Posttraumatic stress disorder (PTSD) and traumatic brain injury (TBI) are heterogenous disorders with overlapping symptoms that are overrepresented in active-duty military and Veteran populations. According to the 2008 Rand Report, comorbid PTSD and TBI were seen at a rate of 7%−76% in military members returning from Iraq and Afghanistan (Tanielian et al., 2008), and those with a TBI were three times more likely to be diagnosed with PTSD (Carlson et al., 2011). These conditions are associated with similar symptoms such as memory impairment (Dolan et al., 2012), sleep disturbance/fatigue (Gilbert et al., 2015), emotional lability (O'Neil et al., 2017), depressed mood (Isaac et al., 2015), increased substance use (Miles et al., 2015), and suicidal ideation (Bahraini et al., 2013). Additionally, the triad of depression, history of military mild TBI (mTBI), and PTSD is associated with greater functional disability and unemployment (Lippa et al., 2015; Amick et al., 2018). There are pathophysiological similarities in mTBI and PTSD in that they are both linked to neuroinflammatory, excitotoxic, and oxidative processes that are associated with white matter changes in the brain (Kaplan et al., 2018). With such symptom convergence, similar pathophysiology, and high co-occurrence, it is challenging to disentangle effects of TBI and PTSD and further whether and how TBI contributes to the expression of PTSD. Mechanisms underlying interactions in these conditions remain unclear.

Factor analyses of self-reported symptoms in TBI and PTSD have underscored the difficulty in dissociating effects of co-morbid conditions. The Neurobehavioral Symptom Inventory (NSI; Cicerone and Kalmar, 1995), a self-report measure of post-concussive symptoms, has been found to assess four underlying symptom constructs (i.e., affective, cognitive, somatosensory, and vestibular), based off best fit models from a sample of National Guard members and Veterans with and without history of mTBI and TBI (all severities), respectively, even when controlling for PTSD (Benge et al., 2009; Vanderploeg et al., 2015). However, correlations between post-concussive symptoms and PTSD symptoms (i.e., re-experiencing, avoidance, hyperarousal, and negative alterations in cognition or mood) are high, possibly due to both overlapping pathophysiology and item characteristics as −15% of items have direct crosswalks to items on common PTSD self-report measures (O'Neil et al., 2021; Scimeca et al., 2021). Furthermore, military service members diagnosed with both PTSD and TBI tend to have greater PTSD symptom severity and higher rates of disability in comparison to those with PTSD only (Lippa et al., 2015). The overlap in symptom presentation and heightened severity of PTSD is possibly due, in part, to commonly affected brain networks in TBI that may contribute to similar symptoms to PTSD as well to increased expression of symptoms of PTSD.

In mild-to-moderate TBI, the observed variability in white matter is likely due to some combination of both premorbid vulnerability and secondary injuries, as direct, primary injuries (e.g., hemorrhage) are relatively less common (Blennow et al., 2012). However, diffuse axonal injury (DAI) via shearing injury is still possible in this population. Notably, frontolimbic pathways responsible for executive functioning and emotion regulation are some of the most vulnerable areas for damage (Wright et al., 2017; Badea et al., 2018; Kulkarni et al., 2019). Following injury, a multi-faceted neurometabolic cascade (i.e., secondary injuries) occurs in which there is a depolarization of the neuron followed by dysregulated glutamate release, thereby triggering a surge in energy demand and a temporary metabolic crisis, contributing to further axonal damage (Giza and Hovda, 2014); this is detectable in blood, even in mTBI (Wang et al., 2021). This is the proposed substrate for acute post-mTBI symptoms with other factors such as age, history of prior concussion, and comorbid neurological and psychiatric conditions moderating recovery (Bonfield et al., 2013; Karr et al., 2014).

While PTSD is thought to be a predominately psychiatric condition, there have been reports of white matter alteration in PTSD, although the mechanisms for these changes are not fully understood. In recently deployed military service members, reduced fractional anisotropy (FA; the most common diffusion-based indicator of underlying white matter structure) of frontolimbic white matter tracts (e.g., the uncinate fasciculus) was associated with startle and subthreshold PTSD symptomology (Costanzo et al., 2016). Further, in a study comparing civilians and Veterans with mTBI, Davenport et al. (2016) found that environment in which mTBI occurred was not determinant of weaker FA connections but that lifetime history of PTSD significantly interacted with deployment mTBI. Contrasting these findings, numerous diffusion-weighted imaging (DWI) studies of people with comorbid PTSD and TBI found no or little difference among common DWI metrics, Bazarian et al. (2013); Petrie et al. (2014); Lopez et al. (2017); Yeh et al. (2017); Bolzenius et al. (2018), and Santhanam et al. (2019), inconsistent significant regions of interest (ROI), and little evidence tying an association between DWI measures and PTSD with either small effect size or low predictive variance (Bazarian et al., 2013; Lepage et al., 2018). However, it is widely acknowledged in the literature that the high cooccurrence of PTSD and mTBI can exacerbate one another both neurophysiologically and symptomatically, but the mechanisms are not yet fully understood (Daniels et al., 2013; Davenport et al., 2016; O'Doherty et al., 2018; Santhanam et al., 2019).

Studies investigating the neurobehavioral outcomes following mTBI have consistently found that the presence of comorbid PTSD is linked to poorer results on a range of assessments, including self-reports of emotional regulation and global functioning (e.g., social engagement, disability) (Pietrzak et al., 2009; Macera et al., 2012; MacDonald et al., 2014; Combs et al., 2015; Haagsma et al., 2015; Jackson et al., 2016). Following TBI, dysfunction within fronto-limbic pathways may impact the ability to inhibit down-up neurophysiological reactivity (i.e., fight-flight, behavioral immobilization) expressed in PTSD (Williamson et al., 2013). This failure of inhibition may lead to disruptions in autonomic systems (Shah et al., 2013), increasing the effects of chronic stress on overall health and daily functioning. Autonomic dysregulation is a core feature of PTSD, criterion D. Hyperarousal symptoms are characterized by increased restlessness, startle response/stressor reactivity, and decreased sleep quality. Chronic hyperarousal is associated with accelerated physiological aging as evidenced by increased DNA methylation (Wolf et al., 2016). While criterion D is typically assessed with self-report, objective autonomic measures are different in patients with PTSD. High frequency heart rate variability, an index of vagal contributions to interbeat variability of the heart, is lower in patients with PTSD (Schneider and Schwerdtfeger, 2020), and, furthermore, responsive to successful treatment with psychotherapy (Shah et al., 2013). High frequency heart rate variability is associated with a variety of health outcomes; lower heart rate variability is associated with metabolic risk (Wulsin et al., 2015), pulmonary disease (Alqahtani et al., 2023), cardiovascular disease, and all-cause mortality (Jarczok et al., 2022). Parasympathetic control is also a critical system in inhibiting sympathetic nervous system response to stress, aka the vagal brake (Porges, 2001; Thayer and Lane, 2009). Frontally mediated networks of critical in autonomic control (Beissner et al., 2013) and disruption in key fronto-limbic white matter pathways may bias autonomic mobilization toward a maladaptive response to stressors; i.e., disrupting this system may be a critical factor in the amplification of PTSD symptoms associated with TBI. Understanding the relationship between symptom severity, autonomic functioning, and symptoms domains in fronto-limbic pathways will help to identify potential treatment targets when dissociating the effects of potentially interacting conditions.

The primary aim of this study was to determine if there was regional variability in the relationship between white matter integrity in major tracts and symptom severity in a mixed sample of Veterans with PTSD and/or mTBI (i.e., are some white matter tracts more closely associated with symptom severity in PTSD and mTBI compared to other white matter tracts?). We assessed symptom severity as a global construct, as well as the potential moderating effect of symptom domain (e.g., trauma-related re-experiencing, post-concussive vestibular complaints) and initial injury severity [i.e., loss of consciousness (LoC) following TBI]. A secondary aim of this study was to contrast analyses of symptom severity with objective measures of autonomic functioning [i.e., respiratory sinus arrythmia (RSA) and baroreceptor sensitivity (BRS)], in order to explore potentially overlapping pathophysiological mechanisms in PTSD and TBI. Examining the discriminant associations between tracts, autonomic functioning, and symptom severity in PTSD and TBI will provide a better understanding of the relationship between these two, often overlapping, conditions.

This study demonstrated regional specificity in the relationship between white matter integrity and global symptom severity, as well as between white matter and autonomic functioning in combat exposed Veterans with and without history of mild TBI and PTSD. These relationships were further moderated by LoC in those with mTBI, suggesting that injury severity during TBI may interact with chronic stress via impact on critical structures in regulating stress (although causality cannot be determined by this study). Consistent with the growing body of literature surrounding symptom typology in PTSD and TBI, our findings did not suggest that symptom domain (i.e., NSI vs. PCL-M) explained additional variance in white matter. These findings may have important implications for the clinical management of PTSD and TBI in Veterans.

Across the full sample, global symptom severity was positively associated with FA in the CAB and negatively associated with FA in the UNC. It is possible that the positive association between symptom severity and FA in the CAB reflects the loss of a sub-section crossing fibers (Bazarian et al., 2007), resulting in a paradoxical increase in FA as white matter degrades. Within the ventro-limbic system of the brain, the CAB and UNC are two pathways implicated in emotional regulation, memory, and autonomic feedback, that may play a role in the maintenance of symptoms (Williamson et al., 2013; Averill et al., 2018; Bottari et al., 2021). However, it is important to note that there did not appear to be regional specificity for symptom domain. This is consistent with previous literature suggesting that “post-concussive” symptoms are not specific to mild TBI (Lagarde et al., 2014; Santhanam et al., 2019) and that PTSD and mild TBI may share common pathophysiological mechanisms (Kaplan et al., 2018), underscoring the difficulty in symptom attribution. Mild TBI, particularly with LoC, may interact with the expression of symptoms of PTSD. Thus, comprehensive assessment and treatment of TBI, particularly in Veteran populations, should include assessment of PTSD.

There was also a strong positive relationship between FA and RSA and BRS difference scores (supine 30° minus standing 90°), in all tracts but the CAB, across the full sample. It is typical for RSA and BRS to increase in the supine position compared with standing (90°), reflecting greater parasympathetic control (Lellamo et al., 1996; Laude et al., 2004). In our sample, lower values of FA in a key fronto-limbic pathway (the UNC), were associated with an abnormal autonomic response (lower RSA and BRS in the supine position). Increased BRS, as a measure of autonomic functioning, is associated with dampening of the locus-coeruleus-noradrenergic pathway, as would be expected in a relaxed state (e.g., when lying down). Dysfunction in this pathway is critical in the neurophysiology of anxiety related disorders. The locus-coeruleus projects to the basolateral amygdala, an important structure in the expression of PTSD and is within the same circuit regulated by prefrontal cortical-limbic projections (Daviu et al., 2019). Our findings suggest that this autonomic dysfunction is associated with reduced FA, possibly reflecting the effects of chronic stress on the brain due to poor management of allostatic load within the autonomic system. While the exact mechanisms underlying the relationship between white matter changes and chronic stress remain unclear, animal models suggest that chronic stress may alter the expression of genes responsible for pre-programmed cell death in oligodendrocytes due to changing cellular energy demands (Antontseva et al., 2020). Longitudinal investigations of the interaction between autonomic function and variability in white matter after TBI are necessary to draw conclusions about potential directionality and identify possible treatment targets. Increased stress after TBI may exacerbate physiological effects of injury with increased blood brain barrier permeability, amplified autonomic response to stressors, and poor sleep which may prolong recovery or perhaps increase the likelihood of chronic symptom presentation. It has been previously demonstrated that indicators of reduced autonomic function prior to exposure to a criterion A event (e.g., combat environments) are predictive of development of PTSD (Minassian et al., 2015).

It is notable that the strongest relationships between FA and stress measures (i.e., both global symptom severity and RSA/BRS difference scores) were found in the ATR and UNC across the full sample. Connecting orbitofrontal cortex to anterior temporal lobes, the UNC is particularly vulnerable to shearing injuries in TBI (Seo et al., 2012), and has been identified as a potential treatment target for deep brain stimulation treatment of PTSD (Hamani et al., 2022). The ATR, part of the anterior limb of the internal capsule, connects dorsolateral prefrontal cortex to the thalamus. Reduced white matter integrity within the fronto-subcortical pathways (ATR and UNC) may impair the ability to regulate emotional and autonomic functions (disinhibition of the amygdala), leading to increased overall stress burden (Williamson et al., 2013). Furthermore, the association between FA in the ATR/UNC and stress measures was strengthened in individuals with LoC in TBI, suggests that severity of TBI (LoC, compared with AoC) has an effect, as postulated by Williamson et al., 2013. Our findings underscore the importance of these limbic pathways in understanding the mechanisms underlying possible interactions between PTSD and TBI. Addressing this interaction with differential approaches to treatment is an important area of future research.

There also appeared to be a dissociation between stress measures and FA in the CAB. Global symptom severity was positively associated with FA in the CAB (possibly related to the loss of crossing fibers) in the LoC group and negatively associated with FA in the group without LoC. In contrast, there was no relationship between the CAB and autonomic functioning across both groups. This may reflect the relatively smaller role that parahippocampal cingulate projections play in autonomic regulation, despite their importance in maintenance of anxiety-avoidance learning, although the differing contributions of subdivisions within the cingulum remain unclear (Jones et al., 2013; Bubb et al., 2018).

The findings in the FMinor, the major interhemispheric connection between the frontal lobes, are more difficult to interpret. While the relationship between FA in the FMinor was significant in the group with LoC across both stress measures, the direction of effects is in contrast to the pattern observed in the UNC and CAB; higher values of FA were associated with both higher global symptom severity and more positive RSA/BRS difference scores (i.e., higher RSA/BRS during the 30° position). This is counter intuitive. On the one hand, it makes sense that higher FA is associated with normalized autonomic responses between standing and supine positions. On the other hand, greater severity of global symptoms is not consistent with this interpretation. It is possible that these findings represent different portions of variance across tracts and stress measures, reflecting heterogeneity in the manifestation of anxiety symptoms and non-emotional elements of autonomic control. However, this may reflect reduced interhemispheric coordination of frontal-limbic and autonomic regions (Xavier et al., 2018), and further research is needed to fully explore this possibility.

There are several important limitations to consider. Our sample is predominantly male, reflecting the demographics of the Veteran population in combat roles during the conflicts experienced by our cohort. Therefore, our findings may have limited generalizability for other populations, particularly for women with co-occurring TBI and PTSD, as is commonly seen in the context of inter-personal violence. This population is understudied and represents an important target of future research. An additional limitation is in our method for categorizing participants in mTBI and PTSD diagnostic groups. While the PCL-M has demonstrated high levels of diagnostic sensitivity, we did not use the gold-standard CAPS interview to assess for PTSD symptomatology due to time constraints, instead using a custom interview to establish criterion A occurrence. Similarly, while the OSU-TBI Identification method has been shown to have strong validity and reliability, there are many challenges to accurately diagnosing TBI using retrospective self-report particularly in military samples (Davenport et al., 2016). This issue is also true in civilian contexts as many people who experience mild TBI do not seek treatment. Third, like most studies of TBI and PTSD in humans, we have limited ability to determine the relative contributions of pre-morbid vulnerability, injury-related, and trauma-related factors, due to the lack of pre-injury assessment. In particular, pre-morbid characteristics such as cardiovascular disease burden and lifetime substance use history have been shown to impact the characteristics of white matter and is an important source of inter-subject variability in this sample. Longitudinal investigations are necessary to disentangle multifactorial contributions, including possible interactions with cerebrovascular risk factors and substance use, as well as possible sources of resiliency (e.g., social support), which would be salient to the Veteran population. Fourth, because FA is observed on the tract level for specific tracts, there may be more granular variations in microstructure within tract that were not explored, including other metrics of variability such as mean diffusivity, that could explain some of the counter-intuitive relationships observed in this study. The use of a pre-existing dataset (and older DWI sequence) precludes our ability to examine this more closely. The possibility of reduced crossing fibers contributing to higher values of FA warrants additional investigation, and has been previously reported (Figley et al., 2022). Finally, sample size, though relatively large given the integration of neuroimaging, autonomic, and neurobehavioral symptoms, is limited and replication of these findings is necessary. Though, it should be noted that others have previously reported differences in ventral-limbic pathway integrity in mild TBI and PTSD (Santhanam et al., 2019). In particular, our sample size limits our ability to fully parse the differences between mTBI with LoC, ambiguous mTBI (mTBI with alterations of consciousness alone), and those without a history of head strikes. Future research further probing heterogeneity within mTBI is necessary.

In conclusion, although this study cannot make determinations regarding causality, our findings add to the growing body of literature suggesting that history of TBI may interact with the symptoms of PTSD, regardless of the type of symptom (e.g., PTSD-related or post-concussive). Furthermore, we found evidence to suggest autonomic dysfunction and variability in ventral-limbic white matter pathways play a role in this relationship. Future research should explore autonomic functioning within this population to identify possible targets for intervention, public health strategies, and increase our understanding of the underlying mechanisms.

The data analyzed in this study is subject to the following licenses/restrictions: Limited datasets are available upon request. Requests to access these datasets should be directed to john.williamson@ufl.edu.

The studies involving humans were approved by the University of Florida Institutional Review Board. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

AW: Conceptualization, Formal analysis, Methodology, Supervision, Visualization, Writing – original draft, Writing – review & editing. SB: Data curation, Formal analysis, Writing – original draft, Writing – review & editing. LJ: Data curation, Formal analysis, Writing – original draft, Writing – review & editing. DL: Methodology, Writing – review & editing. GL: Data curation, Formal analysis, Methodology, Writing – review & editing. JW: Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing – review & editing.