Altered brain activity and functional connectivity after MDMA-assisted therapy for post-traumatic stress disorder

CONCLUSION

We report signatures of brain response during rest and audio listening task in eight veterans and one first-responder with clinically diagnosed chronic and severe PTSD before and two-months after MDMA-assisted therapy. We found a significant reduction in CAPS-IV total severity scores after therapy (Figure), indicating our sub-study participants mirrored the results from the parent study. We found a trend suggesting that RSFC between the amygdala and hippocampus was strengthened post-therapy, particularly in the left hemisphere (Figure). Prior work suggests that modulation of amygdalae-hippocampal RSFC may be an important component of MDMA-AT for PTSD, thus investigating this connection in future studies is warranted. We also found participants had increased activation in areas involved with self-referential processing and autobiographical memory while listening to traumatic versus neutral memory narrations pre-therapy (Figures), and that no significant contrast existed after MDMA-AT (Figure). Comparing trauma versus neutral contrasts between baseline and followup revealed a significant decrease in cuneus contrast after MDMA-AT (Figure). Finally, the pre-to post-therapy reductions in FC during the script listening task between the left amygdala and right PCC, left PCC, and left insula, as well as FC between the left isthmus cingulate and left hippocampal tail strongly and significantly correlated with PTSD symptom improvement (Figure). These results begin to shed light on the neurological mechanisms that may drive MDMA-AT for PTSD. Previous work quantifying functional connectivity in PTSD and in stress exposureand acute MDMA administration in controlssuggests one mechanism of MDMA-AT may be to increase pathologically lowered amygdala-hippocampal RSFC. The amygdala is associated with fear expression, threat recognition, and heightened response to emotional memories and is often dysregulated in patients with PTSD. The hippocampus also plays a central role in PTSD as it is thought to provide contextual information important for cognitive-affect during memory recollection. Sripada et al.found combat veterans with PTSD had reduced amygdala-hippocampal RSFC compared to combat-exposed controls, leading them to speculate that this may relate to an inability to contextualize affective information in PTSD. Carhart-Harris et al.demonstrated that amygdala-hippocampal RSFC is increased acutely in MDMA administration compared to placebo and this increase occurred in a manner that correlated with the drug's subjective effects at a near-significant level, leading these researchers to propose that this functional connection was a primary target of MDMA-AT. Increased amygdalahippocampal RSFC has also been linked to intranasal oxytocin administration after stress exposureand this effect was mediated by serotonin signaling (55)-two neuromodulators that play a significant role in the pro-social and fear extinction effects of MDMA. Prior to our analysis (although after the study was designed and the data collected), we hypothesized that the RSFC between the amygdala and hippocampus would be higher after MDMA-AT compared to pre-therapy levels. Only the RSFC between the left amygdala and left hippocampus was significantly increased (Figure) however, this finding no longer met thresholds for significance after multiple comparisons correction (pFDR = 0.09). We also found that the amount of increased right amygdalaleft hippocampal RSFC after MDMA-AT positively correlated with PTSD symptom improvement at a near-significant level (Supplementary Figure; R = -0.820, uncorrected p = 0.046, pFDR = 0.183). Our current findings, though inconclusive, are suggestive of a trend toward moderate increases in amygdalahippocampal RSFC two-months after MDMA-AT. It is possible that more significant changes would have been observed with a larger sample size, longer resting-state scans, or imaging performed closer to MDMA administration. These findings justify the continued investigation of amygdala-hippocampal RSFC in the therapeutic mechanisms of MDMA-AT in future studies. We next sought to study brain response during autobiographical memory listening to draw additional conclusions about MDMA-AT's effects in individuals with PTSD. Before therapy, participants had larger activation in four areas during an individualized trauma script listening task compared to neutral script listening: the right and left isthmus cingulate and precuneus, the left caudal middle frontal gyrus, the right medial prefrontal cortex, and the left rostral middle frontal gyrus (Figures). These regions are broadly involved in self-processing operations (e.g., first-person perspective taking), episodic memory retrieval, visual-spatial imagery, auto-biographical memory recollection, and are included in or interact with the default mode network. The retrosplenial cortex-located within the isthmus cingulate-is also found to be consistently activated by emotionally salient stimuli, and has been proposed to play a role in the interaction between emotion and memory. We conjecture that increased activation in these regions during traumatic compared to neutral audio listening (Figures) could be related to an increased intensity of the recollection or re-experiencing of traumatic memories compared to neutral ones for patients before therapy. At 2-month follow-up to MDMA-AT, there was no significant difference in the trauma vs neutral script activation (Figure). The longitudinal comparison of these two time points indicated that the contrast between trauma and neutral was larger at baseline, particularly in a significant cluster in the right and left cuneus/lingual gyrus (Figure). Cuneus activity during autobiographical memory tasks often coincides with activity in the frontal regions highlighted by the baseline contrast, and has been found to correlate with memory recall accuracy. Cuneus activity is thought to enhance the visual imagery of autobiographical memory recollection, therefore decreased contrast in this area at follow-up suggests that intensity of visual imagery contrast between trauma and neutral memories may be decreased after MDMA-AT. Larger studies may allow more statistical power to identify additional longitudinal differences. Other longitudinal studies of individuals with PTSD have found that decreases in precuneus, isthmus cingulate, and middle frontal gyrus activation during symptom provocation is correlated with reductions in PTSD symptom severity. PTSD is often associated with hyperactivity in the amygdala (36); the acute administration of MDMA in healthy volunteers decreases blood flow to the amygdala during restand attenuates its response to angry faces. We had hypothesized that we would observe hyperactivity of the amygdala to trauma versus neutral scripts at baseline and that MDMA-AT would attenuate this response, however we observed neither. It is important to note inconsistencies in the literature here. Amygdala hyperactivity in PTSD is not always observed, possibly due to differences in subtypes, sex, cultural representations, or choice of paradigm. Additionally, while MDMA did suppress amygdala activity during rest and in response to angry faces as previously mentioned, there was no observed impact on its response to autobiographical memories. While activation-based analyses deserve continued attention in future studies to rectify these inconsistencies, functional connectivity is a complimentary approach we can use to extract additional information from fMRI. Functional connectivity analyses in individuals with PTSD have revealed aberrant connectivity between several regions within default mode, limbic, and salience networks, and more broadly, regions involved in emotional and selfreferential processing, and, further, the administration of MDMA in healthy volunteers has been shown to disrupt the functional integrity of these networks. Here, we measured functional connectivity during scriptdriven autobiographical memory recall between the right and left hippocampus head, hippocampus tail, amygdala, precuneus, caudal ACC, rostral ACC, PCC, isthmus cingulate, and the insula. Our ROIs were defined and labeled based on the Desikan-Killiany (DK) brain atlas (Supplementary Figures). We chose to segment the hippocampus into anterior (head) and posterior (tail) ROIs based on recent work showing that the two portions' FC are differentially effected by PTSD. We assessed group-level changes in the strength of these functional connections and found no significant differences between baseline and follow-up after corrections for multiple comparisons (Figure). However, we did find that greater recovery (larger decreases in CAPS-IV at follow-up) was associated with reductions in FC between the left amygdala and the right and left PCC, as well as the left insula (Figure). The acute effects of MDMA in healthy volunteers has been shown to decrease the FC of the PCCand insula, and alter amygdala and hippocampus FC, highlighting the potential relevance of our current findings. Amygdala to posterior and mid-cingulate cortex FC has been shown to be associated with PTSD severity at different stages of disease progression, although differing patient populations and assessment timelines lead to conflicting results. One finding in healthy adults shows increased amygdala-PCC FC following the acute exposure to stress (103), thus the association between recovery and reduced amygdala-PCC task FC at follow-up possibly relates to reduced stress response to trauma memories (although the finding by Veer and colleagues is more posterior to the ROI used here). Amygdala and insula RSFC is increased in PTSD[except in one study which finds the opposite], and reduced amygdala-insula FC during negative image reappraisal is associated with larger improvements in PTSD symptoms (101). The strength of left amygdala-insula FC also positively correlates with the amount of acute anxiety measured in participants just before scanning (104). Attenuated functional connectivity of these two regions at follow-up in the present study possibly suggests a decreased intensity of recalled events, less "re-experiencing, " or reduced anxiety during the scriptdriven memory recall. Lastly, we found that reductions in CAPS-IV at follow-up were associated with reduced FC between the left isthmus cingulate and left hippocampal tail (Figure). The isthmus cingulate labeled here consists of the most posterior potions of the PCC (Supplementary Figure). Increased FC between these two regions has previously been reported in PTSD patients compared to trauma-exposed health controls, and the present finding is possibly indicative of changes in memory contextualization and reduced threat sensitivity at two-month follow-up to MDMA-AT compared to baseline (100). PTSD is characterized by decreased fear extinction in response to trauma-related stimuli. One possible mechanism through which MDMA-AT operates is enhanced reconsolidation and/or fear extinction processes. Several studies with MDMA implicate reconsolidation or fear extinction processes, and while it is currently unclear whether MDMA acts on only one or both, it is important to note that the two interact (105). Rodent models have demonstrated that the administration of MDMA prior to extinction learning enhances extinction retention (tested 48 h after learning) and this effect is blocked by acute and chronic treatment with a serotonin transporter inhibitor (20, 106). Hake et al. (107) found that MDMA administered during extinction learning phases did not enhance fear extinction memory, while MDMA administration during reconsolidation phases resulted in prolonged reductions in conditioned fear. In addition, MDMA administered prior to trauma-cue exposure (reconsolidation phase) in rodents resulted in reduced stress-related behavioral responses 7 days later (108). Two recent trials in healthy humans found that MDMA (100 and 125 mg, respectively) administered prior to extinction learning resulted in improved extinction learning at extinction recall phases (48 and 24 h later, respectively) compared to the placebo group. Doss et al.found that 1 mg/kg of MDMA in healthy humans attenuated the encoding and retrieval of salient details from positive and negative stimuli (but not neutral stimuli), suggesting an ability for MDMA to alter emotional memory representation. Interestingly, a fMRI study in healthy humans found decreased activation in the precuneus/PCC during fear extinction learning (109), regions highlighted by our present study and others in PTSD.