Clinical and Magnetic Resonance Spectroscopic Imaging Findings in Veterans With Blast Mild Traumatic Brain Injury and Post-Traumatic Stress Disorder

Abstract

Objectives

To compare magnetic resonance spectroscopic imaging (MRSI) findings from the hippocampal regions of military veterans with blast-related mild traumatic brain injury (blast mTBI) and post-traumatic stress disorder (PTSD) to those with PTSD only; and to examine the relationship of MRSI findings to cognitive and neuromotor impairment.

Methods

35 military veterans—23 with blast mTBI and PTSD (blast mTBI/PTSD) and 12 with PTSD only participated in the study. Whole plane MRSI data including N-acetyl aspartate (NAA) and choline (Ch) were acquired at 7T for the hippocampus. Concurrent cognitive and neuromotor data were collected using established assessments. General linear models (GLMs) with Bonferroni correction were used to compare the two groups on NAA/Ch ratios across regions of the hippocampus. Spearman’s correlations were used to examine correlations between NAA/Ch and cognitive and neuromotor impairment.

Results

The NAA/Ch results for the left hippocampus were lower in the blast mTBI/PTSD group than the PTSD-only group. The blast mTBI/PTSD group also scored worse on the WAIS-IV-vocabulary. Significant correlations between NAA/Ch and neuromotor outcomes—including vestibular impairment—were supported.

Conclusions

Combined MRSI and cognitive and neuromotor data may help inform more objective and accurate diagnoses and effective treatments for patients with blast mTBI and PTSD.

INTRODUCTION

Since 2000, approximately 280,000 U.S. military personnel have sustained a mild traumatic brain injury (mTBI).¹ It is estimated that annual health care costs for U.S. military personnel with mTBI are $12,990 compared to $7,377 for non-TBI controls.² Nearly one in five veterans experience lingering symptoms or impairment related to their mTBI.³ Between 20% and 40% of military personnel experience post-traumatic stress (PTS) symptoms or post-traumatic stress disorder (PTSD) following exposure to mTBI.^4,5 Long-term effects of mTBI and PTSD include residual mental and cognitive impairment, increased suicide risk, poor physical health, unemployment, homelessness, and substance abuse.⁶ Exposure to blast forces such as those from improvised explosive devices, exploding munitions, and other sources is particularly problematic for military personnel, and may result in complicated polytrauma presentations that are difficult to diagnose and treat.^7,8 It is challenging for clinicians to distinguish PTSD from lingering mTBI symptoms in this at-risk group, as there is considerable overlap between them. Both PTSD and mTBI symptoms include anxiety, irritability, memory problems, and sleep disturbances.⁹ This overlap in symptoms presents a challenge in determining the etiology of mTBI versus PTSD symptoms following exposure to mTBI. Misattribution of symptoms to PTSD or mTBI may result in ineffective treatment and prolonged symptoms and morbidity. For example, military personnel misdiagnosed as having mTBI without PTSD may not receive the indicated trauma-focused exposure therapy that is effective with PTSD.¹⁰ In contrast, military personnel misdiagnosed with PTSD only may engage in this therapy with limited success as a result of mTBI-related impaired consciousness.¹¹ If an incorrect diagnosis and treatment were provided, military personnel may return to duty while still experiencing symptoms that were not properly treated, thereby increasing their risk for continued morbidity associated with the untreated condition. This misidentification may, in part, influence the residual PTS symptoms associated with prior exposure to blast mTBI reported in the literature.¹² A better approach to differentiating PTSD from mTBI is warranted, as there is currently no objective method for extricating the effects of one from the other.

Researchers have reported that magnetic resonance spectroscopic imaging (MRSI) may be sensitive to detecting localized neuronal injury in the hippocampus associated with blast-related brain injury.¹³ Previously, we demonstrated that in veterans with a history of mTBI due to blast with self-reported persistent memory impairment, significant declines in the ratios of N-acetyl aspartate to choline (NAA/Ch) were seen in the anterior portions of both hippocampi more than 1 year postinjury in comparison to age-matched control subjects.¹⁴ These changes are consistent with both neuronal loss and impairment (NAA is synthesized only in neuronal mitochondria)¹⁵ and axonal injury (Ch is known to be increased following TBI¹⁶ and multiple sclerosis).¹⁷ The changes seen were independent of the presence or absence of PTSD and depression. In an extension study that added a small group of veterans with PTSD without a history of mTBI, de Lanerolle et al¹³ reported that hippocampal levels of NAA/Ch were generally intermediate between control and veterans with blast-related mTBI. These preliminary findings suggest that MRSI may offer a useful neuroimaging tool to help differentiate mTBI from PTSD and corroborate current clinical delineations based on functional assessments.

The primary purpose of the current study was to identify clinical and imaging biomarkers that might help differentiate veterans with blast mTBI and PTSD from those with PTSD only. To that end, we compared MRSI findings from the hippocampal regions of military veterans with blast mTBI and PTSD to those with PTSD only. We expected that NAA/Ch ratios would be lower in veterans with blast mTBI and PTSD compared to PTSD. A secondary purpose of this study was to examine the relationship of cognitive and neuromotor impairment to MRSI findings across the two groups. We expected a negative correlation between cognitive and neuromotor impairment and NAA/Ch ratios.

METHODS

Participants

U.S. military veterans between the ages of 18 and 65 were identified for the study through clinics within the Veterans Administration Pittsburgh Healthcare System (VAPHS) and referred for screening by their providers. All participants were deemed cognitively able to provide informed consent and provided written informed consent before the initiation of study procedures. Exclusion criteria included 1) ferrous metal in their body and 2) claustrophobia or a body mass index in excess of 34. A total of 50 veterans from the VAPHS met study exclusion criteria and were enrolled in the study. However, only 35/50 (70%) met the following study inclusion criteria: 1) a complete 7T MRI scan, 2) complete cognitive and neuromotor testing, and 3) a current medical diagnosis (per VA medical records) of either: a) blast mTBI and PTSD (n = 23), or b) PTSD only (i.e., no current, symptomatic mTBI) (n= 12). The sample ranged in age from 25 to 65 years. A comparison of select characteristics for the two groups is provided in Table I. Please note that TBI history data presented in Table I are self-reported. As expected, significantly more of the participants in the blast mTBI/PTSD group reported a history of mTBI than in the PTSD-only group. No other group differences were noted.

TABLE I.

Comparison of Select Characteristics for the Mild Traumatic Brain Injury/Post-Traumatic Stress Disorder (mTBI/PTSD) (n = 23) and PTSD only (n = 12) Groups

Group	Age Mean (SD)	Sex n (%)	mTBI History n (%)
mTBI/PTSD	35.00 (9.66)	23 M (100)	23 (100)*
PTSD Only	37.67 (13.18)	10 M (83) 2 F (11)	7 (58)
Total	35.91 (10.87)	33 M (94)2 F (6)	30 (86)

F, female; M, male.

^*p < 0.001.

Measures

Magnetic Resonance Spectroscopic Imaging

Whole-plane MRSI data including NAA and Ch were acquired at 7T using a single-slice acquisition (10 mm thickness) with 24 × 24 encodes over the entire field of view of 240 × 240 mm. A transceiver array (Resonance Research Inc)¹⁸ was used for both transmission and reception. Extra-cerebral lipid contamination was minimized using radio frequency-based outer volume suppression.¹⁹ A moderate echo time (40 milliseconds) was used to minimize contributions from macromolecule and amino acid resonances underlying the NAA and creatine (Cr) resonances. With TR = 1,500 ms, the acquisition time was 14.5 minutes. To maximize spectral quality, a very high order shim insert²⁰ was used, which provided third- and fourth-degree shims. The MRSI data from each channel of the transceiver array was phased and scaled using a gradient echo reference image before summation of the individual channel data.²¹ To minimize the effects of metabolite heterogeneity along the hippocampal formation, the data were reconstructed using single voxel reconstructions.¹⁹ Briefly, hippocampi were manually outlined on T1-weighted images, and a midline running along the length of the hippocampi was calculated. The aqueduct was identified and spectra starting at the level of the aqueduct were reconstructed using voxel shifting methods to provide six loci along the hippocampal formation. The loci were numbered from 1 (most posterior) to 6 (most anterior) with three loci anterior to the aqueduct and two posterior (Fig. 1). Spectra were analyzed by fitting the data with three Gaussian lines (NAA, Cr, Ch) and calculating the ratio Ch/NAA from the integrated areas of Ch and NAA.²²

FIGURE 1.

Sample magnetic resonance spectroscopic imaging data from a veteran with blast mTBI and PTSD from the current study. (Ch, choline; Cr, creatine; mTBI, mild traumatic brain injury; NAA, N-acetyl asparatate; PTSD, post-traumatic stress disorder).

Spectra were reconstructed from six loci spanning the hippocampi. The right and left hippocampi were manually delineated (green and red tracings) and a midline (blue) was automatically calculated. The aqueduct was identified (white cross hairs) and spectra were reconstructed in 9-mm increments starting at the intersection of the midline and the white horizontal line moving along the midline (three anterior and two posterior). The loci were numbered 1 (most posterior) to 6 (most anterior).

PTSD Assessment

The PTSD Checklist (PCL) was used to assess self-reported PTSD symptoms. The PCL encompasses 17 items and requires individuals to indicate on a scale of 1 (not at all) to 5 (extremely) how much each item bothered them during the past month. The PCL yields a PTSD symptom score that ranges from 17 to 85. Higher scores represent higher levels of PTS symptomology.

Cognitive and Neuromotor Assessments

Cognitive assessments included 1) Wechsler Adult Intelligence Scale-IV Vocabulary (WAIS-IV Vocabulary) test, which measures a subjects’ ability to explain the meaning of words; 2) Wechsler Test of Adult Reading (WTAR), which measures intelligence using word pronunciation; 3) Wisconsin Card Sorting Test (WCST), which measures executive function using a card sorting task based on numbers, form, and color categories; 4) Delis-Kaplan Executive Function (D-KEFS) verbal fluency test, which measures letter and category fluency; and 5) Trails A and B, which measure attention and task switching during a timed scanning task. Neuromotor assessments included 1) grooved pegboard test, which measures manipulative dexterity in a timed dominant and nondominant hand pegboard task and 2) the Vestibular/Ocular Motor Screening (VOMS) tool, which measures vestibular and oculomotor symptoms and impairment following the performance of smooth pursuit, horizontal and vertical saccades, vestibular ocular reflex (VOR), and visual motion sensitivity (VMS) tasks.²³ The VOMS also includes a measurement of near-point convergence (NPC) distance in centimeters.

Procedures

Institutional Review Board and Office of Research and Development approvals were obtained from the VAPHS and University of Pittsburgh before study initiation. Following these approvals and written informed consent procedures, all participants completed the cognitive and neuromotor tests in a dedicated interview room at the VAPHS. The cognitive and neuromotor tests were administered by a member of the research team trained in neuropsychological test administration in the following order: PCL-C, Trails A&B, WAIS-IV, Grooved Peg Board, VOMS, WCST, D-KEFS, and WTAR. Other clinical tests were administered subsequently as part of an ongoing National Institutes of Health-funded study (R01NS081772), but are not reported here. The cognitive and neuromotor components required approximately 2.5 to 3 hours to complete. After participants completed the cognitive and neuromotor portions of the study, they were given a 60-minute break and then escorted from the VAPHS to the University of Pittsburgh Magnetic Resonance Research Center for the 7T MRSI scan. The MRSI scan process required an additional hour to complete. The combined total testing time for each participant was approximately 4.5 to 5 hours. Participants were compensated for their participation.

Analysis

We compared the mTBI/PTSD- and PTSD-only groups on NAA/Ch across the six different loci of both the right and left hippocampus and both cognitive and neuromotor assessments using a series of general linear models (GLMs) with Bonferroni correction. The GLMs were used to account for the unequal cell sizes. We also conducted a series of Spearman correlations among ratios of NAA/Ch in the hippocampus and cognitive and neuromotor assessments. Spearman correlations were used because the data were not normally distributed. All statistical analyses were performed using Statistical Analysis System, version 9.3 (SAS Institute, Inc, Cary, North Carolina), with p < 0.05.

RESULTS

Mean Comparisons of Blast mTBI/PTSD and PTSD Groups

The results of the GLMs with Bonferroni correction supported significant differences in the ratio of NAA/Ch in three left anterior loci of the hippocampus (Fig. 2). Specifically, the ratio of NAA/Ch in the left anterior locus (Number 4) was significantly (F = 8.24, p = 0.007) lower in the blast mTBI/ PTSD group (M = 1.20, SD = 0.22) compared to the PTSD-only group (M = 1.41, SD = 0.20). Likewise, the ratio of NAA/Ch in the left anterior locus (#5) was significantly (F = 7.74, p = 0.009) lower in the blast mTBI/PTSD group (M = 1.14, SD = 0.17) compared to the PTSD-only group (M = 1.29, SD = 0.11), and the ratio in the left anterior locus (#6) was significantly (F = 6.82, p = 0.01) lower in the blast mTBI/PTSD group (M = 1.01, SD = 0.17) compared to the PTSD-only group (M = 1.15, SD = 0.10).With regard to cognitive function, the blast mTBI/PTSD group (M = 29.91, SD = 9.69) scored significantly (F = 5.99, p = 0.02) lower on the WAIS-IV vocabulary than the PTSD-only group (M = 38.00, SD = 8.39). No other significant group differences were noted.

FIGURE 2.

NAA/Ch ratios for blast mTBI/PTSD (n = 23) and PTSD-only (n = 12) groups in left (L) and right (R) hippocampal loci. (Ch = choline; mTBI = mild traumatic brain injury; NAA = N-acetyl aspartate; PTSD = post-traumatic stress disorder).

Correlations Between Cognitive and Neuromotor Impairment and MRSI Findings

Results of a series of Spearman correlations for NAA/Ch in the hippocampus and cognitive impairment did not support any significant relationships between the WAIS-IV vocabulary and arithmetic, WTAR, WCST, D-KEFS verbal fluency, Trails A and B, and NAA/Ch ratios for any of the loci of the hippocampus. Results of a series of Spearman correlations for NAA/Ch in the hippocampus and neuromotor impairment revealed several significant relationships (Table II). Specifically, higher symptom provocation on smooth pursuits, saccades, VOR, and VMS were all correlated with decreased NAA/Ch ratios across multiple left and right loci of the hippocampus. Slower performance on the grooved peg board test nondominant hand was correlated with decreased NAA/Ch in the left #2 locus (Spearman’s rho = −0.36, p = 0.04) of the hippocampus. The average NPC distance from the VOMS was not related to any of the NAA/Ch ratios in the loci of the hippocampus.

TABLE II.

Correlations Between Vestibular/Ocular Motor Screening Scores and N-Acetyl Aspartate/Choline (NAA/Ch) Ratios in Loci of the Left and Right Hippocampus (N = 35)

NAA/Ch Loci	Smooth Pursuits	Horizontal Sacc	Vertical Sacc	Horizontal VOR	Vertical VOR	VMS	NPC
Right #6Left #6	−0.62***−0.38*	−0.61***−0.37*	−0.61***−0.34*	−0.58***−0.39*	−0.58***−0.32	−0.07−0.33	0.170.10
Right #5Left #5	−0.41*−0.34*	−0.42*−0.35*	−0.42**−0.35*	−0.58***−0.34*	−0.58***−0.31	−0.21−0.25	−0.080.01
Right #4Left #4	−0.36*−0.16	−0.38*−0.17	−0.36*−0.14	−0.38*−0.16	−0.38*−0.10	−0.23−0.36*	−.060.14
Right #3Left #3	−0.27−0.14	−0.26−0.15	−0.24−0.12	−0.22−0.10	−0.22−0.10	−0.35*−0.45**	0.050.10
Right #2Left #2	−0.45**−0.14	−0.44**−0.15	−0.43**−0.11	−0.42**−0.10	−0.42**−0.07	−0.29−0.40*	−0.100.16
Right #1Left #1	−0.43−0.27	−0.43*−0.29	−0.42**−0.28	−0.46**−0.32	−0.46**−0.25	−0.12−0.17	−0.190.15

NPC, near point convergence; Sacc, saccades; VMS, visual motion sensitivity; VOR, vestibular ocular reflex).

^*p < 0.05;

^**p < 0.01;

^***p < 0.001.

DISCUSSION

To our knowledge the current study is the first to compare MRSI findings and cognitive and neuromotor impairments between military veterans with blast mTBI/PTSD and PTSD only. Consistent with previous findings¹³ and our hypothesis, we reported lower NAA/Ch ratios in anterior regions of the hippocampus in the blast mTBI/PTSD group compared to the PTSD-only group. The current findings support significant differences in the left anterior hippocampus, whereas, the previous findings, although reporting reductions in NAA/ Ch in the anterior loci (#5, #6) of both hippocampi, did not reach statistical significance. This discrepancy may be due to differences between the two veteran populations studied. The previous blast mTBI and PTSD population included symptomatic patients who had subjective cognitive complaints (e.g., memory problems).¹³ In contrast, participants in the current study were not selected based on subjective cognitive complaints. In addition, the current sample completed testing at a later time with respect to blast mTBI than the previous study. The PTSD-only population was older in the deLanerolle et al study, (43 ± 9 years), as compared to the current study (38 + 14 years). However, the blast mTBI/ PTSD groups in the two studies were nearly identical in age (deLanerolle et al sample = 34.1 ± 8 years; current sample = 34 ± 9 years).¹³ Also, given that NAA levels in the hippocampus are known to decline with age in normal aging,¹⁹ additional research focusing on both younger and older age groups is warranted to examine the effect of ageing on the current findings.

The blast mTBI/PTSD group in the current study scored worse on the WAIS-IV vocabulary test than the PTSD-only group. However, surprisingly, our findings did not support any differences in neuromotor impairment between the groups. Researchers have previously reported cognitive deficits, including poor performance on visual memory and reaction time,¹² as well as increased vestibular impairment and related symptoms²⁴ following blast mTBI. However, some researchers have reported no differences on neuropsychological tests following blast-related mTBI between soldiers with and without PTSD,^25,26 suggesting the presence of PTSD may not impact neuropsychological performance. The current study expands this previous research by exploring the impact of PTSD on neuropsychological function in those with a history of blast-related mTBI. The findings supported differences in the WAIS-IV vocabulary test that suggest that military veterans with blast mTBI/PTSD performed significantly worse than those with PTSD only. However, no other cognitive or neuromotor differences were noted in the current study, supporting the previous research that did not find an exacerbating effect of PTSD on cognitive functioning following mTBI.^25,26 The current findings with regard to limited cognitive and no neuromotor impairment differences between the groups highlights the importance of combining neuroimaging such as MRSI with clinical assessments to help distinguish the effects of blast mTBI/PTSD from those of PTSD only.

In the current study, we also explored the relationship between cognitive and neuromotor impairment and MRSI findings. Our results supported correlations between decreased performance on the D-KEFS verbal fluency and category tests and decreased NAA/Cr ratios in the left anterior hippocampus. Our findings also supported strong correlations between decreased NAA/Ch ratios in the left anterior hippocampus and increased vestibular impairment and symptoms as measured by the VOMS. The magnitude of the correlations between the VOMS items and MRSI findings was moderate to high, suggesting a more robust relationship than for cognitive outcomes. These preliminary findings suggest that MRSI findings may be useful in corroborating vestibular impairment and may hold promise to image improvement over time. Future research should include repeated assessments of MRSI and vestibular impairment during the recovery process.

Strengths and Limitations

The current study examined two groups of patients from the VAPHS that represent a substantial challenge for VA and other clinicians to assess and subsequently treat. As a result, the current findings may help to inform clinical practice with these patients. The study also combined MRSI with cognitive and neuromotor data to allow for direct comparison of neuroimaging and clinical findings. The addition of assessments of vestibular/oculomotor impairment and symptoms, which had not been examined in this population before, represents another strength of the current study. We assumed that all participants completed the cognitive and neuromotor assessments in an honest and accurate manner and with full effort. To ensure genuine effort on the tests, we used the Test of Memory Malingering, the results of which indicated all participants gave adequate effort during the study. The sample was relatively small and included veterans from a wide age range. The sample may have also included veterans with undocumented co-morbid diseases and disorders such as substance abuse and personality disorders that might have influenced the results. In addition, very few females were included in the study, limiting the generalizability of the findings to male veterans. Another limitation of the study is that details regarding the nature of the blast mTBI, including severity of symptoms at the time of injury, were not available due to limitations in the information that was provided in veterans’ medical records. A final limitation of the study is that the MRSI data were limited to the hippocampal regions, and as such, our inferences about the effects of blast mTBI and PTSD are limited to this region of the brain.

CONCLUSION

A reduced ratio of NAA/Ch, particularly in left anterior regions of the hippocampus, combined with certain neuromotor impairments may be useful in differentiating military personnel with blast mTBI and PTSD from those with PTSD only. Neuromotor and, in particular, vestibular/oculomotor impairments and symptoms were correlated with decreased ratios of NAA/Ch, suggesting that functional impairment in these areas may correspond to localized injury in the hippocampus following blast injury.¹⁴ Additional research is needed to determine if the combined MRSI, neuromotor, and cognitive measurements in the current study can be used to monitor recovery and the effectiveness of interventions following exposure to blast mTBI as in epilepsy.²⁷ Regardless, MRSI, a novel neuroimaging technique, again supports a biochemical difference in hippocampal regions of the veterans with blast mTBI/PTSD. Specific clinical tests such as cognitive and neuromotor assessments may also serve as indirect assessments of brain dysfunction related to blast mTBI and PTSD. In the future, the use of MRSI in combination with clinical assessments of neuromotor and cognitive function may help inform more accurate diagnosis of mTBI in specific veteran patient populations especially those exposed to blast forces.