Graph theory-based analysis reveals neural anatomical network alterations in chronic post-traumatic stress disorder

Abstract Multimodal imaging using network connectivity techniques shows promise for investigating neuropathology influencing Post-Traumatic Stress Disorder (PTSD) symptom maintenance and course. We recruited World Trade Center (WTC) responders who continued to suffer from chronic PTSD into a diffusion tensor neuroimaging protocol (n=100), along with nine unexposed controls without PTSD from other sources. Using a graph theory approach to probe network alterations in brain diffusion images, we calculated weighted characteristics path length as a surrogate marker for the effective neuroanatomical distance between anatomical nodes. The sample (N=109; 47 with chronic PTSD) was in their mid-fifties, and the majority were male. Responders were matched in terms of cognitive performance, occupation, and demographics. The anatomical connectivity graph was constructed for each participant using deterministic diffusion tractography. We identified a significant difference in weighted Characteristic Path Length (wCPL) between trauma-exposed WTC responders (Cohen’s d = 0.42, P<0.001) that was highest in people with PTSD, and not explained by WTC exposure severity or duration. We also found that wCPL was associated with PTSD symptom severity in responders with PTSD. In the largest study to date to examine the relationship between chronic PTSD and anatomy, we examined the anatomical topography of neural connections and found that wCPL differed between the PTSD+ and PTSD- diagnostic categories.