Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Original Research n Neuroradiology
Evaluation of White Matter Injury Patterns Underlying Neuropsychiatric Symptoms after Mild Traumatic Brain Injury1 Lea M. Alhilali, MD Joseph A. Delic, MD Serter Gumus, MD Saeed Fakhran, MD
Purpose:
To determine if a central axonal injury underlies neuropsychiatric symptoms after mild traumatic brain injury (mTBI) by using tract-based spatial statistics analysis of diffusion-tensor images.
Materials and Methods:
The institutional review board approved this study, with waiver of informed consent. Diffusion-tensor imaging and serial neurocognitive testing with the Immediate PostConcussion Assessment and Cognitive Testing evaluation were performed in 45 patients with mTBI (38 with irritability, 32 with depression, and 18 with anxiety). Control subjects consisted of 29 patients with mTBI without neuropsychiatric symptoms. Fractional anisotropy and diffusivity maps were analyzed by using tract-based spatial statistics with a multivariate general linear model. Diffusion-tensor imaging findings were correlated with symptom severity, neurocognitive test scores, and time to recovery with the Pearson correlation coefficient.
Results:
Compared with control subjects, patients with mTBI and depression had decreased fractional anisotropy in the superior longitudinal fasciculus (P = .006), white matter around the nucleus accumbens (P = .03), and anterior limb of the internal capsule (P = .02). Patients with anxiety had diminished fractional anisotropy in the vermis (P = .04). No regions of significantly decreased fractional anisotropy were seen in patients with irritability relative to control subjects. Injury in the region of the nucleus accumbens inversely correlated with recovery time in patients with depression (r = 20.480, P = .005).
Conclusion:
Unique white matter injury patterns were seen for two major posttraumatic neuropsychiatric symptoms. Injury to the cerebellar vermis in patients with mTBI and anxiety may indicate underlying dysfunction in primitive fear conditioning circuits in the cerebellum. Involvement of the nucleus accumbens in depression after mTBI may suggest an underlying dysfunctional reward circuit that affects the prognosis in these patients.
1
From the Department of Radiology, Division of Neuroradiology, UPMC Presbyterian Hospital, University of Pittsburgh Medical Center, 200 Lothrop St, Presby South Tower, 3rd Floor, Suite 3950, Pittsburgh, PA 15213. Received December 28, 2014; revision requested February 13, 2015; revision received March 24; accepted March 31; final version accepted April 9. Address correspondence to L.M.A. (e-mail: [email protected]).
RSNA, 2015
q
RSNA, 2015
q
Radiology: Volume 000: Number 0— 2015 n radiology.rsna.org
1
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
M
ild traumatic brain injury (mTBI), referred to as “concussion,” affects nearly 3.8 million people in the United States annually (1). While awareness of the cognitive consequences of concussive injuries has grown (2), the association with psychiatric disorders has failed to garner the attention given to associated neurodegenerative syndromes, such as Alzheimer dementia and chronic traumatic encephalopathy (3). However, posttraumatic neuropsychiatric disorders can be as disabling as posttraumatic cognitive deficits, leading to decreased quality of life, social isolation, and poor outcomes (4). Depression is by far the most common neuropsychiatric sequela of mTBI (5). Costs associated with depression after mTBI include increased obesity, suicide, substance abuse (6), and school
Advances in Knowledge nn Injured white matter regions that underlie depression after mild traumatic brain injury (mTBI) (difference in fractional anisotropy [FA] in the superior longitudinal fasciculus, P = .006; difference in FA in the nucleus accumbens, P = .03) strongly resemble both those associated with nontraumatic major depressive disorder and those seen with depression after moderate to severe traumatic brain injury. nn White matter abnormalities that underlie anxiety after mTBI (difference in FA in the cerebellar vermis, P = .04) suggest injury to a more primitive fear circuit involving the vermis, compared with the dysfunction in the frontostriatal regions seen in nontraumatic anxiety disorders. nn Injury to the white matter around the nucleus accumbens inversely correlates with recovery in patients with mTBI and depression (r = 20.480, P = .005), suggesting a dysfunctional reward circuit similar to that seen in nontraumatic major depressive disorder. 2
Alhilali et al
failure (7), as well as more severe symptoms, poorer outcomes (5), and prolonged recovery after combat (8). The disability associated with posttraumatic depression is compounded by high comorbidity with posttraumatic anxiety. Posttraumatic anxiety results in greater functional impairment, prolonged recovery, and clinically significant cognitive impairment. In fact, cognitive deficits after mTBI may in part be due to the effects of posttraumatic anxiety (9). In addition to depression and anxiety, irritability is a common neurobehavioral change that accompanies mTBI, ranking among the most commonly selfreported symptoms (10). Posttraumatic irritability has been linked with decreased processing speed and damages relationships with caregivers (10); as a result, this is perhaps why it is associated with poorer outcomes (11). Unfortunately, the mechanisms that underlie neurobehavioral changes after mTBI are not sufficiently understood. In penetrating brain injury, focal injuries in the right orbitofrontal cortex are associated with posttraumatic anxiety and depression (12), while frontal lobe injuries have resulted in posttraumatic irritability (13). However, the relationship between mild trauma and behavioral changes is less clear. Since most patients with mTBI undergo normal conventional magnetic
Implications for Patient Care nn Detection of a similar injury in patients with mTBI and depression and patients with major depressive disorder, with similar prognostic importance regarding abnormalities in the nucleus accumbens, may suggest a common pathophysiology in both traumatic and nontraumatic depression that may help guide treatment. nn Determining a different site of abnormality in patients with posttraumatic anxiety than that seen in anxiety disorders in the general population may indicate that different treatment targets are required for patients with anxiety after trauma.
resonance (MR) imaging, it is uncertain if neuropsychiatric changes in these patients are from discrete injury beneath the resolution of conventional MR imaging or comorbid circumstantial stressors that mimic organic brain injury. Stressors such as combat situations, inability to return to work, and orthopedic injuries are common after mTBI and may influence mood and social functioning (14). However, recent studies have shown that many other postconcussive symptoms previously thought to be unrelated to a focal brain injury have correlated with a central axonal injury, including vestibulopathy, ocular dysfunction, and sleepwake disturbances (3,15). We therefore hypothesize that a central traumatic axonal injury is also a primary feature of the injury pattern that underlies posttraumatic neuropsychiatric symptoms. Thus, the purpose of this study was to determine if a central axonal injury underlies neuropsychiatric symptoms after mTBI by using tract-based spatial statistics (TBSS) analysis of MR images acquired with diffusion-tensor imaging (DTI).
Materials and Methods Study Population Our institutional review board approved this retrospective study, with waiver of Published online before print 10.1148/radiol.2015142974 Content codes: Radiology 2015; 000:1–8 Abbreviations: DTI = diffusion-tensor imaging FA = fractional anisotropy MNI = Montreal Neurological Institute mTBI = mild traumatic brain injury ROI = region of interest TBSS = tract-based spatial statistics Author contributions: Guarantors of integrity of entire study, L.M.A., S.F.; study concepts/study design or data acquisition or data analysis/ interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, all authors; clinical studies, L.M.A., S.F.; experimental studies, L.M.A.; statistical analysis, L.M.A.; and manuscript editing, all authors Conflicts of interest are listed at the end of this article.
radiology.rsna.org n Radiology: Volume 000: Number 0— 2015
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
informed consent. All studies were performed as standard of care, and images were reviewed retrospectively. We searched our electronic medical record to identify MR imaging studies with DTI performed for mTBI. MR imaging reports from January 1, 2006, to January 25, 2014, were searched by using keywords “concussion,” “mild traumatic brain injury,” and “diffusion tensor imaging.” Inclusion criteria were age of 10–50 years, witnessed closed head trauma, no focal neurologic deficit, loss of consciousness of less than 1 minute, posttraumatic amnesia of less than 30 minutes, and English language proficiency. Ninety-five patients were initially identified. Exclusion criteria were history of a neuropsychiatric illness (two patients) or substance abuse (two patients), abnormal computed tomographic (CT) or conventional MR imaging findings (three patients), lack of DTI (four patients) or neurocognitive assessment (six patients), or a total symptom score of zero (three patients). Neuropsychological assessment and neurocognitive testing were performed by a neuropsychologist with more than 14 years of experience in treating patients with mTBI or by a fellow under supervision (both nonauthors). Patients were classified as having depression or anxiety according to the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders, or DSM-5, criteria for a depressive or anxiety disorder due to another medical condition (mTBI), as follows: (a) persistent depressed mood or markedly diminished interest or pleasure in almost all activities that predominates in the clinical picture (for depressive disorder) or panic attacks or anxiety that are predominant in the clinical picture (for anxiety disorder); (b) evidence from the patient history, physical examination, or laboratory findings that the disturbance is a pathophysiological consequence of another medical condition (mTBI); (c) symptoms not better explained by another mental disorder; (d) symptoms not occurring exclusively during delirium; and (e) symptoms causing clinically significant distress or impairment in social,
occupational, or other important areas of functioning. Patients were classified as having irritability on the basis of the DSM5 criteria for the following behavioral findings: (a) temper outbursts, on average three or more times per week; (b) irritable mood, defined as being easily annoyed and provoked to anger most of the day, nearly every day; and (c) never having had a distinct period of 1 day or more of meeting full criteria (other than the duration criteria) for a manic or hypomanic episode. The Immediate Post-Concussion Assessment Cognitive Test was used for neurocognitive testing, which is a computerized test used to measure cognitive function as well as symptoms by using a seven-point Likert scale over 22 different categories. Data recorded for all individuals included sex, age, type of trauma, date of injury, clinical evaluation, Immediate Post-Concussion Assessment Cognitive Test results, prior concussions, imaging results, and clinical management. Time to recovery was defined as when either the patient stated he or she was asymptomatic or the total symptom score was zero.
DTI and Conventional MR Imaging Conventional MR and DTI were performed with a 1.5-T system (Signa; GE Healthcare, Milwaukee, Wis) with an eight-channel head coil. Despite the extended time period over which this study was conducted, all individuals underwent an identical imaging protocol with the same MR system, as follows: sagittal and axial T1-weighted imaging (repetition time [msec]/echo time [msec], 600/minimum setting; section thickness, 5 mm; number of acquisitions, one), axial fast spin-echo proton density imaging (2000–2500/minimum setting; echo train length, four; section thickness, 5 mm; matrix, 320 3 224; number of acquisitions, one), T2-weighted imaging (2000–2500/84– 102; echo train length, 28; section thickness, 5 mm; matrix, 480 3 480; number of acquisitions, one), fluid attenuation inversion-recovery imaging (9000–10 000/149; inversion time, 2200
Radiology: Volume 000: Number 0— 2015 n radiology.rsna.org
Alhilali et al
msec), diffusion-weighted imaging (single-shot echo planar imaging; 10 000/ minimum setting; section thickness, 5 mm; matrix, 128 3 128), and either T2* gradient-recalled-echo imaging (4400/21; number of acquisitions, one; flip angle, 90°; section thickness, 3 mm) or susceptibility-weighted imaging (37/23; number of acquisitions, one; flip angle, 15°; section thickness, 2.4 mm). Field of view ranged from 200 to 240 mm. DTI was performed with a singleshot echo-planar sequence (4000/80; number of acquisitions, two; section thickness, 5 mm; matrix, 128 3 128; field of view, 260 mm). Diffusion gradients were used in 25 noncollinear directions with b values of 0 and 1000 sec/mm2.
TBSS Analysis TBSS from the Functional Magnetic Resonance Imaging of the Brain, or FMRIB, Software Library (version 1.1, http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/) was used to analyze the white matter (16). TBSS is well suited for studies of both adult and pediatric populations, where head size and field of view variation may confound traditional voxelwise methods. As a result, TBSS has been used in combined pediatric and adult populations to evaluate normal white matter development (17,18), as well as pathologic processes, including mTBI (3,15,19). Analysis was as follows. (a) Fractional anisotropy (FA) or diffusivity images, including mean diffusivity, axial diffusivity, and radial diffusivity, were aligned to a target image. (b) This image was affine aligned into 1 3 1 3 1-mm Montreal Neurological Institute (MNI) 152 space. The higher resolution of the MNI space than that of the DTI MR images prevents any marked interpolation blurring when the nonlinear warp plus standard-space affine transformation is applied to the data of each individual. (c) Every image was transformed to MNI 152 space. (d) A mean image was thinned to create a mean FA skeleton. (e) The skeleton was thresholded to remove regions with considerable interindividual variability. 3
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
Every patient’s FA or diffusivity data were projected onto this skeleton for voxel-based cross-subject statistical analysis. We applied a Monte Carlo permutation test by using threshold-free cluster enhancement with a significance level of P less than .05 (5000 permutations), fully corrected for multiple comparisons (family-wise error corrected). A multivariate general linear model to evaluate for two-group differences was created for each neuropsychiatric symptom, adjusted for covariates of age, sex, type of injury (sports related vs non– sports related), duration of symptoms, prior concussion, and presence or absence of each of the other two neuropsychiatric symptoms. This model was applied to the FA, mean diffusivity, radial diffusivity, and axial diffusivity maps for each neuropsychiatric symptom. (Processing was performed by two neuroradiologists with 3 years of image analysis experience [L.M.A., S.F.]).
Region of Interest Analysis Automated region of interest (ROI) analysis was used to quantify FA or diffusivity values in regions of significant differences detected with TBSS by using the mean FA or diffusivity skeleton overlaid with the regions of significant differences between patients and control subjects (corrected, voxelwise). A single mask was created to correspond with the location (in MNI space) and size of the regions with significant differences. FA or diffusivity values of patients and control subjects were then extracted in an automated fashion by using the identical ROI mask along the individual skeletons that were aligned onto common space during the TBSS processing and subsequently compared with two-sample t test. The Cohen d index was used to assess effect size. Correlation of FA or diffusivity values extracted from the ROI and continuous variables was evaluated with Pearson correlation coefficients. Analysis was performed by two neuroradiologists with 3 years of image analysis experience (L.M.A., S.F.). Comparison of proportions and means in the demographic data between each group of neuropsychiatric 4
Alhilali et al
patients and control subjects was conducted by using a Fisher exact test or unpaired t test, respectively. P values less than .05 were considered to indicate a statistically significant difference for demographic factors and voxel-wise analysis, after correcting for multiple comparisons by using a family-wise error correction. P values of .025 were considered to indicate a statistically significant difference for the ROI correlation analysis, given the Bonferroni adjustment described in the power analysis. Analysis was performed by a physician with postgraduate statistics training (L.M.A.). Since we hypothesized relationships between two brain regions with each of the continuous variables for the ROI correlation analysis on the basis of our previous work (15), our a level was determined to be 05/2 = .025, based on Bonferroni conservative adjustment. For each comparison, 18 individuals in each of the two groups (group 1, patient with a symptom; group 2, patients without the symptom) achieve 80% power to detect an effect size of 1.08 (considered a large effect size) at a = .025. Power was computed on the basis of the mean effect size from prior studies, since the effect sizes from our previous data range anywhere from small to very large (1.34 in absolute value) (15,19).
Results Patients Forty-five patients with mTBI who had neuropsychiatric symptoms and 29 control subjects with mTBI who did not have neuropsychiatric symptoms were included (51 male, 23 female; mean age, 18 years; range, 10–47 years). The median time from injury to clinical presentation was 20 days (range, 0–506 days). The most common mechanism of trauma was sports injury (43 patients, 58%). The second most common mechanism was motor vehicle accident (nine patients, 12%). Of the patients with mTBI who had neuropsychiatric symptoms, 38 had irritability, 32 had depression, and 18
had anxiety. Of these, 13 patients had only irritability, one had only depression, 14 had comorbid irritability and depression, seven had comorbid anxiety and depression, and one had comorbid irritability and anxiety. Ten patients fulfilled the criteria for all three conditions. No significant difference was seen in demographic or clinical factors among the different groups of patients with neuropsychiatric symptoms and control subjects. Demographics and clinical characteristics are summarized in Table 1.
Assessment of DTI MR Images with TBSS Depression.—Patients with mTBI who had depression had significantly lower FA in the region of the right nucleus accumbens, anterior limb of the internal capsule, and superior longitudinal fasciculus (Fig 1). There were no regions where control subjects with mTBI had lower FA than patients with depression. No significant differences in mean diffusivity, axial diffusivity, or radial diffusivity were seen between patients with depression and control subjects; however, mean diffusivity, axial diffusivity, and radial diffusivity values in the left aspect of the body of the corpus callosum trended toward a positive correlation with age (P = .10) in those with depression. Anxiety.—Patients with mTBI who had anxiety had significantly lower FA values in the cerebellar vermis (Fig 2). There were no regions where control subjects with mTBI had lower FA values than patients with anxiety. No significant differences in mean diffusivity, axial diffusivity, or radial diffusivity were seen between patients with mTBI who had anxiety and control subjects. Irritability.—No regions of significantly increased or decreased FA, mean diffusivity, radial diffusivity, or axial diffusivity were seen in patients with irritability compared with control subjects at voxel-wise analysis. Assessment of DTI MR Images with ROI Analysis In patients with depression, FA in the region of the nucleus accumbens
radiology.rsna.org n Radiology: Volume 000: Number 0— 2015
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
Alhilali et al
Table 1 Comparison of Demographic and Clinical Characteristics among Patients with mTBI Who Have Irritability, Depression, and Anxiety and Control Subjects with mTBI but No Neuropsychiatric Symptoms Comparison between Patients with Irritability and Control Subjects Parameter Mean age (y) No. of male subjects† Median time to presentation (d) No. of individuals with prior concussion† No. of individuals with sports injury† Mean ImPACT‡ total symptom score percentile Mean verbal memory score percentile Mean visual memory score percentile Mean processing speed percentile Mean reaction time percentile Median time to recovery (wk)
Comparison between Patients with Depression and Control Subjects
Comparison between Patients with Anxiety and Control Subjects
Control Subjects (n = 29)
Patients with Irritability (n = 38)
P Value*
Patients with Depression (n = 32)
P Value*
Patients with Anxiety (n = 18)
P Value*
16.9 (10–28) 21 (72) 16 (1–275) 13 (45) 20 (69) 29.9 (0–84)
18.2 (11–38) 27 (71) 32 (4–506) 15 (39) 20 (53) 34.3 (2–97)
.36 ..99 .22 .80 .21 .46
19.3 (11–47) 20 (62) 33 (0–506) 11 (34) 15 (47) 36.7 (3–97)
.11 .43 .99 .44 .13 .27
20.1 (11–47) 14 (78) 17 (3–361) 7 (39) 10 (56) 35.2 (4–97)
.08 .74 .13 .77 .37 .48
31.0 (0–84) 23.5 (1–90) 36.6 (1–98) 40.4 (1–94) 42.9 (1–252)
30.6 (1–99) 31.9 (1–97) 38.2 (1–95) 30.6 (1–97) 30.2 (2–194)
.69 .24 .84 .22 .31
30.3 (1–99) 28.6 (1–97) 39.1 (1–95) 27.3 (1–95) 30.8 (2–194)
.65 .46 .75 .11 .41
39.7 (1–80) 18.8 (1–64) 36.7 (1–93) 28.3 (1–91) 30.8 (1–189)
.66 .51 .97 .19 .62
Note.—Numbers in parentheses are ranges, unless indicated otherwise. ImPACT = Immediate Post-Concussion Assessment and Cognitive Test. * P values represent comparisons with control subjects with mTBI who did not have neuropsychiatric symptoms. All P values were two tailed and calculated with an unpaired Student t test for continuous variables and a Fisher exact test for categorical variables. †
Numbers in parentheses are percentages.
‡
Scores are percentiles determined by means of normative data obtained from baseline testing of more than 17 000 athletes as part of their participation before sports, with percentile information accounting for both patient sex and age.
Figure 1
Figure 1: Coronal images derived from TBSS results and rendered on T1-weighted MR images from the MNI atlas show how patients with depression have significantly decreased FA in the region of A, the right nucleus accumbens (arrow) or “reward center,” as well as B, the anterior limb of the internal capsule (arrow) and C, the superior longitudinal fasciculus (arrow)—regions that are involved in depressive disorders in nontraumatic and traumatic settings. Voxels with significant differences detected with TBSS (P , .05 corrected for multiple comparisons) were thickened by using the TBSS fill function into local tracts (red) and overlaid on the white matter skeleton (blue).
identified with TBSS inversely correlated with time to recovery (r = 20.480, P = .005). In patients without depression, no correlation was seen (r = 0.238, P = .21). Summaries of the ROI analysis are provided in Tables 2 and 3.
Discussion By using voxel-based analysis of DTI MR images, we found a central axonal injury pattern underlying posttraumatic depression and anxiety but not irritability. This suggests that posttraumatic
Radiology: Volume 000: Number 0— 2015 n radiology.rsna.org
depression and anxiety may be the result of discrete injuries not visible with the resolution of conventional MR imaging. Injury in the region of the nucleus accumbens inversely correlated with recovery time in patients with depression, while injury to the vermis 5
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
Alhilali et al
Figure 2
Figure 2: A, Axial and B, sagittal images derived from TBSS results and rendered on T1-weighted MR images from the MNI atlas indicate that significant white matter differences in patients with mTBI and anxiety involve the cerebellar vermis, the region responsible for fear conditioning. Voxels with significant differences detected with TBSS (P , .05 corrected for multiple comparisons) were thickened by using the TBSS fill function into local tracts (red) and overlaid on the white matter skeleton (blue).
in patients with anxiety trended toward correlation with neurocognitive test performance. Recognition of this injury pattern on images may help predict symptoms on the basis of the location of damage, as well as provide a quantitative imaging biomarker that may ultimately be used in the determination of prognosis and in the measurement of response to treatment. Previous studies on the evaluation of white matter injury in depression and anxiety after mTBI have been focused mainly on blast injury (20–22) or professional athletes (23,24). Unfortunately, these studies have limited applicability to the general mTBI population. Blast injuries are theorized to have different underlying mechanisms than civilian mTBI injuries (22) and are seen in military populations with unique emotional stressors, which may affect neuropsychological test results (22). Similarly, professional athletes have a unique lifestyle and risk factors not typically seen in the general population (24). Imaging studies in which post-mTBI depression and anxiety were evaluated in the general population have either had very small sample sizes (25) or only involved the comparison between patients with mTBI who had 6
neuropsychiatric symptoms and healthy control subjects, rather than between patients with mTBI who had neuropsychiatric symptoms and other patients with mTBI who did not have neuropsychiatric symptoms (26,27). This is important, because numerous white matter abnormalities seen in patients with mTBI when compared with healthy control subjects are either clinically silent or remote abnormalities, which are likewise not symptomatic (3). These background lesions create noise that may obscure the true lesions that underlie neuropsychiatric symptoms. No imaging studies have yet been conducted to evaluate the white matter injuries that underlie depression, irritability, or anxiety in a large general mTBI population in which other patients with mTBI are used as control subjects. Most importantly, however, the clinical significance of these white matter abnormalities with respect to the prognosis of neuropsychiatric patients with mTBI has not been addressed. In our study, injured regions in depression consisted of both regions known to be abnormal in nontraumatic major depressive disorder and regions injured in posttraumatic depression after severe or repetitive trauma. The three abnormal regions (nucleus
accumbens, anterior limb of the internal capsule, and superior longitudinal fasciculus) in our study have all been found to be abnormal in patients with nontraumatic major depressive disorder (28–30). Interestingly, both the nucleus accumbens and the superior longitudinal fasciculus, but not the anterior limb of the internal capsule, have also been found to be injured in patients with depression after moderate to severe traumatic brain injury (31) and repetitive mTBI (23,24). Injuries to regions exclusively seen in nontraumatic depression, as well as regions seen in posttraumatic depression, suggest that development of post-mTBI depression is multifactorial. Depression after mTBI may be the result of injury to regions known to result in depression after trauma (nucleus accumbens, superior longitudinal fasciculus), combined with additional preexisting abnormalities in regions that render these patients susceptible to depression outside of any traumatic setting (anterior limb of the internal capsule). Additionally, the fact that abnormalities are seen in both the nucleus accumbens and the superior longitudinal fasciculus in both nontraumatic major depressive disorder and posttraumatic depression may indicate that perhaps unrecognized trauma plays a role in the development of major depressive disorder in the general population, as well. Severity of injury to the nucleus accumbens inversely correlated with time to recovery in patients with depression. This correlates well with findings in nontraumatic major depressive disorder, in which a dysfunctional reward circuit is believed to prohibit recovery, and deep brain stimulation of the nucleus accumbens results in improvement in depression that was treatment resistant previously (29,32). Unlike patients with mTBI and depression, patients with mTBI and anxiety demonstrated injured regions distinct from abnormal regions seen in nontraumatic anxiety disorders. Anxiety disorders in the general population are associated with frontostriatal abnormalities (9). In our mTBI population, however, abnormalities were seen in the vermis. The vermis is integral
radiology.rsna.org n Radiology: Volume 000: Number 0— 2015
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
Alhilali et al
Table 2 Comparison of FA Values in ROIs for Patients with mTBI with and without Depression and Those with and without Anxiety Parameter Mean FA value for the region of the nucleus accumbens Mean FA value for the anterior limb of the internal capsule Mean FA value for the superior longitudinal fasciculus Mean FA value for the cerebellum
Patients with mTBI and Depression
Patients with mTBI and Anxiety
Control Subjects with mTBI
P Value*
0.357 (0.233, 0.481) 0.508 (0.429, 0.588) 0.435 (0.383, 0.487) …
… … … 0.265 (0.215, 0.315)
0.391 (0.287, 0.495) 0.532 (0.440, 0.625) 0.456 (0.411, 0.501) 0.296 (0.236, 0.356)
.026 (0.616) .020 (0.530) .006 (0.863) .043 (0.437)
Note.—Numbers in parentheses are 95% confidence intervals, unless indicated otherwise. * P values were two tailed and calculated with an unpaired Student t test. Numbers in parentheses are the Cohen d index.
Table 3 Correlation of FA with Clinical Findings for ROIs in Patients with mTBI and Depression and Those with mTBI and Anxiety Parameter Correlation* with symptom severity score† Correlation with verbal memory score Correlation with visual memory score Correlation with processing speed score Correlation with reaction time score Correlation with time to recovery
Mean FA Value for the Region of the Nucleus Accumbens ROI
Mean FA Value for the Anterior Limb of the Internal Capsule ROI
Mean FA Value for the Superior Longitudinal Fasciculus ROI
Mean FA Value for the Cerebellum ROI
20.097 (.60) 20.038 (.84) 0.037 (.84) 20.013 (.94) 0.109 (.55) 20.480 (.005)‡
20.203 (.27) 20.181 (.32) 20.146 (.43) 20.286 (.11) 20.061 (.74) 20.167 (.36)
20.119 (.52) 20.265 (.14) 0.168 (.36) 0.245 (.18) 0.048 (.79) 20.133 (.47)
0.149 (.55) 0.054 (.83) 0.231 (.36) 0.020 (.94) 20.086 (.74) 20.130 (.61)
Note.—Numbers in parentheses are P values. * Correlation performed with the Pearson correlation coefficient. †
Two-tailed P value for the Pearson correlation coefficient.
‡
Statistically significant correlation.
to fear conditioning and controls the autonomic and emotional response to aversive conditioning, with connections to both the catecholaminergic centers (locus coeruleus and ventral tegmental area) and emotional centers (amygdala, septum, and locus coeruleus) (33). Patients’ vermian lesions are unable to calm autonomic and emotional responses during fear conditioning (33,34). This suggests that anxiety in patients with mTBI results from dysfunction of a more primitive fear circuit than the higher-order frontostriatial dysfunction seen with anxiety outside of trauma. There were several limitations to our study. Our study was retrospective in nature, with a moderate sample size. This moderate sample size prevented detailed subgroup analysis of possible distinct injury patterns in patients with more than one neuropsychiatric symptom. Furthermore, there is no
generally accepted method to evaluate for overfitting of our voxelwise statistical model, which could limit the capacity to generalize the findings. As such, the findings should be confirmed in a prospective study with a larger cohort. Additionally, there is likely selection bias toward more seriously injured patients, since many patients with mTBI do not undergo imaging. However, we would contend that these more symptomatic patients are the type of patients in which biomarkers of injury are most needed. Finally, in our study, we used a 1.5-T magnet, which results in decreased image resolution and has poorer signal-to-noise ratio than the 3.0-T system used in DTI (35). It is unclear if additional white matter injuries may be detected by using a 3.0-T magnet, and future work should include higher field strengths. In summary, detection of the central white matter injuries that underlie
Radiology: Volume 000: Number 0— 2015 n radiology.rsna.org
depression and anxiety but not irritability indicates that not all neuropsychiatric symptoms after mTBI are the result of discrete white matter injuries, but in those with corresponding injuries, the injured regions provide insight into the underlying pathophysiology and prognosis. Detection of similar injuries in both patients with mTBI who have depression and those who have major depressive disorder, with similar prognostic importance to abnormalities in the nucleus accumbens, may suggest a common pathophysiology in both traumatic and nontraumatic depression that may help guide treatment. Determining a different site of abnormality in patients with posttraumatic anxiety than in those with anxiety disorders in the general population may indicate that different treatment targets are required for patients with anxiety after trauma. 7
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
Disclosures of Conflicts of Interest: L.M.A. disclosed no relevant relationships. J.A.D. disclosed no relevant relationships. S.G. disclosed no relevant relationships. S.F. disclosed no relevant relationships.
References 1. Finkelstein EA, Corso PS, Miller TR. The incidence and economic burden of injuries in the United States. New York, NY: Oxford University Press, 2006. 2. Petraglia AL, Plog BA, Dayawansa S, et al. The spectrum of neurobehavioral sequelae after repetitive mild traumatic brain injury: a novel mouse model of chronic traumatic encephalopathy. J Neurotrauma 2014;31(13):1211–1224. 3. Fakhran S, Yaeger K, Alhilali L. Symptomatic white matter changes in mild traumatic brain injury resemble pathologic features of early Alzheimer dementia. Radiology 2013;269(1):249–257.
Alhilali et al
idence of right brain injury in referrals to a neuropsychiatric brain injury unit? Brain Inj 2001;15(1):65–69. 13. Grafman J, Schwab K, Warden D, Pridgen A, Brown HR, Salazar AM. Frontal lobe injuries, violence, and aggression: a report of the Vietnam Head Injury Study. Neurology 1996; 46(5):1231–1238. 14. Parker RS. The spectrum of emotional distress and personality changes after minor head injury incurred in a motor vehicle accident. Brain Inj 1996;10(4):287–302. 15. Alhilali LM, Yaeger K, Collins M, Fakhran S. Detection of central white matter injury underlying vestibulopathy after mild traumatic brain injury. Radiology 2014;272(1):224– 232. 16. Smith SM, Jenkinson M, Johansen-Berg H, et al. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 2006;31(4):1487–1505.
4. Vanderploeg RD, Curtiss G, Luis CA, Salazar AM. Long-term morbidities following self-reported mild traumatic brain injury. J Clin Exp Neuropsychol 2007;29(6):585– 598.
17. Peters BD, Szeszko PR, Radua J, et al. White matter development in adolescence: diffusion tensor imaging and meta-analytic results. Schizophr Bull 2012;38(6):1308– 1317.
5. Moldover JE, Goldberg KB, Prout MF. Depression after traumatic brain injury: a review of evidence for clinical heterogeneity. Neuropsychol Rev 2004;14(3):143–154.
18. Giorgio A, Watkins KE, Chadwick M, et al. Longitudinal changes in grey and white matter during adolescence. Neuroimage 2010;49(1):94–103.
6. Saluja G, Iachan R, Scheidt PC, Overpeck MD, Sun W, Giedd JN. Prevalence of and risk factors for depressive symptoms among young adolescents. Arch Pediatr Adolesc Med 2004;158(8):760–765.
19. Fakhran S, Yaeger K, Collins M, Alhilali L. Sex differences in white matter abnormalities after mild traumatic brain injury: localization and correlation with outcome. Radiology 2014;272(3):815–823.
7. Fletcher JM. Adolescent depression: diagnosis, treatment, and educational attainment. Health Econ 2008;17(11):1215–1235. 8. Hoge CW, McGurk D, Thomas JL, Cox AL, Engel CC, Castro CA. Mild traumatic brain injury in U.S. Soldiers returning from Iraq. N Engl J Med 2008;358(5):453–463. 9. Moore EL, Terryberry-Spohr L, Hope DA. Mild traumatic brain injury and anxiety sequelae: a review of the literature. Brain Inj 2006; 20(2):117–132. 10. Yang CC, Hua MS, Lin WC, Tsai YH, Huang SJ. Irritability following traumatic brain injury: divergent manifestations of annoyance and verbal aggression. Brain Inj 2012; 26(10):1185–1191. 11. Yang CC, Tu YK, Hua MS, Huang SJ. The association between the postconcussion symptoms and clinical outcomes for patients with mild traumatic brain injury. J Trauma 2007; 62(3):657–663. 12. Borek LL, Butler R, Fleminger S. Are neuropsychiatric symptoms associated with ev-
8
20. Matthews SC, Strigo IA, Simmons AN, O’Connell RM, Reinhardt LE, Moseley SA. A multimodal imaging study in U.S. veterans of Operations Iraqi and Enduring Freedom with and without major depression after blast-related concussion. Neuroimage 2011;54(Suppl 1):S69–S75. 21. Matthews SC, Spadoni AD, Lohr JB, Strigo IA, Simmons AN. Diffusion tensor imaging evidence of white matter disruption associated with loss versus alteration of consciousness in warfighters exposed to combat in Operations Enduring and Iraqi Freedom. Psychiatry Res 2012;204(2-3):149–154. 22. Morey RA, Haswell CC, Selgrade ES, et al. Effects of chronic mild traumatic brain injury on white matter integrity in Iraq and Afghanistan war veterans. Hum Brain Mapp 2013; 34(11):2986–2999. 23. Strain J, Didehbani N, Cullum CM, et al. Depressive symptoms and white matter dysfunction in retired NFL players with concussion history. Neurology 2013;81(1):25–32.
24. Hart J Jr, Kraut MA, Womack KB, et al. Neuroimaging of cognitive dysfunction and depression in aging retired National Football League players: a cross-sectional study. JAMA Neurol 2013;70(3):326–335. 25. Rao V, Mielke M, Xu X, et al. Diffusion tensor imaging atlas-based analyses in major depression after mild traumatic brain injury. J Neuropsychiatry Clin Neurosci 2012; 24(3):309–315. 26. Guskiewicz KM, Marshall SW, Bailes J, et al. Association between recurrent concussion and late-life cognitive impairment in retired professional football players. Neurosurgery 2005;57(4):719–726; discussion 719–726. 27. Levin HS, Wilde E, Troyanskaya M, et al. Diffusion tensor imaging of mild to moderate blast-related traumatic brain injury and its sequelae. J Neurotrauma 2010;27(4):683–694. 28. Hall GB, Milne AM, Macqueen GM. An fMRI study of reward circuitry in patients with minimal or extensive history of major depression. Eur Arch Psychiatry Clin Neurosci 2014;264(3):187–198. 29. Morishita T, Fayad SM, Higuchi MA, Nestor KA, Foote KD. Deep brain stimulation for treatment-resistant depression: systematic review of clinical outcomes. Neurotherapeutics 2014;11(3):475–484. 30. Lyden H, Espinoza RT, Pirnia T, et al. Electroconvulsive therapy mediates neuroplasticity of white matter microstructure in major depression. Transl Psychiatr 2014;4:e380. 31. Maller JJ, Thomson RH, Pannek K, The (eigen)value of diffusion tensor ing to investigate depression after matic brain injury. Hum Brain Mapp 35(1):227–237.
et al. imagtrau2014;
32. Millet B, Jaafari N, Polosan M, et al. Limbic versus cognitive target for deep brain stimulation in treatment-resistant depression: accumbens more promising than caudate. Eur Neuropsychopharmacol 2014;24(8):1229–1239. 33. Sacchetti B, Scelfo B, Strata P. Cerebellum and emotional behavior. Neuroscience 2009; 162(3):756–762. 34. Maschke M, Schugens M, Kindsvater K, et al. Fear conditioned changes of heart rate in patients with medial cerebellar lesions. J Neurol Neurosurg Psychiatry 2002;72(1):116–118. 35. Alexander AL, Lee JE, Wu YC, Field AS. Comparison of diffusion tensor imaging measurements at 3.0 T versus 1.5 T with and without parallel imaging. Neuroimaging Clin N Am 2006;16(2):299–309, xi.
radiology.rsna.org n Radiology: Volume 000: Number 0— 2015
Original Research n Neuroradiology
Evaluation of White Matter Injury Patterns Underlying Neuropsychiatric Symptoms after Mild Traumatic Brain Injury1 Lea M. Alhilali, MD Joseph A. Delic, MD Serter Gumus, MD Saeed Fakhran, MD
Purpose:
To determine if a central axonal injury underlies neuropsychiatric symptoms after mild traumatic brain injury (mTBI) by using tract-based spatial statistics analysis of diffusion-tensor images.
Materials and Methods:
The institutional review board approved this study, with waiver of informed consent. Diffusion-tensor imaging and serial neurocognitive testing with the Immediate PostConcussion Assessment and Cognitive Testing evaluation were performed in 45 patients with mTBI (38 with irritability, 32 with depression, and 18 with anxiety). Control subjects consisted of 29 patients with mTBI without neuropsychiatric symptoms. Fractional anisotropy and diffusivity maps were analyzed by using tract-based spatial statistics with a multivariate general linear model. Diffusion-tensor imaging findings were correlated with symptom severity, neurocognitive test scores, and time to recovery with the Pearson correlation coefficient.
Results:
Compared with control subjects, patients with mTBI and depression had decreased fractional anisotropy in the superior longitudinal fasciculus (P = .006), white matter around the nucleus accumbens (P = .03), and anterior limb of the internal capsule (P = .02). Patients with anxiety had diminished fractional anisotropy in the vermis (P = .04). No regions of significantly decreased fractional anisotropy were seen in patients with irritability relative to control subjects. Injury in the region of the nucleus accumbens inversely correlated with recovery time in patients with depression (r = 20.480, P = .005).
Conclusion:
Unique white matter injury patterns were seen for two major posttraumatic neuropsychiatric symptoms. Injury to the cerebellar vermis in patients with mTBI and anxiety may indicate underlying dysfunction in primitive fear conditioning circuits in the cerebellum. Involvement of the nucleus accumbens in depression after mTBI may suggest an underlying dysfunctional reward circuit that affects the prognosis in these patients.
1
From the Department of Radiology, Division of Neuroradiology, UPMC Presbyterian Hospital, University of Pittsburgh Medical Center, 200 Lothrop St, Presby South Tower, 3rd Floor, Suite 3950, Pittsburgh, PA 15213. Received December 28, 2014; revision requested February 13, 2015; revision received March 24; accepted March 31; final version accepted April 9. Address correspondence to L.M.A. (e-mail: [email protected]).
RSNA, 2015
q
RSNA, 2015
q
Radiology: Volume 000: Number 0— 2015 n radiology.rsna.org
1
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
M
ild traumatic brain injury (mTBI), referred to as “concussion,” affects nearly 3.8 million people in the United States annually (1). While awareness of the cognitive consequences of concussive injuries has grown (2), the association with psychiatric disorders has failed to garner the attention given to associated neurodegenerative syndromes, such as Alzheimer dementia and chronic traumatic encephalopathy (3). However, posttraumatic neuropsychiatric disorders can be as disabling as posttraumatic cognitive deficits, leading to decreased quality of life, social isolation, and poor outcomes (4). Depression is by far the most common neuropsychiatric sequela of mTBI (5). Costs associated with depression after mTBI include increased obesity, suicide, substance abuse (6), and school
Advances in Knowledge nn Injured white matter regions that underlie depression after mild traumatic brain injury (mTBI) (difference in fractional anisotropy [FA] in the superior longitudinal fasciculus, P = .006; difference in FA in the nucleus accumbens, P = .03) strongly resemble both those associated with nontraumatic major depressive disorder and those seen with depression after moderate to severe traumatic brain injury. nn White matter abnormalities that underlie anxiety after mTBI (difference in FA in the cerebellar vermis, P = .04) suggest injury to a more primitive fear circuit involving the vermis, compared with the dysfunction in the frontostriatal regions seen in nontraumatic anxiety disorders. nn Injury to the white matter around the nucleus accumbens inversely correlates with recovery in patients with mTBI and depression (r = 20.480, P = .005), suggesting a dysfunctional reward circuit similar to that seen in nontraumatic major depressive disorder. 2
Alhilali et al
failure (7), as well as more severe symptoms, poorer outcomes (5), and prolonged recovery after combat (8). The disability associated with posttraumatic depression is compounded by high comorbidity with posttraumatic anxiety. Posttraumatic anxiety results in greater functional impairment, prolonged recovery, and clinically significant cognitive impairment. In fact, cognitive deficits after mTBI may in part be due to the effects of posttraumatic anxiety (9). In addition to depression and anxiety, irritability is a common neurobehavioral change that accompanies mTBI, ranking among the most commonly selfreported symptoms (10). Posttraumatic irritability has been linked with decreased processing speed and damages relationships with caregivers (10); as a result, this is perhaps why it is associated with poorer outcomes (11). Unfortunately, the mechanisms that underlie neurobehavioral changes after mTBI are not sufficiently understood. In penetrating brain injury, focal injuries in the right orbitofrontal cortex are associated with posttraumatic anxiety and depression (12), while frontal lobe injuries have resulted in posttraumatic irritability (13). However, the relationship between mild trauma and behavioral changes is less clear. Since most patients with mTBI undergo normal conventional magnetic
Implications for Patient Care nn Detection of a similar injury in patients with mTBI and depression and patients with major depressive disorder, with similar prognostic importance regarding abnormalities in the nucleus accumbens, may suggest a common pathophysiology in both traumatic and nontraumatic depression that may help guide treatment. nn Determining a different site of abnormality in patients with posttraumatic anxiety than that seen in anxiety disorders in the general population may indicate that different treatment targets are required for patients with anxiety after trauma.
resonance (MR) imaging, it is uncertain if neuropsychiatric changes in these patients are from discrete injury beneath the resolution of conventional MR imaging or comorbid circumstantial stressors that mimic organic brain injury. Stressors such as combat situations, inability to return to work, and orthopedic injuries are common after mTBI and may influence mood and social functioning (14). However, recent studies have shown that many other postconcussive symptoms previously thought to be unrelated to a focal brain injury have correlated with a central axonal injury, including vestibulopathy, ocular dysfunction, and sleepwake disturbances (3,15). We therefore hypothesize that a central traumatic axonal injury is also a primary feature of the injury pattern that underlies posttraumatic neuropsychiatric symptoms. Thus, the purpose of this study was to determine if a central axonal injury underlies neuropsychiatric symptoms after mTBI by using tract-based spatial statistics (TBSS) analysis of MR images acquired with diffusion-tensor imaging (DTI).
Materials and Methods Study Population Our institutional review board approved this retrospective study, with waiver of Published online before print 10.1148/radiol.2015142974 Content codes: Radiology 2015; 000:1–8 Abbreviations: DTI = diffusion-tensor imaging FA = fractional anisotropy MNI = Montreal Neurological Institute mTBI = mild traumatic brain injury ROI = region of interest TBSS = tract-based spatial statistics Author contributions: Guarantors of integrity of entire study, L.M.A., S.F.; study concepts/study design or data acquisition or data analysis/ interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, all authors; clinical studies, L.M.A., S.F.; experimental studies, L.M.A.; statistical analysis, L.M.A.; and manuscript editing, all authors Conflicts of interest are listed at the end of this article.
radiology.rsna.org n Radiology: Volume 000: Number 0— 2015
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
informed consent. All studies were performed as standard of care, and images were reviewed retrospectively. We searched our electronic medical record to identify MR imaging studies with DTI performed for mTBI. MR imaging reports from January 1, 2006, to January 25, 2014, were searched by using keywords “concussion,” “mild traumatic brain injury,” and “diffusion tensor imaging.” Inclusion criteria were age of 10–50 years, witnessed closed head trauma, no focal neurologic deficit, loss of consciousness of less than 1 minute, posttraumatic amnesia of less than 30 minutes, and English language proficiency. Ninety-five patients were initially identified. Exclusion criteria were history of a neuropsychiatric illness (two patients) or substance abuse (two patients), abnormal computed tomographic (CT) or conventional MR imaging findings (three patients), lack of DTI (four patients) or neurocognitive assessment (six patients), or a total symptom score of zero (three patients). Neuropsychological assessment and neurocognitive testing were performed by a neuropsychologist with more than 14 years of experience in treating patients with mTBI or by a fellow under supervision (both nonauthors). Patients were classified as having depression or anxiety according to the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders, or DSM-5, criteria for a depressive or anxiety disorder due to another medical condition (mTBI), as follows: (a) persistent depressed mood or markedly diminished interest or pleasure in almost all activities that predominates in the clinical picture (for depressive disorder) or panic attacks or anxiety that are predominant in the clinical picture (for anxiety disorder); (b) evidence from the patient history, physical examination, or laboratory findings that the disturbance is a pathophysiological consequence of another medical condition (mTBI); (c) symptoms not better explained by another mental disorder; (d) symptoms not occurring exclusively during delirium; and (e) symptoms causing clinically significant distress or impairment in social,
occupational, or other important areas of functioning. Patients were classified as having irritability on the basis of the DSM5 criteria for the following behavioral findings: (a) temper outbursts, on average three or more times per week; (b) irritable mood, defined as being easily annoyed and provoked to anger most of the day, nearly every day; and (c) never having had a distinct period of 1 day or more of meeting full criteria (other than the duration criteria) for a manic or hypomanic episode. The Immediate Post-Concussion Assessment Cognitive Test was used for neurocognitive testing, which is a computerized test used to measure cognitive function as well as symptoms by using a seven-point Likert scale over 22 different categories. Data recorded for all individuals included sex, age, type of trauma, date of injury, clinical evaluation, Immediate Post-Concussion Assessment Cognitive Test results, prior concussions, imaging results, and clinical management. Time to recovery was defined as when either the patient stated he or she was asymptomatic or the total symptom score was zero.
DTI and Conventional MR Imaging Conventional MR and DTI were performed with a 1.5-T system (Signa; GE Healthcare, Milwaukee, Wis) with an eight-channel head coil. Despite the extended time period over which this study was conducted, all individuals underwent an identical imaging protocol with the same MR system, as follows: sagittal and axial T1-weighted imaging (repetition time [msec]/echo time [msec], 600/minimum setting; section thickness, 5 mm; number of acquisitions, one), axial fast spin-echo proton density imaging (2000–2500/minimum setting; echo train length, four; section thickness, 5 mm; matrix, 320 3 224; number of acquisitions, one), T2-weighted imaging (2000–2500/84– 102; echo train length, 28; section thickness, 5 mm; matrix, 480 3 480; number of acquisitions, one), fluid attenuation inversion-recovery imaging (9000–10 000/149; inversion time, 2200
Radiology: Volume 000: Number 0— 2015 n radiology.rsna.org
Alhilali et al
msec), diffusion-weighted imaging (single-shot echo planar imaging; 10 000/ minimum setting; section thickness, 5 mm; matrix, 128 3 128), and either T2* gradient-recalled-echo imaging (4400/21; number of acquisitions, one; flip angle, 90°; section thickness, 3 mm) or susceptibility-weighted imaging (37/23; number of acquisitions, one; flip angle, 15°; section thickness, 2.4 mm). Field of view ranged from 200 to 240 mm. DTI was performed with a singleshot echo-planar sequence (4000/80; number of acquisitions, two; section thickness, 5 mm; matrix, 128 3 128; field of view, 260 mm). Diffusion gradients were used in 25 noncollinear directions with b values of 0 and 1000 sec/mm2.
TBSS Analysis TBSS from the Functional Magnetic Resonance Imaging of the Brain, or FMRIB, Software Library (version 1.1, http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/) was used to analyze the white matter (16). TBSS is well suited for studies of both adult and pediatric populations, where head size and field of view variation may confound traditional voxelwise methods. As a result, TBSS has been used in combined pediatric and adult populations to evaluate normal white matter development (17,18), as well as pathologic processes, including mTBI (3,15,19). Analysis was as follows. (a) Fractional anisotropy (FA) or diffusivity images, including mean diffusivity, axial diffusivity, and radial diffusivity, were aligned to a target image. (b) This image was affine aligned into 1 3 1 3 1-mm Montreal Neurological Institute (MNI) 152 space. The higher resolution of the MNI space than that of the DTI MR images prevents any marked interpolation blurring when the nonlinear warp plus standard-space affine transformation is applied to the data of each individual. (c) Every image was transformed to MNI 152 space. (d) A mean image was thinned to create a mean FA skeleton. (e) The skeleton was thresholded to remove regions with considerable interindividual variability. 3
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
Every patient’s FA or diffusivity data were projected onto this skeleton for voxel-based cross-subject statistical analysis. We applied a Monte Carlo permutation test by using threshold-free cluster enhancement with a significance level of P less than .05 (5000 permutations), fully corrected for multiple comparisons (family-wise error corrected). A multivariate general linear model to evaluate for two-group differences was created for each neuropsychiatric symptom, adjusted for covariates of age, sex, type of injury (sports related vs non– sports related), duration of symptoms, prior concussion, and presence or absence of each of the other two neuropsychiatric symptoms. This model was applied to the FA, mean diffusivity, radial diffusivity, and axial diffusivity maps for each neuropsychiatric symptom. (Processing was performed by two neuroradiologists with 3 years of image analysis experience [L.M.A., S.F.]).
Region of Interest Analysis Automated region of interest (ROI) analysis was used to quantify FA or diffusivity values in regions of significant differences detected with TBSS by using the mean FA or diffusivity skeleton overlaid with the regions of significant differences between patients and control subjects (corrected, voxelwise). A single mask was created to correspond with the location (in MNI space) and size of the regions with significant differences. FA or diffusivity values of patients and control subjects were then extracted in an automated fashion by using the identical ROI mask along the individual skeletons that were aligned onto common space during the TBSS processing and subsequently compared with two-sample t test. The Cohen d index was used to assess effect size. Correlation of FA or diffusivity values extracted from the ROI and continuous variables was evaluated with Pearson correlation coefficients. Analysis was performed by two neuroradiologists with 3 years of image analysis experience (L.M.A., S.F.). Comparison of proportions and means in the demographic data between each group of neuropsychiatric 4
Alhilali et al
patients and control subjects was conducted by using a Fisher exact test or unpaired t test, respectively. P values less than .05 were considered to indicate a statistically significant difference for demographic factors and voxel-wise analysis, after correcting for multiple comparisons by using a family-wise error correction. P values of .025 were considered to indicate a statistically significant difference for the ROI correlation analysis, given the Bonferroni adjustment described in the power analysis. Analysis was performed by a physician with postgraduate statistics training (L.M.A.). Since we hypothesized relationships between two brain regions with each of the continuous variables for the ROI correlation analysis on the basis of our previous work (15), our a level was determined to be 05/2 = .025, based on Bonferroni conservative adjustment. For each comparison, 18 individuals in each of the two groups (group 1, patient with a symptom; group 2, patients without the symptom) achieve 80% power to detect an effect size of 1.08 (considered a large effect size) at a = .025. Power was computed on the basis of the mean effect size from prior studies, since the effect sizes from our previous data range anywhere from small to very large (1.34 in absolute value) (15,19).
Results Patients Forty-five patients with mTBI who had neuropsychiatric symptoms and 29 control subjects with mTBI who did not have neuropsychiatric symptoms were included (51 male, 23 female; mean age, 18 years; range, 10–47 years). The median time from injury to clinical presentation was 20 days (range, 0–506 days). The most common mechanism of trauma was sports injury (43 patients, 58%). The second most common mechanism was motor vehicle accident (nine patients, 12%). Of the patients with mTBI who had neuropsychiatric symptoms, 38 had irritability, 32 had depression, and 18
had anxiety. Of these, 13 patients had only irritability, one had only depression, 14 had comorbid irritability and depression, seven had comorbid anxiety and depression, and one had comorbid irritability and anxiety. Ten patients fulfilled the criteria for all three conditions. No significant difference was seen in demographic or clinical factors among the different groups of patients with neuropsychiatric symptoms and control subjects. Demographics and clinical characteristics are summarized in Table 1.
Assessment of DTI MR Images with TBSS Depression.—Patients with mTBI who had depression had significantly lower FA in the region of the right nucleus accumbens, anterior limb of the internal capsule, and superior longitudinal fasciculus (Fig 1). There were no regions where control subjects with mTBI had lower FA than patients with depression. No significant differences in mean diffusivity, axial diffusivity, or radial diffusivity were seen between patients with depression and control subjects; however, mean diffusivity, axial diffusivity, and radial diffusivity values in the left aspect of the body of the corpus callosum trended toward a positive correlation with age (P = .10) in those with depression. Anxiety.—Patients with mTBI who had anxiety had significantly lower FA values in the cerebellar vermis (Fig 2). There were no regions where control subjects with mTBI had lower FA values than patients with anxiety. No significant differences in mean diffusivity, axial diffusivity, or radial diffusivity were seen between patients with mTBI who had anxiety and control subjects. Irritability.—No regions of significantly increased or decreased FA, mean diffusivity, radial diffusivity, or axial diffusivity were seen in patients with irritability compared with control subjects at voxel-wise analysis. Assessment of DTI MR Images with ROI Analysis In patients with depression, FA in the region of the nucleus accumbens
radiology.rsna.org n Radiology: Volume 000: Number 0— 2015
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
Alhilali et al
Table 1 Comparison of Demographic and Clinical Characteristics among Patients with mTBI Who Have Irritability, Depression, and Anxiety and Control Subjects with mTBI but No Neuropsychiatric Symptoms Comparison between Patients with Irritability and Control Subjects Parameter Mean age (y) No. of male subjects† Median time to presentation (d) No. of individuals with prior concussion† No. of individuals with sports injury† Mean ImPACT‡ total symptom score percentile Mean verbal memory score percentile Mean visual memory score percentile Mean processing speed percentile Mean reaction time percentile Median time to recovery (wk)
Comparison between Patients with Depression and Control Subjects
Comparison between Patients with Anxiety and Control Subjects
Control Subjects (n = 29)
Patients with Irritability (n = 38)
P Value*
Patients with Depression (n = 32)
P Value*
Patients with Anxiety (n = 18)
P Value*
16.9 (10–28) 21 (72) 16 (1–275) 13 (45) 20 (69) 29.9 (0–84)
18.2 (11–38) 27 (71) 32 (4–506) 15 (39) 20 (53) 34.3 (2–97)
.36 ..99 .22 .80 .21 .46
19.3 (11–47) 20 (62) 33 (0–506) 11 (34) 15 (47) 36.7 (3–97)
.11 .43 .99 .44 .13 .27
20.1 (11–47) 14 (78) 17 (3–361) 7 (39) 10 (56) 35.2 (4–97)
.08 .74 .13 .77 .37 .48
31.0 (0–84) 23.5 (1–90) 36.6 (1–98) 40.4 (1–94) 42.9 (1–252)
30.6 (1–99) 31.9 (1–97) 38.2 (1–95) 30.6 (1–97) 30.2 (2–194)
.69 .24 .84 .22 .31
30.3 (1–99) 28.6 (1–97) 39.1 (1–95) 27.3 (1–95) 30.8 (2–194)
.65 .46 .75 .11 .41
39.7 (1–80) 18.8 (1–64) 36.7 (1–93) 28.3 (1–91) 30.8 (1–189)
.66 .51 .97 .19 .62
Note.—Numbers in parentheses are ranges, unless indicated otherwise. ImPACT = Immediate Post-Concussion Assessment and Cognitive Test. * P values represent comparisons with control subjects with mTBI who did not have neuropsychiatric symptoms. All P values were two tailed and calculated with an unpaired Student t test for continuous variables and a Fisher exact test for categorical variables. †
Numbers in parentheses are percentages.
‡
Scores are percentiles determined by means of normative data obtained from baseline testing of more than 17 000 athletes as part of their participation before sports, with percentile information accounting for both patient sex and age.
Figure 1
Figure 1: Coronal images derived from TBSS results and rendered on T1-weighted MR images from the MNI atlas show how patients with depression have significantly decreased FA in the region of A, the right nucleus accumbens (arrow) or “reward center,” as well as B, the anterior limb of the internal capsule (arrow) and C, the superior longitudinal fasciculus (arrow)—regions that are involved in depressive disorders in nontraumatic and traumatic settings. Voxels with significant differences detected with TBSS (P , .05 corrected for multiple comparisons) were thickened by using the TBSS fill function into local tracts (red) and overlaid on the white matter skeleton (blue).
identified with TBSS inversely correlated with time to recovery (r = 20.480, P = .005). In patients without depression, no correlation was seen (r = 0.238, P = .21). Summaries of the ROI analysis are provided in Tables 2 and 3.
Discussion By using voxel-based analysis of DTI MR images, we found a central axonal injury pattern underlying posttraumatic depression and anxiety but not irritability. This suggests that posttraumatic
Radiology: Volume 000: Number 0— 2015 n radiology.rsna.org
depression and anxiety may be the result of discrete injuries not visible with the resolution of conventional MR imaging. Injury in the region of the nucleus accumbens inversely correlated with recovery time in patients with depression, while injury to the vermis 5
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
Alhilali et al
Figure 2
Figure 2: A, Axial and B, sagittal images derived from TBSS results and rendered on T1-weighted MR images from the MNI atlas indicate that significant white matter differences in patients with mTBI and anxiety involve the cerebellar vermis, the region responsible for fear conditioning. Voxels with significant differences detected with TBSS (P , .05 corrected for multiple comparisons) were thickened by using the TBSS fill function into local tracts (red) and overlaid on the white matter skeleton (blue).
in patients with anxiety trended toward correlation with neurocognitive test performance. Recognition of this injury pattern on images may help predict symptoms on the basis of the location of damage, as well as provide a quantitative imaging biomarker that may ultimately be used in the determination of prognosis and in the measurement of response to treatment. Previous studies on the evaluation of white matter injury in depression and anxiety after mTBI have been focused mainly on blast injury (20–22) or professional athletes (23,24). Unfortunately, these studies have limited applicability to the general mTBI population. Blast injuries are theorized to have different underlying mechanisms than civilian mTBI injuries (22) and are seen in military populations with unique emotional stressors, which may affect neuropsychological test results (22). Similarly, professional athletes have a unique lifestyle and risk factors not typically seen in the general population (24). Imaging studies in which post-mTBI depression and anxiety were evaluated in the general population have either had very small sample sizes (25) or only involved the comparison between patients with mTBI who had 6
neuropsychiatric symptoms and healthy control subjects, rather than between patients with mTBI who had neuropsychiatric symptoms and other patients with mTBI who did not have neuropsychiatric symptoms (26,27). This is important, because numerous white matter abnormalities seen in patients with mTBI when compared with healthy control subjects are either clinically silent or remote abnormalities, which are likewise not symptomatic (3). These background lesions create noise that may obscure the true lesions that underlie neuropsychiatric symptoms. No imaging studies have yet been conducted to evaluate the white matter injuries that underlie depression, irritability, or anxiety in a large general mTBI population in which other patients with mTBI are used as control subjects. Most importantly, however, the clinical significance of these white matter abnormalities with respect to the prognosis of neuropsychiatric patients with mTBI has not been addressed. In our study, injured regions in depression consisted of both regions known to be abnormal in nontraumatic major depressive disorder and regions injured in posttraumatic depression after severe or repetitive trauma. The three abnormal regions (nucleus
accumbens, anterior limb of the internal capsule, and superior longitudinal fasciculus) in our study have all been found to be abnormal in patients with nontraumatic major depressive disorder (28–30). Interestingly, both the nucleus accumbens and the superior longitudinal fasciculus, but not the anterior limb of the internal capsule, have also been found to be injured in patients with depression after moderate to severe traumatic brain injury (31) and repetitive mTBI (23,24). Injuries to regions exclusively seen in nontraumatic depression, as well as regions seen in posttraumatic depression, suggest that development of post-mTBI depression is multifactorial. Depression after mTBI may be the result of injury to regions known to result in depression after trauma (nucleus accumbens, superior longitudinal fasciculus), combined with additional preexisting abnormalities in regions that render these patients susceptible to depression outside of any traumatic setting (anterior limb of the internal capsule). Additionally, the fact that abnormalities are seen in both the nucleus accumbens and the superior longitudinal fasciculus in both nontraumatic major depressive disorder and posttraumatic depression may indicate that perhaps unrecognized trauma plays a role in the development of major depressive disorder in the general population, as well. Severity of injury to the nucleus accumbens inversely correlated with time to recovery in patients with depression. This correlates well with findings in nontraumatic major depressive disorder, in which a dysfunctional reward circuit is believed to prohibit recovery, and deep brain stimulation of the nucleus accumbens results in improvement in depression that was treatment resistant previously (29,32). Unlike patients with mTBI and depression, patients with mTBI and anxiety demonstrated injured regions distinct from abnormal regions seen in nontraumatic anxiety disorders. Anxiety disorders in the general population are associated with frontostriatal abnormalities (9). In our mTBI population, however, abnormalities were seen in the vermis. The vermis is integral
radiology.rsna.org n Radiology: Volume 000: Number 0— 2015
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
Alhilali et al
Table 2 Comparison of FA Values in ROIs for Patients with mTBI with and without Depression and Those with and without Anxiety Parameter Mean FA value for the region of the nucleus accumbens Mean FA value for the anterior limb of the internal capsule Mean FA value for the superior longitudinal fasciculus Mean FA value for the cerebellum
Patients with mTBI and Depression
Patients with mTBI and Anxiety
Control Subjects with mTBI
P Value*
0.357 (0.233, 0.481) 0.508 (0.429, 0.588) 0.435 (0.383, 0.487) …
… … … 0.265 (0.215, 0.315)
0.391 (0.287, 0.495) 0.532 (0.440, 0.625) 0.456 (0.411, 0.501) 0.296 (0.236, 0.356)
.026 (0.616) .020 (0.530) .006 (0.863) .043 (0.437)
Note.—Numbers in parentheses are 95% confidence intervals, unless indicated otherwise. * P values were two tailed and calculated with an unpaired Student t test. Numbers in parentheses are the Cohen d index.
Table 3 Correlation of FA with Clinical Findings for ROIs in Patients with mTBI and Depression and Those with mTBI and Anxiety Parameter Correlation* with symptom severity score† Correlation with verbal memory score Correlation with visual memory score Correlation with processing speed score Correlation with reaction time score Correlation with time to recovery
Mean FA Value for the Region of the Nucleus Accumbens ROI
Mean FA Value for the Anterior Limb of the Internal Capsule ROI
Mean FA Value for the Superior Longitudinal Fasciculus ROI
Mean FA Value for the Cerebellum ROI
20.097 (.60) 20.038 (.84) 0.037 (.84) 20.013 (.94) 0.109 (.55) 20.480 (.005)‡
20.203 (.27) 20.181 (.32) 20.146 (.43) 20.286 (.11) 20.061 (.74) 20.167 (.36)
20.119 (.52) 20.265 (.14) 0.168 (.36) 0.245 (.18) 0.048 (.79) 20.133 (.47)
0.149 (.55) 0.054 (.83) 0.231 (.36) 0.020 (.94) 20.086 (.74) 20.130 (.61)
Note.—Numbers in parentheses are P values. * Correlation performed with the Pearson correlation coefficient. †
Two-tailed P value for the Pearson correlation coefficient.
‡
Statistically significant correlation.
to fear conditioning and controls the autonomic and emotional response to aversive conditioning, with connections to both the catecholaminergic centers (locus coeruleus and ventral tegmental area) and emotional centers (amygdala, septum, and locus coeruleus) (33). Patients’ vermian lesions are unable to calm autonomic and emotional responses during fear conditioning (33,34). This suggests that anxiety in patients with mTBI results from dysfunction of a more primitive fear circuit than the higher-order frontostriatial dysfunction seen with anxiety outside of trauma. There were several limitations to our study. Our study was retrospective in nature, with a moderate sample size. This moderate sample size prevented detailed subgroup analysis of possible distinct injury patterns in patients with more than one neuropsychiatric symptom. Furthermore, there is no
generally accepted method to evaluate for overfitting of our voxelwise statistical model, which could limit the capacity to generalize the findings. As such, the findings should be confirmed in a prospective study with a larger cohort. Additionally, there is likely selection bias toward more seriously injured patients, since many patients with mTBI do not undergo imaging. However, we would contend that these more symptomatic patients are the type of patients in which biomarkers of injury are most needed. Finally, in our study, we used a 1.5-T magnet, which results in decreased image resolution and has poorer signal-to-noise ratio than the 3.0-T system used in DTI (35). It is unclear if additional white matter injuries may be detected by using a 3.0-T magnet, and future work should include higher field strengths. In summary, detection of the central white matter injuries that underlie
Radiology: Volume 000: Number 0— 2015 n radiology.rsna.org
depression and anxiety but not irritability indicates that not all neuropsychiatric symptoms after mTBI are the result of discrete white matter injuries, but in those with corresponding injuries, the injured regions provide insight into the underlying pathophysiology and prognosis. Detection of similar injuries in both patients with mTBI who have depression and those who have major depressive disorder, with similar prognostic importance to abnormalities in the nucleus accumbens, may suggest a common pathophysiology in both traumatic and nontraumatic depression that may help guide treatment. Determining a different site of abnormality in patients with posttraumatic anxiety than in those with anxiety disorders in the general population may indicate that different treatment targets are required for patients with anxiety after trauma. 7
NEURORADIOLOGY: Evaluation of White Matter Injury Patterns after Mild Traumatic Brain Injury
Disclosures of Conflicts of Interest: L.M.A. disclosed no relevant relationships. J.A.D. disclosed no relevant relationships. S.G. disclosed no relevant relationships. S.F. disclosed no relevant relationships.
References 1. Finkelstein EA, Corso PS, Miller TR. The incidence and economic burden of injuries in the United States. New York, NY: Oxford University Press, 2006. 2. Petraglia AL, Plog BA, Dayawansa S, et al. The spectrum of neurobehavioral sequelae after repetitive mild traumatic brain injury: a novel mouse model of chronic traumatic encephalopathy. J Neurotrauma 2014;31(13):1211–1224. 3. Fakhran S, Yaeger K, Alhilali L. Symptomatic white matter changes in mild traumatic brain injury resemble pathologic features of early Alzheimer dementia. Radiology 2013;269(1):249–257.
Alhilali et al
idence of right brain injury in referrals to a neuropsychiatric brain injury unit? Brain Inj 2001;15(1):65–69. 13. Grafman J, Schwab K, Warden D, Pridgen A, Brown HR, Salazar AM. Frontal lobe injuries, violence, and aggression: a report of the Vietnam Head Injury Study. Neurology 1996; 46(5):1231–1238. 14. Parker RS. The spectrum of emotional distress and personality changes after minor head injury incurred in a motor vehicle accident. Brain Inj 1996;10(4):287–302. 15. Alhilali LM, Yaeger K, Collins M, Fakhran S. Detection of central white matter injury underlying vestibulopathy after mild traumatic brain injury. Radiology 2014;272(1):224– 232. 16. Smith SM, Jenkinson M, Johansen-Berg H, et al. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 2006;31(4):1487–1505.
4. Vanderploeg RD, Curtiss G, Luis CA, Salazar AM. Long-term morbidities following self-reported mild traumatic brain injury. J Clin Exp Neuropsychol 2007;29(6):585– 598.
17. Peters BD, Szeszko PR, Radua J, et al. White matter development in adolescence: diffusion tensor imaging and meta-analytic results. Schizophr Bull 2012;38(6):1308– 1317.
5. Moldover JE, Goldberg KB, Prout MF. Depression after traumatic brain injury: a review of evidence for clinical heterogeneity. Neuropsychol Rev 2004;14(3):143–154.
18. Giorgio A, Watkins KE, Chadwick M, et al. Longitudinal changes in grey and white matter during adolescence. Neuroimage 2010;49(1):94–103.
6. Saluja G, Iachan R, Scheidt PC, Overpeck MD, Sun W, Giedd JN. Prevalence of and risk factors for depressive symptoms among young adolescents. Arch Pediatr Adolesc Med 2004;158(8):760–765.
19. Fakhran S, Yaeger K, Collins M, Alhilali L. Sex differences in white matter abnormalities after mild traumatic brain injury: localization and correlation with outcome. Radiology 2014;272(3):815–823.
7. Fletcher JM. Adolescent depression: diagnosis, treatment, and educational attainment. Health Econ 2008;17(11):1215–1235. 8. Hoge CW, McGurk D, Thomas JL, Cox AL, Engel CC, Castro CA. Mild traumatic brain injury in U.S. Soldiers returning from Iraq. N Engl J Med 2008;358(5):453–463. 9. Moore EL, Terryberry-Spohr L, Hope DA. Mild traumatic brain injury and anxiety sequelae: a review of the literature. Brain Inj 2006; 20(2):117–132. 10. Yang CC, Hua MS, Lin WC, Tsai YH, Huang SJ. Irritability following traumatic brain injury: divergent manifestations of annoyance and verbal aggression. Brain Inj 2012; 26(10):1185–1191. 11. Yang CC, Tu YK, Hua MS, Huang SJ. The association between the postconcussion symptoms and clinical outcomes for patients with mild traumatic brain injury. J Trauma 2007; 62(3):657–663. 12. Borek LL, Butler R, Fleminger S. Are neuropsychiatric symptoms associated with ev-
8
20. Matthews SC, Strigo IA, Simmons AN, O’Connell RM, Reinhardt LE, Moseley SA. A multimodal imaging study in U.S. veterans of Operations Iraqi and Enduring Freedom with and without major depression after blast-related concussion. Neuroimage 2011;54(Suppl 1):S69–S75. 21. Matthews SC, Spadoni AD, Lohr JB, Strigo IA, Simmons AN. Diffusion tensor imaging evidence of white matter disruption associated with loss versus alteration of consciousness in warfighters exposed to combat in Operations Enduring and Iraqi Freedom. Psychiatry Res 2012;204(2-3):149–154. 22. Morey RA, Haswell CC, Selgrade ES, et al. Effects of chronic mild traumatic brain injury on white matter integrity in Iraq and Afghanistan war veterans. Hum Brain Mapp 2013; 34(11):2986–2999. 23. Strain J, Didehbani N, Cullum CM, et al. Depressive symptoms and white matter dysfunction in retired NFL players with concussion history. Neurology 2013;81(1):25–32.
24. Hart J Jr, Kraut MA, Womack KB, et al. Neuroimaging of cognitive dysfunction and depression in aging retired National Football League players: a cross-sectional study. JAMA Neurol 2013;70(3):326–335. 25. Rao V, Mielke M, Xu X, et al. Diffusion tensor imaging atlas-based analyses in major depression after mild traumatic brain injury. J Neuropsychiatry Clin Neurosci 2012; 24(3):309–315. 26. Guskiewicz KM, Marshall SW, Bailes J, et al. Association between recurrent concussion and late-life cognitive impairment in retired professional football players. Neurosurgery 2005;57(4):719–726; discussion 719–726. 27. Levin HS, Wilde E, Troyanskaya M, et al. Diffusion tensor imaging of mild to moderate blast-related traumatic brain injury and its sequelae. J Neurotrauma 2010;27(4):683–694. 28. Hall GB, Milne AM, Macqueen GM. An fMRI study of reward circuitry in patients with minimal or extensive history of major depression. Eur Arch Psychiatry Clin Neurosci 2014;264(3):187–198. 29. Morishita T, Fayad SM, Higuchi MA, Nestor KA, Foote KD. Deep brain stimulation for treatment-resistant depression: systematic review of clinical outcomes. Neurotherapeutics 2014;11(3):475–484. 30. Lyden H, Espinoza RT, Pirnia T, et al. Electroconvulsive therapy mediates neuroplasticity of white matter microstructure in major depression. Transl Psychiatr 2014;4:e380. 31. Maller JJ, Thomson RH, Pannek K, The (eigen)value of diffusion tensor ing to investigate depression after matic brain injury. Hum Brain Mapp 35(1):227–237.
et al. imagtrau2014;
32. Millet B, Jaafari N, Polosan M, et al. Limbic versus cognitive target for deep brain stimulation in treatment-resistant depression: accumbens more promising than caudate. Eur Neuropsychopharmacol 2014;24(8):1229–1239. 33. Sacchetti B, Scelfo B, Strata P. Cerebellum and emotional behavior. Neuroscience 2009; 162(3):756–762. 34. Maschke M, Schugens M, Kindsvater K, et al. Fear conditioned changes of heart rate in patients with medial cerebellar lesions. J Neurol Neurosurg Psychiatry 2002;72(1):116–118. 35. Alexander AL, Lee JE, Wu YC, Field AS. Comparison of diffusion tensor imaging measurements at 3.0 T versus 1.5 T with and without parallel imaging. Neuroimaging Clin N Am 2006;16(2):299–309, xi.
radiology.rsna.org n Radiology: Volume 000: Number 0— 2015
