Increased low-frequency brain responses to music after psilocybin therapy for depression

Wall and colleagues situate the study within renewed clinical interest in classic psychedelics such as psilocybin for disorders including depression. They note that music is an integral component of modern psychedelic-assisted therapy and that music acts as a hedonic, emotionally potent stimulus whose neural processing—particularly in auditory cortex and reward/limbic circuits—can index anhedonia, a core symptom of depression. Previous studies have reported reduced neural responses to music in depression and enhanced subjective and neural responses to music under classic psychedelics, but the specific impact of psilocybin therapy on brain responses to music in depressed patients remained incompletely characterised. This study set out to test whether psilocybin therapy alters low-frequency brain activity evoked by music in patients with treatment-resistant major depression. Using functional MRI and an ALFF (Amplitude of Low-Frequency Fluctuations) analysis to quantify slow spontaneous oscillations, the investigators compared music-listening and resting-state scans acquired one week before the first dosing and the day after the second (high-dose) dosing session. They also examined relationships between any ALFF changes and subjective measures obtained during and after therapy, aiming to assess whether acute drug-state experiences related to post-treatment changes in music-evoked brain responsiveness.

Nineteen participants (13 male, 6 female; mean age 41.3, SD 10.5) meeting criteria for treatment-resistant major depression (HAM-D ≥ 17 and non-response to at least two pharmacological treatments) were recruited. Standard medical and psychiatric screening procedures were applied; exclusion criteria included psychotic disorders, family history of psychosis, prior suicide attempts requiring hospitalisation, pregnancy, substance dependence, mania history, contraindicated medications, and MRI contraindications. The study had an open-label design with no placebo or control group; all participants received active psilocybin with full disclosure. Intervention comprised two supervised dosing sessions one week apart: a low-dose session (10 mg) followed by a high-dose session (25 mg). During dosing participants lay with eyes closed and listened to music; two therapists provided non-directive supportive supervision. MRI visits occurred one week prior to the first dosing and the day after the second dosing session. Primary clinical outcome measures included the Quick Inventory of Depressive Symptoms (QIDS) administered at baseline and serially to three months, and additional psychometrics included the 5D-ASC (5-dimension Altered States of Consciousness questionnaire), administered 6–7 hours post-dosing, and the GEMS-3 (Geneva Emotional Music Scale), administered after each MRI session to capture subjective music responses. Functional MRI data were acquired at 3T with eyes closed during both resting-state and continuous music-listening scans. Music stimuli were two different solo piano tracks (edited compositions), one played at the first scan visit and the other at the second, with track order counterbalanced across subjects. Data preprocessing and statistical analyses used FSL (v6.03) and AFNI. Preprocessing included motion correction, spatial smoothing (6 mm Gaussian), pre-whitening, and regression of white matter, CSF and six motion parameters to denoise the data; residuals from these first-level models were carried forward. ALFF images were computed after band-pass filtering (0.01–0.1 Hz) using AFNI and entered into group-level Ordinary Least Squares mixed-effects models in FSL. Thresholding used Z = 2.3 and cluster-corrected p < 0.05. The investigators ran a mean ALFF map, a main task contrast (music vs rest), and within-scan comparisons of pre- versus post-therapy for music and rest separately. Functional ROIs were defined from task and treatment contrast clusters and used for 2×2 ANOVAs (scan type × treatment) and Pearson correlations between ALFF change scores and psychometric measures; ROI data were Z-normalised prior to analysis. Details on certain preprocessing parameters were taken from the extracted text; if additional acquisition parameters are required they are not clearly reported in this extract.

The group mean ALFF distribution matched typical ALFF patterns, with high values in ventral regions and in cingulate, medial/lateral frontal lobes and insula. The main task contrast (music versus rest, collapsed across sessions) revealed relatively higher ALFF in bilateral superior temporal gyrus (primary auditory cortex) during music, and higher ALFF in the lingual gyrus during rest. ROI analyses focused on these task-defined clusters. A 2×2 ANOVA examining effects of scan type (rest vs music) and therapy (before vs after) found in the superior temporal gyrus (STG) ROI a main effect of scan type (F[1,18] = 32.89, p < 0.001) and a significant interaction with treatment (F[1,18] = 4.90, p = 0.04). Post-hoc tests showed that ALFF in the STG increased after psilocybin therapy, but this increase was specific to the music-listening scan (t[31] = 2.09, p = 0.045). In the lingual gyrus ROI there was only a main effect of scan type (F[1,18] = 38.45, p < 0.001) and no treatment effects. Analyses testing for track-order effects found no significant main effects or interactions (all p > 0.1). Voxelwise pre-versus-post comparisons within each scan type indicated decreased ALFF after therapy in medial frontal regions during rest (before > after). In contrast, the music scan showed after > before increases in ALFF in lateral superior temporal areas and the supramarginal gyrus in the left hemisphere, and a smaller right-hemisphere cluster centred near the inferior precentral gyrus; a small left lingual gyrus cluster was also observed. Change (delta) scores from these treatment-derived ROIs were correlated with psychometric measures: increased ALFF to music after therapy correlated with several 5D-ASC subscales obtained after the high-dose session — DED (ego-dissolution), VRS (visionary restructuralization), AUA (auditory alterations), and VIR (vigilance reduction) — whereas the OCEAN (oceanic boundlessness) subscale was not significant. The mean of all five subscales correlated with the ROI data (r = 0.621, p = 0.005), and the VIR subscale alone correlated at r = 0.613, p = 0.005; both the global mean and VIR survived family-wise correction with a corrected alpha of p = 0.008. No significant relationships were found between ALFF changes and other subjective questionnaire measures or clinical rating scales, and the rest-scan data showed no significant correlations with psychometrics. The extracted text does not present full numerical details for every cluster (e.g., cluster sizes and coordinates) in this summary, though thresholding criteria and key statistics above are reported.

Using ALFF, Wall and colleagues identified auditory cortical regions (superior temporal lobe, including Heschl's gyrus and planum temporale) where low-frequency fluctuations were higher during music listening than during rest, and where ALFF increased after psilocybin therapy specifically for the music scans. Separate analyses indicated decreased ALFF in medial frontal regions during rest after therapy. The authors interpret the pattern as evidence that psilocybin therapy enhanced responsiveness to a hedonic auditory stimulus in recognised music and musical-emotion processing areas. The investigators highlight that although prior literature often emphasises reward and limbic regions in music-evoked emotion, growing evidence supports a role for auditory cortex in affective processing and its functional coupling with limbic structures; a recent meta-analysis has implicated Heschl's gyrus in music-evoked emotions. The observed correlations between post-therapy increases in music-related ALFF and higher scores on several 5D-ASC subscales suggest that the intensity or quality of the acute psychedelic experience—including ego-dissolution, visionary restructuralization, auditory alterations and vigilance reduction—was related to the enhanced neural responsiveness to music, implying the acute drug experience may have contributed to the changes. The decreased medial frontal ALFF after therapy is discussed in the context of medial prefrontal dysfunction in depression and prior findings linking medial PFC alterations to both antidepressant effects and acute psychedelic action. The authors cite convergent findings from earlier analyses of this cohort showing changes in connectivity of emotion/reward regions and enhanced responses to emotional face stimuli post-treatment. They also justify the use of continuous, naturalistic music stimuli and ALFF as suitable for capturing low-frequency dynamics, noting ALFF's high test–retest reliability and that the within-subject design mitigates physiological confounds. Key limitations acknowledged include the open-label design, small sample size, and absence of a placebo or control group, all of which limit causal inferences and generalisability. The authors also note that physiological noise likely contributed to ALFF values in ventral regions despite denoising, but argue that within-subject contrasts reduce the impact of such effects. They recommend future studies examine music-evoked responses at longer intervals after dosing, since neural changes may differ between the day after dosing and later time points. Overall, the authors conclude these data provide an initial indication that psilocybin therapy for depression can enhance neural responsiveness to music, related to the subjective acute drug experience, and that music is a useful experimental probe in psychedelic-therapy research.