Psilocybin for treatment-resistant depression: fMRI-measured brain mechanisms

Abstract

Psilocybin with psychological support is showing promise as a treatment model in psychiatry but its therapeutic mechanisms are poorly understood. Here, cerebral blood flow (CBF) and blood oxygen-level dependent (BOLD) resting-state functional connectivity (RSFC) were measured with functional magnetic resonance imaging (fMRI) before and after treatment with psilocybin (serotonin agonist) for treatment-resistant depression (TRD). Quality pre and post treatment fMRI data were collected from 16 of 19 patients. Decreased depressive symptoms were observed in all 19 patients at 1-week post-treatment and 47% met criteria for response at 5 weeks. Whole-brain analyses revealed post-treatment decreases in CBF in the temporal cortex, including the amygdala. Decreased amygdala CBF correlated with reduced depressive symptoms. Focusing on a priori selected circuitry for RSFC analyses, increased RSFC was observed within the default-mode network (DMN) post-treatment. Increased ventromedial prefrontal cortex-bilateral inferior lateral parietal cortex RSFC was predictive of treatment response at 5-weeks, as was decreased parahippocampal-prefrontal cortex RSFC. These data fill an important knowledge gap regarding the post-treatment brain effects of psilocybin, and are the first in depressed patients. The post-treatment brain changes are different to previously observed acute effects of psilocybin and other 'psychedelics' yet were related to clinical outcomes. A 'reset' therapeutic mechanism is proposed.

Figure 1

Whole-brain cerebral blood flow maps for baseline versus one-day post-treatment, plus the difference map (cluster-corrected, p < 0.05, n = 16). Correlation chart shows post-Treatment changes in bilateral amygdala CBF versus changes in depressive symptoms (r = 0.59, p = 0.01). One patient failed to completed the scan 2 QIDS-SR16 rating, reducing the sample size to n = 15 for the correlation analysis. In all of the images, the left of the brain is shown on the left.

Figure 2

Top two rows = sgACC (purple) RSFC before and after psilocybin treatment (hot colours = regions of significantly positive coupling). Bottom row reveals regions where there was a significant increase in sgACC RSFC post-treatment (hot colours). All maps are cluster-corrected, p < 0.05, Z > 2.3.

Figure 3

Top two rows = vmPFC (purple) RSFC before and after psilocybin treatment (hot colours = regions of significantly positive coupling). Bottom row reveals regions where there was a significant increase in vmPFC RSFC post-treatment (hot colours). All maps are cluster-corrected, p < 0.05, Z > 2.3. Increased coupling between the vmPFC and the displayed regions (bottom row) was predictive of clinical response at 5-weeks post-treatment. Chart shows mean values and positive standard errors.

Figure 4

Top two rows = Bilateral PH (purple) RSFC before and after psilocybin treatment (hot colours = regions of significantly positive coupling). Bottom row reveals regions where there was a significant decrease in PH RSFC post-treatment (cold colours). All maps are cluster-corrected, p < 0.05, Z > 2.3. Decreased coupling between the PH and the displayed regions (bottom row) was predictive of clinical response at 5-weeks post-treatment (t = −1.9, p = 0.04). Chart shows mean values and negative standard errors.

Figure 5

Differences in between-RSN RSFC or RSN ‘segregation’ before and after therapy. Each square in the matrix represents the strength of functional connectivity (positive = red, negative = blue) between a pair of different RSNs (beta values). The matrix on the far right displays the between-condition differences in covariance (t values). The RSNs are: 1) medial visual network, 2) lateral visual network, 3) occipital pole network, 4) auditory network, 5) sensorimotor network, 6) DMN, 7) parietal cortex network, 8) the dorsal attention network, 9) the salience network, 10) posterior opercular network, 11) left frontoparietal network, 12) right frontoparietal network. White asterisks represent significant differences (P < 0.05, non-corrected). Both of the significant differences did not survive FDR correction for multiple comparisons.