Post-Traumatic Stress Disorder symptom sub-cluster severity predicts gray matter volume changes better than overall symptom severity

Although Post-Traumatic Stress Disorder (PTSD) can be understood as dysregulation in fear extinction circuitry, it is not a unitary, homogeneous disorder; some symptoms of PTSD do not have direct fear-related correlates (anhedonia, emotional numbing, associated substance abuse). This study examines the relationship between the severity of PTSD symptom sub-clusters and changes in gray matter volume (GMV). 11 patients who had experienced interpersonal violence (females, mean age 38.4 yrs) participated in a MRI study and were assessed using the PTSD Checklist (PCL) (Weathers 1993). Individuals with current substance dependence, lifetime or current psychosis, bipolar disorder, current suicidality, and current psychoactive medications were excluded. A T1-weighted 3D MPRAGE volume was acquired (Siemens 3T Tim Trio, 176 sagittal slices, TR/TE=1900/2.52ms, effective resolution 1mm 3). Voxel-Based Morphometry (VBM) analysis was conducted in SPM5 (Ashburner 2000). Symptom subclusters of Avoidance, Dysphoria, Hyperarousal, and Intrusions were calculated according to Krause et al. (2008). Associations between GMV and subclusters were tested with linear regression, with age as covariate of no interest. Results were thresholded at p=0.0025, with cluster-level correction. Overall, symptom subcluster severity negatively correlated with GMV in regions previously implicated in PTSD. Severity of Avoidance symptoms (avoidant thinking, behavior) was negatively correlated with bilateral mid-cingulate (BA32) and anterior cingulate (BA24) cortices. Severity of Intrusions (flashbacks, nightmares, reliving) was negatively correlated with left-lateralized insula (BA13, BA47), inferior orbitofrontal cortex (BA47), postcentral gyrus (BA3), lingual gyrus, and calcarine sulcus (BA17). Hyperarousal (hypervigilant, “jumpy”) severity was not associated with GMV. Dysphoria symptom severity exhibited a trend of positive correlation with GMV in middle frontal gyrus (p=0.012). Total PCL score tended to correlate negatively (p=0.08) with GMV in right parahippocampal gyrus. These results support the hypothesis that PTSD is an imbalance of brain systems involved in approach and avoidance (Stein and Paulus 2009). Use of symptom sub-clusters revealed, to a greater degree than total PCL score, subtle changes in GMV. Specifically, severity of Intrusions was associated with decreased GMV in visual, motor, and pain processing areas while Dysphoria was positively correlated with GMV in middle frontal gyrus as observed in depressed patients. Greater knowledge of the structural correlates of symptom clusters may enhance understanding and treatment of PTSD.