The purpose of this study is to examine the relation between the microstructural architecture of white matter, as measured by diffusion tensor imaging (DTI), and postconcussion symptom reporting 6–8 weeks following mild traumatic brain injury (MTBI). Participants were 108 patients prospectively recruited from a Level 1 Trauma Center (Vancouver, BC, Canada) following an orthopedic injury [i.e., 36 trauma controls (TCs)] or MTBI (n = 72). DTI of the whole brain was undertaken using a Phillips 3T scanner at 6–8 weeks postinjury. Participants also completed a 5 h neurocognitive test battery and a brief battery of self-report measures (e.g., depression, anxiety, and postconcussion symptoms). The MTBI sample was divided into two groups based on ICD-10 criteria for postconcussional syndrome (PCS): first, PCS-present (n = 20) and second, PCS-absent (n = 52). There were no significant differences across the three groups (i.e., TC, PCS-present, and PCS-absent) for any of the neurocognitive measures (p = .138–.810). For the self-report measures, the PCS-present group reported significantly more anxiety and depression symptoms compared with the PCS-absent and TC groups (p < .001, d = 1.63–1.89, very large effect sizes). For the DTI measures, there were no significant differences in fractional anisotropy, axial diffusivity, radial diffusivity, or mean diffusivity when comparing the PCS-present and PCS-absent groups. However, there were significant differences (p < .05) in MD and RD when comparing the PCS-present and TC groups. There were significant differences in white matter between TC subjects and the PCS-present MTBI group, but not the PCS-absent MTBI group. Within the MTBI group, white-matter changes were not a significant predictor of ICD-10 PCS.

Introduction

Postconcussion symptoms are common in the acute recovery phase following mild traumatic brain injury (MTBI). Patients who sustain MTBI report more acute symptoms than healthy control and trauma control (TC) subjects in the initial days and weeks postinjury in most (Iverson, Silverberg, Lange, & Zasler, 2013), but not all studies (e.g., Landre, Poppe, Davis, Schmaus, & Hobbs, 2006). Fortunately, in the majority of cases, these symptoms resolve within weeks to a few months (Iverson et al., 2013). However, the diagnostic significance of postconcussion symptom reporting following MTBI is controversial. Clinicians often assume that postconcussion symptoms reported many months or years postinjury are directly related to the biological consequences of that brain injury. However, there are many factors unrelated to brain injury that can potentially account for these symptoms (e.g., other medical problems (Lees-Haley & Brown, 1993), personal injury litigation (Lees-Haley & Brown, 1993), posttraumatic stress disorder (Foa, Cashman, Jaycox, & Perry, 1997), orthopedic injuries (Meares et al., 2008; Mickeviciene et al., 2004), chronic pain (Iverson & McCracken, 1997), depression (Iverson, 2006; Suhr & Gunstad, 2002b), and myofascial injury (whiplash) (Sullivan, Hall, Bartolacci, Sullivan, & Adams, 2002). Complicating matters further, the perception and reporting of symptoms can be influenced by a diverse range of psychological and social-psychological factors that may cause, maintain, or worsen the myriad of symptoms that are reported postinjury (e.g., premorbid personality [Hibbard et al., 2000], clinical methods used to elicit symptoms [Iverson, Brooks, Ashton, & Lange, 2010], nocebo effect [Hahn, 1997], “expectation as etiology” [Gunstad & Suhr, 2001], diagnosis threat [Suhr & Gunstad, 2002a], and “good old days” bias [Gunstad & Suhr, 2001; Iverson, Lange, Brooks, & Rennison, 2010]). Some studies have illustrated that preinjury mental health problems and postinjury difficulties with anxiety are strongly related to symptoms that persist for many months following injury (Ponsford et al., 2012).

To date, there is limited evidence to support a direct causal link between macrostructural brain injury and symptom reporting. In general, neuroimaging studies have not documented a relationship between greater postconcussion symptom reporting and intracranial abnormalities identified on structural computed tomography (CT) or magnetic resonance imaging (MRI) scans within the first few days (Hughes et al., 2004; Kurca, Sivak, & Kucera, 2006), 3–6 months (Hofman et al., 2001; Lannsjo, Backheden, Af Johansson, Geijerstam, & Borg, 2013; Sigurdardottir, Andelic, Roe, Jerstad, & Schanke, 2009; Taylor et al., 2010), or 6–12 months (Deepika, Munivenkatappa, Devi, & Shukla, 2013; Sigurdardottir et al., 2009; Taylor et al., 2010) following injury. In fact, de Guise and colleagues (2010) found, paradoxically, greater postconcussion symptom reporting 20 weeks following MTBI in those patients who had normal day-of-injury CT scans (rather than abnormal scans). Of course, the lack of association between postconcussion symptoms and intracranial abnormalities on structural CT or MRI scan may be due, in part, to the lack of sensitivity of these methods to detect diffuse microstructural changes often associated with MTBI (Bruns & Jagoda, 2009; Gaetz & Weinberg, 2000; e.g., Hofman et al., 2001; Lewine et al., 2007).

Diffusion tensor imaging (DTI) is useful for detecting white-matter changes in the brain following TBI (Belanger, Vanderploeg, Curtiss, & Warden, 2007; Xu, Rasmussen, Lagopoulos, & Haberg, 2007) and it has become an important tool for the evaluation of MTBI in the acute, subacute, and chronic phases of the recovery trajectory (Inglese, Bomsztyk, et al., 2005; Inglese, Makani, et al., 2005; Kraus et al., 2007; Kumar et al., 2009; Lipton et al., 2009; Rutgers et al., 2008; Shenton et al., 2012). To date, a number of researchers have established the sensitivity of DTI to differentiate healthy controls from those individuals who have sustained MTBI and who report postconcussion symptoms 3–6 years following injury (Lipton et al., 2008; Niogi et al., 2008). The methodology of these small, cross-sectional, correlational studies, however, does not allow one to assume a causal link between microstructural white-matter findings and postconcussion symptoms. In recent years, the number of DTI studies that have examined the relationship between postconcussion symptom reporting and MTBI has increased steadily. These studies can be characterized into two categories: correlational studies and group studies.

Results from correlational studies have found a relatively consistent relationship between postconcussion symptom reporting and various DTI metrics. For example, researchers have reported that increased postconcussion symptom reporting is associated with (a) increased fractional anisotropy (FA) and decreased apparent diffusion coefficients (ADCs) and radial diffusivity (RD) in the corpus callosum in 10 adolescents within the first 6 days following MTBI (Wilde et al., 2008); (b) low trace values in the whole brain at 72 h and 1 month postinjury in six patients following MTBI (Bazarian et al., 2007); (c) pre–post season changes in FA and mean diffusivity (MD) in nine high school athletes who sustained repeated multiple subconcussive blows (Bazarian, Zhu, Blyth, Borrino, & Zhong, 2012); and (d) increased MD in the inferior fronto-occipital fasciculus, inferior longitudinal fasciculus, and superior longitudinal fasciculus; and decreased FA in uncinate fasciculus, inferior fronto-occipital fasciculus, internal capsule, corpus callosum, and the parietal and frontal subcortical white matter in 20 patients 1 month postinjury (Smits et al., 2011). One exception to this is a study by Levin and colleagues (2010) who found no relationship between FA and ADC and postconcussion symptoms in 37 mild–moderate TBI service members 2.3 years postinjury using fiber-tracking and region of interest (ROI) methodology. It is important to appreciate, however, that despite the significant correlations found between postconcussion symptom reporting and some DTI metrics in the above studies, there were many other DTI metrics that were not associated with postconcussion symptom reporting.

In contrast to correlational studies, there are mixed findings in group studies that have compared DTI metrics across subgroups defined by postconcussion symptom reporting. Messe and colleagues (Messé et al., 2011,, 2012) found some DTI differences in a small sample of patients with MTBI divided into poor outcome (Postconcussion Syndrome [PCS] present [i.e., PCS+]) versus good outcome (PCS absent [i.e., PCS-]) groups. These differences were found for MD only, and not for FA, axial diffusivity (AD), or RD (Messé et al., 2011). At 3 months postinjury (n = 23), participants in the PCS+ group had higher MD compared with the PCS− group in the forceps major and minor, the corpus callosum, the inferior-occipital fasciculus bilaterally, and inferior longitudinal fasciculus bilaterally. The PCS+ group (but not the PCS− group) also had higher MD compared with a healthy control sample in some ROIs (i.e., superior longitudinal fasciculus, corticospinal tract bilaterally, and left anterior thalamic radiation). In a follow-up longitudinal study of 53 MTBI patients evaluated in the acute phase of recovery (8–21 days [T1]) and subacute (6 months [T2]), Messe and colleagues (2012) (Messe et al., 2012) reported that the PCS+ group had a greater number of abnormal brain regions on DTI than the PCS− group when compared with healthy controls. These authors did not provide detailed information regarding a direct comparison between PCS+ vs. PCS− groups; however, they concluded that “significant structural differences were found between PCS+ and PCS− groups at T1 and less at T2 for the MD and AD measures, that involved primarily the corpus callosum, the longitudinal fasciculus, and the internal capsules.” (p. 288). There were no differences found for FA or RD measures. Similarly, Hartikainen and colleagues (2010) examined 18 patients with mild–moderate TBIs at 6 months postinjury and found that those who reported persistent postconcussion symptoms (n = 7) had lower ADC (d = 2.33) and higher FA (d = 1.86) compared with those who reported no symptoms (n = 11) in the midbrain structures (i.e., thalamus, internal capsule, centrum semiovale, and mesencephalon).

In contrast to these studies, Waljas et al. (2014) examined 48 patients with uncomplicated MTBI and found no differences across 16 ROIs for FA, or 10 ROIs for ADC, between those who did and did not report postconcussion symptoms at 3 weeks following injury. Similarly, Lange, Iverson, Brubacher, Madler, and Heran (2012) examined the relation between DTI of the corpus callosum and reporting in 60 patients who had sustained MTBI, and 34 patients with orthopedic/soft-tissue injuries (i.e., TCs). At 6–8 weeks postinjury, there were no significant differences between the MTBI and TC groups for FA or for MD in the genu, body, and splenium of the corpus callosum. Similarly, in the MTBI sample, there were no significant differences on any DTI measures between those who did and did not meet ICD-10 symptom criteria for postconcussion syndrome (PCS). In the MTBI sample, Cohen's effect sizes ranged from very small (d = 0.02) to small (d = 0.16) for all DTI measures; with the exception of small–medium effect size found for FA in the genu (d = 0.39) and body (d = 0.37) of the corpus callosum.

The purpose of this study is to expand the study by Lange and colleagues (2012) that examined the relation between DTI and postconcussion symptom reporting by: first, using DTI to examine the whole brain; second, include a larger number of DTI metrics such FA, MD, RD, and AD; third, use more advanced and comprehensive DTI analysis methods including tract-based spatial statistics (TBSS) and standard brain atlas generated ROI analyses; fourth, include neurocognitive tests and more symptom scales as outcome measures; fifth, include a sample of individuals in the MTBI group that do not have intracranial abnormalities that predate their injury; and finally, include a sample of participants in the TC group that do not include minor intracranial abnormalities (e.g., T2 hyperintensities).

The hypotheses are threefold. First, compared with the TC group, patients with MTBI will (a) report a greater number of postconcussion, depression, and anxiety symptoms; (b) perform comparably on neurocognitive measures; and (c) have reduced white-matter integrity in the whole brain as measured by DTI metrics. Second, in the MTBI group, patients who meet ICD-10 criteria for PCS (i.e., PCS-present) will report a greater number of depression and anxiety symptoms compared with those patients who do not meet ICD-10 PCS criteria (i.e., PCS-absent). Third, in the MTBI group, there will be no differences across all neurocognitive measures and DTI metrics when comparing the PCS-present vs. PCS-absent groups.

Method

Participants

Participants were 108 patients (72 MTBI and 36 TC) prospectively recruited from the Emergency Department of Vancouver General Hospital (Level 1 Trauma Center) between June 2007 and April 2012. Patients were identified for potential inclusion in the study via daily reviews of consecutive Emergency Department admissions. Patients were targeted for recruitment and consent if they presented to the Emergency Department after sustaining a MTBI (i.e., MTBI group), or they had sustained a soft-tissue or orthopedic injury without brain injury (i.e., TC group).

All participants were enrolled in the study if they were first, between 19 and 55 years of age, second, injured as a result of a traumatic injury (e.g., fall, motor vehicle accident, assault, etc.), and third, had a blood alcohol level (BAL) obtained at the time of injury by hospital staff on admission to the Emergency Department. General exclusion criteria included first, lack of proficiency in conversational English; second, educated in a language other than English after age 10 (Note that the second and more stringent criterion for ESL was only introduced in 2009 after the evaluation of a series of patients whose neurocognitive profile was considered to be affected by a lack of English-based education (e.g., intact scores on all measures except language-based measures).); (c) history of a neurological disorder (e.g., stroke or multiple sclerosis), TBI, learning disability, or psychiatric illness requiring hospitalization; (d) presence of any contraindication to MRI, (e) history of significant drug abuse other than alcohol; (f) presence of upper body injuries restricting the use of hands or arms; or (g) difficulties with eyesight.

Participants were included in the TC group if (a) they sustained a soft-tissue or orthopedic injury below the neck; (b) there was no evidence of an altered state of consciousness as indicated by a reduction in Glasgow Coma Scale (GCS) score, or presence of a loss of consciousness (LOC), posttraumatic amnesia (PTA), or posttraumatic confusion; and (c) there was no evidence of physical head trauma, whiplash, or cervical strain based on medical chart review (e.g., absence of lacerations/contusions to the head, absence of complaints of head, neck, or back pain). In a small number of cases (n = 2), TC participants had undergone a head CT but had no evidence of intracranial abnormality.

Participants were included in the MTBI group if they (a) presented to the Emergency Department following head trauma, and (b) had evidence of MTBI as indicated by at least one of the following: (i) witnessed LOC of at least 1 min duration but not >30 min, (ii) PTA of at least 15 min duration but not >24 h, and/or (iii) GCS score of 13 or 14. Patients were further classified into uncomplicated and complicated MTBI groups as follows: (a) uncomplicated MTBI: GCS = 13–15, PTA < 24 h, LOC <30 min, and no trauma-related intracranial abnormality on day-of-injury CT or 6–8 weeks structural MRI scan; (b) complicated MTBI: GCS = 13–15, PTA <24 h, LOC <30 min, and trauma-related intracranial abnormality on day-of-injury CT scan or 6–8 week structural MRI scan. Of the 72 MTBI patients included in this study, there were 38 (52.8%) with uncomplicated MTBI and 34 (47.2%) with complicated MTBI.

For the purpose of this study, the MTBI group was divided into two groups based on ICD-10 criteria for PCS: PCS-present (n = 20) and PCS-absent (n = 52)—see Self-reported symptoms for classification criteria.

Participant Selection

Participants were selected from a larger sample of 170 patients enrolled in the study (105 MTBI and 65 TC) in a three-step manner. First, participants were initially included in the sample if they met the following criteria: (1) completed the entire neuropsychological test battery (99.4% of sample), (2) neurocognitive test performance was not considered influenced by reduced English language proficiency (95.9% of sample), (3) scored above the recommended cutoff for good effort on the Test of Memory Malingering (100% of entire sample), (4) behavioral observations during the neuropsychological evaluation did not provide suspicion of questionable motivation or attention that may have negatively influenced test performance (98.8% of sample), (5) successfully completed MRI scanning (99.4% of sample), (6) there was no evidence of obvious incidental neurological abnormalities on MRI such as meningioma, cistern mass, venous anomaly (95.8% of sample), and (7) structural MRI scans were considered complete and readable by a neuroradiologist (98.8% of sample). In addition, in the MTBI group, participants were only included if (8) there was no evidence of intracranial abnormality that was considered to predate the current injury (97.1% of the MTBI sample). Similarly, in the TC group, participants were only included if (9) they had ≤1 white-matter hyperintensity (WMHIs [76.2% of TC sample]). Researchers have reported that healthy adults with two or more WMHIs have abnormalities on DTI metrics (distal from the hyperintensities) compared with adults with zero or one of hyperintensity (Iverson et al., 2011; Lange, Brickell, et al., 2013; Lange, Shewchuk, et al., 2013). Application of these criteria resulted in retaining 130 participants (39 TC, 91 MTBI [26 PCS-present, 65 PCS-absent]).

Second, in order to control for the influence of age, handedness, and preinjury alcohol consumption, participants were further excluded if they were 50 years or older, left-handed, or whose preinjury alcohol consumption was considered to be unusually high (defined as drinking >30 drinks per week and/or drinking five or more drinks per occasion during four or more days per week). These additional criteria were applied following an exploratory frequency analysis that identified a number of outliers that were considered to bias the analyses. Application of these criteria resulted in retaining 120 participants (39 TC and 81 MTBI [i.e., 22 PCS-present, 59 PCS-absent]).

Third, prior to DTI postprocessing, the MRI scans were inspected for image quality. Scans were excluded if they exhibited significant motion artifact or brain coverage was inadequate. Application of these criteria resulted in retaining a final sample of 108 participants (36 TC and 72 MTBI [i.e., 20 PCS-present, 52 PCS-absent]).

Some of the participants in this study were included in a previous study by Lange and colleagues (2012). In the current MTBI group, 54.2% (n = 39) were included in the previous study. In the current TC group, 50.0% (n = 18) were included in the original study. There were two primary reasons why participants from the original study were not included here. First, the inclusion of the neuropsychological variables in this study resulted in the exclusion of some participants due to unsuccessful test completion, poor English proficiency, and/or fluctuating attention (i.e., criteria 1–4 above). Second, more stringent criteria was applied to the TC and MTBI groups in this study with regard to the structural MRI scans, in order to create “cleaner” groups (i.e., criteria 8 and 9 above).

Measures and Procedure

Participants completed a 1-h MRI brain scan and a 5-h neuropsychological assessment battery that included measures of neurocognitive functioning and self-reported mental health and postconcussion symptoms, ∼6–8 weeks postinjury (M = 46.6 days, SD = 6.0). All participants gave written informed consent in accordance with the Clinical Research Ethics Board at the University of British Columbia, Vancouver, Canada.

Self-Reported symptoms

Participants completed the Beck Depression Inventory-Second Edition (BDI-II), Beck Anxiety Inventory (BAI), and the British Columbia Postconcussion Symptom Inventory (BC-PSI). The BDI-II (Beck, Steer, & Brown, 1996) and BAI (Beck & Steer, 1993) are widely used measures designed to assess depressive and anxiety symptoms, respectively. Total scores for each measure were obtained by summing the individual responses to all items on each scale separately. In addition, for the purposes of this study, an additional total score was calculated for the BDI-II using 10 selected symptoms that are believed to have the least overlap with symptoms of MTBI, and that are considered most representative of depression (i.e., BDI-II 10-item total score). These items included sadness, loss of interest, loss of pleasure, pessimism, past failure, guilt feelings, punishment feelings, self-criticalness, crying, and suicidal thoughts or wishes. The BDI-II 10-item total score was calculated by summing the responses to these 10 individual items.

The BC-PSI (Iverson, Zasler, & Lange, 2007) is a symptom inventory designed based on ICD-10 (American Psychiatric Association, 2000) symptom criteria for PCS that requires the test taker to rate the frequency (0 = “not at all” to 5 = “constantly”) and intensity (0 = “not at all” to 5 = “very severe problem”) of 13 symptoms (e.g., headaches, dizziness, nausea, fatigue, sensitivity to noises, etc.) as well as the effect of three co-occurring life problems on daily living (i.e., effects of alcohol consumption, worrying/ dwelling on symptoms, and self-perception of brain damage). For each of the 13 symptoms, the two ratings are multiplied together (frequency × intensity) to create a single score for each item. These product-based scores were then converted to item scores that reflect both the frequency and intensity of symptom endorsement (range = 0–4). Item scores of 3 are interpreted as falling in the moderate range.

Responses on the BC-PSI were classified into two categories based on ICD-10 criteria for PCS: (a) PCS-present and (b) PCS-absent. Participant responses were classified as PCS-present if they endorsed BC-PSI items as moderate or greater on ≥3 of the six ICD-10 Category C criteria. Similarly, participant responses were classified as PCS-absent if they endorsed BC-PSI items as a moderate or greater on <3 of the six ICD-10 Category C criteria. The percentages of patients who endorsed each individual symptom domain were as follows: physical symptoms = 28.7%, emotional symptoms = 15.7%, cognitive symptoms = 16.7%, poor sleep = 20.4%, sensitivity to alcohol = 18.5%, and being overly focused on and concerned about their symptoms = 35.2%.

It is important to appreciate that symptom endorsement on the BC-PSI cannot be used in isolation to evaluate whether a person meets ICD-10 criteria for PCS. The diagnosis of PCS is complicated, challenging, and requires a comprehensive clinical evaluation (see Iverson et al., 2013 for a comprehensive discussion). This was not done here. Rather, endorsement of BC-PSI items that meet ICD-10 symptom criteria for PCS simply reflects clinically elevated postconcussion symptom reporting and not a diagnosis per se.

Neurocognitive measures

The neurocognitive measures consisted of 16 tests from the Neuropsychological Assessment Battery (NAB; Stern & White, 2003). The NAB is a comprehensive, co-normed (across all tests) neuropsychological test battery that consists of 24 individual tests designed to assess cognitive functioning across five domains: attention, language, memory, spatial, and executive functioning. The normative sample is large and the coverage of neuropsychological abilities assessed is broad. The NAB can be used in a fixed or flexible manner. Only 16 of the 24 tests were selected for use in order to reduce administration time. The 16 selected tests result in 23 scores of interest. In order to reduce the number of cognitive variables for the analyses, the 23 scores of interest were used to generate index scores for each of the five cognitive domains. The Attention and Memory indexes were generated as per the instructions in the manual. For the Language, Spatial, and Executive Functioning Indexes, however, not all tests included in these indexes were administered. As such, these indexes were prorated. These three indexes were calculated by generating a prorated “sum of T-scores” which was converted to a standard score using the look-up table (i.e., Table 7.1) in the NAB normative manual (Stern & White, 2003). The prorated sum of T-scores for the three indexes was calculated by averaging the demographically adjusted T-scores across all available tests that are included in each index. The mean T-score was then multiplied by the number of subtests that are used to generate the full version of the index. For the Language Index, two of five possible subtests were used (i.e., Oral Production and Naming). For the Spatial Index, two of four possible subtests were used (i.e., Visual Discrimination and Design Construction). For the Executive Functioning Index, three of four possible subtests were used (Categories, Mazes, and Word Generation). A NAB Total Score Index was also generated using the standard scores derived for the five indexes above as per the instructions in the manual.

As part of the larger neurocognitive test battery, participants were also administered the (a) Test of Memory Malingering (Tombaugh, 1996) to evaluate the possibility of poor effort during neuropsychological testing, (b) Reynolds Intellectual Screening Test (Reynolds & Kamphaus, 2003) to assess current level of intellectual ability, and (c) Wechsler Test of Adult Reading (The Psychological Corporation, 2001) to assess premorbid level of intellectual ability. In addition, participants completed a semi-structured interview designed to gather information about their past and present medical, psychiatric, and personal history. Detailed information regarding the duration of PTA was obtained during this interview and reconstructed, when possible, using collateral information obtained from hospital records.

Neuroimaging

All MRI data were acquired on a Philips Achieva 3T scanner equipped with Dual Nova Gradients (maximum gradient strength 80 mT/m, max. slew rate 200 mT/m/s) and an eight-channel head coil. Partial parallel imaging was performed using sensitivity encoding (SENSE) (Pruessmann, Weiger, Scheidegger, & Boesiger, 1999). The total data acquisition time was 43 min. The MRI protocol included: (a) axial T2-weighted turbo spin echo scan (TR = 3000 ms, TE = 80 ms, flip angle = 90°, acquisition matrix = 320 × 245, field of view = 240 × 192 × 139 mm3, acquired voxel size = 0.75 × 0.78 × 4 mm3, reconstructed voxel size = 0.47 × 0.47 × 4 mm3, SENSE factor of 1.2 along the left–right direction, 2 averages); (b) axial T2-weighted fluid attenuated inversion recovery [FLAIR] scan, TR = 10,000 ms, TE = 10 ms, acquisition matrix = 304 × 194, field of view = 240 × 193 × 139 mm3, acquired voxel size = 0.79 × 0.99 × 4.00 mm3, reconstructed voxel size = 0.47 × 0.47 × 4.00 mm3, SENSE factor of 1.6 along the left–right direction); and (c) DTI scan (TR = 5618 ms, TE = 75 ms, flip angle = 90°, acquisition matrix = 96 × 95, field of view = 240 × 240 × 125 mm3, acquired voxel size = 2.50 × 2.50 × 2.50 mm3, reconstructed voxel size = 1.88 × 1.88 × 2.50 mm3, SENSE factor of 2.4 along the anterior–posterior direction, 15 diffusion directions, 3 averages, b factor = 1,000 s/mm2). Sagittal 3D T1-weighted, axial T1-weighted spin echo, and axial and coronal 2D T2*-weighted gradient echo scans were also obtained but are not directly relevant to this study.

The structural T1 and FLAIR images were reviewed by a board certified radiologist (JRS or MKSH) who determined whether any visible trauma-related intracranial pathology was present. Common pertinent MRI pathologies that were coded included number and location of susceptibility and T2 hyperintense foci, encephalomalacia, and extra-axial collections.

Tract-based spatial statistics and randomize

The TBSS (Smith et al., 2006) and randomize (Winkler, Ridgway, Webster, Smith, & Nichols, 2014) part of the FMRIB Software Library (FSL, Analysis Group, FMRIB, Oxford, UK) (Smith et al., 2004) were used for comparisons of FA, MD, AD, and RD. DTI sequences with poor field-of-view resulting in significant deficits in whole brain coverage or artifacts were not considered in the analysis. Motion and artifacts in the diffusion data were corrected using affine registration of all gradient volumes with the first b = 0 volume (FLIRT; FMRIB Software Library, Oxford, UK), and gradient directions were compensated for rotations (Landman et al., 2007). This was followed by creation of a manual brain mask based on the first b = 0 image using 3D Slicer version 2.7 (Gering et al., 2001). Voxel-based comparisons (on the TBSS skeleton using randomize) of FA, MD, AD, and RD were performed using FSL Version 4.1.9 (Smith et al., 2006). Briefly, individual FA maps were nonlinearly registered via FSL-FNIRT to the Montreal Neurological Institute International Consortium Brain Mapping (ICBM)-152 template provided by FSL, followed by creation of a mean FA skeleton. Individual FA values were then projected onto this mean skeleton. MD, AD, and RD values were projected onto the same template using the previously created FA transformation. FA threshold was set to 0.35 and the number of permutations was set at 5,000. The conservative threshold of 0.35 was chosen based on a visual inspection of the mean TBSS skeleton and represented the best qualitative compromise between skeletal continuity and peripheral white-matter inclusion. Threshold-free cluster enhancement with correction for family wise error rate was employed. The p < .05 T-contrast maps were projected on the mean FA images. To identify which specific anatomical areas were implicated, the mean FA skeleton and contrast map were overlaid with the Johns Hopkins University (JHU) ICBM-DTI-81 white-matter label atlas (Hua et al., 2008; Mori, Wakana, Nagai-Poetscher, & va Zigl, 2005; Wakana et al., 2007).

FSL statistics were then used to compute the mean FA values for each individual for each of the ROIs on the mean FA skeleton. Mean values for MD, AD, and RD were done similarly. Forty-eight individual ROIs were identified according to the JHU ICBM-DTI-81 white-matter labels atlas (Hua et al., 2008; Mori et al., 2005; Wakana et al., 2007).

The ROIs from the ICBM-DTI-81 atlas included the (a) genu, body, and splenium of corpus callosum; (b) pontine crossing tract, fornix, and middle cerebellar peduncle; and (c) two unilateral symmetrical ROIs (left/right) each for the corticospinal tract, medial lemniscus, inferior cerebellar peduncle, superior cerebellar peduncle, cerebral peduncle, anterior limb of internal capsule, posterior limb of internal capsule, retrolenticular part of internal capsule, anterior corona radiata, superior corona radiata, posterior corona radiata, posterior thalamic radiation, sagittal stratum, external capsule, cingulum (cingulate gyrus), cingulum (hippocampus), fornix/stria terminalis, superior longitudinal fasciculus, superior fronto-occipital fasciculus, uncinate fasciculus, and tapetum.

Due to the large number of ROIs, four summary scores were calculated for each participant and used in all statistical analyses. The four summary scores represent the number of ROIs with FA, MD, AD, and RD values that fell below/above a specified cutoff score for each participant. Cutoff scores were identified by calculating the means and SD for FA, MD, AD, and RD values in each of the 48 ROIs based on DTI of 36 participants who were recruited for this study, that had sustained an orthopedic injury, but not a TBI. FA values that were >2 SDs below the mean, and MD, AD, and RD values that were >2 SDs above the mean, were classified as reflecting an ROI with “reduced white-matter integrity” (i.e., abnormal score).

Results

Demographic and Injury Characteristics

Descriptive statistics and group comparisons (using ANOVA and χ2 analyses) of demographic and injury characteristics are presented in Table 1 (continuous variables) and Table 2 (categorical variables). In the MTBI group, there were no significant differences between the PCS-present and PCS-absent group for GCS scores, duration of LOC, duration of PTA, intracranial abnormality on day-of-injury CT scans, or intracranial abnormality on MRI scans 6–8 weeks postinjury (range: p = .233–.999). When comparing the remaining variables across the three groups (i.e., PCS-present, PCS-absent, and TC), there were no significant effects for age, education, gender, ethnicity, preinjury alcohol use, day-of-injury BAL, mechanism of injury, days tested postinjury, current intellectual ability, or estimated premorbid intellectual ability (all p > .05).

Table 1.

Descriptive statistics, group comparisons, and effect sizes of demographic and injury severity characteristics (continuous variables) by group

 1. Mild TBI: PCS-absent
 
2. Mild TBI: PCS-present
 
3. Trauma control
 
 Cohen's effect size
 
M SD M SD M SD p 1 versus 2 1 versus 3 2 versus 3 
Age (years) 34.1 11.3 34.1 10.4 31.6 10.2 .521 0.01 0.23 0.25 
Education (years) 15.3 2.1 14.4 2.3 14.2 2.0 .055 0.40 0.51a 0.09 
Day-of-injury BAL 17.5 25.6 17.8 26.5 6.7 16.1 .074 0.01 0.50a 0.56 
Days tested postinjury 46.8 6.5 46.7 5.5 46.2 5.7 .920 0.02 0.09 0.08 
Lowest GCS >30 min 14.4 0.7 14.4 0.7 — — .999 0.01 — — 
RIST Index 110.3 9.6 107.8 8.0 106.9 10.9 .271 0.27 0.33 0.09 
Premorbid IQ (WTAR) 107.6 7.1 106.2 7.4 107.6 6.9 .713 0.20 0.00 0.21 
 1. Mild TBI: PCS-absent
 
2. Mild TBI: PCS-present
 
3. Trauma control
 
 Cohen's effect size
 
M SD M SD M SD p 1 versus 2 1 versus 3 2 versus 3 
Age (years) 34.1 11.3 34.1 10.4 31.6 10.2 .521 0.01 0.23 0.25 
Education (years) 15.3 2.1 14.4 2.3 14.2 2.0 .055 0.40 0.51a 0.09 
Day-of-injury BAL 17.5 25.6 17.8 26.5 6.7 16.1 .074 0.01 0.50a 0.56 
Days tested postinjury 46.8 6.5 46.7 5.5 46.2 5.7 .920 0.02 0.09 0.08 
Lowest GCS >30 min 14.4 0.7 14.4 0.7 — — .999 0.01 — — 
RIST Index 110.3 9.6 107.8 8.0 106.9 10.9 .271 0.27 0.33 0.09 
Premorbid IQ (WTAR) 107.6 7.1 106.2 7.4 107.6 6.9 .713 0.20 0.00 0.21 

Notes: N = 108 (mild TBI PCS-absent, n = 52; mild TBI PCS-present, n = 20; TC, n = 36). BAL = blood alcohol level; GCS = Glasgow Coma Scale; RIST = Reynolds Intellectual Screening Test; IQ = intelligence quotient; WTAR = Wechsler Test of Adult Reading.

Significant pairwise comparisons: ap < .05.

Table 2.

Descriptive statistics and group comparisons of demographic and injury severity characteristics (categorical variables) by group

 Mild TBI: PCS-absent
 
Mild TBI: PCS-present
 
Trauma control
 
χ2 
N n n 
Gender 
 Men 41 78.8a 11 55.0a 26 72.2 0.129 
 Women 11 21.2a 45.0a 10 27.8  
Ethnicity 
 Caucasian 44 84.6 15 75.0 30 83.3 0.634 
 Asian/East-Indian/other 15.4 25.0 16.7  
Mechanism of injury 
 MVA 21 40.4 40.0 14 38.9 0.990 
 Non-MVA 31 59.6 12 60.0 22 61.1  
Preinjury alcohol1 
 Low-moderate 17 32.7 15.0 12 33.3 0.283 
 Heavy 35 67.3 17 85.0 24 66.7  
Day-of-injury BAL 
 Sober (<21 mmol/l) 33 63.5b 13 65.0 30 83.3b 0.113 
 Intoxicated (≥21 mmol/l) 19 36.5b 35.0 16.7b  
LOC 
 None 5.8 10.0 — — 0.527 
 Transient 14 26.9 10.0 — —  
 1–30 min 35 67.3 16 80.0 — —  
PTA 
 <15 min 9.6 20.0 — — 0.233 
 >15 min 47 90.4 16 80.0 — —  
GCS 
 15 27 51.9 10 50.0 — — 0.884 
 13–14 25 48.1 10 50.0 — —  
DOI CT scan 
 Normal 39 75.0 15 75.0 — — 0.826 
 Abnormal 12 23.1 20.0 — —  
 Not ordered 1.9 5.0 — —  
DOI CT or 6-week MRI 
 Normal 26 50.0 12 60.0 — — 0.446 
 Abnormal 26 50.0 40.0 — —  
#MRI Sus foci 
 None 29 55.8 14 70.0 — — 0.443 
 1 9.6 5.0 — —  
 2–5 13.5 5.0 — —  
 >5 11 21.2 20.0 — —  
#MRI T2 foci 
 None 37 71.2 15 75.0 — — 0.968 
 1 13.5 10.0 — —  
 2–5 5.8 10.0 — —  
 >5 9.6 5.0 — —  
 Mild TBI: PCS-absent
 
Mild TBI: PCS-present
 
Trauma control
 
χ2 
N n n 
Gender 
 Men 41 78.8a 11 55.0a 26 72.2 0.129 
 Women 11 21.2a 45.0a 10 27.8  
Ethnicity 
 Caucasian 44 84.6 15 75.0 30 83.3 0.634 
 Asian/East-Indian/other 15.4 25.0 16.7  
Mechanism of injury 
 MVA 21 40.4 40.0 14 38.9 0.990 
 Non-MVA 31 59.6 12 60.0 22 61.1  
Preinjury alcohol1 
 Low-moderate 17 32.7 15.0 12 33.3 0.283 
 Heavy 35 67.3 17 85.0 24 66.7  
Day-of-injury BAL 
 Sober (<21 mmol/l) 33 63.5b 13 65.0 30 83.3b 0.113 
 Intoxicated (≥21 mmol/l) 19 36.5b 35.0 16.7b  
LOC 
 None 5.8 10.0 — — 0.527 
 Transient 14 26.9 10.0 — —  
 1–30 min 35 67.3 16 80.0 — —  
PTA 
 <15 min 9.6 20.0 — — 0.233 
 >15 min 47 90.4 16 80.0 — —  
GCS 
 15 27 51.9 10 50.0 — — 0.884 
 13–14 25 48.1 10 50.0 — —  
DOI CT scan 
 Normal 39 75.0 15 75.0 — — 0.826 
 Abnormal 12 23.1 20.0 — —  
 Not ordered 1.9 5.0 — —  
DOI CT or 6-week MRI 
 Normal 26 50.0 12 60.0 — — 0.446 
 Abnormal 26 50.0 40.0 — —  
#MRI Sus foci 
 None 29 55.8 14 70.0 — — 0.443 
 1 9.6 5.0 — —  
 2–5 13.5 5.0 — —  
 >5 11 21.2 20.0 — —  
#MRI T2 foci 
 None 37 71.2 15 75.0 — — 0.968 
 1 13.5 10.0 — —  
 2–5 5.8 10.0 — —  
 >5 9.6 5.0 — —  

Notes: N = 108 (mild TBI PCS-absent, n = 52; mild TBI PCS-present, n = 20; TC, n = 36). CT = computed tomography; GCS = Glasgow Coma Scale; PTA = posttraumatic amnesia; LOC loss of consciousness; TBI = traumatic brain injury; MVA = motor vehicle accident; BAL = blood alcohol level.

1Defined based on criteria for heavy drinking established by the National Institute on Alcohol Abuse and Alcoholism: (a) females: 8 or more drinks per week or 4 or more drinks on a single occasion >52 times per year; (b) males: 15 or more drinks per week or 5 or more drinks on a single occasion >52 times per year.

ap = .043; bp = .042; 2 × 2 comparison completed by collapsing “Transient” and “1–30 min” categories together; 2 × 2 comparison completed by collapsing “None” and “1” categories together, and “2–5” and “>5” categories together.

There was a significantly greater proportion of women in the PCS-present group (45.0%) compared with men (21.2%; p = .043). There was a significantly greater proportion of individuals who were intoxicated at the time of injury in the PCS-absent group (76.0%) compared with those who were sober (52.4%; p = .042). The effect sizes (phi-coefficient; φ) for these two findings were relatively small (i.e., gender: φ = 0.24; intoxication: φ = 0.22). Further, the PCS-absent group had a significantly higher level of education, and higher BALs at the time of injury, compared with the TC group (both p < .05, d = 0.51 and 0.50, respectively).

Neuropsychological Measures

Descriptive statistics, group comparisons, and Cohen's effect sizes (Cohen, 1988) for the self-reported symptoms and neurocognitive measures, by group, are presented in Table 3. For the self-reported symptoms, there were significant main effects (using Kruskal–Wallis H tests due to non-normal distribution) for the BC-PSI total, BAI total, and BDI-II total scores. As expected, pairwise comparisons (using Mann–Whitney U-tests) revealed that the PCS-present group had higher total scores on the BC-PSI, BAI, and BDI-II compared with both the PCS-absent group (d = 1.63–1.89 [Of course, the significant difference between the PCS-present and PCS-absent group for the BC-PSI total score is not meaningful. This group difference is a direct consequence of group classification. As such, the Cohen's effect size for this comparison is not reported here.], very large effect sizes) and TC group (d = 1.22–2.46, very large effect sizes). There were no significant differences, or substantial effect sizes, when comparing the PCS-absent group and TC group across the three measures.

Table 3.

Descriptive statistics, group comparisons, and effect sizes for neuropsychological measures: self-report measures and NAB Indexes

 1. Mild TBI: PCS-absent
 
2. Mild TBI: PCS-present
 
3. Trauma control
 
 Effect size
 
M SD M SD M SD p 1 versus 2 1 versus 3 2 versu 3 
Symptom questionnaires1 
 BC-PSI total 6.5 4.6 24.3 10.4 4.9 6.5 <.001 n/a 0.30 2.46c 
 BAI total 5.2 4.8 15.5 10.2 5.9 6.6 <.001 1.63c 0.13 1.22c 
 BDI-II total 6.1 5.0 19.2 11.7 6.5 7.4 <.001 1.89c 0.07 1.41c 
Neurocognitive2 
 Total Index 105.5 13.5 107.3 13.6 107.4 13.3 .761 0.13 0.15 0.01 
 NAB Attention Index 105.7 13.8 104.5 10.3 102.1 13.6 .455 0.10 0.26 0.19 
 NAB Memory Index 101.5 13.2 103.1 13.4 104.7 11.9 .526 0.12 0.25 0.13 
 NAB Language Index* 100.1 14.8 108.6 22.8 105.7 18.8 .138 0.50 0.34 0.14 
 NAB Spatial Index* 107.5 16.0 107.3 15.2 110.0 14.2 .714 0.01 0.16 0.19 
 NAB Executive Index* 106.9 15.9 104.5 14.9 106.8 13.7 .810 0.16 0.01 0.16 
 1. Mild TBI: PCS-absent
 
2. Mild TBI: PCS-present
 
3. Trauma control
 
 Effect size
 
M SD M SD M SD p 1 versus 2 1 versus 3 2 versu 3 
Symptom questionnaires1 
 BC-PSI total 6.5 4.6 24.3 10.4 4.9 6.5 <.001 n/a 0.30 2.46c 
 BAI total 5.2 4.8 15.5 10.2 5.9 6.6 <.001 1.63c 0.13 1.22c 
 BDI-II total 6.1 5.0 19.2 11.7 6.5 7.4 <.001 1.89c 0.07 1.41c 
Neurocognitive2 
 Total Index 105.5 13.5 107.3 13.6 107.4 13.3 .761 0.13 0.15 0.01 
 NAB Attention Index 105.7 13.8 104.5 10.3 102.1 13.6 .455 0.10 0.26 0.19 
 NAB Memory Index 101.5 13.2 103.1 13.4 104.7 11.9 .526 0.12 0.25 0.13 
 NAB Language Index* 100.1 14.8 108.6 22.8 105.7 18.8 .138 0.50 0.34 0.14 
 NAB Spatial Index* 107.5 16.0 107.3 15.2 110.0 14.2 .714 0.01 0.16 0.19 
 NAB Executive Index* 106.9 15.9 104.5 14.9 106.8 13.7 .810 0.16 0.01 0.16 

Notes: N = 108 (mild TBI PCS-absent, n = 52; mild TBI PCS-present, n = 20; TC, n = 36); *Cohen's (1988) effect size (d): small (0.20), medium (0.50), and large (0.80). BAI = Beck Anxiety Inventory; BDI-II = Beck Depression Inventory-Second Edition; BC-PSI = British Columbia Postconcussion Inventory; NAB = Neuropsychological Assessment Battery.

1Group comparisons for the neurobehavioral measures were undertaken using Kruskal–Wallis H tests (across three groups) and Mann–Whitney U tests (pairwise comparisons) due to non-normal distribution of scores. 2Group comparisons for the neurocognitive measures were undertaken using ANOVA.

Significant pairwise comparisons: ap < .05; bp < .01; cp ≤ .001; Data not reported because comparison is not meaningful. This variable was used for group classification. *Prorated Index scores.

For the neurocognitive variables, there were no significant main effects (using ANOVA) for the NAB Total Index (p = .761) or for the five NAB Indexes (range: p = .138–.810). Pairwise comparisons revealed no significant differences between groups, though there were a number of meaningful effect sizes. There was a trend for the PCS-present group to have better scores on the Language Index compared with the PCS-absent group (d = 0.50, medium effect size). There was a trend for the PCS-absent group to have better scores on the Attention, Memory, and Language Indexes compared with the TC group (d = 0.25–0.34, small–medium effect sizes).

Comparison of the prevalence of the number of low neurocognitive scores was undertaken by considering all 23 individual NAB measures simultaneously. The cumulative percentages of the number of low scores (using <16th and <10th percentile as cutoff scores) by group are presented in Table 4. Using χ2 analyses, there were no significant differences in the percentage of patients that had multiple low scores across groups. However, there were a number of notable comparisons that approached significance. Using a <16th percentile cutoff score: (a) 15.0% of PCS-present group had 3 or more low scores, compared with 38.5% of PCS-absent group (p = .056), and 33.3% of TC group; and (b) 40.0% of PCS-present group had 2 or more low scores, compared with 66.7% of TC group (p = .053), and 57.7% of PCS-absent group. Therefore, there was a trend for the PCS-present group to perform better across the battery of neuropsychological tests than the other two groups.

Table 4.

Base rate of low neurocognitive scores by group

Number of Low Scores <16th percentile
 
<10th percentile
 
1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1 versus 2 1 versus 3 2 versus 3 1 versus 2 1 versus 3 2 versus 3 
9 low scores — — 2.8 — 2.8 2.8 — — — — — — 
8 or more 1.9 — 2.8 1.9 0.9 2.8 — — 2.8 — 2.8 2.8 
7 or more 3.8 5.0 2.8 1.2 1.0 2.2 — 5.0 2.8 5.0 2.8 2.2 
6 or more 5.8 5.0 2.8 0.8 3.0 2.2 — 5.0 2.8 5.0 2.8 2.2 
5 or more 13.5 10.0 11.1 3.5 2.4 1.1 1.9 5.0 2.8 3.1 0.9 2.2 
4 or more 28.8 10.0 25.0 18.8 3.8 15.0 9.6 5.0 8.3 4.6 1.3 3.3 
3 or more 38.5 15.0 33.3 23.5 5.2 18.3 23.1 10.0 16.7 13.1 6.4 6.7 
2 or more 57.7 40.0 66.7 17.7 9.0 26.7 46.2 25.0 38.9 21.2 7.3 13.9 
1 or more 76.9 85.0 77.8 8.1 0.9 7.2 65.4 55.0 58.3 10.4 7.1 3.3 
0 scores 100 100 100 — — — 100 100 100 — — — 
Number of Low Scores <16th percentile
 
<10th percentile
 
1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1 versus 2 1 versus 3 2 versus 3 1 versus 2 1 versus 3 2 versus 3 
9 low scores — — 2.8 — 2.8 2.8 — — — — — — 
8 or more 1.9 — 2.8 1.9 0.9 2.8 — — 2.8 — 2.8 2.8 
7 or more 3.8 5.0 2.8 1.2 1.0 2.2 — 5.0 2.8 5.0 2.8 2.2 
6 or more 5.8 5.0 2.8 0.8 3.0 2.2 — 5.0 2.8 5.0 2.8 2.2 
5 or more 13.5 10.0 11.1 3.5 2.4 1.1 1.9 5.0 2.8 3.1 0.9 2.2 
4 or more 28.8 10.0 25.0 18.8 3.8 15.0 9.6 5.0 8.3 4.6 1.3 3.3 
3 or more 38.5 15.0 33.3 23.5 5.2 18.3 23.1 10.0 16.7 13.1 6.4 6.7 
2 or more 57.7 40.0 66.7 17.7 9.0 26.7 46.2 25.0 38.9 21.2 7.3 13.9 
1 or more 76.9 85.0 77.8 8.1 0.9 7.2 65.4 55.0 58.3 10.4 7.1 3.3 
0 scores 100 100 100 — — — 100 100 100 — — — 

Notes: N = 108 (mild TBI PCS-absent, n = 52; mild TBI PCS-present, n = 20; TC, n = 36).

Diffusion Tensor Imaging

Tract-based spatial statistics analysis

The TC group was compared with the PCS-absent group on DTI measures of FA, MD, AD, and RD using TBSS. There were no statistically significant differences (using t-tests) between the groups in FA, MD, AD, or RD at the p < .05 level.

TBSS comparisons between the PCS-present group and the TC group revealed multiple areas of significant differences (using t-tests) in MD and RD. RD was significantly increased (p < .05) in the PCS-present group in widespread areas including the genu, body, and splenium of the corpus callosum; bilateral anterior, superior, and posterior corona radiata; bilateral superior longitudinal fasciculus; bilateral cingulum; bilateral posterior thalamic radiations; bilateral anterior limb of the internal capsule; and bilateral external capsule (see Fig. 1). MD values were also significantly increased (p < .05) in the PCS-present group in the genu and body of the corpus callosum, bilateral superior corona radiata, right anterior corona radiata, right anterior and posterior limb of the internal capsule, and the right superior longitudinal fasciculus (see Fig. 2). No significant differences were observed for AD values, although a trend (p < .10) of decreased FA in the PCS+ group was noted in multiple regions (data not shown).

Fig. 1.

Tract-based spatial statistics lightbox comparing MTBI-PCS-present patients and controls on RD. Voxels highlighted in yellow/orange indicated areas of significant increase in RD at the p< .05 level in the MTBI-PCS-present patients. Neuroanatomical areas where significant increases were noted include the genu, body, and splenium of the corpus callosum, bilateral anterior, superior and posterior corona radiata, bilateral superior longitudinal fasciculus, bilateral cingulum, bilateral posterior thalamic radiations, bilateral anterior limb of the internal capsule, and bilateral external capsule.

Fig. 1.

Tract-based spatial statistics lightbox comparing MTBI-PCS-present patients and controls on RD. Voxels highlighted in yellow/orange indicated areas of significant increase in RD at the p< .05 level in the MTBI-PCS-present patients. Neuroanatomical areas where significant increases were noted include the genu, body, and splenium of the corpus callosum, bilateral anterior, superior and posterior corona radiata, bilateral superior longitudinal fasciculus, bilateral cingulum, bilateral posterior thalamic radiations, bilateral anterior limb of the internal capsule, and bilateral external capsule.

Fig. 2.

Tract-based spatial statistics lightbox comparing MTBI-PCS-present patients and controls on MD. Voxels highlighted in yellow/orange indicated areas of significant increase in MD at the p < .05 level in the MTI-PCS-present group. Neuroanatomical areas where significant increases were observed included genu and body of the corpus callosum, bilateral superior corona radiata, right anterior corona radiata, right anterior and posterior limb of the internal capsule and the right superior longitudinal fasciculus.

Fig. 2.

Tract-based spatial statistics lightbox comparing MTBI-PCS-present patients and controls on MD. Voxels highlighted in yellow/orange indicated areas of significant increase in MD at the p < .05 level in the MTI-PCS-present group. Neuroanatomical areas where significant increases were observed included genu and body of the corpus callosum, bilateral superior corona radiata, right anterior corona radiata, right anterior and posterior limb of the internal capsule and the right superior longitudinal fasciculus.

The PCS-absent group was then compared with the PCS-present group on the four DTI indices. There were no statistically significant differences in FA, MD, AD, or RD. In summary, TBSS revealed significant white-matter differences between the PCS-present group and the TC group. There were no significant differences between the PCS-absent group and the TC group, or between the two MTBI groups.

Region of interest analyses

An atlas-based multivariate ROI analysis, based on 48 regions, was conducted by determining the number of abnormal areas of white matter in each individual as defined by mean DTI values that were >2 SDs from the means of the TC group. Descriptive statistics, group comparisons, and Cohen's effect sizes (Cohen, 1988) for the number of abnormal FA, AD, RD, and MD brain regions, by group, are presented in Table 5. There were no significant main effects (Kruskal–Wallis H tests) for the number of regions with abnormal FA, AD, RD, or MD across the three groups. However, pairwise comparisons (Mann–Whitney U tests) revealed that participants in the PCS-present group had a statistically (p < .05) greater number of locations with abnormal RD and MD compared with the TC group (d = 0.36 [small–medium effect size] and d = 0.56 [medium effect size], respectively). Although not statistically significant (p = 0.056), there was also a medium effect size (d = 0.45) found for the number of regions with abnormal FA between these two groups (i.e., PCS-present >TC). Similarly, there was a medium effect size (d = 0.49) found for the number of regions of abnormal MD between the TC group and the PCS-absent group. There was a nonsignificant trend for the PCS-absent group to have a greater number of abnormal MD regions compared with the TC group (p = .073). There were no significant differences when comparing the number of abnormal FA, AD, RD, or MD scores across the PCS-present and PCS-absent groups. The largest effect sizes were found for FA and AD (d = 0.25 and 0.23, respectively) between these groups.

Table 5.

Descriptive statistics, group comparisons1, and effect sizes for the number of abnormal ROIs

 1. Mild TBI: PCS-absent
 
2. Mild TBI: PCS-present
 
3. Trauma control
 
 Effect size
 
M SD M SD M SD p 1 versus 2 1 versus 3 2 versus 3 
Fractional anisotropy 1.5 2.7 2.2 2.9 1.1 2.3 .153 0.25 0.17 0.45 
Axial diffusivity 1.4 2.1 1.0 1.5 0.9 1.4 .812 0.23 0.25 0.00 
Radial diffusivity 2.1 4.0 2.7 4.9 1.3 3.3 .064 0.13 0.21 0.36* 
Mean diffusivity 2.0 3.3 2.2 4.5 0.7 1.5 .084 0.05 0.49 0.56* 
 1. Mild TBI: PCS-absent
 
2. Mild TBI: PCS-present
 
3. Trauma control
 
 Effect size
 
M SD M SD M SD p 1 versus 2 1 versus 3 2 versus 3 
Fractional anisotropy 1.5 2.7 2.2 2.9 1.1 2.3 .153 0.25 0.17 0.45 
Axial diffusivity 1.4 2.1 1.0 1.5 0.9 1.4 .812 0.23 0.25 0.00 
Radial diffusivity 2.1 4.0 2.7 4.9 1.3 3.3 .064 0.13 0.21 0.36* 
Mean diffusivity 2.0 3.3 2.2 4.5 0.7 1.5 .084 0.05 0.49 0.56* 

Notes: N = 108 (mild TBI PCS-absent, n = 52; mild TBI PCS-present, n = 20; TC, n = 36); *Cohen's (1988) effect size (d): small (0.20), medium (0.50), and large (0.80). ROI = region of interest; TBI = traumatic brain injury.

1Group comparisons for the neurobehavioral and neuroimaging measures were undertaken using Kruskal–Wallis H tests (across three groups) and Mann–Whitney U tests (pairwise comparisons) due to non-normal distribution of scores: *p < .05.

The cumulative percentages of the number of abnormal FA, AD, RD, and MD regions by group are presented in Tables 6 and 7. Using χ2 analyses, there were significant pairwise comparisons when comparing the PCS-present and TC groups on FA, RD, and MD, but not AD. For FA, the proportion of the PCS-present group that had two or more (50.0%) abnormal areas was significantly greater (p = .033) compared with the TC group (22.2%). Similarly, the proportion of the PCS-present group that had three or more (35.0%) abnormal areas was significantly greater (p = .041; Fisher's exact test) compared with the TC group (11.1%). For RD, there was a statistically greater proportion of the PCS-present group that had one or more abnormal areas (65.0%, p = .038) and two or more abnormal areas (50.0%, p = .008) when compared with the TC group (36.1% and 16.7%, respectively). For MD, the proportion of the PCS-present group that had one or more (60.0%) abnormal brain regions was significantly greater (p = .032) compared with the TC group (30.6%).

Table 6.

Base rate of the number of abnormal DTI-ROI scores by group—FA and AD

Number of low scores Fractional anisotropy
 
Axial diffusivity
 
1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1 versus 2 1 versus 3 2 versus 3 1 versus 2 1 versus 3 2 versus 3 
15 or morea       — — — — — — 
14 or more 1.9 — — — — — — — — — — — 
13 or more 1.9 — — — — — — — — — — — 
12 or more 1.9 — 2.8 — 0.9 — — — — — — — 
11 or more 1.9 5.0 2.8 3.1 0.9 2.2 1.9 — — — — — 
10 or more 1.9 5.0 2.8 3.1 0.9 2.2 1.9 — — — — — 
9 or more 3.8 5.0 2.8 1.2 1.0 2.2 1.9 — — — — — 
8 or more 5.8 5.0 2.8 0.8 3.0 2.2 1.9 — — — — — 
7 or more 7.7 10.0 2.8 2.3 4.9 7.2 1.9 — — — — — 
6 or more 7.7 10.0 2.8 2.3 4.9 7.2 1.9 5.0 — 3.1 — — 
5 or more 9.6 15.0 8.3 5.4 1.3 6.7 9.6 5.0 5.6 4.6 4.0 0.6 
4 or more 13.5 20.0 8.3 6.5 5.2 11.7 17.3 5.0 8.3 12.3 9.0 3.3 
3 or more 15.4 35.0 11.1 19.6 4.3 23.9 23.1 10.0 8.3 13.1 14.8 1.7 
2 or more 32.7 50.0 22.2 17.3 10.5 27.8 30.8 25.0 25.0 5.8 5.8 
1 or more 46.2 60.0 38.9 13.8 7.3 21.1 48.1 45.0 47.2 3.1 0.9 2.2 
0 scores 100.0 100.0 100.0 100.0 100.0 100.0 
Number of low scores Fractional anisotropy
 
Axial diffusivity
 
1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1 versus 2 1 versus 3 2 versus 3 1 versus 2 1 versus 3 2 versus 3 
15 or morea       — — — — — — 
14 or more 1.9 — — — — — — — — — — — 
13 or more 1.9 — — — — — — — — — — — 
12 or more 1.9 — 2.8 — 0.9 — — — — — — — 
11 or more 1.9 5.0 2.8 3.1 0.9 2.2 1.9 — — — — — 
10 or more 1.9 5.0 2.8 3.1 0.9 2.2 1.9 — — — — — 
9 or more 3.8 5.0 2.8 1.2 1.0 2.2 1.9 — — — — — 
8 or more 5.8 5.0 2.8 0.8 3.0 2.2 1.9 — — — — — 
7 or more 7.7 10.0 2.8 2.3 4.9 7.2 1.9 — — — — — 
6 or more 7.7 10.0 2.8 2.3 4.9 7.2 1.9 5.0 — 3.1 — — 
5 or more 9.6 15.0 8.3 5.4 1.3 6.7 9.6 5.0 5.6 4.6 4.0 0.6 
4 or more 13.5 20.0 8.3 6.5 5.2 11.7 17.3 5.0 8.3 12.3 9.0 3.3 
3 or more 15.4 35.0 11.1 19.6 4.3 23.9 23.1 10.0 8.3 13.1 14.8 1.7 
2 or more 32.7 50.0 22.2 17.3 10.5 27.8 30.8 25.0 25.0 5.8 5.8 
1 or more 46.2 60.0 38.9 13.8 7.3 21.1 48.1 45.0 47.2 3.1 0.9 2.2 
0 scores 100.0 100.0 100.0 100.0 100.0 100.0 

Notes: N = 108 (mild TBI PCS-absent, n = 52; mild TBI PCS-present, n = 20; TC, n = 36).

aTable has been limited to 15 or more scores.

Table 7.

Base rate of the number of abnormal DTI-ROI scores by group—RD and MD

Number of low scores Radial diffusivity
 
Mean diffusivity
 
 1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1 versus 2 1 versus 3 2 versus 3 1 verusu 2 1 versus 3 2 versus 3 
15 or morea 1.9 5.0 2.8 3.1 0.9 2.2 1.9 5.0 — 3.1 1.9 — 
14 or more 1.9 5.0 2.8 3.1 0.9 2.2 1.9 5.0 — 3.1 1.9 — 
13 or more 1.9 5.0 2.8 3.1 0.9 2.2 1.9 5.0 — 3.1 1.9 — 
12 or more 1.9 5.0 2.8 3.1 0.9 2.2 1.9 5.0 — 3.1 1.9 — 
11 or more 5.8 5.0 2.8 0.8 3.0 2.2 3.8 5.0 — 1.2 3.8 — 
10 or more 5.8 5.0 5.6 0.8 0.2 0.6 3.8 5.0 — 1.2 3.8 — 
9 or more 5.8 5.0 5.6 0.8 0.2 0.6 5.8 5.0 — 0.8 5.8 — 
8 or more 9.6 5.0 5.6 4.6 4.0 0.6 7.7 5.0 — 2.7 7.7 — 
7 or more 9.6 5.0 5.6 4.6 4.0 0.6 7.7 5.0 2.8 2.7 4.9 2.2 
6 or more 11.5 10.0 5.6 1.5 5.9 4.4 13.5 10.0 2.8 3.5 10.7 7.2 
5 or more 15.4 15.0 8.3 0.4 7.1 6.7 19.2 10.0 2.8 9.2 16.4 7.2 
4 or more 17.3 15.0 8.3 2.3 9.0 6.7 23.1 10.0 5.6 13.1 17.5 4.4 
3 or more 19.2 30.0 11.1 10.8 8.1 18.9 25.0 15.0 11.1 10.0 13.9 3.9 
2 or more 34.6 50.0 16.7 15.4 17.9 33.3 32.7 40.0 16.7 7.3 16.0 23.3 
1 or more 50.0 65.0 36.1 15.0 13.9 28.9 46.2 60.0 30.6 13.8 15.6 29.4 
0 scores 100.0 100.0 100.0 100.0 100.0 100.0 
Number of low scores Radial diffusivity
 
Mean diffusivity
 
 1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1. MTBI PCS-absent (%) 2. MTBI PCS-present (%) 3. Trauma ctrl (%) Percent diff. (%)
 
1 versus 2 1 versus 3 2 versus 3 1 verusu 2 1 versus 3 2 versus 3 
15 or morea 1.9 5.0 2.8 3.1 0.9 2.2 1.9 5.0 — 3.1 1.9 — 
14 or more 1.9 5.0 2.8 3.1 0.9 2.2 1.9 5.0 — 3.1 1.9 — 
13 or more 1.9 5.0 2.8 3.1 0.9 2.2 1.9 5.0 — 3.1 1.9 — 
12 or more 1.9 5.0 2.8 3.1 0.9 2.2 1.9 5.0 — 3.1 1.9 — 
11 or more 5.8 5.0 2.8 0.8 3.0 2.2 3.8 5.0 — 1.2 3.8 — 
10 or more 5.8 5.0 5.6 0.8 0.2 0.6 3.8 5.0 — 1.2 3.8 — 
9 or more 5.8 5.0 5.6 0.8 0.2 0.6 5.8 5.0 — 0.8 5.8 — 
8 or more 9.6 5.0 5.6 4.6 4.0 0.6 7.7 5.0 — 2.7 7.7 — 
7 or more 9.6 5.0 5.6 4.6 4.0 0.6 7.7 5.0 2.8 2.7 4.9 2.2 
6 or more 11.5 10.0 5.6 1.5 5.9 4.4 13.5 10.0 2.8 3.5 10.7 7.2 
5 or more 15.4 15.0 8.3 0.4 7.1 6.7 19.2 10.0 2.8 9.2 16.4 7.2 
4 or more 17.3 15.0 8.3 2.3 9.0 6.7 23.1 10.0 5.6 13.1 17.5 4.4 
3 or more 19.2 30.0 11.1 10.8 8.1 18.9 25.0 15.0 11.1 10.0 13.9 3.9 
2 or more 34.6 50.0 16.7 15.4 17.9 33.3 32.7 40.0 16.7 7.3 16.0 23.3 
1 or more 50.0 65.0 36.1 15.0 13.9 28.9 46.2 60.0 30.6 13.8 15.6 29.4 
0 scores 100.0 100.0 100.0 100.0 100.0 100.0 

Notes: N = 108 (mild TBI PCS-absent, n = 52; mild TBI PCS-present, n = 20; TC, n = 36)

aTable has been limited to 15 or more scores.

There were some significant differences when comparing the PCS-absent and TC groups on MD, but not FA, RD, and AD. There was a statistically greater proportion of the PCS-absent group that had four or more abnormal areas on MD (23.1%, p = .037; Fisher's exact test) and five or more abnormal areas on MD (19.2%, p = .024; Fisher's exact test) compared with the TC group (5.6% and 2.8%, respectively). There was a nonsignificant trend (p = .063) towards a greater proportion of the PCS-absent group to have two or more (34.6%) brain areas with abnormal RD compared with the TC group (16.7%).

There were no significant differences when comparing the PCS-present and PCS-absent groups for FA, RD, MD, and AD. There was, however, a nonsignificant trend (p = .066) towards a greater proportion of the PCS-present group to have three or more (35.0%) abnormal areas on FA compared with the PCS-absent group (15.4%).

Extended region of interest analyses

Based on the results of the TBSS analyses, further exploratory multivariate ROI analyses were undertaken using only those DTI metrics and ROIs that were identified as significantly different between some groups (i.e., 19 ROIs for RD [see Fig. 1], and 8 ROIs for MD [see Fig. 2]). The number of abnormal areas of white matter was calculated for each individual across the select ROIs and was compared across the three groups. There were significant main effects (Kruskal–Wallis H tests) for the number of abnormal RD regions (p = .029), but not MD regions (p = .131). Pairwise comparisons (Mann–Whitney U tests) revealed that participants in both the MTBI groups had a significantly greater (p < .05) number of abnormal RD regions (i.e., PCS-absent: M = 0.9, SD = 2.1; PCS-present: M = 1.5, SD = 3.6) compared with the TC group (M = 0.5, SD = 1.8; d = 0.22 and 0.39, respectively). There were no significant differences between the PCS-present and PCS-absent groups (p > .05, d = 0.20, small effect size).

Examination of the cumulative percentages of the number of brain regions with abnormal RD and MD by group (data not shown; available on request from RTL) revealed no significant pairwise differences across the three groups for the percentage of patients that had multiple abnormal MD or RD regions. The one exception to this was for RD. The proportion of the PCS-present group (45.0%) and PCS-absent group (30.8%) that had one or more abnormal RD regions was significantly higher (p < .05) compared with the TC group (11.1%). There were no significant differences between the PCS-present and PCS-absent groups.

Prediction of PCS-Present versus PCS-Absent Groups

To examine which factors provide the most unique contribution towards the prediction of postconcussion symptom reporting, logistic regression analyses were undertaken to determine if select factors could identify participants in the PCS-present versus PCS-absent groups. Factors were selected were as follows: (i) gender, (ii) education, (iii) BDI-II 10-item total score, (iv) BAI total score, and (v) the total number of abnormal FA and MD scores combined. For the logistic regression analysis, all five factors were entered simultaneously as the independent variables and group membership as the dependent variable.

Only two of the six variables were significant predictors of group membership (depression [p = .043] and BAI total [p = .043]) accounting for 52.1% of the variance (Nagelkerke R2; p < .001). Overall, 83.3% of participants were correctly classified using depression and BAI total scores alone. However, classification rates were substantially higher when predicting the PCS-absent group (94.2%) compared with the PCS-present group (55.0%). No DTI metrics entered the prediction model.

Discussion

The purpose of this study was to examine the relationship between DTI and postconcussion symptom reporting following MTBI. There were three hypotheses. First, compared with the TC group, patients in both the MTBI groups would (a) report a greater number of depression, anxiety, and postconcussion symptoms; (b) perform comparably on neurocognitive measures; and (c) have reduced white-matter integrity in the brain as measured by DTI metrics. Second, in the MTBI group, patients who met ICD-10 criteria for PCS would report a greater number of depression and anxiety symptoms compared with those who did not meet criteria for PCS. Third, in the MTBI group, there would be no differences across all neurocognitive measures and DTI metrics when comparing those patients who are classified as PCS-present versus PCS-absent.

MTBI versus Trauma Controls

Consistent with the first hypothesis, compared with the TC group, the PCS-present group (a) reported a greater number of depression, anxiety, and postconcussion symptoms; (b) had multiple regions of the brain with significant differences in MD and RD using TBSS (but not AD or FA); and (c) had a greater number of brain regions with abnormal RD and MD (but not AD or FA) using ROI analyses. In contrast, these differences were not found when comparing the PCS-absent and TC groups. Although the PCS-absent group had a greater number of abnormal MD regions compared with the TC group, there were no group differences when comparing (a) all DTI measures using TBSS, (b) FA, AD, and RD measures using ROI analyses, and (c) measures of depression, anxiety, and postconcussion symptoms.

Fractional anisotropy is a representation of the directional coherence (anisotropy) of water diffusion, and MD represents the average total diffusion of water within a particular voxel, irrespective of direction. Mean diffusivity values generally increase, and FA values decrease in response to white-matter injury (Hulkower, Poliak, Rosenbaum, Zimmerman, & Lipton, 2013). Radial diffusivity represents the diffusion of water perpendicular to the axon, and thus is thought to be sensitive to processes such as demyelination (Mac Donald, Dikranian, Bayly, Holtzman, & Brody, 2007). AD represents the diffusivity parallel to the axon and hence changes in AD are thought to indicate axonal compromise (Mac Donald et al., 2007). Therefore, an explanation for our results is that the PCS+ group showed evidence of white-matter compromise (increased MD) mainly via a membrane integrity mechanism (increased RD). This is consistent with other research showing increased RD in the subacute period (Hulkower et al., 2013; Mac Donald et al., 2007).

TBSS revealed a nonsignificant trend to lower FA values in the PCS+ groups. Given that increases in mean and RD generally result in lower FA (because they are interrelated mathematical derivations), the nonsignificant findings for FA values could result from insufficient sample size, or simply that MD is a more sensitive measure of white-matter integrity during the subacute period, as has been observed in other studies (Hulkower et al., 2013).

It is possible that some of the differences in DTI metrics in this study between the PCS-present and TC groups could be related, at least in part, to the influence of depression and anxiety on physiology, rather than differences in the microstructure of white-matter attributable to differences in biomechanical forces applied to the brain. Numerous studies have shown that depression and other affective disorders are associated with white-matter abnormalities as identified by DTI. In a systematic review of DTI studies in affective disorders, Sexton, Mackay, and Ebmeier (2009) concluded that “21 of 27 studies found significantly lower [fractional] anisotropy in subjects with affective disorders compared to control subjects” (p. 814).

Nonetheless, the DTI results in this study are relatively consistent with some past studies that have compared controls with MTBI groups divided into PCS-present/absent groups. For example, Messé et al. (2011) found that MTBI patients who were classified as having persistent PCS from <1 month to 3 months postinjury (but not those whose symptoms had improved) had higher MD scores compared with a healthy controls in some ROIs (i.e., superior longitudinal fasciculus, corticospinal tract bilaterally, and left anterior thalamic radiation). In another study, Messe et al. (2012) reported that compared with healthy controls, MTBI patients who met criteria for ICD-10 PCS <1 month and 6 months postinjury had a greater number of abnormal brain regions compared with those who did not meet ICD-10 PCS criteria.

Also consistent with the first hypothesis, there were no significant differences on the neurocognitive measures when comparing the TC group with both MTBI groups. This finding was not surprising or unexpected. Patients with MTBI perform poorly on neuropsychological tests in the initial days (e.g., Hughes et al., 2004; Lovell, Collins, Iverson, Johnston, & Bradley, 2004) and up to the first month following the injury (e.g., Levin et al., 1987; Ponsford et al., 2000), but cognitive deficits typically are not seen after 1–3 months (e.g., Gentilini et al., 1985; Ponsford et al., 2000) in prospective group studies. This has been illustrated repeatedly in reviews and meta-analyses (e.g., McCrea et al., 2009; Rohling, Larrabee, & Millis, 2012). In our study, we examined patients at 6–8 weeks postinjury. Although this time period falls in the middle of when we might expect most, but certainly not all, patients to have normal performance on neuropsychological tests; it is our expectation that there would be no measurable group differences on neuropsychological testing at this time.

PCS-Present versus PCS-Absent

In support of the second hypothesis, those individuals in the MTBI group who were classified as PCS-present at 6–8 weeks postinjury reported a greater number of depression and anxiety symptoms on the BAI and BDI-II compared with those patients who were classified as PCS-absent. The effect sizes were very large (d = 1.63 and 1.89, respectively). Of course, the relationship between postconcussion symptoms and depression and anxiety found here is not new. Many researchers have reported a strong relation between postconcussion symptoms and other self-reported mental health problems. In a military context, the relation between postconcussion symptoms and depression or PTSD symptoms following MTBI is well established (e.g., Lange, Brickel, et al., 2013; Lange, Shewchuk, et al., 2013; Lange et al., 2014). Similarly, in a civilian context, the relationship between postconcussion symptoms and depression and anxiety following MTBI is also well established (Bryant & Harvey, 1999; Garden & Sullivan, 2010).

Consistent with the third hypothesis, there were no significant differences between those patients in the MTBI group that were PCS-present versus PCS-absent across all neurocognitive measures and DTI metrics. The DTI results reported in this study are consistent with some research that has also found no group differences in DTI metrics when an MTBI sample is divided into ICD-10 PCS-present versus PCS-absent groups at 3-week postinjury (Waljas et al., 2014) and 6–8 weeks postinjury (Lange et al., 2012). Of course, these results are not consistent with other studies that have reported differences between MTBI groups when divided into PCS-present versus PCS-absent groups at 3 months (Messé et al., 2011) and 6 months (Hartikainen et al., 2010; Messe et al., 2012) postinjury. Collectively, based on these results, one might argue that differences between PCS-present versus PCS-absent groups are more likely to be detectable after 3 months following injury. However, contrary to this, in a longitudinal study by Messe et al. (2012), these authors concluded that there were “significant structural differences” found between PCS-present versus PCS-absent groups at <1 month postinjury, but less at 6-month postinjury.

Limitations

This study has several methodological limitations. First, TBSS is a conservative DTI approach that we chose because we had no a priori hypothesis that some regions would be affected over others. Other approaches that interrogate specific regions, such as user-defined track selection or probabilistic tractography may indeed be more sensitive than a full brain analysis, if driven by a well-informed hypothesis. Second, this study does not address medium- or long-term postconcussion symptom reporting, nor did it address the course of recovery following MTBI. Longitudinal studies would assist in understanding the role of microstructural white-matter abnormality on the short-, medium-, and long-term outcome from MTBI. There are no such studies reported in adults to date. Third, although the focus of this study was on the relationship between postconcussion symptom reporting and DTI, it is not possible to exclude the influence of other factors that may have had a contributory role (e.g., persistent pain [Smith-Seemiller, Fow, Kant, & Franzen, 2003], “good old days” bias [Gunstad & Suhr, 2001], diagnosis threat [Suhr & Gunstad, 2002a], etc.). However, we did include TC patients (rather than a healthy control group) because these patients tend to share more preinjury characteristics to their head-injured counterparts (Dikmen, Machamer, Winn, & Temkin, 1995). As such, the potential bias of these factors influencing the results in one group, but not the other, is mitigated. Fourth, there were a greater proportion of women in this sample who met ICD-10 criteria for PCS compared with men. Nearly, 1 in 2 women met symptom criteria for PCS compared with 1 in 5 men. This is consistent with research that has found increased postconcussion symptom reporting in women compared with men (e.g., Meares et al., 2008). It is possible that sex differences influenced the results of this study. Fifth, it would be a mistake to assume that the results from this study generalize to all people who sustain an MTBI. Our patients were recruited from a Level 1 trauma center. The majority of people who sustain an MTBI in daily life are not evaluated in the Emergency Department of a hospital (Sosin, Sniezek, & Thurman, 1996). Our patients had clear evidence of an MTBI. The vast majority of our patients underwent day-of-injury CT scanning and a substantial minority had a day-of-injury intracranial abnormality. Therefore, the patient sample used in this study represents a minority of the more seriously injured MTBI population.

Conclusion

Following an MTBI, there is no doubt that many people experience a range of postconcussion symptoms (e.g., headaches, dizziness, fatigue, and cognitive difficulties). These symptoms typically resolve within the first few days, weeks, or months for most people. However, for some people, these symptoms are reported many months or years postinjury. In this study, symptoms of depression and anxiety differentiated patients with MTBIs who met criteria for the postconcussion symptom versus those who did not. In contrast, these groups did not differ on diverse metrics of DTI. These findings once again highlight that there are many factors that can potentially account for postconcussion symptom reporting, even in the first 6–8 weeks of recovery. It has been recommended that postconcussion symptom reporting be viewed from a biopsychosocial perspective (Iverson et al., 2013). That is, there are a diverse range of biopsychosocial factors (i.e., pre-, peri-, and postinjury factors) that can influence the perception and reporting of symptoms long after an MTBI.

Funding

This work was primarily supported by the Canadian Institutes of Health Research (200903MOP-200377-BSB-CAAA-161276 to R.L., G.I., J.B., and M.H.); and in part by the Alcohol Beverage Medical Research Foundation (to R.L., G.I., J.B., and M.H.). GLI and MES note that they were supported in part by the INTRuST Posttraumatic Stress Disorder and Traumatic Brain Injury Clinical Consortium funded by the Department of Defense Psychological Health/Traumatic Brain Injury Research Program (X81XWH-07-CC-CSDoD). MES further notes that she was supported in part by the Veterans Affairs Merit Review Program. The views expressed in this article are those of the authors and do not reflect the official policy of the Department of Defense or U.S. Government. GLI has received past research funding from several test publishing companies, including ImPACT Applications, Inc., CNS Vital Signs, and Psychological Assessment Resources (PAR, Inc.). He receives royalties for two books in neuropsychology and one test (WCST-64). He has a clinical practice in forensic neuropsychology involving individuals who have sustained mild TBIs (including athletes). He is a co-investigator, collaborator, or consultant on grants relating to mild TBI funded by several organizations.

Conflict of Interest

None declared.

Acknowledgments

Portions of these data were presented at the National Academy of Neuropsychology annual conference, San Diego, October 2013. The authors thank Jan Buchanen, Liz Holland, Lisa Casagrande Hoshino, and Angela Aquino for assistance with patient recruitment and testing; and the staff at the University of British Columbia MRI center for the provision of neuroimaging services. We also thank Burkhard Madler, PhD for assistance with choosing the initial image acquisition parameters and with past DTI metrics.

References

American Psychiatric Association
Diagnostic and statistical manual of mental disorders
 
2000
Washington, DC
Author
(4th, text rev. ed.)
Bazarian
J. J.
Zhong
J.
Blyth
B.
Zhu
T.
Kavcic
V.
Peterson
D.
Diffusion tensor imaging detects clinically important axonal damage after mild traumatic brain injury: A pilot study
Journal of Neurotrauma
 
2007
24
9
1447
1459
Bazarian
J. J.
Zhu
T.
Blyth
B.
Borrino
A.
Zhong
J.
Subject-specific changes in brain white matter on diffusion tensor imaging after sports-related concussion
Magnetic Resonance Imaging
 
2012
30
2
171
180
Beck
A. T.
Steer
R. A.
Beck anxiety inventory manual
 
1993
San Antonio, TX
The Psychological Corporation
Beck
A. T.
Steer
R. A.
Brown
G. K.
Manual for the beck depression inventory-II (finnish version)
 
1996
San Antonio, TX
The Psychological Corporation
Belanger
H. G.
Vanderploeg
R. D.
Curtiss
G.
Warden
D. L.
Recent neuroimaging techniques in mild traumatic brain injury
Journal of Neuropsychiatry and Clinical Neuroscience
 
2007
19
1
5
20
Bruns
J. J.
Jr.
Jagoda
A. S.
Mild traumatic brain injury
Mt Sinai Journal of Medicine
 
2009
76
2
129
137
Bryant
R. A.
Harvey
A. G.
Postconcussive symptoms and posttraumatic stress disorder after mild traumatic brain injury
Journal of Nervous and Mental Disease
 
1999
187
5
302
305
Cohen
J.
Statistical power analysis for the behavioral sciences
 
1988
2nd ed.
Hillsdale, NJ
Lawrence Erlbaum Associates
Deepika
A.
Munivenkatappa
A.
Devi
B. I.
Shukla
D.
Does isolated traumatic subarachnoid hemorrhage affect outcome in patients with mild traumatic brain injury?
Journal of Head Trauma Rehabilitation
 
2013
28
6
442
445
de Guise
E.
Lepage
J. F.
Tinawi
S.
LeBlanc
J.
Dagher
J.
Lamoureux
J.
et al.  
Comprehensive clinical picture of patients with complicated vs uncomplicated mild traumatic brain injury
Archives of Clinical Neuropsychology
 
2010
24
7
1113
1130
Dikmen
S. S.
Machamer
J. E.
Winn
R.
Temkin
N. R.
Neuropsychological outcome 1-year post head injury
Neuropsychology
 
1995
9
80
90
Foa
E. B.
Cashman
L.
Jaycox
L.
Perry
K.
The validation of a self-report measure of posttraumatic stress disorder: The Posttraumatic Diagnostic Scale
Psychological Assessment
 
1997
9
4
445
451
Gaetz
M.
Weinberg
H.
Electrophysiological indices of persistent post-concussion symptoms
Brain Injury
 
2000
14
9
815
832
Garden
N.
Sullivan
K. A.
An examination of the base rates of post-concussion symptoms: The influence of demographics and depression
Applied Neuropsychology
 
2010
17
1
1
7
Gentilini
M.
Nichelli
P.
Schoenhuber
R.
Bortolotti
P.
Tonelli
L.
Falasca
A.
et al.  
Neuropsychological evaluation of mild head injury
Journal of Neurology, Neurosurgery and Psychiatry
 
1985
48
2
137
140
Gering
D. T.
Nabavi
A.
Kikinis
R.
Hata
N.
O'Donnell
L. J.
Grimson
W. E.
et al.  
An integrated visualization system for surgical planning and guidance using image fusion and an open MR
Journal of Magnetic Resonance Imaging
 
2001
13
6
967
975
Gunstad
J.
Suhr
J. A.
“Expectation as etiology” versus “the good old days:” Postconcussion syndrome symptom reporting in athletes, headache sufferers, and depressed individuals
Journal of the International Neuropsychological Society
 
2001
7
3
323
333
Hahn
R. A.
The nocebo phenomenon: Concept, evidence, and implications for public health
Preventive Medicine
 
1997
26
5 Pt 1
607
611
Hartikainen
K. M.
Waljas
M.
Isoviita
T.
Dastidar
P.
Liimatainen
S.
Solbakk
A. K.
et al.  
Persistent symptoms in mild to moderate traumatic brain injury associated with executive dysfunction
Journal of Clinical and Experimental Neuropsychology
 
2010
32
7
767
774
Hibbard
M. R.
Bogdany
J.
Uysal
S.
Kepler
K.
Silver
J. M.
Gordon
W. A.
et al.  
Axis II psychopathology in individuals with traumatic brain injury
Brain Injury
 
2000
14
1
45
61
Hofman
P. A.
Stapert
S. Z.
van Kroonenburgh
M. J.
Jolles
J.
de Kruijk
J.
Wilmink
J. T.
MR imaging, single-photon emission CT, and neurocognitive performance after mild traumatic brain injury
American Journal of Neuroradiology
 
2001
22
3
441
449
Hua
K.
Zhang
J.
Wakana
S.
Jiang
H.
Li
X.
Reich
D. S.
et al.  
Tract probability maps in stereotaxic spaces: Analyses of white matter anatomy and tract-specific quantification
Neuroimage
 
2008
39
1
336
347
Hughes
D. G.
Jackson
A.
Mason
D. L.
Berry
E.
Hollis
S.
Yates
D. W.
Abnormalities on magnetic resonance imaging seen acutely following mild traumatic brain injury: Correlation with neuropsychological tests and delayed recovery
Neuroradiology
 
2004
46
7
550
558
Hulkower
M. B.
Poliak
D. B.
Rosenbaum
S. B.
Zimmerman
M. E.
Lipton
M. L.
A decade of DTI in traumatic brain injury: 10 years and 100 articles later
American Journal of Neuroradiology
 
2013
34
11
2064
2074
Inglese
M.
Bomsztyk
E.
Gonen
O.
Mannon
L. J.
Grossman
R. I.
Rusinek
H.
Dilated perivascular spaces: Hallmarks of mild traumatic brain injury
American Journal of Neuroradiology
 
2005
26
4
719
724
Inglese
M.
Makani
S.
Johnson
G.
Cohen
B. A.
Silver
J. A.
Gonen
O.
et al.  
Diffuse axonal injury in mild traumatic brain injury: A diffusion tensor imaging study
Journal of Neurosurgery
 
2005
103
298
303
Iverson
G. L.
Misdiagnosis of the persistent postconcussion syndrome in patients with depression
Archives of Clinical Neuropsychology
 
2006
21
4
303
310
Iverson
G. L.
Brooks
B. L.
Ashton
V. L.
Lange
R. T.
Interview versus questionnaire symptom reporting in people with the postconcussion syndrome
Journal of Head Trauma Rehabilitation
 
2010
25
1
23
30
Iverson
G. L.
Hakulinen
U.
Waljas
M.
Dastidar
P.
Lange
R. T.
Soimakallio
S.
et al.  
To exclude or not to exclude: White matter hyperintensities in diffusion tensor imaging research
Brain Injury
 
2011
25
13–14
1325
1332
Iverson
G. L.
Lange
R. T.
Brooks
B. L.
Rennison
V. L.
“Good old days” bias following mild traumatic brain injury
The Clinical Neuropsychologist
 
2010
24
1
17
37
Iverson
G. L.
McCracken
L. M.
‘Postconcussive’ symptoms in persons with chronic pain
Brain Injury
 
1997
11
11
783
790
Iverson
G. L.
Silverberg
N.
Lange
R. T.
Zasler
N.
Zasler
N.
Katz
D.
Zafonte
R.
Arciniegas
D. B.
Bullock
M. R.
Kreutzer
J. S.
Conceptualizing outcome from mild traumatic brain injury
Brain injury medicine: Principles and practice
 
2013
2nd ed.
New York
Demos Medical Publishing
470
497
Iverson
G. L.
Zasler
N. D.
Lange
R. T.
Zasler
N. D.
Katz
D.
Zafonte
R. D.
Post-concussive disorder
Brain injury medicine: Principles and practice
 
2007
New York
Demos Medical Publishing
373
405
Kraus
M. F.
Susmaras
T.
Caughlin
B. P.
Walker
C. J.
Sweeney
J. A.
Little
D. M.
White matter integrity and cognition in chronic traumatic brain injury: A diffusion tensor imaging study
Brain
 
2007
130
10
2508
2519
Kumar
R.
Gupta
R. K.
Husain
M.
Chaudhry
C.
Srivastava
A.
Saksena
S.
et al.  
Comparative evaluation of corpus callosum DTI metrics in acute mild and moderate traumatic brain injury: Its correlation with neuropsychometric tests
Brain Injury
 
2009
23
7
675
685
Kurca
E.
Sivak
S.
Kucera
P.
Impaired cognitive functions in mild traumatic brain injury patients with normal and pathologic magnetic resonance imaging
Neuroradiology
 
2006
48
9
661
669
Landman
B. A.
Farrell
J. A.
Jones
C. K.
Smith
S. A.
Prince
J. L.
Mori
S.
Effects of diffusion weighting schemes on the reproducibility of DTI-derived fractional anisotropy, mean diffusivity, and principal eigenvector measurements at 1.5T
Neuroimage
 
2007
36
4
1123
1138
Landre
N.
Poppe
C. J.
Davis
N.
Schmaus
B.
Hobbs
S. E.
Cognitive functioning and postconcussive symptoms in trauma patients with and without mild TBI
Archives of Clinical Neuropsychology
 
2006
21
4
255
273
Lange
R. T.
Brickell
T.
French
L. M.
Ivins
B.
Bhagwat
A.
Pancholi
S.
et al.  
Risk factors for postconcussion symptom reporting after traumatic brain injury in U.S. military service members
Journal of Neurotrauma
 
2013
30
4
237
246
Lange
R. T.
Brickell
T. A.
Kennedy
J. E.
Bailie
J. M.
Sills
C.
Asmussen
S.
et al.  
Factors influencing postconcussion and posttraumatic stress symptom reporting following military-related concurrent polytrauma and traumatic brain injury
Archives of Clinical Neuropsychology
 
2014
29
4
329
347
Lange
R. T.
Iverson
G. L.
Brubacher
J. R.
Madler
B.
Heran
M. K.
Diffusion tensor imaging findings are not strongly associated with postconcussional disorder 2 months following mild traumatic brain injury
Journal of Head Trauma Rehabilitation
 
2012
27
3
188
198
Lange
R. T.
Shewchuk
J. R.
Heran
M. K.
Rauscher
A.
Jarrett
M.
Brubacher
J. R.
et al.  
To exclude or not to exclude: Further examination of the influence of white matter hyperintensities in diffusion tensor imaging research
Journal of Neurotrauma
 
2014
31
2
198
205
Lannsjo
M.
Backheden
M.
Johansson
U.
Af Geijerstam
J. L.
Borg
J.
Does head CT scan pathology predict outcome after mild traumatic brain injury?
European Journal of Neurology
 
2013
20
1
124
129
Lees-Haley
P. R.
Brown
R. S.
Neuropsychological complain base rates of 170 personal injury claimants
Archives of Clinical Neuropsychology
 
1993
8
3
203
209
Levin
H. S.
Mattis
S.
Ruff
R. M.
Eisenberg
H. M.
Marshall
L. F.
Tabaddor
K
et al.  
Neurobehavioral outcome following minor head injury: A three-center study
Journal of Neurosurgery
 
1987
66
2
234
243
Levin
H. S.
Wilde
E.
Troyanskaya
M.
Petersen
N.
Scheibel
R.
Newsome
M.
et al.  
Diffusion tensor imaging of mild to moderate blast-related traumatic brain injury and its sequelae
Journal of Neurotruma
 
2010
27
683
694
Lewine
J. D.
Davis
J. T.
Bigler
E. D.
Thoma
R.
Hill
D.
Funke
M.
et al.  
Objective documentation of traumatic brain injury subsequent to mild head trauma: Multimodal brain imaging with MEG, SPECT, and MRI
Journal of Head Trauma Rehabilitation
 
2007
22
3
141
155
Lipton
M. L.
Gellella
E.
Lo
C.
Gold
T.
Ardekani
B. A.
Shifteh
K.
et al.  
Multifocal white matter ultrastructural abnormalities in mild traumatic brain injury with cognitive disability: A voxel-wise analysis of diffusion tensor imaging
Journal of Neurotrauma
 
2008
25
11
1335
1342
Lipton
M. L.
Gulko
E.
Zimmerman
M. E.
Friedman
B. W.
Kim
M.
Gellella
E.
et al.  
Diffusion-tensor imaging implicates prefrontal axonal injury in executive function impairment following very mild traumatic brain injury
Radiology
 
2009
252
3
816
824
Lovell
M. R.
Collins
M. W.
Iverson
G. L.
Johnston
K. M.
Bradley
J. P.
Grade 1 or “ding” concussions in high school athletes
American Journal of Sports Medicine
 
2004
32
1
47
54
Mac Donald
C. L.
Dikranian
K.
Bayly
P.
Holtzman
D.
Brody
D.
Diffusion tensor imaging reliably detects experimental traumatic axonal injury and indicates approximate time of injury
The Journal of Neuroscience
 
2007
27
44
11869
11876
McCrea
M.
Iverson
G. L.
McAllister
T. W.
Hammeke
T. A.
Powell
M. R.
Barr
W. B.
et al.  
An integrated review of recovery after mild traumatic brain injury (MTBI): Implications for clinical management
Clinical Neuropsychology
 
2009
23
8
1368
1390
Meares
S.
Shores
E. A.
Taylor
A. J.
Batchelor
J.
Bryant
R. A.
Baguley
I. J.
et al.  
Mild traumatic brain injury does not predict acute postconcussion syndrome
Journal of Neurology, Neurosurgery and Psychiatry
 
2008
79
3
300
306
Messé
A.
Caplain
S.
Paradot
G.
Garrigue
D.
Mineo
J.-F.
Soto Ares
G.
et al.  
Diffusion tensor imaging and white matter lesions at the subacute stage in mild traumatic brain injury with persistent neurobehavioral impairment
Human Brain Mapping
 
2011
32
6
999
1011
Messe
A.
Caplain
S.
Pelegrini-Issac
M.
Blancho
S.
Montreuil
M.
Levy
R.
et al.  
Structural integrity and postconcussion syndrome in mild traumatic brain injury patients
Brain Imaging and Behavior
 
2012
6
2
283
292
Mickeviciene
D.
Schrader
H.
Obelieniene
D.
Surkiene
D.
Kunickas
R.
Stovner
L. J.
et al.  
A controlled prospective inception cohort study on the post-concussion syndrome outside the medicolegal context
European Journal of Neurology
 
2004
11
6
411
419
Mori
A.
Wakana
S.
Nagai-Poetscher
I. M.
va Zigl
P. C.
MRI atlas of human white matter
 
2005
Amsterdam
Elsevier
Niogi
S. N.
Mukherjee
P.
Ghajar
J.
Johnson
C.
Kolster
R. A.
Sarkar
R.
et al.  
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
American Journal of Neuroradiology
 
2008
29
5
967
973
Ponsford
J.
Cameron
P.
Fitzgerald
M.
Grant
M.
Mikocka-Walus
A.
Schonberger
M.
Predictors of postconcussive symptoms 3 months after mild traumatic brain injury
Neuropsychology
 
2012
26
3
304
313
Ponsford
J.
Willmott
C.
Rothwell
A.
Cameron
P.
Kelly
A. M.
Nelms
R.
et al.  
Factors influencing outcome following mild traumatic brain injury in adults
Journal of the International Neuropsychological Society
 
2000
6
5
568
579
Pruessmann
K. P.
Weiger
M.
Scheidegger
M. B.
Boesiger
P.
SENSE: Sensitivity encoding for fast MRI
Magnetic Resonance in Medicine
 
1999
42
5
952
962
Reynolds
C. R.
Kamphaus
R. W.
Reynolds intellectual assessment scales and reynolds intellectual screening test professional manual
 
2003
Lutz, FL
Psychological Assessment Resources
Rohling
M. L.
Larrabee
G. J.
Millis
S. R.
The “Miserable Minority” following mild traumatic brain injury: Who are they and do meta-analyses hide them?
Clinical Neuropsychology
 
2012
26
2
197
213
Rutgers
D. R.
Fillard
P.
Paradot
G.
Tadie
M.
Lasjaunias
P.
Ducreux
D.
Diffusion tensor imaging characteristics of the corpus callosum in mild, moderate, and severe traumatic brain injury
American Journal of Neuroradiology
 
2008
29
9
1730
1735
Sexton
C. E.
Mackay
C. E.
Ebmeier
K. P.
A systematic review of diffusion tensor imaging studies in affective disorders
Biological Psychiatry
 
2009
66
9
814
823
Shenton
M. E.
Hamoda
H. M.
Schneiderman
J. S.
Bouix
S.
Pasternak
O.
Rathi
Y.
et al.  
A review of magnetic resonance imaging and diffusion tensor imaging findings in mild traumatic brain injury
Brain Imaging and Behavior
 
2012
6
2
137
192
Sigurdardottir
S.
Andelic
N.
Roe
C.
Jerstad
T.
Schanke
A. K.
Post-concussion symptoms after traumatic brain injury at 3 and 12 months post-injury: A prospective study
Brain Injury
 
2009
23
6
489
497
Smith
S. M.
Jenkinson
M.
Johansen-Berg
H.
Rueckert
D.
Nichols
T. E.
Mackay
C. E.
et al.  
Tract-based spatial statistics: Voxelwise analysis of multi-subject diffusion data
Neuroimage
 
2006
31
4
1487
1505
Smith
S. M.
Jenkinson
M.
Woolrich
M. W.
Beckmann
C. F.
Behrens
T. E.
Johansen-Berg
H.
et al.  
Advances in functional and structural MR image analysis and implementation as FSL
Neuroimage
 
2004
23
Suppl. 1
S208
S219
Smith-Seemiller
L.
Fow
N. R.
Kant
R.
Franzen
M. D.
Presence of post-concussion syndrome symptoms in patients with chronic pain vs mild traumatic brain injury
Brain Injury
 
2003
17
3
199
206
Smits
M.
Houston
G. C.
Dippel
D. W.
Wielopolski
P. A.
Vernooij
M. W.
Koudstaal
P. J.
et al.  
Microstructural brain injury in post-concussion syndrome after minor head injury
Neuroradiology
 
2011
53
8
553
563
Sosin
D. M.
Sniezek
J. E.
Thurman
D. J.
Incidence of mild and moderate brain injury in the United States, 1991
Brain Injury
 
1996
10
1
47
54
Stern
R. A.
White
T.
Neuropsychological assessment battery: Administration, scoring, and interpretation manual
 
2003
Lutz, FL
Psychological Assessment Resources
Suhr
J. A.
Gunstad
J.
“Diagnosis Threat:” The effect of negative expectations on cognitive performance in head injury
Journal of Clinical and Experimental Neuropsychology
 
2002a
24
4
448
457
Suhr
J. A.
Gunstad
J.
Postconcussive symptom report: The relative influence of head injury and depression
Journal of Clinical and Experimental Neuropsychology
 
2002b
24
8
981
993
Sullivan
M. J.
Hall
E.
Bartolacci
R.
Sullivan
M. E.
Adams
H.
Perceived cognitive deficits, emotional distress and disability following whiplash injury
Pain Research and Management
 
2002
7
3
120
126
Taylor
A. G.
Deitrich
A.
Nuss
K.
Wright
M.
Rusin
J.
Bangert
B.
et al.  
Post-concussive symptoms in children with mild traumatic brain injury
Neuropsychology
 
2010
24
2
148
159
The Psychological Corporation
Wechsler test of adult reading: Manual
 
2001
San Antonio, TX
Author
Tombaugh
T. N.
Test of memory malingering
 
1996
North Tonawanda, NY
Multi-Health Systems
Wakana
S.
Caprihan
A.
Panzenboeck
M. M.
Fallon
J. H.
Perry
M.
Gollub
R. L.
et al.  
Reproducibility of quantitative tractography methods applied to cerebral white matter
Neuroimage
 
2007
36
3
630
644
Waljas
M.
Lange
R. T.
Hakulinen
U.
Huhtala
H.
Dastidar
P.
Hartikainen
K.
et al.  
Biopsychosocial outcome after uncomplicated mild traumatic brain injury
Journal of Neurotrauma
 
2014
31
1
108
124
Wilde
E. A.
McCauley
S. R.
Hunter
J. V.
Bigler
E. D.
Chu
Z.
Wang
Z. J.
et al.  
Diffusion tensor imaging of acute mild traumatic brain injury in adolescents
Neurology
 
2008
70
12
948
955
Winkler
A. M.
Ridgway
G. R.
Webster
M. A.
Smith
S. M.
Nichols
T. E.
Permutation inference for the general linear model
Neuroimage
 
2014
92
381
397
Xu
J.
Rasmussen
I. A.
Lagopoulos
J.
Haberg
A.
Diffuse axonal injury in severe traumatic brain injury visualized using high-resolution diffusion tensor imaging
Journal of Neurotrauma
 
2007
24
5
753
765