Individual Patterns of Abnormality in Resting-State Functional Connectivity Reveal Two Data-Driven PTSD Subgroups

Abstract

Objective: A major challenge in understanding and treating posttraumatic stress disorder (PTSD) is its clinical heterogeneity, which is likely determined by various neurobiological perturbations. This heterogeneity likely also reduces the effectiveness of standard group comparison approaches. The authors tested whether a statistical approach aimed at identifying individual-level neuroimaging abnormalities that are more prevalent in case subjects than in control subjects could reveal new clinically meaningful insights into the heterogeneity of PTSD.

Resting-state functional network models for posttraumatic stress disorder

“…It is thus important to consider contributions from both approaches and to replicate divergent findings to test the boundaries of existing theories. The benefits of data-driven methods are especially clear in studies like that of Maron-Katz et al (5), whereby subtypes of brain network activations corresponded with different PTSD presentations. Due to the diagnostic heterogeneity inherent to PTSD, it is fathomable that an evaluation of networks in a pool of individuals with PTSD may wash out differential brain activity that may have corresponded with different PTSD presentations.…”

“…Despite its debated interpretation, the finding of lower global centrality in treatment responders may be a possible biomarker indicative of treatment success (6). Also, the finding of heterogeneous clusters of networks affirmed the need for continued investigation into different PTSD presentations and for individualized medicine (5). Future research could benefit from larger sample sizes to detect additional possible connectivity.…”

“…This may be expected given that Maron-Katz et al ( 5) used a purely data-driven approach, whereas the other three studies were more hypothesis-driven in selection of regions. Unfortunately, Maron-Katz et al (5) did not provide a mechanistic interpretation of these findings, opening the door to further investigation of these novel regions and their role in PTSD symptomatology.…”

“…Another way to explore networks for PTSD is to consider how heterogeneous presentations might differ. In a study by Maron-Katz et al (5), two clusters of PTSD symptom presentations were detected compared with TECs; these clusters were also validated by electroencephalogram. Cluster 1 was characterized by reduced connectivity between the sensorimotor (i.e., somatomotor) network and the visual network, which notably shares regions with the SN.…”

“…Possible explanations might include limited sample size, heterogeneity of symptom presentations, or the possibility that these symptoms are a product of widely distributed changes in brain function rather than changes within discrete networks (8). As the finding of PTSD clusters also demonstrates (5), it is possible that there is no homogeneous presentation of PTSD, which would limit the ability of between-group studies to find average differences in symptom severity by network.…”