Emotions and brain function are altered up to one month after a single high dose of psilocybin

Earlier research has shown that psilocybin, a classic 5-HT2A receptor partial agonist, can produce acute reductions in negative mood, increases in positive mood, and decreased amygdala responses to negative emotional stimuli, and small clinical studies suggest antidepressant and anti-addiction effects that can persist for months after one or a few administrations. However, the mechanisms that might underlie any enduring therapeutic effects remain unclear. Two candidate, potentially interacting mechanisms are modulation of negative affect (and related affective biases) and changes in large-scale brain network plasticity; both amygdala reactivity and anterior cingulate cortex (ACC) function figure prominently in models of mood and substance use disorders and in prior psychedelic neuroimaging findings. Barrett and colleagues therefore conducted an open-label, within-subjects pilot study to test whether a single high dose of psilocybin (25 mg/70 kg) produces enduring changes in affect, personality, task-evoked neural responses to emotional stimuli, and resting-state functional connectivity. Participants were assessed one day before, one week after, and one month after psilocybin using a battery of self-report affect and personality measures and functional MRI during three emotion tasks and resting state. The study aimed to determine whether post-acute changes in affect and brain function persist after drug elimination and could point to trans-diagnostic mechanisms relevant to mood and substance use disorders.

This was an open-label, within-subjects longitudinal pilot study with twelve healthy right-handed volunteers (7 female; mean age 32.1 ± 7.5 years). Inclusion required medical and psychiatric health (Structured Clinical Interview for DSM-IV), and exclusion criteria included MRI contraindications, personal or family history of psychotic or bipolar disorder, recent moderate-to-severe substance use disorder, and use of psychoactive medications. Participants reported limited lifetime hallucinogen use. The protocol included preparatory meetings with session monitors, a single supervised psilocybin administration session (25 mg/70 kg in a capsule), and post-session debriefing; safety monitoring (blood pressure, heart rate, staff ratings) occurred repeatedly over the roughly 7-hour session. Behavioural and questionnaire assessments targeted state and trait affect and personality. Questionnaires administered one day before, one week after, and one month after dosing included PANAS-X, POMS, DPES, DASS, and STAI. The Big Five Inventory and the Tellegen Absorption Scale were completed at screening and again at one month. Mixed-effects repeated-measures one-way ANOVAs evaluated changes in self-report measures, with Tukey correction for post-hoc pairwise comparisons; paired t-tests assessed personality change. Neuroimaging took place one day before, one week after, and one month after dosing on a Philips 3T scanner. Each session included three affective tasks (emotion discrimination, emotion recognition, and an emotional conflict Stroop) interleaved with eyes-open resting-state scans (two 8-minute runs, 16 minutes total). Task-based BOLD data were preprocessed (slice timing, motion correction, normalization, smoothing) and analysed focussing on four a priori ROIs: left and right amygdala and left and right ACC. First eigenvariates for each ROI were extracted and entered into subject-level GLMs with motion regressors and scrubbing; group-level one-way ANOVAs assessed time effects with Holm–Bonferroni correction for post-hoc comparisons. Whole-brain voxel-wise exploratory GLMs were also run (threshold p < 0.0005 uncorrected; cluster-forming p < 0.05 uncorrected). Resting-state data were preprocessed similarly, bandpass filtered (0.009–0.08 Hz), nuisance-regressed (including white matter/CSF eigenvectors and motion scrubbing), and parcellated with the Shen 268-node atlas. Static functional connectivity was computed as Pearson correlations between node time series, Fisher z-transformed, and one-sample t-tests (Bonferroni-corrected across 35,778 edges) identified reliable edges at each time point. Edges significant at any time point were then contrasted across time (paired t-tests). One subject was excluded from resting-state analyses for missing the 1-week scan. Network-level summaries were obtained by averaging edges within and between eight canonical networks defined from the Shen atlas.

Self-report affect and personality: Several measures of negative affect and stress were reduced at one week post-psilocybin and moved back toward baseline by one month. Specifically, DASS stress, PANAS negative affect, STAI state anxiety, and POMS tension, depression, and total mood disturbance scores were significantly lower at 1 week versus baseline and tended to return toward baseline at 1 month. Trait anxiety was reduced at 1 month relative to baseline. Positive dispositional affect increased: DPES joy and content subscales showed a main effect of time (DPES joy F[2,20] = 6.03, p = 0.009, partial η2 = 0.36), with scores higher at both 1 week and 1 month than at baseline. Personality measures showed significant increases in conscientiousness (t = 2.33, p = 0.042, d = 0.738) and in absorption (TAS; t = 3.55, p = 0.005, d = 1.122) from baseline/screening to 1 month; other Big Five changes were numerical but not statistically emphasised. Emotion tasks and ROI fMRI: Performance in the emotion discrimination task was near ceiling at baseline (92.04% ± SEM 1.7%) and improved modestly to 94.7% (SEM 1.4%) at 1 week (t = 3.089, p = 0.006) and 94.1% (SEM 1.5%) at 1 month (t = 2.669, p = 0.014). No significant time-related changes in amygdala or ACC responses were observed for the emotion discrimination task, and whole-brain analyses were null. For the emotion recognition task, no whole-brain voxel-wise effects were found; supplementary materials report exploratory associations between self-report and amygdala change. In ROI analyses, the authors report decreased amygdala responses to emotional stimuli at 1 week post-psilocybin that rebounded by 1 month. Emotional conflict Stroop and whole-brain fMRI: Behavioural performance was at ceiling across sessions (98.9% ± SEM 0.21%) but a classic Stroop interference effect was evident (accuracy and response-time main effects, F[3,2575] = 5.704, p = 0.00069 for accuracy; F[3,2575] = 7.019, p = 0.00011 for response time). ROI analyses did not show amygdala or ACC effects for the Stroop contrasts. However, whole-brain voxel-wise analyses for the high-demand incongruent versus low-demand congruent contrast (CI > CC) revealed increased BOLD response from baseline to 1 week in dorsal lateral prefrontal cortex (DLPFC) and medial orbitofrontal cortex (MOFC), and from baseline to 1 month in somatosensory cortex and fusiform gyrus. Additionally, greater BOLD responses were observed at 1 week compared to 1 month in a mainly left-hemisphere network including inferior frontal gyrus, anterior insula, parietal lobule, and fusiform gyrus. Resting-state functional connectivity: Of 35,778 possible edges in the Shen atlas, 695 edges were significant (Bonferroni-corrected) for at least one time point. Overall connectivity strength increased from baseline to 1 week (38 edges showing greater connectivity, 10 showing less) and this pattern largely persisted at 1 month (29 edges greater, 18 edges less). Seven of the 29 edges that increased at 1 month overlapped with those that increased at 1 week, and these were distributed across lobes and networks. Changes did not follow a clear, single-network pattern; numerically there were more increases than decreases both within and between networks at 1 week and 1 month. Measures of dispersion of connectivity strengths within and between networks were unchanged across time points.

The authors highlight four main sustained effects following a single high dose of psilocybin in this healthy volunteer sample: an acute reduction in negative affect at 1 week that returned toward baseline at 1 month; reduced amygdala responsiveness to emotional stimuli at 1 week with a rebound by 1 month; increased recruitment of reward-learning, attention, and decision-making regions (DLPFC and MOFC) during high-demand emotional conflict at 1 week, with other sensory and fusiform increases evident at 1 month; and widespread increases in resting-state functional connectivity at both post-dose time points. The investigators interpret these findings as more consistent with a transient neuroplastic period than with residual pharmacology, noting that psilocybin and its active metabolite have short half-lives (roughly 3 hours) and that 5-HT2A receptor internalisation would be expected to resolve within days. In terms of mechanism, increased DLPFC and MOFC recruitment during emotionally conflicting trials is discussed as a possible top-down modulation of affect — DLPFC is implicated in executive control and emotion regulation and can down-regulate amygdala activity, while anterior MOFC may signal abstract positive value and support altered reward learning. The authors therefore propose that greater top-down control could shift the salience balance between positive and negative information, contributing to reduced negative affect and increased positive affect. Resting-state connectivity increases across networks are framed as consistent with other reports of post-acute network change and with preclinical evidence for psychedelics promoting neuritogenesis and spinogenesis, suggesting a domain-general cortical plasticity process. The authors acknowledge limitations that temper conclusions: the small sample size, absence of placebo or active-control conditions, open-label design, and limited pre-dosing baseline assessments leave open the possibility of expectancy effects, demand characteristics, or practice/habituation influences. They note, however, that several effects (for example, the return of negative affect and task-based fMRI effects toward baseline at 1 month) are not easily explained by simple practice effects. Differences between the current healthy-volunteer findings and prior reports in patient samples are emphasised as potentially driven by population differences and methodological variations (e.g. eyes-open versus eyes-closed resting-state, scan length). Finally, the authors suggest that observed short-term reductions in negative affect and increases in connectivity point to negative affect as a plausible trans-diagnostic target for psilocybin and that psilocybin may initiate a time-limited neuroplastic window which could be harnessed in clinical interventions. Replication in larger samples with rigorous control conditions is recommended.

Barrett and colleagues conclude that a single high dose of psilocybin produced preliminary evidence of sustained shifts in affect and in neural correlates of affective processing that extended beyond the acute pharmacological window. The authors propose that these changes—particularly reductions in negative affect and broad increases in functional connectivity—identify negative affect as a potential therapeutic target and are consistent with the hypothesis that psilocybin initiates a transient neuroplastic period that may facilitate longer-term emotional and cognitive change. They emphasise that these findings are preliminary and call for replication in larger, controlled studies.