White Matter Integrity Following Traumatic Brain Injury: The Association with Severity of Injury and Cognitive Functioning

ORIGINAL PAPER

White Matter Integrity Following Traumatic Brain Injury:The Association with Severity of Injury and Cognitive Functioning

Gershon Spitz • Jerome J. Maller •

Richard O’Sullivan • Jennie L. Ponsford

Received: 16 November 2012 / Accepted: 19 March 2013

� Springer Science+Business Media New York 2013

Abstract Traumatic brain injury (TBI) frequently results

in impairments of memory, speed of information process-

ing, and executive functions that may persist over many

years. Diffuse axonal injury is one of the key pathologies

following TBI, causing cognitive impairments due to the

disruption of cortical white matter pathways. The current

study examined the association between injury severity,

cognition, and fractional anisotropy (FA) following

TBI. Two diffusion tensor imaging techniques—region-

of-interest tractography and tract-based spatial statistics—

were used to assess the FA of white matter tracts. This

study examined the comparability of these two approaches

as they relate to injury severity and cognitive performance.

Sixty-eight participants with mild-to-severe TBI, and 25

healthy controls, underwent diffusion tensor imaging

analysis. A subsample of 36 individuals with TBI also

completed cognitive assessment. Results showed reduction

in FA values for those with moderate and severe TBI,

compared to controls and individuals with mild TBI.

Although FA tended to be lower for individuals with mild

TBI no significant differences were found compared to

controls. Information processing speed and executive

abilities were most strongly associated with the FA of

white matter tracts. The results highlight similarities and

differences between region-of-interest tractography and

tract-based spatial statistics approaches, and suggest that

they may be used together to explore pathology following

TBI.

Keywords Cognition � MRI � DTI � Traumatic brain

injury

Introduction

Traumatic brain injury (TBI) results in long-lasting chan-

ges to individuals’ independence, work capacity, and social

functioning (Ponsford et al. 1995a, b). The most common

impairments are those of memory, speed of information

processing, and executive functions. Although cognitive

impairments dissipate over the initial months in the

majority of individuals with mild injuries, those with more

severe injuries may display difficulties for years following

injury (Dikmen et al. 1995; Draper and Ponsford 2008).

Traumatic axonal injury (TAI) is one of the key patholo-

gies following TBI—occurring as a result of acceleration,

deceleration, and rotational forces—causing cognitive

deficits due to the disruption of cortical and subcortical

white matter pathways (Buki and Povlishock 2006; Gen-

narelli et al. 1998; Povlishock and Christman 1994).

The detection of TAI can be investigated using con-

ventional MRI scans, in the form of haemorrhagic and non-

G. Spitz (&) � J. L. Ponsford

School of Psychology and Psychiatry, Monash University,

Clayton, Melbourne, VIC 3800, Australia

e-mail: [email protected]

G. Spitz � J. L. Ponsford

Monash-Epworth Rehabilitation Research Centre,

Epworth Hospital, Melbourne, VIC, Australia

J. J. Maller

Central Clinical School, Monash University, Melbourne,

VIC, Australia

J. J. Maller

Monash Alfred Psychiatry Research Centre, Alfred Hospital,

Melbourne, VIC, Australia

R. O’Sullivan

Healthcare Imaging Services, Epworth Hospital,

Melbourne, VIC, Australia

123

Brain Topogr

DOI 10.1007/s10548-013-0283-0

haemorrhagic lesions (Scheid et al. 2003). However, con-

ventional scans tend to underestimate the extent of damage

(Arfanakis et al. 2002; Huisman et al. 2003; Rugg-Gunn

et al. 2001; Shenton et al. 2012; Toth et al. 2013). DTI is a

technique that is inherently well-suited for investigating

damage to white matter by quantifying the diffusion of

water molecules, providing a more accurate representation

of white matter disruption (Basser 1995; Basser et al.

1994). The directionality of water is used to infer the

integrity of white matter tracts; that is, diffusion should be

greater along the primary orientation of the tract, remaining

relatively restricted perpendicular to the fibre. Whilst DTI

enables the extraction of several meaningful indices of

fibre integrity, fractional anisotropy (FA) is the most

widely utilised for demonstrating the degree of direction-

ality of water diffusion.

Using DTI, widespread disruption of white matter has

been found following TBI. Even following mild TBI, indi-

viduals display white matter disruption both acutely and as

long as 7 years post-injury (Lipton et al. 2008; Lo et al. 2009;

Mayer et al. 2010; Miles et al. 2008; Niogi et al. 2008a). Not

surprisingly, significantly greater white matter disruption is

apparent in those with moderate and severe injuries (Benson

et al. 2007; Bigler et al. 2010; Kennedy et al. 2009; Kraus

et al. 2007; Kumar et al. 2009; Nakayama et al. 2006;

Newcombe et al. 2007; Salmond et al. 2006; Sidaros et al.

2008; Xu et al. 2007). Interestingly, whilst Kumar et al.

(2009) found a decrease in FA in the corpus callosum

between 9 days and 6 months post-injury, Sidaros et al.

(2008) reported increases in FA in several white matter tracts

between 8 weeks and 12 months post-injury. These findings

suggest that neural disruption may occur initially, followed

by axonal reorganisation or even regenerative processes at

later stages post-injury that may reflect change in function

following injury (Levin 2003).

Indeed, the integrity of white matter has been shown to

relate to cognitive processes following TBI (Bigler et al.

2010; Kraus, et al. 2007; Kumar, et al. 2009; Levin et al.

2010; Lo et al. 2009; Little et al. 2010; Niogi et al. 2008b;

Salmond, et al. 2006; Sugiyama et al. 2009). These studies

demonstrate that FA in several of the projection, com-

missural, and association white matter tracts is associated

with memory, whilst Niogi et al. (2008a, b) specifically

implicated the uncinate fasciculus. In addition, Lo et al.

(2009) found that reduced FA in several regions, particu-

larly the dorsolateral prefrontal cortex, was found to predict

executive functioning. In another study, Little et al. (2010)

found that FA in the thalamocortical projection fibres may

play a role in modulating executive ability, as well as

attention. The corona radiata has also been implicated in

attention (Kraus et al. 2007; Niogi et al. 2008b).

Techniques used to examine DTI largely fall under

region-of-interest (ROI) analyses, in which specific brain

regions are chosen a priori based on previous literature, and

voxel-based morphometry (VBM) approaches, which often

assess changes over the entire brain. ROI analysis provides

the opportunity for fine-grained analysis of tracts—both

quantitatively and qualitatively—that may be specifically

implicated in pathology and subsequent outcome. How-

ever, relying solely on ROI techniques poses the risk of

missing abnormal tracts, especially considering the heter-

ogeneous nature of TBI. Although VBM approaches offer

analysis of the entire brain, issues have been noted

regarding image alignment and the variable effect of spatial

smoothing (Ashburner and Friston 2000; Bookstein 2001;

Davatzikos 2004; Jones et al. 2002, 2005; Park et al. 2003).

Tract-based spatial statistics (TBSS) is a recently devel-

oped voxel-based technique that attempts to circumvent

these issues through the development of a white matter

skeleton, representing tracts common to all individuals in a

given study, which is applied to each individual’s image to

correct for residual misalignments (Smith et al. 2006).

TBSS has been used widely and utilised to demonstrate, for

example, that injury severity and cognitive functioning was

related to the integrity of white matter (Kinnunen et al.

2011). However, it is not known whether TBSS provides

results which are comparable with ROI approaches.

The current study examined whether white matter

changes following TBI were associated with injury sever-

ity, defined using days of post-traumatic amnesia (PTA). In

addition, the relationship between the integrity of white

matter and cognitive functions, including memory, infor-

mation processing speed, and executive functions was

investigated. Lastly, this is the first study to examine the

comparability of TBSS—providing an analysis of white

matter voxels over the entire brain—and ROI tractography

as they relate to the assessment of injury severity and

cognition following TBI.

Materials and Methods

Participants

Seventy-eight participants with mild to severe TBI were

recruited for the current study. Individuals with TBI were

excluded if they had previous history of TBI or other neuro-

logical disorder. Two TBI participants were excluded due to

severe brain pathology to avoid problems with image align-

ment. One TBI participant was excluded due to a large sub-

dural hematoma causing brain displacement to the left,

whereas the second participant was excluded due to wide-

spread hydrocephalus. In addition, eight individuals with TBI

were excluded due to large focal lesions, including four

individuals with large right frontal lobe contusions, one with

large left frontal contusion, two with bilateral frontal

Brain Topogr

123

contusions, and the last with an extensive fronto-temporal-

parietal contusion. The final TBI sample comprised 68 TBI

participants. Moreover, MRI scanning for the TBI group was

conducted on average 18.0 months post-injury (SD = 14.01,

range = 2.30–52.67). Twenty-five healthy controls—recrui-

ted by disseminating study flyers—with no prior neurological

or psychological disorders, were also included in the study.

Demographic information for TBI and control participants is

presented in Table 1. The overall TBI group was found to be

less educated and older, with a large proportion of males,

compared to controls. Therefore, subsequent comparisons

between TBI control and TBI participants controlled for,

gender, age, and education. A subsample of 36 TBI partici-

pants also completed cognitive tests, and did not differ from

the TBI participants for which cognitive assessment was not

present in terms of age, education, PTA, or gender. For this

subsample cognitive assessments and MRI scans were con-

ducted on average 4.6 (SD = 10.87, range = 0–55) days

apart.

Neuropsychological Assessment

Participants completed a number of neuropsychological

tests that collectively assessed memory, information pro-

cessing speed and executive functions, and which have

been shown to be sensitive to cognitive impairments fol-

lowing TBI (Draper and Ponsford 2008).

BIRT Memory and Information Processing Battery-List

Learning Subtest

The BIRT list learning test (Coughhlan et al. 2007)

assesses verbal recall and recognition. Participants recalled

a list of words over five trials. Total correctly recalled

words over the five learning trials were used as a measure

of ‘Total Recall’.

The Doors Test from the Doors and People Test

The doors test (Baddeley et al. 1994) is designed to assess

visual memory. This test required participants to recognise a

target door they have previously seen among three other

distractor doors not seen before. There were two trials, of 12

target doors, with the second trial using more similar dis-

tractors. The score for each trial ranges from 0 to 12 with an

overall score for the test ranging from 0 to 24. The two trials

were aggregated to form a measure of ‘Visual Memory’.

The Hayling Sentence Completion Test from the Hayling

and Brixton Tests

The Hayling sentence completion test (Burgess and Shallice

1997) assesses participants’ speed of initiation and response

inhibition abilities. This test comprises two sections, each

comprising 15 sentences with the last word missing. In the

first section (Hayling A) participants were required to finish

the sentence with a word that logically completed the sen-

tence as quickly as possible. In the second section (Hayling

B) participants were asked to complete the sentences with an

unrelated word. The overall scaled score—which incorpo-

rates the total time taken in parts A and B of the test, as well

as the number of errors—was used in the current study.

Controlled Oral Word Association Tests

The COWAT (Benton et al. 1994) is a measure of verbal

fluency and idea generation. Participants were required to

generate as many words as they could, beginning with a

specific letter, within a time span of one minute. The total

number of words generated was recorded for the letters C,

F and L.

Trail Making Test—Parts A and B

The TMT (Reitan and Wolfson 1988) measures processing

speed, attention, and mental flexibility, requiring partici-

pants to join numbers in sequence (Trails A) and to switch

between numbers and letters in sequence (Trails B) as

quickly as possible. In addition to the raw times for Trails

A and Trails B, a third variable was created by subtracting

Trails A time from Trails B, to quantify the executive

component of the TMT (Trails B–Trails A).

Digit Span Subtest of the Wechsler Memory Scale—Third

and Fourth Editions

The forwards and backwards digit span (Wechsler 1997,

2008) was used to assess participants’ immediate attention

span and working memory capacity. The number of correct

trials was recorded for the forward and backward

conditions.

Symbol Digit Modalities Test—Oral Version (SDMT)

The SDMT (Smith 1973) assesses processing speed. Par-

ticipants are provided with 90 s to verbalise to the exper-

imenter the correct number that belongs under a specific

symbol, based on the way they are paired in an index at the

top of the page. Total number of correct entries was used in

the current study.

Neuroimaging Acquisition

Neuroimaging was performed on a Siemens Magnetom

Verio 3 Tesla MRI scanner (Siemens Medical Imaging,

Erlangen, Germany). After a scout scan, an anterior

Brain Topogr

123

commissure posterior commissure (AC-PC) aligned axial

fluid attenuated inversion recovery (FLAIR) sequence was

acquired (repetition time = 8,500, echo time = 81, slice

thickness = 3.0 mm, field of view = 240 mm, matrix =

192 9 256, in-plane = 0.94 9 0.94 mm), followed by a

DTI sequence (TR = 5,500, TE = 110, 20 diffusion

encoding directions, number of excitations = 2, slice

thickness = 4.0 mm, field of view = 240 mm, matrix =

122 9 122, in-plane = 1.97 9 1.97 mm, b value = 1,000

s/mm2, one acquisition for each run with b = 0 s/mm2). In

addition, a 3D T1-weighted sequence was acquired in the

sagittal orientation [inversion time = 1,100 ms, repetition

Table 1 Descriptive statistics for and individuals with TBI, categorised by injury severity, and controls

Total TBI sample

(n = 68)

Mild TBI

(n = 10)

Moderate TBI

(n = 14)

Severe TBI

(n = 44)

TBI cognitive subsample

(n = 36)

Control

(n = 25)

N (%) N (%) N (%) N (%) N (%) N (%)

Gender (male) 56 (81.2) 7 (70) 12 (85.7) 37 (84.1) 29 (80.5) 11 (44)

Cause of injury

Car accident 35 (51.5) 4 (28.6) 4 (28.6) 27 (61.4) 17 (47.2)

Motorcycle 12 (17.6) 1 (10) 2 (14.3) 9 (20.5) 8 (22.2)

Pedestrian 6 (8.8) 2 (20.0) 2 (14.3) 2 (4.5) 2 (5.6)

Bicycle 6 (8.8) 1 (10) 4 (28.6) 1 (12.3) 5 (13.9)

Fall 6 (8.8) 1 (10) 1 (7.1) 4 (9.1) 4 (11.1)

Other 3 (4.5) 1 (10) 1 (7.1) 2 (2.3) 0 (0)

PTA categories

Mild (0–1 day) 10 (14.7) 4 (11.1)

Moderate

([1 to B7 days)

14 (20.6) 8 (22.2)

Severe ([7 days) 44 (64.7) 24 (66.7)

Worst GCS categories

Mild (13–15) 20 (31.7) 9 (27.3)

Moderate (9–12) 14 (22.2) 8 (24.2)

Severe (3–8) 29 (46.0) 16 (48.5)

Abnormal CT 57 (83.8) 10 (100) 8 (57.1) 39 (88.6) 29 (80.6)

Scan

Contusion 27 (39.7) 6 (60) 1 (7.1) 20 (45.5) 13 (36.1)

DAI 11 (16.2) 0 (0) 1 (7.1) 10 (22.7) 8 (22.2)

SAH 23 (33.8) 4 (40) 2 (14.3) 17 (38.6) 10 (27.8)

SDH 14 (20.6) 2 (20.0) 3 (21.4) 9 (20.5) 5 (13.9)

EDH 5 (7.4) 2 (20.0) 2 (14.3) 1 (2.3) 3 (8.3)

ICH 14 (20.6) 3 (30.0) 2 (14.3) 9 (20.5) 8 (22.2)

IVH 0 (0) 0 (0) 0 (0) 9 (20.5) 1 (2.8)

Mdn, M (SD)

range

Mdn, M (SD)

range

Mdn, M (SD)

range

Mdn, M (SD)

range

Mdn, M (SD)

range

Mdn, M (SD)

range

Age 35.41, 38.61 (14.64)

17–78

40.12, 39.63 (14.52)

18–62

33.55, 37.69 (14.06)

23–65

35.07, 38.67 (15.15)

17–78

41.76, 49.92 (13.36)

17–69

25.95, 30.26 (12.48)

18–69

Education 12.00, 12.36 (2.82)

6–20

12.00, 12.40 (2.68)

10–18

12.00, 12.69 (2.70)

8–18

12.00, 12.25 (2.94)

6–20

12.00, 12.94 (3.16)

6–20

15,00, 14.68 (1.51)

12–17

PTA (days) 12.00, 17.56 (19.29)

.1–105

10.50, 15.07

(15.45)

.1–82

Worst GCS 9.00, 8.76

3–15

10.00, 8.58

3–15

PTA post traumatic amnesia, GCS glasgow coma score, DAI diffuse axonal injury, SAH subarachnoid haemorrhage, SDH subdural haemorrhage, EDH

epidural haemorrhage, ICH intracerebral haemorrhage, IVH intraventricular haemorrhage

Brain Topogr

123

time = 1,930 ms echo time = 2.48 ms, resolution = 256

9 256 9 160, flip angle = 12�, field of view = 256 mm,

slice thickness = 1.00 mm (160 slices)].

Data Analysis

Tract-Based Spatial Statistics

The FMRIB Software Library (Smith et al. 2004) was used to

perform TBSS. Analyses were undertaken using the steps

outlined by Smith et al. (2006). Using both linear and non-linear

alignment, participants’ FA images were registered into stan-

dard space using the FMRIB58 FA template. The aligned

images were then averaged to reveal the white matter tracts

common to all participants; based on this image an FA skeleton

was created representing the centre of all white matter tracts. An

FA value of .2 was used as a threshold for the FA skeleton to

exclude tracts with high inter-individual variability as well as

those containing a high level of partial volume. Each partici-

pant’s aligned FA image was projected onto the FA skeleton to

correct for residual misalignments by calculating the difference

between the skeletonised tracts and the white matter tract

centres in each individual image. Voxelwise statistics were then

undertaken using the general linear model to compare differ-

ences in white matter FA between controls and TBI participants

and to regress the performance of each cognitive test on the FA

values of white matter tracts. The ‘randomise’ tool was used to

conduct significance testing, applying a threshold-free cluster

enhancement (Smith and Nichols, 2009) with 5,000 permuta-

tions, before a threshold of p\ .05 was applied. Comparisons

between TBI and control participants were controlled for age,

education, and gender, whereas analyses correlating FA values

with neuropsychological performance in the TBI group

covaried for gender, age, years of education, and time post injury.

Region-of-Interest Tractography

Diffusion images were reconstructed and directionally

encoded colour FA maps calculated using DTI studio (Jiang

et al. 2006). Eight ROIs were manually chosen a priori based

on previous studies: the corpus callosum (CC), left and right

superior longitudinal fasciculus (SLF), inferior longitudinal

fasciculus (ILF), posterior limb of the internal capsule

(PLIC), and the cingulate bundle. Representation of these

tracts is presented in Fig. 1. As shown in Fig. 2, ellipses

were placed on seed location in order to track each ROI

individually. Depending on the specific tract, ellipses were

placed when viewing the directionally encoded anisotropy

maps from sagittal, axial, or coronal orientations. As

explained by Mori and van Zijl (2002) and Mori et al.

(2005), tract reconstruction was performed using fiber

assignment by continuous tracking (FACT) method. Tracts

were linearly propagated based on the orientation of the

largest principal axis. Line propagation was discontinued

based on the standard criteria: that is, using the standard

anisotropy threshold of FA [.2 and an angle of change[75�between pixels. Lastly, the white matter atlas provided by

Mori et al. (2005) was used to assess the accuracy of tracked

ROIs. Average FA values for each of the a priori chosen

tracts were derived for each participant. Independent-sam-

ples t tests were undertaken to compare average FA of each

tract between control and TBI participants controlling for

age, education, and gender. To control for Type I error rate,

a Bonferroni correction was applied to the t tests performed

for each tract, such that the Bonferroni correction value was

now .008. For TBI participants only, Pearson correlations

were conducted, individually for each tract, to examine the

relationship between tract integrity and neuropsychological

performance. Correlation analyses controlled for gender,

age, level of education, and time post injury.

Results

Fractional Anisotropy is Associated with Injury

Severity

Both ROI tractography and TBSS approaches detected

differences in FA between control and TBI participants. As

Fig. 1 Representation of fibers assessed in DTIstudio in a sagittal, b coronal, and c axial planes. Red corpus callosum, Yellow cingulate bundles,

Brown superior longitudinal fasciuli, Green inferior longitudinal fasciculi, Blue internal capsules (Color figure online)

Brain Topogr

123

Fig. 3 shows, several of the ROIs differentiated signifi-

cantly between control and TBI participants. Moderate and

severe TBI participants displayed significantly lower CC

FA compared to controls and those with mild TBI. Severe

TBI participants were the only TBI group to display a

reduction in cingulate bundle FA, compared to control and

moderate TBI participants. Similarly, severe TBI partici-

pants were the only group to show reduced FA in the left

and right SLF compared to controls and mild TBI partici-

pants. However, this difference was not significant when

the Bonferroni-corrected p value was applied. Participants

with moderate and severe TBI displayed lower FA in the

left ILF, whereas in the right ILF severe TBI participants

displayed FA reductions compared to all other groups.

Moderate and severe TBI participants displayed reduced

FA in the left and right internal capsules compared to

controls and those with mild TBI. Although individuals

with mild TBI displayed a tendency to have lower FA

compared to controls, such as in the cingulate bundle and

internal capsules, none of the differences were statistically

significant.

The results of TBSS display similar patterns of white

matter injury compared to ROI tractography. That is, no

significant differences in FA were found between control

and mild TBI participants, whereas moderate and severely

injured TBI participants displayed significant widespread

reductions in FA. The reduction in FA as injury severity

increased can be seen in panel A of Fig. 4, wherein PTA

treated continuously is examined. That is, with greater

days of PTA reduced FA is present almost throughout the

entire brain. As displayed in panels C and D of Fig. 4,

compared to controls, individuals with moderate and

severe TBI displayed reduced FA in the CC, ILF, SLF,

cingulate bundle, as well as the PLIC. However, the TBSS

results show that reductions in FA may be present in only

specific parts of tracts, and may not be found over the

entire white matter bundle. Severe TBI participants were

found to have the most extensive reductions in FA, with

significant differences found compared to control, mild,

as well as moderate groups. However, as can be noted in

Fig. 4, the differences between moderate and severe TBI

groups was not as extensive as when comparing severe

TBI participants with controls or indivduals with mild

TBI (Fig. 5).

Neuropsychological Performance

Comparisons were made between TBI and control partic-

ipants, co-varying for level of education. After controlling

for education, participants with TBI displayed significantly

Fig. 2 Initial seed point

locations used to track fibers in

DTIStudio. a Initial seed point

for the posterior limb of the

internal capsule. b Initial seed

point for the corpus callosum.

c Initial seed point for the

inferior longitudinal fasciculus.

d Initial seed point for the

superior longitudinal fasciculus

(brown) and the cingulum

bundle (yellow) (Color figure

online)

Brain Topogr

123

poorer performance in total recall, and the SDMT. Table 2

displays the results of these comparisons.

Neuropsychological Performance is Related

to Fractional Anisotropy

Several associations were found between neuropsycho-

logical measures and FA, controlling for age, years of

education, time post injury, and gender. Using ROI trac-

tography, the COWAT verbal fluency was found to be

positively correlated with FA in the left SLF (r = .35,

p \ .05). Worse performance on the Hayling sentence

completion test part B was correlated with lower FA in the

left (r = -.39, p \ .05) and right (r = -.38, p \ .05)

SLF. Greater Hayling overall scaled score was significantly

correlated with higher FA in the left (r = .52, p \ .01) and

right (r = .50, p \ .01) SLF. Poorer performance on the

Trails B was related to lower FA in the right internal

capsule (r = -.41, p \ .05). Lastly, worse performance on

the SDMT correlated with lower FA in the CC (r = .37,

p \ .05) right SLF (r = .36, p \ .05), and the right ILF

(r = .35, p \ .05).

TBSS analyses examining the relationships between

cognition and FA revealed some similarities to those

obtained using the ROI approach. Lower Hayling overall

scaled score was related to reduced FA in parts of SLF.

Fig. 3 Injury group

comparisons using region-

of-interest tractography. Graphsdisplay scatter dot plots of FA

values for control, mild,

moderate, and severe groups.

p values have been derived from

ANCOVA, controlling for age,

gender, and years of education.

All p values below the .05

threshold were included;

however, only the p values

denoted with a remained

significant at the Bonferroni

corrected level of p \ .008. The

overlayed horizontal line plots

represent the mean and standard

error above and below the mean

Brain Topogr

123

However the Hayling overall scaled score was also related

to FA in the CC, ILF, and cingulum, and was almost

exclusively restricted to the left hemisphere. Moreover, the

SDMT displayed the strongest relationship with the ROI

tractography, which was also found to be the case in the

TBSS analyses. Similar to the Hayling overall scaled score,

performance on the SDMT was found to correlate with

FA in a more extensive network of tracts. These included

the left and right ILF, left and right SLF, CC, as well as the

cingulate bundle. Lastly, worse performance on the

DOORS visual memory test was related to FA in the TBSS

analyses, but not in the ROI analyses. The relationship

between the DOORS test and FA was similar to the Hay-

ling overall scaled score, being restricted to the left

hemisphere and being associated with sections of the ILF,

SLF, and CC.

Discussion

This study had three primary aims: to examine whether FA

of white matter tracts following TBI was associated with

the severity of the injury—defined using days of PTA,

whether cognitive performance was associated white mat-

ter FA, and lastly to investigate the comparability of ROI

and TBSS approaches. Overall, the findings indicated that

significantly lower FA is found in several white matter

tracts following moderate and severe TBI. Individuals with

mild injuries did not show significantly lower FA, com-

pared to controls, although trends were apparent in certain

cases and the lack of statistical significance may be due to

the small size of the mild TBI subsample in the current

study. Information processing and executive cognitive

abilities were found to be associated with several

Fig. 4 Tract-based spatial

statistics showing injury

severity group comparisons.

a Correlation analysis showing

the association between FA and

PTA measured continuously.

b Group comparisons between

controls and the TBI group as a

whole. c Group comparisons

between control and moderate

TBI participants. d Group

comparisons between control

and severe TBI participants.

e Group comparisons between

mild and severe TBI

participants. f Group

comparisons between moderate

and severe TBI participants.

Tract-based spatial statistics

were thresholded at p \ .05.

Higher intensity colours

represent lower p values. Axial

z coordinates are the same for

all slices and are represented

above the top comparison.

Tracts have been ‘thickened’ to

enhance visualisation of results

Brain Topogr

123

prominent white matter tracts. Interestingly, visual memory

was also associated with FA of several white matter tracts.

For the first time this study has compared ROI tractography

and whole-brain TBSS approaches, and results suggests

both similar and divergent patterns.

Demonstrated using both ROI and TBSS approaches,

the present findings showed that FA decreases with greater

injury severity. That is, all regions chosen a priori for

tractography analyses displayed reduced FA in individuals

with severe TBI compared to healthy controls. Although

differences were not as pronounced as in those with severe

injuries, participants with moderate TBI displayed lower

FA in all a priori tracked regions except for the SLF and

cingulate bundle. The TBSS analyses largely support the

ROI tractography approach. Moderate and severe TBI

participants displayed reduced FA compared to controls,

while individuals with severe TBI displayed the most

reduced FA compared to controls, mild, or moderate TBI

participants. Nevertheless, the TBSS results also highlight

that parts of white matter tracts—rather than the entire

bundle—may display reduced FA, which may go unnoticed

when conducting ROI tractography; for example, only

certain parts of the CC differed when comparing control

and moderate TBI participants, or when comparing mod-

erate and severe TBI participants.

Although a trend of lower FA was found between the

mild TBI group and healthy controls, no significant dif-

ferences were found, contrary to several previous studies

(Lipton et al. 2008; Lo et al. 2009; Mayer et al. 2010; Niogi

et al. 2008a, b). This may have been due to the small

sample of individuals with mild TBI which may have

resulted in low statistical power. This inconsistency in

study findings may also be due to differences in the time at

which participants were scanned, and in participant

recruitment criteria. Miles et al. (2008) scanned partici-

pants between one and 10 days post-injury, whereas Mayer

et al. (2010) scanned participants within 21 days following

injury. Since the greatest change to axonal structure and

function occurs in the immediate days following the injury,

scans of acute changes may more readily detect pathology.

Axonal pathology may be absent at follow-up, at least for

mild injuries due to axonal reorganisation (Bigler 2001;

Povlishock and Jenkins 1995; Povlishock and Katz 2005).

Indeed, when Miles et al. (2008) followed up participants

at 6 months post-injury, differences were no longer found

between healthy controls and individuals with mild TBI in

FA measures. Moreover, variability in the spatial distri-

bution of injury across studies, as well as between injury

severity groups, may in part explain these non-significant

group findings.

Furthermore, Mayer et al. (2010), Niogi et al. (2008a, b),

and Lipton et al. (2008) only included individuals that had

reported some form of post-concussive symptoms, includ-

ing cognitive difficulties. Since it has been widely noted

that cognition recovers for most of those with mild TBI

within the first 3 months after injury, it may be that the

Fig. 5 Tract-based spatial statistics showing areas of significant

correlation between FA and neuropsychological performance.

a Tract-based spatial statistics showing significant correlations

between FA and performance on the DOORS Test. b Tract-based

spatial statistics showing significant correlations between FA and

performance on the SDMT. c Tract-based spatial statistics showing

significant correlations between FA and performance on the Hayling

sentence completion test. Tract-based spatial statistics were threshol-

ded at p \ .05. Higher intensity colours represent lower p values.

Axial z coordinates are represented above each correlation. Tracts

have been ‘thickened’ to enhance visualisation of results

Brain Topogr

123

samples involved in these studies reflected the subgroup of

individuals that continue to show impairments (Dikmen

et al. 1995; Schretlen and Shapiro 2003). Indeed, Messe

et al. (2011) found that only those individuals with mild

TBI with residual neurobehavioural complaints demon-

strated poorer white matter integrity compared to healthy

controls. Moreover, recent studies suggest that standard

neuropsychological measures may not be sufficiently sen-

sitive to detect the subtle changes in FA following mild

TBI. For example, in their recent study, Geary et al. (2010)

found that deficits in learning strategy were associated with

reductions in FA following chronic mild TBI.

The relationship between white matter integrity and

cognitive abilities is further strengthened by the findings of

the current study, which show that prominent tracts

including the ILF, internal capsule, SLF, and the CC are

asssociated with information processing and executive

functions. However, memory was largely unrelated to the

integrity of white matter, with the exception of visual

memory, which showed an association with the left ILF,

SLF, and CC in the TBSS analyses. These findings differ

from other studies, which have found an association

between memory and the integrity of several white matter

tracts. For example Bigler et al. (2010) found the visual

memory was associated with FA in the right ILF and left

inferior fronto-occipital fasciculi (IFOF) whereas Kraus

et al. (2007) found that memory was related to FA in the

CC, forceps, sagittal stratum and IFOF.

However, it is also not surprising that memory was not

strongly associated with white matter integrity since tracts

primarily known to modulate memory, such as the fornices

and thalamic nuclei, were not examined in this study

(Aggleton 2008; Aggleton and Brown 1999; Metzler-

Baddeley et al. 2011). Even if these regions had been

considered a priori, the streamline tractography used in the

current study may not have been able to adequately

reconstruct these fibers, unlike what has been shown using

probabilistic tractography methods (Behrens et al. 2003;

Johansen-Berg et al. 2005). In addition, due to the high

level of uncertainty regarding fibre direction, and the dif-

ficulty in separating white and grey matter boundaries in

these regions, it is likely that these tracts were excluded as

part of the TBSS analyses in order to avoid difficulties in

interpretation (Smith et al. 2007). Imaging parameters such

as those used by Kinnunen et al. (2011), in which diffu-

sion-weighted data is collected in a greater number of

directions and using significantly reduced slice thickness

may need to be utilised to more reliably explore these

pathways.

Nevertheless, the significant associations between the

integrity of white matter tracts, information processing, and

executive functions is largely consistent with previous

research in this area. The finding that performance on the

COWAT, Hayling test, and SDMT was related to associ-

ation tracts—connecting posterior to anterior brain

regions—as well as the CC and frontal tracts is not

Table 2 Cognitive test performance comparisons between TBI and control participants

Measure TBI total cognitive

subsample

N = 36

TBI mild

N = 4

TBI moderate

N = 8

TBI severe

N = 24

Control

N = 25

Total TBI

vs control

Mdn, M (SD) Mdn, M (SD) Mdn, M (SD) Mdn, M (SD) Mdn, M (SD) t

Total recall 44.50, 43.72 (10.16) 44.00, 40.25 (14.15) 47.00, 48.12 (9.00) 45.00, 42.17 (9.85) 56.00, 55.36 (8.12) 4.01***

Visual memory 17.00, 16.61 (3.92) 18.00, 18.25 (3.30) 16.00, 17.00 (4.90) 17.00, 16.21 (3.73) 19.00, 19.00 (2.53) 1.87

Hayling—Time

A

6.00, 12.52 (14.06) 8.00, 8.25 (3.30) 7.50, 12.00 (11.46) 6.00, 13.45 (16.05) 6.00, 7.54 (5.18) -0.42

Hayling—Time B 23.00, 32.45 (29.55) 22.00, 20.50 (14.82) 14.50, 20.43 (20.25) 25.00, 38.45 (32.69) 18.50, 19.17 (11.48) -1.49

Hayling—scaled 6.00, 5.67 (1.49) 6.00, 5.75 (1.26) 6.00, 6.00 (1.53) 6.00, 5.55 (1.57) 6.00, 6.13 (.61) -.01

COWAT 36.50, 36.50 (13.32) 39.00, 39.50 (3.00) 36.50, 43.25 (17.08) 36.00, 33.75 (12.45) 43.50, 42.52 (9.33) 1.01

Trails A time 25.38, 28.86 (10.49) 29.92, 33.03 (11.41) 21.28, 24.58 (10.02) 27.52, 29.46 (10.52) 21.71, 23.47 (9.00) -1.01

Trails B time 60.34, 71.18 (32.92) 63.54, 74.19 (45.02) 60.34, 74.58 (53.04) 75.10, 69.48 (22.86) 53.04, 55.99 (23.74) -.97

Trails B–Trails A 36.00, 43.17 (28.33) 27.08, 41.16 (41.13) 39.60, 49.99 (43.50) 36.67, 41.27 (20.05) 29.05, 32.53 (18.45) -.72

Digit span:

forward

10.00, 10.23 (2.93) 9.00, 9.00 (.82) 8.00, 10.00 (3.82) 11.00, 10.52 (2.86) 11.00, 11.24 (2.45) .92

Digit span:

backwards

8.00, 8.23 (2.94) 7.00, 7.50 (3.00) 7.50, 9.00 (4.07) 8.00, 8.09 (2.56) 9.00, 9.68 (2.21) 1.13

SDMT 50.00, 50.67 (12.38) 55.00, 53.50 (7.72) 51.00, 55.75 (18.16) 50.00, 48.50 (10.49) 63.00, 62.84 (9.21) 3.15**

All comparisons between TBI and control participants were controlled for years of education

SDMT symbol digit modalities test, Hayling Hayling overall scaled score

* p \ .05; ** p \ .01; *** p \ .001

Brain Topogr

123

surprising given that these measures require the integration

of information across various modalities, speed of infor-

mation processing as well as executive functions. That is,

previous studies have highlighted the use of the CC—

especially the genu— as well as the association tracts that

connect occipital to frontal regions—including the SLF,

ILF— in modulating speed of information processing in

the general population (Bucur et al. 2008; Fling et al. 2011;

Konrad et al. 2009; Turken et al. 2008). Information is

initially received in the occipital visual areas before being

transmitted to frontal regions via the dorsal and ventral

streams (Foxe and Simpson 2002).

As a whole, the results are promising regarding the

comparability of ROI and TBSS approaches, both in terms

of their relationships with injury severity and cognitive

functioning. That is, both techniques delineated between the

severity of TBI, using PTA, whereby greater reductions in

FA were found with worse TBI severity. In addition, both

techniques indicated associations between FA, the SDMT

and Hayling test. However, there were also some notable

differences between the ROI and TBSS results. Namely,

whereas the ROI tractography was limited to the small

number of a priori selected regions, TBSS analysis offered

greater breadth of analysis. This difference was demon-

strated in the greater number of tracts associated with injury

severity and cognition in the TBSS analyses, compared to

the ROI approach. This was especially noticeable for the

associations with cognition. Whereas, only one or two ROIs

were related to cognitive performance for each measure,

TBSS analysis detected a significantly greater number of

associated tracts. A second difference is that TBSS is

capable of detecting differences anywhere along a tract, or

for only a section of the tract. This is especially critical for

the exploration of cognitive functions, as only sections of

larger tracts may be of interest. Due to the aforementioned

differences between the two techniques, it is suggested that

examination of white matter pathology may be applied in a

two-stage process. As occurred with the current findings, it

is likely that more frequently ROI tractography—when

restricted to whole white matter bundles—underestimates

the extensiveness of FA reductions following TBI. More-

over, specificity within white matter tracts is missing,

whereby only parts of the tracts may display reduced FA.

Whole-brain TBSS may be applied initially as a screening

measure with the intent of highlighting all significant

regions where white matter changes are found. A second

step may then be taken to explore these significant regions in

more depth using quantitative tractography. That is, TBSS

only captures significant change to the centres of white

matter tracts. Visualising regions using tractography pro-

vides more information, both quantitative and qualitative,

regarding the changes that have taken place, including any

disconnection between tracts and changes in structure.

The current study has some limitations regarding the

imaging protocol and methods of analysis. Firstly, due to

the clinical setting in which this study took place it was not

feasible to scan participants for an extended length of time,

which would have ensured diffusion-weighted acquisition

in a greater number of directions and at higher resolution.

To circumvent this limitation, larger white matter tracts

were the focus of study. However, as already discussed,

other tracts, located in more complex regions of the brain

but nevertheless related to function, were not explored. In

addition, the current study focused solely on FA as an

indicator of white matter integrity. Other indices, including

mean, radial, and axial diffusivity may have also revealed

associations of interest with injury severity and cognitive

functioning. The limitation inherent to the DTI method,

namely the difficulty dealing with crossing fibers, is of

concern especially given the DTI resolution used in the

current study. Additionally, although it is inferred in the

current study that reduced FA was specifically indicative of

abnormal connectivity, other changes following TBI may

also have an impact on FA. For example, it is not known to

what extent atrophic or neuroinflammatory changes or

increased extracellular space impacts on FA.

In conclusion, the current study showed significantly

reduced FA following moderate and severe TBI. Informa-

tion processing, executive abilities, and visual memory

were found to be related to FA of white matter tracts con-

necting the posterior to anterior cortices. For the first time,

both ROI tractography and TBSS techniques were used in a

TBI sample to examine how they pertain to injury severity

and cognitive functioning. The results suggested that using

these techniques in combination may derive better results

regarding detection of pathology following TBI.

Acknowledgments We thank all the staff of Healthcare Imaging

MRI facility at the Epworth Hospital, Melbourne, Victoria.

Conflict of interest The authors declare that they have no com-

peting financial interests.

References

Aggleton JP (2008) Understanding anterograde amnesia: disconnec-

tions and hidden lesions. Q J Exp Psychol 61:1441–1471

Aggleton JP, Brown MW (1999) Episodic memory, amnesia and

the hippocampal-anterior thalamic axis. Behav Brain Sci 22:

425–444

Arfanakis K, Haughton VM, Carew JD, Rogers BP, Dempsey RJ,

Meyerand ME (2002) Diffusion tensor mr imaging in diffuse

axonal injury. Am J Neuroradiol 23:794–802

Ashburner J, Friston KJ (2000) Voxel-based morphometry—the

methods. Neuroimage 11:805–821

Baddeley A, Emslie H, Nimmo-Smith I (1994) Doors and people.

Thames Valley Test Company, Bury St Edmunds

Brain Topogr

123

Basser PJ (1995) Inferring microstructural features and the physio-

logical state of tissues from diffusion-weighted images. NMR

Biomed 8:333–344

Basser PJ, Mattiello J, Lebihan D (1994) Mr diffusion tensor

spectroscopy and imaging. Biophys J 66:259–267

Behrens TEJ, Johansen-Berg H, Woolrich MW, Smith SM, Wheeler-

Kingshott CAM, Boulby PA et al (2003) Non-invasive mapping

of connections between human thalamus and cortex using

diffusion imaging. Nat Neurosci 6:750–757

Benson RR, Meda SA, Vasudevan S, Kou ZF, Govindarajan KA,

Hanks RA et al (2007) Global white matter analysis of diffusion

tensor images is predictive of injury severity in traumatic brain

injury. J Neurotrauma 24:446–459

Benton AL, Hamsher K, Rey GJ (1994) Multilingual aphasia

examination, 3rd edn. AJA Associates, Iowa Cita

Bigler ED (2001) The lesion(s) in traumatic brain injury: implica-

tions for clinical neuropsychology. Arch Clin Neuropsychol 16:

95–131

Bigler ED, McCauley SR, Wu TC, Yallampalli R, Shah S, MacLeod

M et al (2010) The temporal stem in traumatic brain injury:

preliminary findings. Brain Imaging Behav 4:270–282

Bookstein FL (2001) ‘‘Voxel-based morphometry’’ should not be used

with imperfectly registered images. Neuroimage 14:1454–1462

Bucur B, Madden DJ, Spaniol J, Provenzale JM, Cabeza R, White LE

et al (2008) Age-related slowing of memory retrieval: contribu-

tions of perceptual speed and cerebral white matter integrity.

Neurobiol Aging 29:1070–1079

Buki A, Povlishock JT (2006) All roads lead to disconnection?

Traumatic axonal injury revisited. Acta Neurochir 148:181–193

Burgess PW, Shallice T (1997) The Hayling and Brixton tests.

Thames Valley Test Company, Bury St Edmunds

Coughhlan AK, Oddy M, Crawford JR (2007) Birt memory and

information processing battery London. Brain injury rehabilita-

tion trust, Wakefield

Davatzikos C (2004) Why voxel-based morphometric analysis should

be used with great caution when characterizing group differ-

ences. Neuroimage 23:17–20

Dikmen SS, Machamer JE, Winn HR, Temkin NR (1995) Neuro-

psychological outcome at 1-year post head injury. Neuropsy-

chology 9:80–90

Draper K, Ponsford J (2008) Cognitive functioning ten years

following traumatic brain injury and rehabilitation. Neuropsy-

chology 22:618–625

Fling BW, Chapekis M, Reuter-Lorenz PA, Anguera J, Bo J, Langan J

et al (2011) Age differences in callosal contributions to cognitive

processes. Neuropsychologia 49:2564–2569

Foxe JJ, Simpson GV (2002) Flow of activation from v1 to frontal

cortex in humans: a framework for defining ‘‘early’’ visual

processing. Exp Brain Res 142:139–150

Geary EK, Kraus MF, Pliskin NH, Little DM (2010) Verbal learning

differences in chronic mild traumatic brain injury. J Int Neuro-

psychol Soc 16:506–516

Gennarelli TA, Thibault LE, Graham DI (1998) Diffuse axonal injury:

an important form of traumatic brain damage. Neuroscientist

4:202–215

Huisman TAGM, Sorensen AG, Hergan K, Gonzalez RG, Schaefer

PW (2003) Diffusion-weighted imaging for the evaluation of

diffuse axonal injury in closed head injury. J Comput Assist

Tomogr 27:5–11

Jiang H, van Zijl PCM, Kim J, Pearlson GD, Mori S (2006) Dtistudio:

resource program for diffusion tensor computation and fiber

bundle tracking. Comput Methods Programs Biomed 81:

106–116

Johansen-Berg H, Behrens TEJ, Sillery E, Ciccarelli O, Thompson

AJ, Smith SM et al (2005) Functional–anatomical validation and

individual variation of diffusion tractography-based segmenta-

tion of the human thalamus. Cereb Cortex 15:31–39

Jones DK, Griffin LD, Alexander DC, Catani M, Horsfield MA,

Howard R et al (2002) Spatial normalization and averaging of

diffusion tensor MRI data sets. Neuroimage 17:592–617

Jones DK, Symms MR, Cercignani M, Howard RJ (2005) The effect

of filter size on VBM analyses of DT-MRI data. Neuroimage

26:546–554

Kennedy MR, Wozniak JR, Muetzel RL, Mueller BA, Chiou HH,

Pantekoek K et al (2009) White matter and neurocognitive

changes in adults with chronic traumatic brain injury. J Int

Neuropsychol Soc 15:130–136

Kinnunen KM, Greenwood R, Powell JH, Leech R, Hawkins PC,

Bonnelle V et al (2011) White matter damage and cognitive

impairment after traumatic brain injury. Brain 134:449–463

Konrad A, Vucurevic G, Musso F, Stoeter P, Winterer G (2009)

Correlation of brain white matter diffusion anisotropy and mean

diffusivity with reaction time in an oddball task. Neuropsycho-

biology 60:55–66

Kraus MF, Susmaras T, Caughlin BP, Walker CJ, Sweeney JA, Little

DM (2007) White matter integrity and cognition in chronic

traumatic brain injury: a diffusion tensor imaging study. Brain

130:2508–2519

Kumar R, Husain M, Gupta RK, Hasan KM, Haris M, Agarwal AK

et al (2009) Serial changes in the white matter diffusion tensor

imaging metrics in moderate traumatic brain injury and corre-

lation with neuro-cognitive function. J Neurotrauma 26:481–495

Levin HS (2003) Neuroplasticity following non-penetrating traumatic

brain injury. Brain Inj 17:665–674

Levin HS, Wilde E, Troyanskaya M, Petersen NJ, Scheibel R,

Newsome M et al (2010) Diffusion tensor imaging of mild to

moderate blast-related traumatic brain injury and its sequelae.

J Neurotrauma 27:683–694

Lipton ML, Gellella E, Lo C, Gold T, Ardekani BA, Shifteh K et al

(2008) Multifocal white matter ultrastructural abnormalities in

mild traumatic brain injury with cognitive disability: a voxel-

wise analysis of diffusion tensor imaging. J Neurotrauma

25:1335–1342

Little DM, Kraus MF, Joseph J, Geary EK, Susmaras T, Zhou XJ et al

(2010) Thalamic integrity underlies executive dysfunction in

traumatic brain injury. Neurology 74:558–564

Lo C, Shifteh K, Gold T, Bello JA, Lipton ML (2009) Diffusion

tensor imaging abnormalities in patients with mild traumatic

brain injury and neurocognitive impairment. J Comput Assist

Tomogr 33:293–297

Mayer AR, Ling J, Mannell MV, Gasparovic C, Phillips JP, Doezema

D et al (2010) A prospective diffusion tensor imaging study in

mild traumatic brain injury. Neurology 74:643–650

Messe A, Caplain S, Paradot G, Garrigue D, Mineo J-F, Ares GS et al

(2011) Diffusion tensor imaging and white matter lesions at the

subacute stage in mild traumatic brain injury with persistent

neurobehavioral impairment. Hum Brain Mapp 32:999–1011

Metzler-Baddeley C, Jones DK, Belaroussi B, Aggleton JP, O’Sul-

livan MJ (2011) Frontotemporal connections in episodic memory

and aging: a diffusion MRI tractography study. J Neurosci

31:13236–13245

Miles L, Grossman RI, Johnson G, Babb JS, Diller L, Inglese M

(2008) Short-term DTI predictors of cognitive dysfunction in

mild traumatic brain injury. Brain Inj 22:115–122

Mori S, van Zijl PCM (2002) Fiber tracking: principles and strategies:

a technical review. NMR Biomed 15:468–480

Mori S, Wakana S, Van Zijl PCM (2005) MRI atlas of human white

matter. Elsevier, Amsterdam

Nakayama N, Okumura A, Shinoda J, Yasokawa YT, Miwa K,

Yoshimura SI et al (2006) Evidence for white matter disruption

Brain Topogr

123

in traumatic brain injury without macroscopic lesions. J Neurol

Neurosurg Psychiatry 77:6

Newcombe VFJ, Williams GB, Nortje J, Bradley PG, Harding SG,

Smielewski P et al (2007) Analysis of acute traumatic axonal

injury using diffusion tensor imaging. Br J Neurosurg

21:340–348

Niogi SN, Mukherjee P, Ghajar J, Johnson C, Kolster RA, Sarkar R

et al (2008a) Extent of microstructural white matter injury in

postconcussive syndrome correlates with impaired cognitive

reaction time: a 3t diffusion tensor imaging study of mild

traumatic brain injury. Am J Neuroradiol 29:967–973

Niogi SN, Mukherjee P, Ghajar J, Johnson CE, Kolster R, Lee H et al

(2008b) Structural dissociation of attentional control and mem-

ory in adults with and without mild traumatic brain injury. Brain

131:3209–3221

Park H-J, Kubicki M, Shenton ME, Guimond A, McCarley RW,

Maier SE et al (2003) Spatial normalization of diffusion tensor

MRI using multiple channels. Neuroimage 20:1995–2009

Ponsford J, Olver JH, Curran C (1995a) A profile of outcome: 2 years

after traumatic brain injury. Brain Inj 9:1–10

Ponsford J, Olver JH, Ng K (1995b) Prediction of employment status

2 years after traumatic brain injury. Brain Inj 9:11–20

Povlishock JT, Christman CW (1994) The pathobiology of traumat-

ically induced axonal injury in animals and humans: a review of

current thoughts. US Government Printing Office, Washington,

Dc

Povlishock JT, Jenkins LW (1995) Are the pathobiological changes

evoked by traumatic brain injury immediate and irreversible?

Brain Pathol 5:415–426

Povlishock JT, Katz DI (2005) Update of neuropathology and

neurological recovery after traumatic brain injury. J Head

Trauma Rehabil 20:76–94

Reitan RM, Wolfson D (1988) The Halstead–Reitan neuropsycho-

logical test battery. Neuropsychology Press, Tucson

Rugg-Gunn FJ, Symms MR, Barker GJ, Greenwood R, Duncan JS

(2001) Diffusion imaging shows abnormalities after blunt head

trauma when conventional magnetic resonance imaging normal.

J Neurol Neurosurg Psychiatry 70:530–533

Salmond CH, Menon DK, Chatfield DA, Williams GB, Pena A,

Sahakian BJ et al (2006) Diffusion tensor imaging in chronic

head injury survivors: correlations with learning and memory

indices. Neuroimage 29:117–124

Scheid R, Preul C, Gruber O, Wiggins C, von Cramon DY (2003)

Diffuse axonal injury associated with chronic traumatic brain

injury: evidence from t2*-weighted gradient-echo imaging at 3 t.

Am J Neuroradiol 24:1049–1056

Schretlen DJ, Shapiro AM (2003) A quantitative review of the effects

of traumatic brain injury on cognitive functioning. Int Rev Psych

15:341–349

Shenton ME, Hamoda HM, Schneiderman JS, Bouix S, Paternak O,

Rathi Y et al (2012) A review of magnetic resonance imaging

and diffusion tensor imaging findings in mild traumatic brain

injury. Brain Imaging Behav 6:137–192

Sidaros A, Engberg AW, Sidaros K, Liptrot MG, Herning M, Petersen

P et al (2008) Diffusion tensor imaging during recovery from

severe traumatic brain injury and relation to clinical outcome: a

longitudinal study. Brain 131:559–572

Smith A (1973) Symbol digit modalities test. Western Psychological

Services, Los Angeles

Smith SM, Nichols TE (2009) Threshold-free cluster enhancement:

addressing problems of smoothing, threshold dependence and

localisation in cluster inference. Neuroimage 44:83–98

Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TEJ,

Johansen-Berg H et al (2004) Advances in functional and

structural mr image analysis and implementation as fsl. Neuro-

image 23:S208–S219

Smith SM, Jenkinson M, Johansen-Berg H, Rueckert D, Nichols TE,

Mackay CE et al (2006) Tract-based spatial statistics: voxelwise

analysis of multi-subject diffusion data. Neuroimage 31:1487–1505

Smith SM, Johansen-Berg H, Jenkinson M, Rueckert D, Nichols TE,

Miller KL et al (2007) Acquisition and voxelwise analysis of

multi-subject diffusion data with tract-based spatial statistics.

Nat Protoc 2:499–503

Sugiyama K, Kondo T, Oouchida Y, Suzukamo Y, Higano S, Endo M

et al (2009) Clinical utility of diffusion tensor imaging for

evaluating patients with diffuse axonal injury and cognitive

disorders in the chronic stage. J Neurotrauma 26:1879–1890

Toth A, Kovacs N, Perlaki G, Orsi G, Aradi M, Komaromy H et al

(2013) Multi-modal magnetic resonance imaging in the acute

and sub-acute phase of mild traumatic brain injury: can we see

the difference? J Neurotrauma 30:2–10

Turken U, Whitfield-Gabrieli S, Bammer R, Baldo JV, Dronkers NF

et al (2008) Cognitive processing speed and the structure of

white matter pathways: convergent evidence from normal

variation and lesion studies. Neuroimage 42:1032–1044

Wechsler D (1997) Wechsler memory scale, 3rd edn. The Psycho-

logical Corporation, San Antonio

Wechsler D (2008) Wechsler adult intelligence scale, 4th edn.

Pearson, San Antonio

Xu J, Rasmussen IA, Lagopoulos J, Haberg A (2007) Diffuse axonal

injury in severe traumatic brain injury visualized using high-

resolution diffusion tensor imaging. J Neurotrauma 24:753–765

Brain Topogr

123