Psilocybin therapy increases cognitive and neural flexibility in patients with major depressive disorder

Abstract

Psilocybin has shown promise for the treatment of mood disorders, which are often accompanied by cognitive dysfunction including cognitive rigidity. Recent studies have proposed neuropsychoplastogenic effects as mechanisms underlying the enduring therapeutic effects of psilocybin. In an open-label study of 24 patients with major depressive disorder, we tested the enduring effects of psilocybin therapy on cognitive flexibility (perseverative errors on a set-shifting task), neural flexibility (dynamics of functional connectivity or dFC via functional magnetic resonance imaging), and neurometabolite concentrations (via magnetic resonance spectroscopy) in brain regions supporting cognitive flexibility and implicated in acute psilocybin effects (e.g., the anterior cingulate cortex, or ACC). Psilocybin therapy increased cognitive flexibility for at least 4 weeks post-treatment, though these improvements were not correlated with the previously reported antidepressant effects. One week after psilocybin therapy, glutamate and N-acetylaspartate concentrations were decreased in the ACC, and dFC was increased between the ACC and the posterior cingulate cortex (PCC). Surprisingly, greater increases in dFC between the ACC and PCC were associated with less improvement in cognitive flexibility after psilocybin therapy. Connectome-based predictive modeling demonstrated that baseline dFC emanating from the ACC predicted improvements in cognitive flexibility. In these models, greater baseline dFC was associated with better baseline cognitive flexibility but less improvement in cognitive flexibility. These findings suggest a nuanced relationship between cognitive and neural flexibility. Whereas some enduring increases in neural dynamics may allow for shifting out of a maladaptively rigid state, larger persisting increases in neural dynamics may be of less benefit to psilocybin therapy.

Fig. 1. The effects of psilocybin therapy on depression and cognitive flexibility.

Depression symptomology (a) as measured by the GRID-Hamilton Depression Rating Scale (GRID-HAMD; also reported in [3]) and cognitive flexibility (b) as measured by perseverative errors on the Penn Conditional Exclusion Test (PCET) were improved from pre- to post-psilocybin therapy. These changes were not found between repeated tests pre-psilocybin therapy in the delayed group (i.e., between −8 weeks and Baseline time points). Each line color represents a unique participant, and the mapping of individual colors to unique participants remains consistent across all figures in which individual participant data are plotted.

Fig. 2. The effects of psilocybin therapy on neurometabolite concentrations and functional connectivity.

Split violin plots with horizontal bars reflecting the mean, error bars reflecting the 95% confidence interval, and “strings” reflecting individual participant data. The color of a given participant is consistent across all figures plotting individual participant data. Psilocybin therapy reduced (a) glutamate and (b) N-acetylaspartate in the anterior cingulate cortex but not in the left or right hippocampus. c Psilocybin therapy increased dFC but not sFC between the anterior and posterior cingulate. *p < 0.050. ACC anterior cingulate cortex, L Hipp left hippocampus, R Hipp right hippocampus, Glu glutamate, tCr total creatine, NAA N-acetylaspartate, PCC posterior cingulate.

Fig. 3. Greater pre- to post-psilocybin therapy increases in neural flexibility were associated with less improvements in cognitive flexibility.

Relationship between changes in anterior to posterior cingulate cortex (ACC and PCC, respectively) dynamics of functional connectivity (dFC) 1 week post-psilocybin therapy and changes in perseverative errors on the Penn Conditional Exclusion Test (PCET) at (a) 1 week and (b) 4 weeks post-psilocybin therapy.

Fig. 4. Connectome-based predictive modeling of pre- to post-psilocybin therapy changes in depression and cognitive flexibility.

Performance of models trained on baseline functional connectivity as a function of the threshold for feature selection predicting 1-week (a) and 4-week (b) changes in depression (ΔHAMD) and 1-week (c) and 4-week (d) changes in cognitive flexibility (ΔPCET). Horizontal gray lines indicate model performance at p = 0.05. Brains on the right side of panels are the best-performing models (see gray circles on performance lines). A brain was not plotted for predicting 1-week changes in depression, as there was no model was that performed at r > 0. Edges that were included in at least 75% of folds were plotted in dark blue, magenta, and light blue, representing sFC edges positively correlated with behavior, sFC edges negatively correlated behavior, and dFC edges positively correlated with behavior, respectively. Brain visualizations created with NeuroMArVL (https://immersive.erc.monash.edu/neuromarvl/). LOO leave-one-out, sFC static functional connectivity, dFC dynamics of functional connectivity, ACC anterior cingulate cortex, PCC posterior cingulate cortex.

Fig. 5. The relationship between model features and cognitive flexibility.

Positively and negatively correlated features that were selected in the best models trained on baseline sFC and dFC predicting (a) 1-week changes and (b) 4-week changes in cognitive flexibility. Positively correlated dFC edges were consistently selected in these models, suggesting that greater baseline dFC was associated with more PCET perseverative errors (i.e., greater cognitive rigidity). In contrast, when models were trained on baseline dFC to predict baseline cognitive flexibility (c), dFC edges were still predictive of cognitive flexibility, but the correlations between dFC edges and cognitive flexibility was reversed (d). The sFC + dFC Full model was plotted here to highlight that regardless of how many edges were allowed into the model, far more dFC edges were negatively correlated with PCET perseverative errors. Pos. positively, Neg. negatively, Corr. correlated, LOO leave-one-out, sFC static functional connectivity, dFC dynamics of functional connectivity, ACC anterior cingulate cortex, PCC posterior cingulate cortex.