PTSD Psychotherapy Outcome Predicted by Brain Activation During Emotional Reactivity and Regulation

Abstract

Objective: Exposure therapy is an effective treatment for posttraumatic stress disorder (PTSD), but many patients do not respond. Brain functions governing treatment outcome are not well characterized. The authors examined brain systems relevant to emotional reactivity and regulation, constructs that are thought to be central to PTSD and exposure therapy effects, to identify the functional traits of individuals most likely to benefit from treatment.

Method: Individuals with PTSD underwent functional MRI (fMRI) while completing three tasks assessing emotional reactivity and regulation. Participants were then randomly assigned to immediate prolonged exposure treatment (N=36) or a waiting list condition (N=30). A random subset of the prolonged exposure group (N=17) underwent single-pulse transcranial magnetic stimulation (TMS) concurrent with fMRI to examine whether predictive activation patterns reflect causal influence within circuits. Linear mixed-effects modeling in line with the intent-to-treat principle was used to examine how baseline brain function moderated the effect of treatment on PTSD symptoms.

Results: At baseline, individuals with larger treatment-related symptom reductions (compared with the waiting list condition) demonstrated 1) greater dorsal prefrontal activation and 2) less left amygdala activation, both during emotion reactivity; 3) better inhibition of the left amygdala induced by single TMS pulses to the right dorsolateral prefrontal cortex; and 4) greater ventromedial prefrontal/ventral striatal activation during emotional conflict regulation. Reappraisal-related activation was not a significant moderator of the treatment effect.

Conclusions: Capacity to benefit from prolonged exposure in PTSD is gated by the degree to which prefrontal resources are spontaneously engaged when superficially processing threat and adaptively mitigating emotional interference, but not when deliberately reducing negative emotionality.

Trial registration: ClinicalTrials.gov NCT01507948.

Keywords: Brain Imaging Techniques; Emotion; Exposure Therapy; Posttraumatic Stress Disorder; Psychotherapy; Transcranial Magnetic Stimulation.

Figure 1. Prefrontal Emotional Reactivity Baseline Activation Moderators Treatment-Related Symptom Change

During emotional reactivity, i.e. conscious processing of fearful vs. neutral facial stimuli (A), greater degree of activation in the dorsal anterior cingulate (B), left dorsolateral prefrontal/frontopolar cortex (C), right dorsolateral prefrontal cortex (D), left anterior insula (E), and right frontopolar cortex (F) predicted better treatment outcomes for individuals randomized to immediate treatment vs. those randomized to waitlist. Separate lines within each group represent individuals above (solid line) or below (dotted line) the median level of activation across the entire sample for the purposes of visualizing disparate symptom change trajectories within each group. ACC = anterior cingulate cortex; BOLD = blood oxygenation level-dependent response; CAPS = Clinician-Administered PTSD Scale for DSM-IV; DLPFC = dorsolateral prefrontal cortex; F = fear; FDR = false discovery rate; Fp = frontopolar cortex; N = neutral; PE = prolonged exposure group; WL = waitlist group; *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 2. Left Amygdala Activation During Emotional Reactivity Moderates Symptom Change Following Prolonged Exposure

In individuals randomized to immediate prolonged exposure vs. those randomized to waitlist, less left amygdala activation (A) to consciously-processed fearful vs. neutral faces (B) predicted a better treatment response. Separate lines within each group represent individuals above (solid line) or below (dotted line) the median level of activation across the entire sample for the purposes of visualizing disparate symptom change trajectories within each group. BOLD = blood oxygenation level-dependent response; CAPS = Clinician-Administered PTSD Scale for DSM-IV; F = fear; N = neutral; PE = prolonged exposure group; WL = waitlist group; *p < 0.05; **p < 0.01.

Figure 3. Emotional Conflict Task Activation Moderators of Treatment Response

During conflict-free trials of congruent fear vs. congruent happy, individuals randomized to immediate treatment (but not to waitlist) displaying greater activation in the dorsal anterior cingulate and the bilateral dorsolateral prefrontal cortices demonstrated greater reductions in symptoms (A). When examining emotional conflict regulation, greater activation in the ventromedial prefrontal cortex in the immediate treatment group predicted greater symptom reduction, but not in those randomized to waitlist (B). This effect continued to hold when contrasting emotional conflict regulation with gender conflict regulation, indicating emotional specificity of the conflict regulation effect (C). Separate lines within each group represent individuals above (solid line) or below (dotted line) the median level of activation across the entire sample for the purposes of visualizing disparate symptom change trajectories within each group. ACC = anterior cingulate cortex; CAPS = Clinician-Administered PTSD Scale for DSM-IV; Con = congruent; DLPFC = dorsolateral prefrontal cortex; FDR = false discovery rate; Inc = incongruent; PE = prolonged exposure group; PFC = prefrontal cortex; WL = waitlist group; *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 4. Single-Pulse TMS and Concurrent fMRI Reveals Right Dorsolateral Prefrontal Causally-Induced Left Amygdala Inhibition at Baseline Predicts Treatment Response to Prolonged Exposure

The right dorsolateral prefrontal region observed to moderate treatment response during emotional reactivity largely overlapped with the right prefrontal node (posterior middle frontal gyrus) of the canonical resting state executive control network (A). A random subset of individuals (N = 17) randomized to immediate treatment underwent concurrent single pulse TMS to this executive control network node as well as another right prefrontal node (anterior middle frontal gyrus) of the canonical resting state salience network. The area of the left amygdala in which less activation during emotional reactivity was found to moderate treatment response was also modulated by single pulse TMS to the right posterior vs. anterior middle frontal gyrus, such that individuals displaying greater inhibition of the left amygdala to right posterior vs. anterior middle frontal gyrus stimulation at rest displayed better treatment outcomes (B). This effect arose entirely from right posterior middle frontal stimulation and not right anterior middle frontal gyrus stimulation (C). The red areas representing the TMS targets in (B) are 8mm spheres centered on the cluster centers of mass for the right executive control and right salience network prefrontal nodes independently-derived from a separate healthy control dataset. Separate lines represent individuals above (solid line) or below (dotted line) the median level of activation for the purposes of visualizing disparate symptom change trajectories within the immediate treatment group. aMFG = anterior middle frontal gyrus; CAPS = Clinician-Administered PTSD Scale for DSM-IV; ECN = executive control network; PE = prolonged exposure group; pMFG = posterior middle frontal gyrus; TMS = transcranial magnetic stimulation; **p < 0.01.