32.2.1. Assessment and Diagnosis

A variety of concussion assessment schemes has been developed in recent decades, such as those developed by the American Congress of Rehabilitation Medicine (ACRM; Kay et al., 1993), World Health Organization (Holm et al., 2005), and American Academy of Neurology (1997). Although subtle variations exist across these schemes, a fundamental tenet is that a diagnosis of concussion is made on the basis of acute stage injury parameters, not on the basis of retrospective self-report of symptoms alone. The ACRM criteria have proven to be especially influential and continue to be relied upon with regularity in both civilian and military/veteran samples. For example, the ACRM criteria were instrumental in the development of the later World Health Organization (2005) criteria. More recently, the VA and Department of Defense (DOD) relied upon the ACRM criteria to define combat-related concussion. According to the VA/DOD Clinical Practice Guideline for Management of Concussion/mTBI (U.S. Department of Veterans Affairs/Department of Defense), the ACRM criteria are “the most widely accepted criteria for mild TBI” in the United States.

Table 32.1 presents the criteria that are relied upon to define TBI severity within the VA and DOD (2009). The criteria included in the mild severity category are essentially the same as those included within the ACRM convention. According to the ACRM, an mTBI consists of “a traumatically induced physiological disruption of brain function” that is accompanied by (1) any period of loss of consciousness (LOC) persisting for approximately 30 minutes or less; (2) any loss of memory for events immediately before or after the accident (posttraumatic amnesia [PTA]), but not greater than 24 hours; (3) any alteration of mental state (e.g., dazed, disoriented, confused); and (4) focal neurological deficit(s) that may or may not be transient. To be regarded as mild traumatic brain injuries, they must result in Glasgow Coma Scale (GCS) status that is not less than 13 (i.e., between 13 and 15). Importantly, these injuries are the result of the head being struck or striking an object or acceleration/deceleration, which excludes other acquired conditions, such as stroke, anoxia, tumor, or encephalitis.

TABLE 32.1

Definition of Traumatic Brain Injury by Severity

Why is it essential to rely upon acute-stage injury parameters, rather than late-stage self-reported symptoms, to define TBI severity? First, an extensive neuropsychology literature has shown that LOC, PTA, and GCS (e.g., time to follow commands) are strongly predictive of long-term recovery outcomes; injuries of greater severity typically result in worse outcomes relative to injuries of lesser severity as defined by these parameters. Indeed, there is strong empirical support for a dose-response relationship between such variables as GCS, and LOC/PTA even more so (Dikmen, Machamer, and Temkin, 2009), and long-term cognitive and adaptive functioning.

Consider, for example, the findings of one of the highest quality prospective studies that has been carried out in civilian TBI samples to date (Dikmen et al., 1995). The authors conducted a prospective examination of neuropsychological outcomes in a group of 436 adults who underwent hospital admission related to varying severities of TBI and compared performances against a group of 121 general-trauma control participants hospitalized for injuries other than TBI. The latter group did not differ significantly to the TBI group with respect to age, educational attainment, or gender, which allowed the authors to control for the potential effects of socioeconomic status as well as the various effects of physical injury itself independent of TBI (e.g., pain, psychosocial stress) on neuropsychological performance. The TBI group was classified according to six different injury severities on the basis of time to follow commands after injury as follows: (1) <1 hour; (2) 1–24 hours; (3) 25 hours to 6 days; (4) 7–13 days; (5) 14–28 days; and (6) >29 days. Performances were then compared across groups at 1 month postinjury and 1 year later.

Results suggested a general dose-response relationship between injury severity and cognition over time across the six TBI groups on a comprehensive neuropsychological test battery, with progressively higher levels of persisting impairment noted by TBI injury severity at 1 year. By contrast, participants who sustained minor brain injury (i.e., <1 hour time to follow command) demonstrated far less evidence of cognitive impairment at 1 year postinjury. In fact, cognitive performances in the minor brain injury group were not significantly different from the non-TBI injury controls. Summarizing the Dikmen et al. results through development of composite cognitive effect sizes, Larrabee (2012) concluded that “the effect size for the Dikmen et al. MTBI group, tested 1 year post-trauma, is essentially zero, consistent with complete overlap of the MTBI and trauma control group score distributions” (p. 235). (See Figure 32.1 for a review of overall effect sizes in cognitive performance in the Dikmen et al., 1995, study between the TBI and injury control groups, as summarized by Larrabee, 2012.)

FIGURE 32.1

Overall impairment in neuropsychological performance at 1 year postinjury by TBI severity (the amount of time after injury until victim could follow verbal commands). Results from Dikmen et al. (1995) as summarized by Larrabee (2012). Effect sizes represented (more...)

A second reason why acute-stage injury parameters are relied upon to define TBI diagnosis, rather than late stage self-reported symptoms, relates to the highly nonspecific quality of the physical, cognitive, and emotional symptoms that may be reported in the postacute phase of recovery. Factors other than concussion itself often better account for the symptoms/impairments that individuals report/exhibit in the months and years after concussions are sustained. Indeed, this issue was emphasized by the developers of the ACRM criteria; although the authors suggested that various physical (e.g., nausea, vomiting, dizziness, headache), cognitive (e.g., inattention, memory difficulty), and behavioral changes (e.g., irritability, emotional lability) could potentially be used as “additional evidence” of mTBI, the developers emphasized that symptoms could be used to support a mTBI diagnosis only to the extent that these symptoms could not be accounted for by causes other than mTBI itself (e.g., peripheral injury, psychological reaction to physical or emotional stress).

The various symptoms included in the so-called “postconcussion syndrome” (PCS) (Diagnostic and Statistical Manual of Mental Disorders IV), for example, are not at all specific to remote history of concussion and cannot be relied upon, in isolation, to infer a history of traumatic brain injury of any severity. Meares et al. (2011) compared symptom reports of 62 mTBI participants and 58 non–brain-injured trauma controls (i.e., physical injuries other than mTBI). Whereas mTBI status was not predictive of PCS symptoms, the authors found that preinjury psychiatric symptoms (depression, anxiety), and acute posttraumatic stress were significantly predictive of PCS symptoms after physical trauma. Pain, regardless of mTBI status, was also predictive of PCS. These and similar findings reported by other independent researchers (Dikmen et al., 1995; Ponsford et al., 2012) highlight the importance of comparing outcomes of concussion samples with those of non-TBI trauma controls to account for various non-TBI confounds (e.g., pain, psychosocial stress) that may impact performances and/or symptom reports following concussion.

The nonspecific nature of PCS symptoms is also illustrated by the fact that a significant proportion of individuals with alternate conditions, but not TBI, also endorse these symptoms with great regularity. PCS symptoms are frequently endorsed in psychiatric groups (Iverson, 2006; Lange et al., 2011), chronic pain patients (Iverson and McCracken, 1997), and healthy community samples (Iverson and Lange, 2003). Moreover, researchers have demonstrated that non–TBI-related factors, such as cognitive bias (Ozen and Fernandes, 2011; Suhr and Gunstad, 2002, 2005) and false expectations regarding recovery after mTBI (Mittenberg et al., 1992), may of themselves reinforce persisting symptoms among those who continue to endorse symptoms in the late stage of recovery. The very belief that concussion results in lasting impairments, rather than concussion itself, may for some individuals maintain symptoms/impairments.

Secondary gain and response validity issues are also known to moderate self-reported symptoms as well as objective “impairments” on performance-based tasks in the months/years that follow concussive injury. In their survey of clinical neuropsychologists practicing within the United States, for example, Mittenberg et al. (2003) found that rates of probable malingering among those who underwent neuropsychological evaluation in the context of litigated mTBI were as high as 39%. In their widely cited meta-analytic review of neuropsychological outcomes in mTBI samples, Belanger et al. (2005b) found litigation status to be the most significant moderator of persisting cognitive limitations. The World Health Organization concluded that litigation and compensation issues were very influential factors to consider in the context of persisting symptoms/impairments following mTBI (Carroll et al., 2004). Recent studies of OEF/OIF concussion samples have also shown that rates of insufficient effort (Nelson et al., 2010) and symptom exaggeration (Nelson et al., 2011a,b) increase when veterans are assessed in a disability context or show indication of active disability claims. Thus, whether a civilian or veteran, response validity assessment is essential to the neuropsychological evaluation of any individual with a claim of disabling concussion, especially in the presence of known incentives for appearing more impaired than may in fact be the case.

32.2.2. “Natural History” of Recovery and Long-Term Neuropsychological Outcomes

If the primary task of concussion assessment is to establish that a given injury event likely resulted in concussion as defined by previously mentioned acute-stage injury parameters, or in other words surpassed a “minimum biomechanical threshold” (McCrea, 2008) of concussive injury, a secondary task is to establish the specific effects caused by the concussion, particularly with respect to onset and course of postinjury symptoms/impairments over time.

David Hume (1739–40, 1748; as cited by Field, 2013) suggested that to infer cause and effect (1) cause and effect must transpire contiguously (close together in time), (2) the cause must transpire before the effect, and (3) the effect should never occur in the absence of the cause. Considering concussion, researchers must demonstrate that (1) a temporal relationship exists between concussion and postinjury symptoms/impairments (contiguity), (2) the concussion transpires before the reported symptoms/impairments (i.e., were not present before the injury event), and (3) symptoms/impairments do not occur in the absence of concussion. Arguably, these conditions of causation have been most closely satisfied in studies conducted among athletes who undergo longitudinal assessments before (i.e., preseason baseline) and at multiple time points after (acute, subacute, and postacute) the times of their sports-related injuries.

As nicely described by McCrea (2008), there are various “unique advantages” inherent to conducting concussion research among athletes, and several researchers (Belanger and Vanderploeg, 2005a; McCrea et al., 2003, 2013) have capitalized on these advantages to elucidate the “natural history” of recovery that follows concussive injury. First, athletes represent a group that is at elevated risk of sustaining concussive injury within a well-defined time period (sports season). Close surveillance of potential concussive injuries can be conducted throughout the period of risk, and researchers are able to conduct their assessments very soon after the time of injury, thus allowing researchers to observe the necessary condition of contiguity. Related, sports-related concussions are most often witnessed (e.g., by teammates, spectators, game film), and this independent observation of concussion allows researchers to summarize acute injury characteristics (e.g., LOC, PTA) with far greater precision relative to retrospective self-report alone. Sports concussion researchers are also able to maintain a continuity of care that allows for direct comparison of acute-stage symptoms/impairments with baseline functioning, thus identifying the magnitude of effect that is attributable to the concussion itself. Systematic follow-up assessments conducted at various postinjury time points elucidate the usual patterns of recovery in subsequent hours, days, weeks, and months.

Furthermore, athletes represent a relatively “clean sample” that is typically healthy and free of premorbid physical (e.g., migraine headache), cognitive (e.g., attention-deficit), and psychiatric (e.g., depression, anxiety) symptoms that might confound an understanding of symptoms/impairments that may follow during the postinjury assessment phase. Sports concussion researchers also have ready access to a matched noninjury control group to further control for any symptoms/impairments that might naturally occur in the absence of concussion. Issues of secondary gain and response validity, which have been identified as important moderators of persisting symptoms/impairments in certain concussion samples (Belanger et al., 2005a; Carroll et al., 2004), are less likely relevant to athletes, a sample that tends to be very motivated to participate in sport (i.e., they typically have less incentive to exaggerate symptoms and/or subvert cognitive performances relative to other civilian concussion samples). By eliminating, or at least minimizing the degree to which these premorbid and postinjury variables have on recovery after concussion, sports concussion researchers are also able to conduct studies that largely uphold Hume’s second (i.e., ensure that the symptoms/impairments were present only after the causal concussion) and third (i.e., control for any symptoms/impairments that might occur in the absence of concussion) conditions of causation.

To illustrate, consider the approach of McCrea et al. (2003), who conducted a prospective cohort study of 1,631 college football players to understand recovery expectations following sports-related concussion. From this sample, the symptoms/performances of 94 athletes with verified concussion (as defined by American Academy of Neurology, 1997, criteria) were compared against those of 56 noninjured controls at eight independent postinjury assessment points: immediately, 3 hours, and 1, 2, 3, 5, 7, and 90 days after injury. The authors conducted various assessments across these time points, including measures of self-reported postconcussive symptoms, postural stability/balance, and cognition (i.e., Standardized Assessment of Concussion) (McCrea et al., 2000). Relative to controls, individuals who sustained sports concussion endorsed significantly more symptoms, imbalance, and cognitive impairment immediately after concussion, which implicates an abrupt onset of considerable brain disturbance. However, these significant symptoms/impairments were followed by a rapid and very favorable course of improvement in subsequent days. Overall, postconcussion symptoms resolved by 7 days, balance difficulties resolved between 3–5 days, and cognitive functioning improved to baseline within 5–7 days. No significant differences were observed across symptom- or performance-based outcomes at 90 days. This pattern of acute-stage disturbance and exponential rate of recovery is remarkably consistent with the pattern of physiological disturbance (e.g., ionic shifts, reduced glucose metabolism) and recovery that has been reported in neuroscience research (Giza and Hovda, 2001).

In a more recent study, which represents the largest prospective study conducted to date on the incidence and course of “prolonged” recovery in sports concussion samples, McCrea et al. (2013) examined outcomes of 570 concussed athletes and 166 controls obtained from an overall sample of 18,531 athletes who had been monitored over a 10-year time frame. Using the same measures of postconcussion symptoms, balance, and cognitive functioning as McCrea et al. (2003), athletes were assessed immediately, 3 hours, and 1, 2, 3, 5, 7, and 45 or 90 days after concussion, and results were compared against preseason baseline. The concussion group was further subdivided according to those who sustained “typical” (i.e., recovery within 7 days; n = 513) and “prolonged” (>7 days; n = 57) recovery (i.e., the authors found that approximately 10% of the sample showed prolonged recovery patterns beyond the usual expectation of recovery within 7 days). The typical recovery group endorsed significantly more symptoms than the control group at day 3, and symptoms were no longer significant at day 5 postinjury. The typical recovery group also showed significantly greater performance-based impairments than controls immediately after injury, but no differences were found at day 2. By contrast, the prolonged recovery group endorsed significantly more symptoms than the typical recovery and control groups at every postinjury recovery point, including the 45/90-day period. The prolonged recovery group also showed significantly greater cognitive impairment relative to controls through the first 7 days postinjury, but no significant differences were observed in cognitive performance (or balance) at the 45/90-day mark relative to controls. Predictors of prolonged recovery in this study included LOC, PTA, and more severe acute-stage symptoms. Only 2.3% of the full injury sample reported symptoms at the 6- to 12-week assessment point. The authors concluded that only a small percentage of concussed athletes experience symptoms and cognitive impairment beyond 1 week, and prolonged recovery was most associated with acute-stage indices that were suggestive of more severe injury.

Results of well-controlled sports concussion studies have detailed the “natural history” of concussive injury (McCrea, 2008), and recovery outcomes following sports concussion. In their meta-analytic review of neuropsychological functioning following sports-related concussion, Belanger and Vanderploeg (2005) found acute-stage (i.e., first 24 hours) cognitive impairments to be greatest, particularly in the areas of learning/memory and global cognition measures. However, the authors found that the acute effect of concussion was “essentially zero beyond 7 days post injury (10 days for delayed memory)” (p. 352). Other meta-analytic reviews (Belanger et al., 2005b; Binder et al., 1997; Frencham et al., 2005; Iverson et al., 2005; Rohling et al., 2011; Schretlen and Shapiro, 2003), including civilian samples that have not been exclusively devoted to sports concussion, have reported very similar findings with respect to long-term outcomes after concussion. Rather than describe these in detail, we refer the reader to Figure 32.2, which summarizes the results of five meta-analyses reporting neuropsychological outcomes at approximately 90 days after concussion. As can be seen, there is no evidence of meaningful cognitive impairment at this 3-month assessment point. This contrasts dramatically with cognitive effects in the case of moderate-severe traumatic brain injury, which may result in significant impairments at 2 years postinjury and beyond (Schretlen and Shapiro, 2003).

FIGURE 32.2

Effect of conventional mTBI, blast mTBI, and moderate-severe TBI on cognitive impairment (positive values reflect worse functioning). Effect sizes, represented as Cohen’s d, for conventional mTBI and moderate-severe TBI are derived from the following (more...)

Two primary lessons can be derived from these studies of neuropsychological outcomes following concussion in primarily civilian samples. First, impairments following a single, uncomplicated (i.e., not accompanied by any form of demonstrable intracranial disturbance), non–blast-related mTBI are most pronounced immediately after the time of injury and may include fairly significant impairments within the first week postinjury (e.g., d = −0.41) (Schretlen et al., 2004). Second, the vast majority of individuals who sustain concussion show a rapid trajectory of recovery; many if not most attain baseline function within 7–10 days, and there is no evidence of persisting impairment at approximately 90 days postinjury at most. In a review of findings like these, Iverson (2005, p. 311) properly concluded “under most circumstances, we should anticipate good recovery following mTBI. Patients and athletes should be reassured.”

32.3.1. Blast as a Unique Injury Mechanism

Blast represents an injury mechanism that is qualitatively distinct from conventional (i.e., non–blast-related) forms of injury. As summarized by Taber et al. (2006), explosive blast is accompanied by an extraordinary change in atmospheric pressure that involves an initial period of peak overpressure, and then a subsequent negative blast phase (underpressure) that may cause various forms of physical injury, including TBI. Theoretically, blast may result in four types of independent mechanisms of injury: primary, secondary, tertiary, and quaternary effects (DePalma et al., 2005; Mayorga et al., 1997; Taber et al., 2006).

Primary blast injury refers to the direct effect of the blast itself. Air-filled organs, such as the lungs, have been identified as especially vulnerable to primary blast effect, and it is not uncommon for primary blast exposure to cause pulmonary hemorrhages or other forms of internal bleeding. The tympanic membrane (TM) is also regarded as especially vulnerable to primary blast, and has in fact been identified as a potential proxy for concussive injury (Xydakis et al., 2007). There is debate as to whether concussion likely results from the primary blast effect itself (Saljo et al., 2011) in the absence of TM perforation or injury to other bodily regions known to be more vulnerable to pressure wave effects than the brain, which is relatively well-protected within the skull. There are several theorized means by which concussion might result from primary blast exposure, such as thoracic, translational/rotational, and direct cranial entry mechanisms (Courtney and Courtney, 2011). Secondary blast effects, which refer to injuries that are sustained as a result of debris being projected toward the body, include such injuries as penetrating injuries from projectiles or fragments (DePalma et al., 2005). Tertiary blast injuries occur as a result of the body being displaced as a result of blast, and in the case of TBI, may occur when an individual is thrown against a wall or other stationary object. Quaternary blast injury refers to burns, toxic exposure, and other potential effects of the explosion. In short, the nature of the blast mechanism itself is novel. As such, some researchers (Cernak et al., 2001) have suggested that blast-related concussion may be accompanied by a pattern of neuropsychological recovery that is distinct relative to more conventional (i.e., non–blast-related) concussion.

Blast-related concussion often occurs in life-threatening circumstances beyond the blast (e.g., military combat). Thus, psychological distress may interact with blast concussion to complicate long-term neuropsychological recovery. Uncomplicated, non–blast-related concussion typically results in rapid and favorable patterns of cognitive recovery (Belanger et al., 2005b; Binder et al., 1997; Frencham et al., 2005; Iverson et al., 2005; Rohling et al., 2011; Schretlen and Shapiro, 2003), but also psychological and emotional recovery. A meta-analytic review of psychological adjustment after conventional concussion (Panayiotou et al., 2010) revealed that mTBI was associated with small overall effect sizes across the domains of depression, anxiety, coping, and psychosocial disability; the authors concluded, “mTBI had a small to negligible effect on emotional symptom reporting” (p. 463).

An unpredicted blast event that results in the death or severe injury of a fellow soldier in a war zone, for example, in many instances represents a more traumatic event relative to a concussion sustained as a result of a sports injury. Indeed, exposure to explosive blast among OEF/OIF personnel has been identified as increasing risk of developing trauma symptoms relative to those who sustain non–blast-related injuries (Belanger et al., 2009). Blasts frequently result in burns and other physical injuries in addition to concussion (Edwards et al., 2012) and also contribute to trauma symptoms. Lopes Cardozo et al. (2012) conducted a mental health survey of Cambodian civilians who survived landmine explosions as part of war. The authors found very high rates of mental health difficulties, including depression, anxiety, and posttraumatic stress disorder (PTSD), despite the fact that only a minority of the sample sustained comorbid head injuries. There is the possibility that some of these psychological consequences were at least partially associated with the general situation of the Cambodian war (i.e., not necessarily a direct effect of blast exposure itself). Nevertheless, findings like these illustrate that further research is needed to identify not only the degree to which blast concussion itself may impact long-term recovery, but also potential independent and interaction effects due to the blast event being a significant psychosocial stressor.

32.3.2. Challenges and Complexities Associated with Assessment and Diagnosis

Although a great deal of research has been devoted to the study of long-term outcomes after blast-related concussion in recent years, it is far less research than conducted in civilian samples (Lamberty et al., 2013). Moreover, blast concussion researchers do not enjoy the previously described “unique advantages” (McCrea, 2008) that exist in conducting research in sports concussion samples (Nelson and Keenan, in press). Like sports concussion samples, OEF/OIF soldiers and other military personnel represent a group that is at increased risk of sustaining mTBI, and researchers can also follow them during a specified period of mTBI risk exposure (e.g., defined period of deployment). However, unlike sports concussion samples, blast concussions sustained by OEF/OIF personnel are only occasionally witnessed, assessments are not typically conducted in a standardized manner and a controlled environment immediately after the blast, and only rarely are personnel systematically followed during acute, subacute, and postacute phases as has been conducted in sports concussion samples (e.g., McCrea et al., 2003, 2013). Most notably, blast concussion researchers do not typically have access to acute-stage injury findings that corroborate the previously mentioned indices of LOC, PTA, GCS, and other parameters that are so essential to establishing injury severity and predicting recovery.

The Military Acute Concussion Evaluation (MACE) (DVBIC, 2006) is a useful tool that may improve assessment of blast concussion. The MACE, which was meant to be completed by medics or other emergency medical personnel as soon as feasible after injury, consists of a history (e.g., incident description, cause of injury), neurological screening (eye, verbal, motor responsivity), and cognitive performance sections. The latter section was modeled after the Standardized Assessment of Concussion (McCrea et al., 2000) and allows the examiner to obtain rapid evaluation of orientation, immediate memory, concentration, and delayed recall performances that result in a composite score ranging from 0 to 30. From this information examiners are to offer an impression of whether a concussion was sustained and whether it resulted in LOC. Integration of MACE results with the accounts provided by OEF/OIF veterans months or years later could be helpful in determining the reliability of postdeployment accounts years after blast events. For example, it would be difficult to support a diagnosis of blast concussion in a veteran if MACE results (i.e., information documenting acute-stage concussion indices) were inconsistent with concussive injury. Unfortunately, researchers rarely have access to MACE information or other records documenting acute-stage injury characteristics following blast exposure. More often than not, researchers who work with soldiers and veterans are forced to rely upon retrospective self-report or results of “TBI screening” instruments months or years postinjury to inform diagnosis of blast concussion.

Concerns have been noted about the reliability of retrospective self-report of blast events and TBI screening in OEF/OIF samples (Belanger et al., 2012; Donnelly et al., 2011; Nelson et al., in press; Van Dyke et al., 2010). Test-retest reliability of the VA TBI Screening Instrument (VATBISI or TBI Clinical Reminder; GAO, 2008), for example, was described as “sobering” in a study conducted by Van Dyke et al. The authors repeatedly administered the VATBISI to 44 OEF/OIF veterans at average 6-month assessment points, and found inconsistency on such important parameters as LOC (36% inconsistency) and amnesia (32% inconsistency). Other more recent studies have raised serious concerns related to false-positive identifications of TBI using screening instruments like the VATBISI (Belanger et al., 2012; Donnelly et al., 2011), and it appears that PTSD and physical symptoms reported during the postdeployment are predictive of inconsistent reports of combat-related TBI over time.

Amidst these challenges and complexities, blast concussion researchers and clinicians are much more restricted in their ability to uphold the conditions necessary to establish cause and effect (Hume, 1739–40, 1748; as cited by Field, 2013) relative to what has been reported in sports concussion samples. With few exceptions (Kennedy et al., 2013; Luethcke et al., 2011), evaluators of blast concussion are not able to corroborate acute-stage injury characteristics and cognitive performances immediately after blast concussions, which impede their ability to uphold the condition of contiguity. Moreover, because of the frequent lack of control for premorbid and postmorbid physical and psychiatric conditions in many blast concussion studies, evaluators are typically unable to uphold the second and third of Hume’s conditions of causality—absent before event and not present when the event does not occur.

Thus, evaluators of blast concussion confront a dilemma: blast concussion has been identified as a common hazard of the recent wars in Iraq and Afghanistan, but the elucidation of the severity, frequency, and anticipated outcomes associated with historical blast concussions is obscured by the unknown reliability and validity of self-report information obtained through contemporary TBI screening methodologies. With these limitations in mind, evaluators are forced to “work with what we have” when determining whether an historical blast exposure likely resulted in blast concussion.

32.3.3. Semistructured Interview Strategies

In light of the important limitations that have been documented of the VA TBI clinical reminder and other contemporary TBI screening instruments, several research groups have developed semistructured interviews for TBI to arrive at a more informed or “clinician-confirmed” assessment and diagnosis of historical blast-related concussion. Examples of such interview approaches include the Structured Interview for TBI Diagnosis (Donnelly et al., 2011), VA TBI Identification Clinical Interview (Vanderploeg et al., 2012), Boston Assessment of Traumatic Brain Injury – Lifetime (Brawn Fortier et al., 2012), and the Minnesota Blast Exposure Screening Tool (MN-BEST) (Nelson et al., 2011). To be clear, none of these methods has been developed with complementary acute-stage injury information for cross-validation, leaving significant questions regarding their own psychometric properties and utility in classifying concussion. But there is little question that these more extended interview-based approaches represent an improvement relative to TBI screening instruments alone, which as mentioned result in unacceptably high false-positive identification rates (Belanger et al., 2012). Here, we briefly describe the rationale and development of just one of these interview-based approaches, the MN-BEST, to illustrate how researchers of the Minneapolis Veterans Administration Health Care System “work with what they have” when confronting the challenge of assessing and classifying reports of blast concussion in the absence of corroborating information from the combat zone.

The MN-BEST, which is reproduced in Appendix 32.A, was developed in 2009 to be used in conjunction with results of the VATBISI (GAO, 2008) and TBI Secondary Evaluation that is the current procedure of screening and further evaluating OEF/OIF veterans who endorse a history of combat-related TBI. Early in our research, it was recognized that many of our OEF/OIF participants reported a history of many previous blast exposures that may or may not have resulted in blast-related concussion. Therefore, a primary rationale in developing the MN-BEST was to develop a composite index of blast-related concussion that could be used as a single indicator of previous blast concussion history. It was also our anecdotal observation that physicians and other clinical providers varied widely in their approach to the assessment and diagnosis of self-reported blast concussion. Review of clinical records revealed that providers did not necessarily assess such parameters as LOC and PTA to the same level of specificity. Variability was also noted in the number of blast-related concussions that were explored with veterans during the TBI interview process. Thus, another rationale for the development of the MN-BEST was to develop a method of maintaining a uniform, standardized approach to the assessment and diagnosis of every case of blast-related concussion that was reported among our participants. More specifically, the MN-BEST was developed to maintain a consistent strategy of assessing the frequency, severity, and plausibility of the three most significant concussions sustained as a result of blast.

Regular consensus meetings are held by a team of experienced clinical neuropsychologists who review self-reported information (and corroborating information if available) among study participants. To administer the MN-BEST, the evaluator first invites participants to estimate the sheer number of blast exposures that they confronted during combat, regardless of whether they perceived that these events resulted in blast concussion. This gives the assessor a general sense of relative risk that each individual participant may have had of sustaining blast concussion. An explosive ordnance disposal worker whose task it is to locate and intentionally detonate improvised explosive devices (IEDs), for example, would be regarded to be at greater risk of sustaining blast-related concussion than a veteran who never received combat training and was never located near the combat zone.

Next, the evaluator invites the veteran to describe the three most significant blast events that were confronted during combat. For each of the three events, it is the assessor’s task to essentially re-create the injury event with as much detail as possible to establish whether it is more likely than not that a given blast exposure resulted in a blast concussion according to a standardized set of diagnostic criteria. We were interested in identifying whether there might be meaningful differences within the mild spectrum of injury (e.g., whether an individual who did not sustain any LOC as a result of blast may show outcomes that are meaningfully different relative to an individual with 20 minutes of LOC). Specifically, the MN-BEST includes four concussion severity classifications, labeled as types 0, I, II, and III (refer to Appendix 32.A) that were adapted from both the ACRM (Kay et al., 1993) criteria and standards proposed by Ruff and Richardson (1999). As shown in Appendix 32.A, type 0 injuries involve no indication of LOC/PTA and are restricted to one or more neurologic symptoms or signs. Type I injuries involve altered mental state or unclear LOC, very brief PTA (1–60 seconds), and one or more neurologic symptoms/signs. Type II concussions are defined by definite LOC persisting less than 5 minutes, PTA of 60 seconds to 12 hours, and one or more neurologic symptoms/signs. Finally, type III injuries are accompanied by LOC from 5 to no more than 30 minutes in duration, 12–24 hours of PTA, and one or more neurologic symptoms or signs. Each concussive injury is then converted into a metric that, if determined to be plausible, may contribute to an overall composite of blast concussion. Type 0 injuries are assigned a severity score of 1, type I injuries a 2, type II injuries a 3, and type III injuries a 4, resulting in a range of 0–12 for the overall composite of blast concussion. The same process is also conducted for all non–blast-related events during the respondent’s lifetime (including those that transpired before, during, or after deployment), allowing for the control of non–blast-related concussions on outcomes of interest.

The classification of blast-related concussions during the frequency and severity rating stages is determined by taking the self-reported symptoms/signs at face value. After classifying the frequency and severity of blast- and non–blast-related concussions on the basis of face validity alone, the final step in the MN-BEST assessment process is to offer an opinion regarding the plausibility that the blast event as described resulted in concussive injury; is it “more likely than not” on the basis of available information that the blast event surpassed the minimal biomechanical threshold of concussive injury? Only those injuries that are determined to be plausible are included in the overall composites of concussion (blast or nonblast).

It is during this final plausibility rating process that the researcher must exercise clinical judgment and knowledge of blast effects to determine whether the blast accounts are coherent and plausible. For example, the level of specificity surrounding reported blast events may determine whether a given event was consistent with concussion. Thus, descriptions of blast events that cannot be located in place or time may be regarded as less plausible relative to events that are described in a specified city on a specified month, date, and year. Other detailed information, such as personnel who accompanied the respondent at the time of the blast, the precise activities being conducted at the time of the blast, type of weaponry (e.g., IED, rocket-propelled grenade), and conveyances (e.g., riding in a Humvee), may also assist the assessor to contextualize the event.

Another common factor that evaluators may rely upon to inform the likelihood that a reported blast exposure resulted in blast concussion relates to the level of anticipation that was associated with the blast event. For instance, a rather common scenario among veterans who report historical blast exposures in our work in Minneapolis is the report of blast exposure as a result of intentional detonation of an IED. More often than not, the consensus team concludes that these intentional detonations are less likely than not to have resulted in concussive injury. Unlike blasts encountered during combat, blast exposures that result from these intentional detonations are known to transpire at predetermined times, which allows soldiers to take steps to insure their safety (e.g., standing behind protective barrier an extended distance away, making use of protective gear).

A veteran’s reported proximity from blast represents another example of an indicator that might inform the plausibility of blast concussion. Howe (2009) indicates that the blast wave dissipates by a cubed root of one’s distance from the blast source; thus, an individual who is 10 feet from a given blast source is exposed to nine times more overpressure than an individual who is 20 feet from the blast source. Thus, a participant who reports having been 1,500 m from a controlled detonation might be regarded as less likely to have sustained blast concussion relative to an individual who reports having encountered a rocket-propelled grenade that was 5 m away. Conversely, the reliability of an event involving the detonation of an IED “1 foot” away from one’s person might be brought into question if he or she denies that this exposure resulted in any immediate concussive effects or physical injury.

The other physical injuries that a given blast event contributed to may further inform the likelihood of blast concussion. As noted previously, the tympanic membrane is especially vulnerable to blast overpressure and may even represent a reliable proxy for concussive injury (Xydakis et al., 2007). Thus, a blast event that contributes to TM perforation may be plausibly regarded as an event that also resulted in concussion. Related, consideration of the injuries of peers who were in the same vicinity of the respondent at the time of the blast may also assist the researcher to determine the credibility of a given blast event and inform the likelihood of concussive injury. Case studies have shown that individuals who are exposed to blast at comparable distances show “strikingly similar” physical injuries (Commandeur et al., 2012). Thus, accounts suggesting that the respondents’ peers sustained similar injuries as the respondent may be more plausible than accounts in which a blast resulted in injury to the respondent only (in spite of the report that his or her peers were located the same distance from the blast).

A benefit of the MN-BEST is that mTBI can be either dimensionally or categorically characterized, as well as collapsing across blast and conventional mTBIs to achieve a total cumulative mTBI for an individual’s lifetime. Given the sources of information and variability in assessor ratings MN-BEST scores are unlikely to be numerically precise but nonetheless provide quantitative estimate of the cumulative degree of brain injury from traumatic events. Therefore, the MN-BEST overall blast concussion composite can be used as a continuous measure to explore questions of interest that are better suited for dimensional data (e.g., dose-response relationship between blast concussion and emotional distress as defined by select Minnesota Multiphasic Personality Inventory-2 clinical scales). Although preliminary work suggests that the inter-rater reliability of the MN-BEST is strong (>0.90 per Nelson et al., 2011), to date there is no validation of the instrument with known injury characteristics obtained during the acute phase of injury. Ideally, blast-related concussion would be assessed on the basis of acute-stage injury characteristics, but in the vast majority of cases this is not possible with military samples. We recognize the MN-BEST as only one potential strategy of enhancing the reliability of self-report information in OEF/OIF veterans with self-reported histories of concussion, but analysis of white matter integrity (number of voxels with low FA) in 116 OEF/OIF military personnel revealed that individuals classified on the MN-BEST as having blast-related mTBI tended to fall in the upper two-fifths of subjects on the low FA voxel count index (Figure 32.3). Thus, there is some preliminary neuroimaging evidence to support use of the MN-BEST.

FIGURE 32.3

Number of individuals with and without blast-related mTBI (i.e., bTBI) as determined by the MN-BEST within each 20th percentile bin for a measure of white matter integrity (count of number of white matter voxels with low FA) in 116 individuals who had (more...)

32.3.4. Neuropsychological Outcomes

As noted, no research group to date has effectively implemented the same longitudinal approach to the assessment of blast concussion as has been conducted in sports concussion samples (e.g., McCrea et al., 2003, 2013). This dramatically limits the ability to reliably determine the “natural history” of recovery after blast-related concussion. Some concussion researchers have examined cognitive performance within the acute stage of combat-related concussion (Coldren et al., 2010, 2012; Kennedy et al., 2012), but investigators have not systematically followed performance trajectories beyond the acute stage to understand long-term recovery patterns. Further, some of these acute stages of injury blast concussion studies (e.g., Kennedy et al., 2012) have mixed blast concussion samples with other forms of combat-related concussion, which eliminates the possibility of understanding differences between blast and nonblast mechanisms of injury.

Researchers have recently studied predeployment cognitive performance as a control for postdeployment cognitive function (Roebeck-Spencer et al., 2012; Vasterling et al., 2012), yet these have not included assessments during the acute stage of concussive injury. Some of these pre-post studies have also mingled blast with other mechanisms of injury (e.g., only 63% of the concussion sample in Roebeck-Spencer et al. were blast-related), which again limits the ability to identify whether blast or other combat-related concussions account for reported findings. Interestingly, in both studies, the majority of individuals who reported having sustained mTBI during deployment failed to show lasting cognitive impairments. However, psychological and emotional difficulties were found to contribute to longer term neuropsychological impairments. Vasterling et al. concluded that “milder TBI reported by deployed service members typically has limited lasting neuropsychological consequences; PTSD and depression are associated with more enduring cognitive compromise” (p. 186). These findings are very much consistent with previous studies showing favorable course of recovery in the case of non–blast-related concussion (Belanger et al., 2005) as well as the impact that psychological conditions, such as PTSD (Oien et al., 2011) and depression (Zakzansis et al., 1999), have on cognitive performance.

Of particular interest to the current review are two threads of neuropsychology literature: (1) studies that have compared blast and nonblast concussion samples on cognitive performances and self-reported psychological symptoms (Belanger et al., 2009, 2011; Cooper et al., 2012; Kontos et al., 2013; Lange et al., 2012; Lippa et al., 2010; Luethcke et al., 2011) and (2) studies that have compared blast concussion samples with deployment controls and psychiatric comparison groups (Nelson et al., 2012; Shandera-Ochsner et al., 2013). With respect to contrasts of blast and nonblast samples, we generated neuropsychological effect size differences across five studies that examined performances in blast and nonblast concussion groups at various postinjury recovery stages (Belanger et al., 2009; Cooper et al., 2012; Kontos et al., 2013; Lange et al., 2012; Luethcke et al., 2011). As shown in Figure 32.4, the composite effect size for neuropsychological performance between blast (n = 1,006) and nonblast concussion (n = 1,806) samples in these studies was minimal (d = −0.06). The similarity of blast and nonblast concussion samples was demonstrated at 72 hours (Luethcke et al., 2011; d = 0.19), 4 months (Lange et al., 2012; d = 0.14), 27 weeks (Cooper et al., 2012; d = −0.22), and more than a year (Belanger et al., 2009; d = 0.13) after blast concussions were sustained. Taking this literature together, there is minimal evidence that blast results in cognitive changes that are distinct from nonblast mechanisms of concussive injury.

FIGURE 32.4

Composite effect sizes of blast versus non–blast-related concussion. Effect sizes, represented as Cohen’s d, with positive values reflecting greater prevalence in blast samples. Effect size (1) derived from cognitive performances obtained (more...)

To contrast psychological consequences of blast and nonblast concussive injury we also generated effect size differences across six studies that included self-report measures of psychological and emotional functioning (Belanger et al., 2011; Cooper et al., 2012; Kontos et al., 2013; Lippa et al., 2010; Luethcke et al., 2011, Lange et al., 2012). As shown in Figure 32.4, the composite effect size of self-reported psychological symptoms between blast (n = 1,404) and non-blast (n = 1,939) samples was small in magnitude (d = 0.28) by conventional standards (Cohen, 1988), though certainly more sizeable than what was observed of neurocognitive outcomes in blast versus nonblast samples. Effect sizes for psychological symptoms were generally small across assessments conducted 72 hours (Luethcke et al., 2011; d = 0.10), and approximately 1 month (Lippa et al., 2010; d = 0.31), 4 months (Lange et al., 2012; d = 0.15), 6 months (Cooper et al., 2012; d = 0.10), and 12 months (Belanger et al., 2011; d = 0.24) after blast concussions were sustained. Thus, individuals who sustain concussion as a result of blast might be at somewhat greater risk of developing trauma symptoms and other forms of emotional difficulty than those who sustain concussion by other means (Belanger et al., 2009).

Studies have also compared OEF/OIF veterans with reported histories of blast concussion to deployment controls on neuropsychological performance (Nelson et al., 2012; Shandera-Ochsner et al., 2013). Nelson et al. (2012) examined neuropsychological outcomes in a sample of 118 OEF/OIF veterans with varied combat histories. Using the Clinician-Administered PTSD Scale (CAPS) and Structured Clinical Interview for DSM Disorders (SCID) to ascertain axis I diagnosis, and the MN-BEST to define self-reported history of blast concussion, the authors compared neuropsychological performances across four groups: (1) mTBI only (n = 18), (2) axis I diagnosis only (n = 24), (3) comorbid mTBI/axis I (n = 34), and (4) deployment control (n = 28). Although a main effect was found for axis I diagnosis on neuropsychological performance, no main effect was found for mTBI status.

Shandera-Ochsner et al. (2013) aimed to extend and replicate the Nelson et al. (2012) study by examining neuropsychological performances in a sample of 81 OEF/OIF veterans in four groups: (1) mTBI only (n = 20), (2) PTSD only (n = 19), (3) comorbid PTSD/ mTBI (n = 21), and (4) deployment control (n = 21). Similar to Nelson et al., the mTBI-only group demonstrated neuropsychological performances that were comparable to the deployment control group, whereas both PTSD groups demonstrated significantly worse cognitive performances relative to deployment controls. The authors concluded that “PTSD seems to be an important variable affecting neuropsychological profiles in the post-deployment time period” (p. 881). They also concluded that the lack of effect between the mTBI and control group was consistent with findings from the civilian mTBI literature, which suggests that mTBI does not of itself contribute to enduring cognitive impairments.

Taken together, the results of Nelson et al. (2012) and Shandera-Ochsner et al. (2013) support two primary conclusions. First, when comparing blast concussion samples with no ongoing indication of psychological and emotional symptoms to deployment controls, there is no evidence that a remote history of blast concussion alone contributes to lasting cognitive impairments months or years later. In spite of the noteworthy limitations of these two studies (e.g., small sample sizes; some proportion of the Nelson et al. mTBI group showed ongoing signs of alcohol dependency), the overall effect size between the 38 OEF/OIF veterans with mTBI (and without axis I conditions) and OEF/OIF controls in these two studies was small (d = 0.19), and comparable to effect sizes obtained in meta-analytic reviews of concussion (Figure 32.2). Second, the results of Nelson et al. (2012) and Shandera-Ochsner et al. (2013) clearly illustrate that PTSD and other forms of axis I psychopathology clearly impact cognitive performances during the postdeployment phase, and their findings comport with those of other researchers who have found psychological distress as having a significant effect on cognitive function in OEF/OIF personnel (e.g., Vasterling et al., 2012). As shown in Figure 32.5, regardless of blast concussion history, the effect sizes of the two samples with ongoing psychopathology from the Nelson et al. and Shandera-Ochsner et al. studies are quite consistent with the moderate effect sizes reported in two meta-analyses that examined neuropsychological functioning among individuals with major depression (Zakzanis et al., 1999) and PTSD (Oien et al., 2011).

FIGURE 32.5

Effect of major depression, PTSD, and other axis I conditions on cognitive impairment. Effect sizes for major depression and PTSD are derived from the following meta-analyses that compared cognitive performances of psychiatric and comparison samples: (more...)

In sum, despite various weaknesses in neuropsychological studies of blast-related mTBI, several preliminary conclusions can be made. First, although blast has been shown to represent a unique injury mechanism (DePalma et al., 2005; Taber et al., 2006), there is essentially no evidence that neuropsychological outcomes vary by injury mechanism. Studies have generally failed to show meaningful differences between blast and non−blast-related mTBI at various points after injury. Second, there is very little evidence that blast concussion per se results in enduring neuropsychological impairment. Available research appears to suggest that long-term recovery after a single blast concussion is quite favorable, consistent with what has been reported in the civilian literature among those who sustain non−blast-related concussion. Third, blast concussion does appear to increase risk of developing PTSD and other forms of psychological distress compared with other causes of concussion, though effect sizes for differences in psychological symptoms are fairly modest based upon studies reviewed here. Finally, PTSD, depression, and other forms of psychological distress, independent of blast concussion history, contribute to cognitive impairments of at least moderate magnitude in OEF/OIF samples, consistent with what has been reported in previous meta-analytic reviews of neuropsychological functioning in PTSD and major depression.

32.4.1. Clinical Neuroimaging Strategies

In civilian settings, neuroimaging is frequently used to determine the presence and extent of any cerebrovascular ruptures (e.g., hematoma) that may require immediate attention and/or acute signs of brain damage. Both X-ray computed tomography (CT) and magnetic resonance imaging (MRI) are able to provide this information, though CT is somewhat preferred in emergency room settings because of its lower cost, shorter duration, and overall higher safety for persons with altered or reduced consciousness who cannot reliably report contraindications to MRI (e.g., implanted medical devices). Although CT can reliably be used to detect gross structural damage, mild TBI is not typically associated with CT abnormalities (Borg et al., 2004). In contrast, MRI is much more sensitive to traumatic lesions such as diffuse axonal injury and cerebral contusions, with up to 77% of mTBI individuals demonstrating abnormalities on MRI; however, the presence of abnormal findings on MRI relates poorly to cognitive impairments or long-term outcomes (Hofman et al., 2001; Hughes et al., 2004; Lee et al., 2008). As David Hughes and colleagues (2004) stated, “Although non-specific abnormalities are frequently seen, standard MRI techniques are not helpful in identifying patients with MTBI who are likely to have delayed recovery” (p. 550). Therefore, the search for neuroimaging measures that can provide a clinically useful, objective marker of mTBI has focused on more specific measures than overall structure.

32.4.2. DTI

Although conventional MRI scans merely reflect the type of tissue (e.g., gray matter, white matter, fat) present at various locations in the brain, DTI has received attention over the past decade for being able to provide information about the underlying microstructural properties of that tissue based on patterns of water diffusion (Basser, 1995; Beaulieu, 2002). By measuring the magnitude of Brownian motion of water (i.e., diffusion) in each of several directions, DTI summarizes the overall diffusion pattern within each volume element (i.e., “voxel”) as a set of three perpendicular vectors of varying magnitudes (Basser et al., 1994). In white matter, diffusion along the orientation of the axon (axial diffusivity) is much more substantial than diffusion perpendicular to the axon (radial diffusivity) because the latter is constrained by the axon membrane, myelin, and cytoskeletal elements (Beaulieu, 2002). This anisotropic (i.e., nonspherical) diffusion pattern is dependent on multiple characteristics of healthy white matter (e.g., high axon density, myelination), and the degree of anisotropy is generally considered a valid measure of overall white matter integrity (Beaulieu, 2002). Fractional anisotropy (FA), the most commonly used DTI measure, varies between 0 (perfect sphere) and 1 (infinite cylinder) to represent the shape of the tensor independent of its size, whereas the complementary measure of mean diffusivity (MD) represents the overall size of the tensor independent of its shape (Basser and Pierpaoli, 1996). Reduced FA and increased MD (i.e., reduced constraints to diffusion) are generally considered to be associated with white matter damage or poor integrity.

In civilian populations, some researchers have reported mTBI to be associated with lower FA and higher MD, especially in frontal association pathways and posterior corpus callosum (Aoki et al., 2012; Niogi and Mukherjee, 2010). However, the nature of military mTBI in general, and those involving exposure to blast in particular, limits the generalizability of the civilian literature. Indeed, among the seven studies that directly compared DTI measures of soldiers who experienced mTBI during their most recent deployment to those of deployed veterans without mTBI (Bazarian et al., 2012; Davenport et al., 2012; Jorge et al., 2012; Levin et al., 2010; MacDonald et al., 2011; Morey et al., 2013; Sorg et al., 2013), only two reported group differences in individual regions (MacDonald et al., 2011) or voxels (Morey et al., 2013), and the affected regions in these studies did not overlap. However, even the studies with negative results have provided valuable insight into the complexity of studying this population. The first DTI study of blast-related mTBI, conducted by Levin and colleagues (2010), compared a group of veterans previously diagnosed with mild or moderate TBI to a group of veterans with neither exposure to blast nor mTBI during deployment, thus maximizing group differences in both mTBI and exposure to blast. As expected, individuals with mTBI had higher rates of PTSD, depression, and postconcussive symptoms as well as higher levels of global distress, but there were no group differences in quality of life or overall cognitive function. No group differences in FA or MD were found for any region of interest (ROI) or in a voxel-wise analysis, indicating that even moderate TBI may not be associated with systematic reductions of white matter integrity of particular regions. This is consistent with a lack of regional FA differences between veterans with mTBI and healthy civilians despite differences in coherence of electrophysiological activity among brain regions (Sponheim et al., 2011). One potential interpretation for the failure to detect group differences in FA using traditional ROI and voxel-wise analyses is that the structural effects of mTBI are spatially heterogeneous across individuals (Ptak et al., 2003). Consistent with this hypothesis, veterans with blast mTBI have been reported to have more ROIs (MacDonald et al., 2011) or voxels (Davenport et al., 2012; Jorge et al., 2012) with abnormally low anisotropy (i.e., several standard deviations below the mean), raising the suggestion of white matter findings associated with mTBI even if no single region is consistently affected.

Given the general lack of differences in DTI measures between mTBI and veteran control groups, several studies have instead focused on subsets of veterans with mTBI. Interestingly, although Jorge and colleagues (2012) reported that the number of voxels with abnormally low FA was similarly elevated in veterans with evidence of LOC or PTA (i.e., probable mTBI) as in veterans with “vague symptoms of confusion and dazedness” (i.e., possible mTBI), Sorg and colleagues (2013) reported that only mTBI involving LOC was associated with lower FA in ventral prefrontal white matter. Moreover, Matthews and colleagues (2012) reported lower FA associated with LOC, compared to alteration of consciousness, in 14 widespread regions of white matter. These results underscore the potential importance of acute symptoms (e.g., LOC, PTA) on long-term neurological changes, making the collection of this information even more critical to the continued study of blast mTBI.

Overall, select experimental investigations that have included DTI have raised the possibility that white matter disruptions associated with blast mTBI are present in the chronic stage after injury and most likely vary in location across individuals. There is also preliminary evidence that certain features of the mTBI event (e.g., loss vs. alteration of consciousness) affect the neurological outcomes, though this does not necessarily comport with clinical observations (e.g., Dikman et al., 1995). Additional studies that can better characterize these acute injury symptoms may lead to the ability for DTI to inform assessments of injury severity and, eventually, reduce the reliance on retrospective self-report.