editing.
et al., 2015), illustrating that such an association is not present
only in healthy individuals. However, some other groups have
found that the relationship between FA and performance is not Role of funding agencies
always inversed. In a later study, results showed that FA in areas of
the cerebellar white matter correlated positively with auditory The funding agencies (VA Rehabilitation, Research and Devel-
reaction time in young healthy adults (Bohr et al., 2007). In a opment & the National Institute of Health) were not involved in the
sample of children diagnosed with Williams syndrome, Hoeft et al. study development, data analysis, writing of report or in the deci-
(Hoeft et al., 2007) showed that worse performance on a visuo- sion to submit this manuscript for publication.
spatial task was associated with greater FA. A similar relationship
was observed in a mixed sample of healthy individuals and patients Acknowledgements
with brain lesions (Turken et al., 2008). Our findings seem to sup-
port the latter observations whereby a lesion, illness or, in our case, The authors would like to thank Mr. Wally Musto for his
excessive stress may be associated with a linear relationship be- championship of our work among military personnel and his tire-
tween FA and performance on cognitive testing, whereas in healthy less recruitment efforts on our behalf. We would further like to
cases, no relationship appears between white matter and perfor- acknowledge the work of Drs. Alexandra Kenna, Catherine Fortier,
mance. Interestingly, for our study, the association was present only Ann Rasmusson, Brad Brummett, Sara Lippa and Colleen Jackson for
for the reaction time to positive targets, and thus to negative dis- the clinical assessments and diagnoses. The authors would also like
tracters. This is in line with studies that have shown that in- to thank Dr. Jennifer Fonda for assistance in the statistical support
dividuals exposed to trauma present an attention bias to negative and database management.
or trauma-related stimuli, potentially interfering with performance This research was support by the Translational Research Center
(Fani et al., 2012; Pannu Hayes et al., 2009; Pine et al., 2005). for TBI and Stress Disorders (TRACTS), a VA Rehabilitation Research
This study does have limitations. Our clinical characterization of and Development Traumatic Brain Injury Center of Excellence
early life trauma did not allow us to measure some dimensions of (B9254-C) provided by the United States (U.S.) Department of
adversity in childhood that could have affected the performance Veterans Affairs, as well as NIH grant #R01NR10827 (Dr. David
and integrity of white matter tracts, such as emotional abuse and Salat). The contents do not represent the views of the U.S.
physical/emotional neglect. Similarly, while there may be an asso- Department of Veterans Affairs or the United States Government.
ciation between the number of trauma exposures/types and
behavioral outcomes, we were unable to test for this. Further, our Appendix A. Supplementary data
study sample was composed of Veterans who have been deployed
and thus re-exposed to trauma. While we controlled statistically for Supplementary data related to this article can be found at http://
the impact of current severity of PTSD symptoms, further studies dx.doi.org/10.1016/j.jpsychires.2016.05.001.
should include a non-deployed sample. Lastly, while the mal-
e:female ratio in this sample was representative of the Veterans References
from recent conflicts, because of the low number of women, it was
impossible to properly assess for the impact of sex on the effects Anda, R.F., Felitti, V.J., Bremner, J.D., Walker, J.D., Whitfield, C., Perry, B.D., et al.,
76 V. Corbo et al. / Journal of Psychiatric Research 79 (2016) 70e77

2006. The enduring effects of abuse and related adverse experiences in child- is not always better: increased fractional anisotropy of superior longitudinal
hood. A convergence of evidence from neurobiology and epidemiology. Eur. fasciculus associated with poor visuospatial abilities in Williams syndrome.
Archives Psychiatry Clin. Neurosci. 256, 174e186. J. Neurosci. 27, 11960e11965.
Andersen, S.L., 2003. Trajectories of brain development: point of vulnerability or Jackowski, A.P., Douglas-Palumberi, H., Jackowski, M., Win, L., Schultz, R.T.,
window of opportunity? Neurosci. Biobehav Rev. 27, 3e18. Staib, L.W., et al., 2008. Corpus callosum in maltreated children with post-
APA, 2004. In: Diagnostic and Statistical Manual of Mental Disorders. 4th, Text traumatic stress disorder: a diffusion tensor imaging study. Psychiatry Res. 162,
Revision Ed. American Psychiatric Association. 256e261.
Basser, P.J., Mattiello, J., LeBihan, D., 1994. MR diffusion tensor spectroscopy and Jauregui-Huerta, F., Ruvalcaba-Delgadillo, Y., Gonzalez-Castaneda, R., Garcia-
imaging. Biophysics J. 66, 259e267. Estrada, J., Gonzalez-Perez, O., Luquin, S., 2010. Responses of glial cells to stress
Beaulieu, C., 2002. The basis of anisotropic water diffusion in the nervous system - a and glucocorticoids. Curr. Immunol. Rev. 6, 195e204.
technical review. NMR Biomed. 15, 435e455. Jenkinson, M., Bannister, P., Brady, M., Smith, S., 2002. Improved optimization for
Benedetti, F., Bollettini, I., Radaelli, D., Poletti, S., Locatelli, C., Falini, A., et al., 2014. the robust and accurate linear registration and motion correction of brain im-
Adverse childhood experiences influence white matter microstructure in pa- ages. Neuroimage 17, 825e841.
tients with bipolar disorder. Psychol. Med. 44, 3069e3082. Jones, D.K., Horsfield, M.A., Simmons, A., 1999. Optimal strategies for measuring
Blake, D.D., Weathers, F.W., Nagy, L.M., Kaloupek, D.G., Gusman, F.D., Charney, D.S., diffusion in anisotropic systems by magnetic resonance imaging. Magn. Reson.
et al., 1995. The development of a clinician administered PTSD scale. J. Trauma. Med. 42, 515e525.
Stress 8, 75e90. Kaplow, J.B., Hall, E., Koenen, K.C., Dodge, K.A., Amaya-Jackson, L., 2008. Dissociation
Bohr, S., Gullmar, D., Knab, R., Reichenbach, J.R., Witte, O.W., Haueisen, J., 2007. predicts later attention problems in sexually abused children. Child Abuse Negl.
Fractional anisotropy correlates with auditory simple reaction time perfor- 32, 261e275.
mance. Brain Res. 1186, 194e202. Koenig, L.B., Jacob, T., Haber, J.R., 2009. Validity of the lifetime drinking history: a
Budde, M.D., Janes, L., Gold, E., Turtzo, L.C., Frank, J.A., 2011. The contribution of comparison of retrospective and prospective quantity-frequency measures.
gliosis to diffusion tensor anisotropy and tractography following traumatic J. Stud. Alcohol Drugs 70, 296e303.
brain injury: validation in the rat using Fourier analysis of stained tissue sec- Konrad, A., Vucurevic, G., Musso, F., Stoeter, P., Winterer, G., 2009. Correlation of
tions. Brain 134, 2248e2260. brain white matter diffusion anisotropy and mean diffusivity with reaction time
Carrion, V.G., Weems, C.F., Eliez, S., Patwardhan, A., Brown, W., Ray, R.D., et al., 2001. in an oddball task. Neuropsychobiology 60, 55e66.
Attenuation of frontal asymmetry in pediatric posttraumatic stress disorder. Kubany, E.S., Haynes, S.N., Leisen, M.B., Owens, J.A., Kaplan, A.S., Watson, S.B., et al.,
Biol. Psychiatry 50, 943e951. 2000. Development and preliminary validation of a brief broad-spectrum
Carrion, V.G., Wong, S.S., 2012. Can traumatic stress alter the brain? Understanding measure of trauma exposure: the traumatic life events questionnaire. Psychol.
the implications of early trauma on brain development and learning. J. Adolesc. Assess. 12, 210e224.
Health 51, S23eS28. Manna, A., Raffone, A., Perrucci, M.G., Nardo, D., Ferretti, A., Tartaro, A., et al., 2010.
Casey, B.J., Giedd, J.N., Thomas, K.M., 2000. Structural and functional brain devel- Neural correlates of focused attention and cognitive monitoring in meditation.
opment and its relation to cognitive development. Biol. Psychol. 54, 241e257. Brain Res. Bull. 82, 46e56.
Choi, J., Jeong, B., Polcari, A., Rohan, M.L., Teicher, M.H., 2012. Reduced fractional McEwen, B.S., 2008. Understanding the potency of stressful early life experiences
anisotropy in the visual limbic pathway of young adults witnessing domestic on brain and body function. Metabolism 57 (Suppl. 2), S11eS15.
violence in childhood. Neuroimage 59, 1071e1079. Navalta, C.P., Polcari, A., Webster, D.M., Boghossian, A., Teicher, M.H., 2006. Effects of
Choi, J., Jeong, B., Rohan, M.L., Polcari, A.M., Teicher, M.H., 2009. Preliminary evi- childhood sexual abuse on neuropsychological and cognitive function in college
dence for white matter tract abnormalities in young adults exposed to parental women. J. Neuropsychiatry Clin. Neurosci. 18, 45e53.
verbal abuse. Biol. Psychiatry 65, 227e234. Nikulina, V., Widom, C.S., 2013. Child maltreatment and executive functioning in
Cohen, R.A., Grieve, S., Hoth, K.F., Paul, R.H., Sweet, L., Tate, D., et al., 2006. Early life middle adulthood: a prospective examination. Neuropsychology 27, 417e427.
stress and morphometry of the adult anterior cingulate cortex and caudate Pannu Hayes, J., Labar, K.S., Petty, C.M., McCarthy, G., Morey, R.A., 2009. Alterations
nuclei. Biol. Psychiatry 59, 975e982. in the neural circuitry for emotion and attention associated with posttraumatic
Corbo, V., Salat, D.H., Amick, M.M., Leritz, E.C., Milberg, W.P., McGlinchey, R.E., 2014. stress symptomatology. Psychiatry Res. 172, 7e15.
Reduced cortical thickness in veterans exposed to early life trauma. Psychiatry Perry, B.D., Pollard, Ronnie A., Blakley, Toi L., Baker, William L., Vigilante, D., 1995.
Res. Neuroimaging 223, 53e60. Childhood trauma, the neurobiology of adaptation, and use-dependent devel-
Corrigan, J.D., Bogner, J., 2007. Initial reliability and validity of the Ohio State Uni- opment of the brain: how states become traits. Infant Ment. Health J. 16,
versity TBI identification method. J. Head Trauma Rehabilitation 22, 318e329. 271e291.
Dannlowski, U., Stuhrmann, A., Beutelmann, V., Zwanzger, P., Lenzen, T., Pine, D.S., Mogg, K., Bradley, B.P., Montgomery, L., Monk, C.S., McClure, E., et al.,
Grotegerd, D., et al., 2012. Limbic scars: long-term consequences of childhood 2005. Attention bias to threat in maltreated children: implications for vulner-
maltreatment revealed by functional and structural magnetic resonance im- ability to stress-related psychopathology. Am. J. Psychiatry 162, 291e296.
aging. Biol. Psychiatry 71, 286e293. Porter, C., Lawson, J.S., Bigler, E.D., 2005. Neurobehavioral sequelae of child sexual
De Bellis, M.D., Zisk, A., 2014. The biological effects of childhood trauma. Child. abuse. Child. Neuropsychol. 11, 203e220.
Adolesc. Psychiatr. Clin. N. Am. 23, 185e222 vii. Reese, T.G., Heid, O., Weisskoff, R.M., Wedeen, V.J., 2003. Reduction of eddy-current-
DePrince, A.P., Weinzierl, K.M., Combs, M.D., 2009. Executive function performance induced distortion in diffusion MRI using a twice-refocused spin echo. Magn.
and trauma exposure in a community sample of children. Child Abuse Negl. 33, Reson. Med. 49, 177e182.
353e361. Richert, K.A., Carrion, V.G., Karchemskiy, A., Reiss, A.L., 2006. Regional differences of
Fani, N., Jovanovic, T., Ely, T.D., Bradley, B., Gutman, D., Tone, E.B., et al., 2012. Neural the prefrontal cortex in pediatric PTSD: an MRI study. Depress Anxiety 23,
correlates of attention bias to threat in post-traumatic stress disorder. Biol. 17e25.
Psychol. 90, 134e142. Roy, A.K., Vasa, R.A., Bruck, M., Mogg, K., Bradley, B.P., Sweeney, M., et al., 2008.
Fischl, B., Sereno, M.I., Tootell, R.B., Dale, A.M., 1999. High-resolution intersubject Attention bias toward threat in pediatric anxiety disorders. J. Am. Acad. Child
averaging and a coordinate system for the cortical surface. Hum. Brain Mapp. 8, Adolesc. Psychiatry 47, 1189e1196.
272e284. Rueckriegel, S.M., Bruhn, H., Thomale, U.W., Hernaiz Driever, P., 2015. Cerebral
Fortier, C.B., Amick, M., Grande, L., McGlynn, S., Kenna, A., Morra, L., et al., 2013. The white matter fractional anisotropy and tract volume as measured by MR im-
boston assessment of traumatic brain injury-lifetime (BAT-L) semi-structured aging are associated with impaired cognitive and motor function in pediatric
interview: evidence of research utility and validity. J. Head Trauma Rehabili- posterior fossa tumor survivors. Pediatr. Blood Cancer 62, 1252e1258.
tation 66, 1373e1382. Sasson, E., Doniger, G.M., Pasternak, O., Tarrasch, R., Assaf, Y., 2012. Structural cor-
Gamble, A.L., Rapee, R.M., 2009. The time-course of attentional bias in anxious relates of cognitive domains in normal aging with diffusion tensor imaging.
children and adolescents. J. Anxiety Disord. 23, 841e847. Brain Struct. Funct. 217, 503e515.
Gould, F., Clarke, J., Heim, C., Harvey, P.D., Majer, M., Nemeroff, C.B., 2012. The effects Smith, S.M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T.E.,
of child abuse and neglect on cognitive functioning in adulthood. J. Psychiatry Mackay, C.E., et al., 2006. Tract-based spatial statistics: voxelwise analysis of
Res. 46, 500e506. multi-subject diffusion data. Neuroimage 31, 1487e1505.
Green, P., 2003. Green’s Medical Symptom Validity Test (MSVT) for Windows: Smith, S.M., Jenkinson, M., Woolrich, M.W., Beckmann, C.F., Behrens, T.E., Johansen-
User’s Manual. Green’s Publishing, Edmonton, Canada. Berg, H., et al., 2004. Advances in functional and structural MR image analysis
Greenberg, L.M., Waldmant, I.D., 2006. Developmental normative data on the test of and implementation as FSL. Neuroimage 23 (Suppl. 1), S208eS219.
variables of attention (T.O.V.A.). J. Child Psychol. Psychiatry 34, 1019e1030. Teicher, M.H., Samson, J.A., Sheu, Y.S., Polcari, A., McGreenery, C.E., 2014. Hurtful
Grieve, S.M., Williams, L.M., Paul, R.H., Clark, C.R., Gordon, E., 2007. Cognitive aging, words: association of exposure to peer verbal abuse with elevated psychiatric
executive function, and fractional anisotropy: a diffusion tensor MR imaging symptom scores and corpus callosum abnormalities. Am. J. Psychiatry 167,
study. Am. J. Neuroradiol. 28, 226e235. 1464e1471.
Gunning-Dixon, F.M., Raz, N., 2000. The cognitive correlates of white matter ab- Thomaes, K., Dorrepaal, E., Draijer, N., de Ruiter, M.B., van Balkom, A.J., Smit, J.H.,
normalities in normal aging: a quantitative review. Neuropsychology 14, et al., 2010. Reduced anterior cingulate and orbitofrontal volumes in child
224e232. abuse-related complex PTSD. J. Clin. Psychiatry 71, 1636e1644.
Hanson, J.L., Chung, M.K., Avants, B.B., Shirtcliff, E.A., Gee, J.C., Davidson, R.J., et al., Tomoda, A., Suzuki, H., Rabi, K., Sheu, Y.S., Polcari, A., Teicher, M.H., 2009. Reduced
2010. Early stress is associated with alterations in the orbitofrontal cortex: a prefrontal cortical gray matter volume in young adults exposed to harsh
tensor-based morphometry investigation of brain structure and behavioral risk. corporal punishment. Neuroimage 47 (Suppl. 2), T66eT71.
J. Neurosci. 30, 7466e7472. Trotter, B.B., Robinson, M.E., Milberg, W.P., McGlinchey, R.E., Salat, D.H., 2015. Mil-
Hoeft, F., Barnea-Goraly, N., Haas, B.W., Golarai, G., Ng, D., Mills, D., et al., 2007. More itary blast exposure, ageing and white matter integrity. Brain 138 (8),
 V. Corbo et al. / Journal of Psychiatric Research 79 (2016) 70e77 77

2278e2292. Waters, A.M., Mogg, K., Bradley, B.P., Pine, D.S., 2008. Attentional bias for emotional
Turken, A., Whitfield-Gabrieli, S., Bammer, R., Baldo, J.V., Dronkers, N.F., faces in children with generalized anxiety disorder. J. Am. Acad. Child Adolesc.
Gabrieli, J.D., 2008. Cognitive processing speed and the structure of white Psychiatry 47, 435e442.
matter pathways: convergent evidence from normal variation and lesion Winkler, A.M., Ridgway, G.R., Webster, M.A., Smith, S.M., Nichols, T.E., 2014. Per-
studies. Neuroimage 42, 1032e1044. mutation inference for the general linear model. Neuroimage 92, 381e397.
Vogt, D.S., Proctor, S.P., King, D.W., King, L.A., Vasterling, J.J., 2008. Validation of Wood, J.N., Grafman, J., 2003. Human prefrontal cortex: processing and represen-
scales from the deployment risk and resilience inventory in a sample of oper- tational perspectives. Nat. Rev. Neurosci. 4, 139e147.
ation Iraqi freedom veterans. Assessment 15, 391e403.