LSD reconfigs cortex thru faster brain rhythms and increased fractal dimension

A study combining source-resolved magnetoencephalography, spectral parameterization, temporal complexity metrics, and interpretable machine learning found that LSD induces robust, spatially structured increases in alpha and beta peak frequencies alongside genuine attenuation of oscillatory power, with partly dissociable cortical patterns. The research, published in Communications Biology, also reports that LSD flattens the aperiodic 1/f spectral slope and increases neural signal fractality and complexity, preferentially affecting sensory, language, emotion, and imagery-related networks while sparing motor cortex. Music did not robustly amplify these neural signatures and instead showed a trend toward attenuation.

Abstract Lysergic acid diethylamide LSD profoundly alters conscious experience, yet the electrophysiological mechanisms by which it reshapes neural dynamics remain incompletely understood. A hallmark of psychedelic states is widespread cortical desynchronization, typically inferred from reductions in spectral power, but whether such effects reflect genuine weakening of neural oscillations or are confounded by shifts in oscillatory peak frequencies remains unresolved. We combine source-resolved magnetoencephalography MEG , spectral parameterization, temporal complexity metrics, and interpretable machine learning in an LSD versus placebo design, with and without music. We show that LSD induces robust, spatially structured increases in alpha and beta peak frequencies alongside genuine attenuation of oscillatory power, with partly dissociable cortical patterns. Beyond rhythmic activity, LSD is associated with flattening of the aperiodic 1/f spectral slope and increased neural signal fractality and complexity, preferentially affecting sensory, language, emotion, and imagery-related networks while sparing motor cortex. Machine-learning analyses identify peak-frequency shifts, aperiodic parameters, and complexity measures as key discriminators of the psychedelic state. Music does not robustly amplify these neural signatures and instead shows a trend toward attenuation. These findings provide an electrophysiological account of how LSD reorganizes large-scale human brain dynamics and highlight features that may differentiate its neural signature from other psychedelics. Similar content being viewed by others Introduction Psychedelics are potent psychoactive compounds that elicit profound alterations in conscious experience, spanning heightened sensory perception, vivid imagery, and distortions in the sense of self<sup 1 https://www.nature.com/articles/s42003-026-10629-7 ref-CR1 </sup . Beyond their characteristic phenomenology, these compounds offer a powerful experimental lever for perturbing large-scale brain dynamics and probing the neural mechanisms by which the brain gives rise to mind<sup 2 ref-CR2 , 3 ref-CR3 , 4 ref-CR4 , 5 https://www.nature.com/articles/s42003-026-10629-7 ref-CR5 </sup . Psychedelics have been used in rituals and cultural settings for thousands of years<sup 6 https://www.nature.com/articles/s42003-026-10629-7 ref-CR6 </sup , often alongside music<sup 7 https://www.nature.com/articles/s42003-026-10629-7 ref-CR7 </sup . Classic psychedelics—including psilocybin, N , N -dimethyltryptamine DMT , and lysergic acid diethylamide LSD —exert their profound effects primarily through agonism at the 5-HT2 A receptor. Among these, LSD is uniquely characterized by its additional affinity for dopaminergic receptors<sup 8 https://www.nature.com/articles/s42003-026-10629-7 ref-CR8 </sup , its extreme potency producing effects at remarkably low doses , and its extended receptor kinetics. Despite these well-known pharmacological properties, the precise manner in which LSD perturbs cortical dynamics—spanning oscillatory rhythms, aperiodic spectral structure, and temporal signal complexity—and how these effects are shaped by sensory stimuli, such as music, remains poorly understood. Modern neuroimaging has played a central role in uncovering how psychedelics alter large-scale brain function and core aspects of conscious experience, a process thought to support their therapeutic effect<sup 5 https://www.nature.com/articles/s42003-026-10629-7 ref-CR5 , 9 https://www.nature.com/articles/s42003-026-10629-7 ref-CR9 , 10 https://www.nature.com/articles/s42003-026-10629-7 ref-CR10 </sup . These mechanisms have been investigated across different modalities, including positron emission tomography PET <sup 11 https://www.nature.com/articles/s42003-026-10629-7 ref-CR11 </sup , functional MRI fMRI <sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 </sup , electroencaphelography EEG <sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 </sup , and magnetoencephalography MEG <sup 14 https://www.nature.com/articles/s42003-026-10629-7 ref-CR14 </sup . fMRI has been especially influential in understanding the psychedelic state e.g., ref. <sup 15 https://www.nature.com/articles/s42003-026-10629-7 ref-CR15 </sup . However, the fMRI signal reflects neural activity only indirectly via blood oxygenation, and psychedelics strongly modulate neurovascular processes<sup 16 https://www.nature.com/articles/s42003-026-10629-7 ref-CR16 </sup . As a result, psychedelic-induced BOLD changes may partly reflect disruption in neurovascular coupling rather than neural dynamics per se<sup 16 https://www.nature.com/articles/s42003-026-10629-7 ref-CR16 </sup . Although simultaneous EEG-fMRI studies are beginning to address these issues<sup 17 https://www.nature.com/articles/s42003-026-10629-7 ref-CR17 </sup , direct electrophysiological measurements remain essential for capturing the fast-changing dynamics that characterize the psychedelic state. M/EEG provides a direct measurement of the brain’s electrophysiological activity at a millisecond scale, allowing precise characterization of oscillatory dynamics<sup 18 https://www.nature.com/articles/s42003-026-10629-7 ref-CR18 , 19 https://www.nature.com/articles/s42003-026-10629-7 ref-CR19 </sup . Early electrophysiological studies revealed a robust broadband desynchronization under psychedelics—including psilocybin<sup 14 https://www.nature.com/articles/s42003-026-10629-7 ref-CR14 </sup , LSD<sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 </sup , and DMT<sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 </sup . However, the spectral power of the neural signals reflects a mixture of periodic oscillatory and aperiodic 1/ f -like components<sup 20 https://www.nature.com/articles/s42003-026-10629-7 ref-CR20 , 21 https://www.nature.com/articles/s42003-026-10629-7 ref-CR21 </sup , and changes in one component can be misattributed to the other, leading to an ambiguous interpretation of spectral power. This has motivated a shift toward explicit spectral parameterization to separately quantify rhythmic and arrhythmic activity<sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 , 22 https://www.nature.com/articles/s42003-026-10629-7 ref-CR22 , 23 https://www.nature.com/articles/s42003-026-10629-7 ref-CR23 </sup . While broadband desynchronization appears as a robust, cross-psychedelic effect<sup 12 ref-CR12 , 13 ref-CR13 , 14 https://www.nature.com/articles/s42003-026-10629-7 ref-CR14 </sup , emerging evidence suggests that LSD may additionally modulate spectral peak frequencies. Carhart-Harris and colleagues<sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 </sup reported subtle shifts in the alpha peak in posterior MEG sensors, a modulation not reported for DMT<sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 </sup and only inconsistently for psilocybin<sup 23 https://www.nature.com/articles/s42003-026-10629-7 ref-CR23 </sup . These observations raise the possibility that frequency-specific perturbations, particularly within the alpha and beta bands, constitute a distinguishing signature of LSD relative to other serotonergic psychedelics. Yet, existing studies have not mapped how these peak-frequency shifts are organized across the cortex, leaving the spatial expression of oscillatory power reduction, as well as alpha- and beta-peak modulation, an open question in psychedelic neuroscience. Importantly, traditional analyses using fixed canonical bands can produce misleading results when oscillatory peaks shift. Because the power spectrum follows a 1/ f -like decay, a peak that shifts upward may appear to lose strength simply because it is measured against a lower baseline, erroneously suggesting a desynchronization when the oscillation has merely changed frequency. To our knowledge, no previous study has analyzed LSD-induced spectral power changes while taking the peak shifts into account. Music is a powerful sensory stimulus that shapes perception, emotion, and internal experience<sup 24 https://www.nature.com/articles/s42003-026-10629-7 ref-CR24 , 25 https://www.nature.com/articles/s42003-026-10629-7 ref-CR25 </sup . Historically and across cultures, it has accompanied ritual practices and psychedelic experiences, suggesting a close relationship between music and altered states of consciousness. The heightened sensory perception and mood characterizing the psychedelic state synergize with and potentiate music’s evocative power. Music, therefore, serves as a potent contextual factor that shapes internal experience and provides a useful framework for investigating altered brain dynamics<sup 26 https://www.nature.com/articles/s42003-026-10629-7 ref-CR26 , 27 https://www.nature.com/articles/s42003-026-10629-7 ref-CR27 </sup . Set Mindset and Setting surrounding factors play a crucial role in a psychedelic trip<sup 28 https://www.nature.com/articles/s42003-026-10629-7 ref-CR28 </sup . Music has emerged as a particularly influential setting variable and its usefulness in psychedelic-assisted therapy is now fully recognized<sup 29 https://www.nature.com/articles/s42003-026-10629-7 ref-CR29 , 30 https://www.nature.com/articles/s42003-026-10629-7 ref-CR30 </sup . Empirical studies have shown that music under psychedelics increases emotional response<sup 31 https://www.nature.com/articles/s42003-026-10629-7 ref-CR31 </sup , enhances mental imagery<sup 32 https://www.nature.com/articles/s42003-026-10629-7 ref-CR32 </sup , supports meaning-making<sup 33 https://www.nature.com/articles/s42003-026-10629-7 ref-CR33 </sup . In addition, converging evidence suggests that emotional breakthrough and psychological insight may be more robust and clinically predictive dimensions of psychedelic experience <sup 34 ref-CR34 , 35 ref-CR35 , 36 ref-CR36 , 37 ref-CR37 , 38 ref-CR38 , 39 https://www.nature.com/articles/s42003-026-10629-7 ref-CR39 </sup . These constructs arguably capture tangible processes, such as affective release and cognitive restructuring, and may therefore offer a more direct bridge between phenomenology and underlying neurobiology. Despite this extensive phenomenological and therapeutic work, the neural mechanisms underlying these effects remain poorly understood. While M/EEG studies have shown attenuated top-down suppression of prediction error under LSD during basic auditory processing <sup 40 https://www.nature.com/articles/s42003-026-10629-7 ref-CR40 </sup —and increases in the richness of the neural signal during music listening <sup 41 https://www.nature.com/articles/s42003-026-10629-7 ref-CR41 </sup , a mechanistic account of how music modulates the oscillatory and temporal dynamics of the psychedelic brain is yet to be established. Beyond their effects on spectral structure, psychedelics are known to alter the temporal complexity of neural activity. Lempel-Ziv complexity LZC , a prominent metric in psychedelic research, quantifies the diversity of neural signals by measuring the rate at which unique temporal patterns appear, effectively indexing the richness of the time series. Numerous studies consistently reported an increased LZC under psilocybin, LSD, and DMT<sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 , 23 https://www.nature.com/articles/s42003-026-10629-7 ref-CR23 , 41 https://www.nature.com/articles/s42003-026-10629-7 ref-CR41 , 42 https://www.nature.com/articles/s42003-026-10629-7 ref-CR42 </sup . In contrast, measures of fractality, such as the Higuchi fractal dimension HFD , quantify the degree to which brain dynamics are self-similar across time scales, capturing scale-free temporal structure. It therefore complements LZC by revealing how complexity is organized across scales rather than how novel the signal is over time. HFD has been applied in other altered states of consciousness<sup 43 https://www.nature.com/articles/s42003-026-10629-7 ref-CR43 </sup , but has received far less attention in psychedelic research. Recent work suggests that psychedelic-induced rise in complexity can be modulated by external context, such as whether one attends to a stimulus or not<sup 41 https://www.nature.com/articles/s42003-026-10629-7 ref-CR41 </sup . This raises another question we address in the present study: to what extent are LSD-induced changes in temporal complexity intrinsic to the drug state, and to what extent are they shaped by structured sensory input, such as music? Machine learning ML offers a data-driven framework for distinguishing brain states and has been successfully applied to EEG and MEG data across a range of altered states, including anesthesia<sup 44 https://www.nature.com/articles/s42003-026-10629-7 ref-CR44 </sup , meditation<sup 43 https://www.nature.com/articles/s42003-026-10629-7 ref-CR43 </sup , sleep modulation<sup 45 https://www.nature.com/articles/s42003-026-10629-7 ref-CR45 </sup , and psychedelic drug effects<sup 46 https://www.nature.com/articles/s42003-026-10629-7 ref-CR46 , 47 https://www.nature.com/articles/s42003-026-10629-7 ref-CR47 </sup . Beyond classification accuracy, ML models can provide insights into the neural features that drive differences in drug states, particularly through feature-importance analyses. Such approaches offer a complementary perspective to traditional group-level statistics by highlighting neurophysiologically meaningful variables that contribute most strongly to discriminability. In this study, we leverage interpretable ML to characterize how LSD—and its interaction with music—manifests in large-scale cortical dynamics. In this study, we ask three questions: Q1 To what extent does LSD produce region-specific modulations of the peak frequency of neural oscillations within the alpha and beta bands? Q2 To what extent does LSD modulate the power of alpha and beta oscillations once the confounding effects of peak frequency shifts are accounted for? Q3 Does music modulate LSD-induced alterations in neural temporal organization, including oscillatory dynamics and signal complexity? To address these questions, we apply spectral parameterization, oscillatory peak-frequency analysis, and temporal complexity metrics LZC and HFD to source-reconstructed MEG data collected under LSD and placebo PLA conditions, with and without music. We further use machine-learning models to quantify the discriminative value of these features. Based on prior work and the mechanistic considerations outlined above, we formulated three hypotheses: H1 LSD induces region-specific shifts in oscillatory peak frequencies; H2 LSD induces genuine attenuation of oscillatory power independent of peak-frequency shifts; H3 music potentiates LSD-induced neural effects by amplifying changes in oscillatory organization and temporal complexity. Materials and Methods Data In this study, we analyzed MEG data from 17 participants collected by Carhart-Harris and colleagues<sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 </sup . The experimental setup followed a within-subject single-blind design, with the order counterbalanced. MEG recordings took place on two different days during which participants under the influence of either intravenous LSD 75 μg or Placebo saline , rested, or performed passive tasks. In this study, we used a subset of the conditions during which the participants either had their eyes-closed NoMusic condition , or eyes-closed with music Music condition . The duration of each recording was 7 minutes. For the music condition, participants listened to two seven-minute-long excerpts from the electro-acoustic ambient album Yearning by Robert Rich and Lisa Moskow. LSD and PLA sessions were 14 days apart to prevent the drug crossover effect. Details about the criteria for the recruitment of the participants are outlined in the original data study<sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 </sup . Approximately 4 h post-injection, MEG was recorded using a CTF-275 system at 1200 Hz 0–300 Hz bandpass , with four excessively noisy sensors turned off. Following each MEG run, participants completed in-scanner subjective rating scales assessing several phenomenological dimensions, such as mood and imagery see supplementary Fig. S1 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 for a complete list and raw scores . Preprocessing Our preprocessing pipeline resembles the approach used in the original study, but incorporates additional automated procedures to improve artifact removal implemented using the Autoreject package<sup 48 https://www.nature.com/articles/s42003-026-10629-7 ref-CR48 </sup Fig. 1 https://www.nature.com/articles/s42003-026-10629-7 Fig1 . The pipeline consists of four steps: 1 filtering, visual inspection, and epoching; 2 first pass of Autoreject; 3 ICA-based artefactual components removal; 4 second pass of Autoreject. In particular, the continuous MEG recordings were bandpass filtered between 1 and 120Hz implemented in MNE v1.8’s Raw.filter , and notch-filtered at 50 and 100 Hz implemented in MNE’s Raw.notch filter . As with the original study, we segmented the continuous MEG signal into 2s epochs 210 epochs per run <sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 </sup . This enabled us to isolate and systematically remove segments containing artifacts. Large, clearly identifiable artifacts were removed through visual inspection. The majority of the epochs passed our visual inspection Mean epochs per subject: 205.5; SD: 9.3 , and epoch retention was rather consistent across the LSD and PLA groups paired t -test; t = −1.61; p < 0.12 . The first pass of Autoreject was run in mask-only mode, i.e., Autoreject was employed to identify and label the bad epochs, which were then excluded prior to ICA. ICA decomposition Picard algorithm, see ref. <sup 49 https://www.nature.com/articles/s42003-026-10629-7 ref-CR49 </sup enabled us to identify and exclude the artefactual components. The remaining components were back-projected to reconstruct the signal across all epochs, including those initially marked as bad or interpolated. To further attenuate residual artifacts, a second pass of Autoreject was applied to the reconstructed data. Epochs labeled as good or interpolated after this step were retained as clean data. The final cleaned dataset contained approximately 200 epochs per run on average SD = 20 , again with no significant difference in epoch retention between LSD and placebo conditions t = − 1.69, p < 0.12 . Source localization Structural T1-weighted MRI scans were available for each participant in both LSD and PLA sessions<sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 </sup . These anatomical images were processed using FreeSurfer’s standard reconstruction pipeline recon-all, v7.4.1 , which performs motion correction, intensity normalization, skull stripping, and white-gray matter surface segmentation. All reconstructions were visually inspected for accuracy prior to source analysis. These subject-specific cortical surfaces were used to estimate the cortical sources for the MEG signal. Source localization requires both a forward model and an inverse solver. For the forward model, we generated Boundary Element Model BEM surfaces using FreeSurfer’s watershed algorithm implemented in MNE . Because MEG is relatively insensitive to conductivity differences across head tissues<sup 19 https://www.nature.com/articles/s42003-026-10629-7 ref-CR19 </sup , we employed the standard single-layer inner-skull, conductivity = 0.3 BEM model commonly used in MEG source estimation. Individual source spaces were defined on the ico5 spacing 20,484 vertices . MEG sensor positions were co-registered to anatomy using three fiducials nasion, left and right preauricular points , yielding a 4 × 4 affine transformation matrix mapping MEG coordinates to each participant’s MRI space. The inverse solution requires specification of the noise covariance matrix. Because empty-room noise recordings were not available for this dataset, we adopted the commonly accepted alternative of using an identity covariance matrix, which assumes uniform noise across sensors. We set the orientation constraints to be mostly perpendicular to the cortex MNE’s loose = 0.2 . Source estimation was performed using the dynamic Statistical Parametric Mapping dSPM inverse operator<sup 50 https://www.nature.com/articles/s42003-026-10629-7 ref-CR50 </sup , yielding time series at each cortical vertex for each epoch. For group-level analyses, individual source estimates were morphed source space morphing to the FreeSurfer fsaverage v5 template. The morphed source time courses were then parcellated onto the HCP-MMP atlas<sup 51 https://www.nature.com/articles/s42003-026-10629-7 ref-CR51 </sup , implemented in MNE’s extract label time course. In particular, we extracted the ROI’s activity by averaging over the corresponding fsaverage vertices, resulting in N = 360 regional time series per participant per condition. These source-projected signals were used to compute temporal metrics, such as LZC and HFD. Power spectral density PSD estimates were obtained at the source level 20484 vertices using the multitaper method implemented in MNE’s compute psd epochs, followed by morphing and parcellation to yield PSDs for each cortical region in the HCP-MMP atlas. Spectral parametrization Electrophysiological signals contain both periodic oscillatory and aperiodic 1/ f -like components<sup 20 https://www.nature.com/articles/s42003-026-10629-7 ref-CR20 </sup . Because these components jointly contribute to the power spectrum, analyzing raw, non-parametrized spectra can confound the interpretation. To disentangle these contributions, we parameterize each power spectrum using the Fitting Oscillations and One-Over-f toolbox<sup 20 https://www.nature.com/articles/s42003-026-10629-7 ref-CR20 </sup . To this end, we fit spectral parametrization models to the epoch-averaged spectra over the frequency range 1–120 Hz using the fixed aperiodic mode, with the maximum number of peaks set to 5. Prior to fitting, we applied a narrow one-dimensional interpolation around the powerline frequencies and the harmonics 50 Hz and 100 Hz; 3 Hz notch width to prevent sharp notch artifacts from distorting the parametrization. Model performance and goodness-of-fit metrics are reported in the Supplementary Materials https://www.nature.com/articles/s42003-026-10629-7 MOESM1 Supplementary Fig. S2 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . The spectral parametrization model decomposes each spectrum into a set of Gaussian peaks representing oscillatory activity and an aperiodic component, characterized by the offset and the slope, which corresponds to the exponent of the 1/ f -like distribution of the power spectrum. We extracted these parameters and tested for the perturbations induced by LSD and by LSD x music interaction. In addition, we examined the pure oscillatory component and the band peaks as identified by the spectral parametrization model. The oscillatory activity was robustly present in alpha and beta bands, and notably present in the high-gamma band Supplementary Fig. S3 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . Consequently, in this study, we focused on alpha, beta and high-gamma activity. We computed the oscillatory power without 1/ f component in the canonical frequency range results presented in Supplementary Fig. S4 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . However, we observed shifts in the alpha and beta peaks that may confound the band power analysis. Thus, as a spectral feature, we also considered the Peak Power, defined as the maximum pure oscillatory activity in the specified band. Also, to assess the robustness of our parametrization pipeline, we also performed the analysis by fitting the model for a narrower range of 1–45 Hz. The corresponding results can be found in the supplementary material https://www.nature.com/articles/s42003-026-10629-7 MOESM1 Supplementary Fig. S2 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . Complexity Lempel-Ziv complexity quantifies the compressibility of the EEG or MEG signals, i.e., how diverse the patterns are in the signal<sup 52 https://www.nature.com/articles/s42003-026-10629-7 ref-CR52 </sup . While it is not an entropy measure per se, increases in LZC are commonly interpreted as reflecting more diverse and less constrained neural dynamics, often associated with higher entropic brain states<sup 53 https://www.nature.com/articles/s42003-026-10629-7 ref-CR53 </sup . Computing LZC involves two steps: First, the continuous time series is binarized, typically using the mean or median as a threshold such that values greater than the threshold are assigned “1” and all others “0”. Second, the number of distinct patterns in the resulting binary sequence is counted, yielding the LZC value. Lower LZC indicates more regular, predictable activity, whereas higher LZC reflects greater temporal diversity. LZC has been successfully applied to electrophysiological data, especially with the altered states of consciousness, including the recent work investigating the effects of external stimulation under LSD<sup 41 https://www.nature.com/articles/s42003-026-10629-7 ref-CR41 </sup . Although Mediano, Rosas et al.<sup 41 https://www.nature.com/articles/s42003-026-10629-7 ref-CR41 </sup examined the effects of LSD on brain complexity across multiple task contexts, we extend this work in two important ways. First, we specifically quantify LZC in the context of interaction between LSD and music, directly probing how sensory context modulates signal richness under psychedelics. Second, we compute LZC using our own preprocessing pipeline, which is appreciated for replicability. In this work, we computed LZC for a broadband signal 1–120 Hz, main text ; b bandlimited signals within each canonical frequency band alpha through high-gamma; Fig. S5 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . LZC was computed by concatenating all artifact-free epochs. The binarization threshold was set to the average of the signal. Because the duration of the concatenated signals varied slightly across participants after preprocessing, we normalized<sup 54 https://www.nature.com/articles/s42003-026-10629-7 ref-CR54 </sup the LZC by ${{{{\rm{LZC}}}}} {{{{\rm{norm}}}}} X =\frac{\log n }{n}\times {{{{\rm{LZC}}}}} {{{{\rm{raw}}}}} X $, where X denotes the signal, and n its length. Fractality Fractal dimension quantifies scale-invariant temporal structure in neural signals. Here, we estimate fractal dimension using HFD<sup 55 https://www.nature.com/articles/s42003-026-10629-7 ref-CR55 </sup , which captures self-similarity across time scales in EEG and MEG data, capturing scale-free temporal structure<sup 56 https://www.nature.com/articles/s42003-026-10629-7 ref-CR56 , 57 https://www.nature.com/articles/s42003-026-10629-7 ref-CR57 </sup . It complements Lempel-Ziv complexity by characterizing how signal complexity is organized across temporal scales, rather than how diverse the signal patterns are over time. The HFD algorithm estimates how the “length” of a signal changes as a function of the scale at which the signal is sampled. HFD computation involves a 4-step process. First, for each integer $k=1,\ldots ,{K} {\max }$, the original time series X 1 , X 2 , …, X N is used to construct k subsampled sequences each beginning at offset m and sampled at interval k . Second, for each sequence ${X} {m}^{ k }$, a curve length L <sub m </sub k is computed using the normalized sum of absolute successive differences, and the mean curve length across all offsets is Third, the set of values { L k } is examined in log–log space, where fractal signals exhibit a systematic decrease in L k with increasing k. Finally, the fractal dimension is obtained as the slope D of the linear regression of $\log L k $ versus $\log 1/k $. HFD ranges between 1 and 2, increasing asymptotically and saturating at some ${K} {max}^{optimal}$, a hyperparameter e.g., ref. <sup 58 https://www.nature.com/articles/s42003-026-10629-7 ref-CR58 </sup . We computed the HFD across ROIs, subjects, drug and task conditions as a function of ${K} {\max }\in$ 2, 50 . To select the ${K} {max}^{optimal}$, we examined the HFD-${K} {\max }$ curve and identified the saddle point by applying a knee-detection algorithm. The resulting knee index yielded an optimal value of ${K} {max}^{optimal}$ = 18, which we adopted for all the HFD computations. We computed the HFDs for each epoch, and subsequently averaged over epochs within each recording. All fractal-dimension estimates were obtained using the higuchi fd function from the antropy Python package. Functional decoding The LSD-induced neurophysiological changes were mapped onto the large-scale cognitive and affective systems using Functional Decoding, with NiMare<sup 59 https://www.nature.com/articles/s42003-026-10629-7 ref-CR59 </sup , following previous studies<sup 60 ref-CR60 , 61 ref-CR61 , 62 https://www.nature.com/articles/s42003-026-10629-7 ref-CR62 </sup . The aim of this analysis is to identify which cognitive or affective processes are most strongly associated with the spatial distribution of LSD-induced effects. Functional decoding relies on the Neurosynth database, a large-scale, automated meta-analytic resource that aggregates activation coordinates from more than 14,000 neuroimaging studies over 500,000 statistical maps . Neurosynth organizes the literature into topics<sup 63 https://www.nature.com/articles/s42003-026-10629-7 ref-CR63 </sup , which are latent semantic components derived from the co-occurrence of terms in article abstracts and their associated activation patterns. Each topic, therefore, represents a broad cognitive or affective construct e.g., visual perception, autobiographical memory, emotion, language rather than a single experimental task or region. For each electrophysiological feature, the unthresholded statistical contrast map LSD and PLA was rank-ordered and segmented into 10 equally sized bins 10% increments . Each decile mask was binarized and submitted to the ROIAssociationDecoder in NiMARE, which computes the spatial correspondence between the input masks and meta-analytic activation maps see ref. <sup 63 https://www.nature.com/articles/s42003-026-10629-7 ref-CR63 </sup for details . The decoder returns correlation coefficients reflecting the degree to which that spatial pattern is associated with each Neurosynth topic. These correlations were transformed into z-statistics, and only associations surviving a threshold of p < 0.001 were retained. Following<sup 60 https://www.nature.com/articles/s42003-026-10629-7 ref-CR60 </sup , topic terms were ordered based on the weighted average of each topic’s z-scores across the ten deciles. This ranking highlights the cognitive and affective domains that best characterize the regions showing the greatest LSD-induced electrophysiological changes. Phenomenology regression analysis We investigated how LSD-induced neurophysiological changes relate to participants’ subjective experience by regressing the changes in the neurophysiological features against the subjective ratings. We quantified the relationship between the two variables using Spearman’s rank correlation ρ , a nonparametric, rank-based measure that captures their monotonic relationship. The resulting strength and direction are captured by the correlation coefficient ρ , which ranges from -1 to 1, indicating a negative or positive relationship between the two variables. Analyses were conducted at two spatial scales: i global scale i.e., region-averaged ; ii region-resolved scale, which enables detection of localized brain-experience relationships that would be obscured in global averages. Given the limited sample size and multiple comparisons inherent to region-resolved analyses, these regressions should be interpreted as exploratory. The corresponding results can be found in the Supplementary Figs. S6 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 and S7 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . Statistics and reproducibility For each feature, we contrasted LSD and PLA using paired statistical tests with a multiple-comparisons correction. In particular, for the global spectral analyses see Fig. 3A https://www.nature.com/articles/s42003-026-10629-7 Fig3 , we performed non-parametric cluster-level tests implemented in MNE’s permutation cluster 1samp test , in which the temporal adjacency was defined along the frequency points Hz . Clusters were formed by grouping consecutive frequency bins exhibiting similar test-statistic values both in sign and magnitude , using a cluster-forming threshold corresponding to a p < 0.05 see ref. <sup 64 https://www.nature.com/articles/s42003-026-10629-7 ref-CR64 </sup for details . For all other features, we employed non-parametric permutation-corrected tests<sup 64 https://www.nature.com/articles/s42003-026-10629-7 ref-CR64 </sup implemented in MNE’s permutation t test . Empirical test statistics were computed from paired LSD-PLA contrasts and evaluated against a null distribution obtained by permuting the regions N = 50,000 permutations . For each region, the p -value corresponded to the percentile rank of the observed statistic T <sub o b s </sub within this surrogate distribution, following the procedure outlined in ref. <sup 64 https://www.nature.com/articles/s42003-026-10629-7 ref-CR64 </sup . Statistical significance was defined at p < 0.05. In cases where the LSD-induced perturbations were broad and spatially extensive throughout the cortex, we additionally highlighted regions showing the strongest effects by marking the strongest decile of the resulting statistical distribution e.g., t-values . For the regression analyses linking neural features to subjective experience, we controlled for multiple comparisons using false discovery rate FDR<sup 65 https://www.nature.com/articles/s42003-026-10629-7 ref-CR65 </sup correction at p < 0.05. We show in the supplementary material https://www.nature.com/articles/s42003-026-10629-7 MOESM1 for the comparisons Supplementary Figs. S6 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 and S7 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . In other scenarios, such as the interaction effect of LSD and music, we report that the effects are nominally significant i.e., p < 0.05 uncorrected , reflected in all features. The Main Effect Drug quantifies the effect of LSD regardless of the task condition, whereas Main Effect Music captures the effect of Music irrespective of the drug state. The Interaction effect LSD x Music tests whether music modulates the way LSD affects the brain. Finally, to further decompose the LSD x Music interaction, we examined the Drug effect under music Simple Effect Music and without music Simple Effect NoMusic to elucidate how it manifests. Machine learning We used a Random Forest classifier, an ensemble tree-based ML model, to decode LSD versus PLA states from the range of temporal and spectral features oscillatory power, peak frequencies, LZC, and HFD . We chose a Random Forest classifier as it can capture potential non-linear relationships and interactions across heterogeneous feature domains, and because it generally provides robust performance without requiring assumptions about linear separability or feature independence. We extract features important for discriminating between the two states, termed Feature Importances. We conducted two complementary analyses by building single-domain and multi-domain models. In the single-domain Model, each classifier operated on one feature family at a time e.g., alpha Peak Power, beta peak frequency, exponent, LZC , using the set of regional values as model inputs. In this formulation, the ML “features” correspond to cortical regions, and the resulting Feature Importances identify the regions whose values most strongly discriminate LSD from PLA. This notion is analogous to classical statistics, where the region is significantly different when the LSD and PLA distributions are not invariant. In contrast, the multi-domain model pooled temporal and spectral features across regions into a single classifier, enabling us to determine which feature families and which regions within them carry the greatest predictive weight overall. Such multi-domain decoding models are increasingly used to investigate altered states of consciousness—including those induced by caffeine during sleep<sup 45 https://www.nature.com/articles/s42003-026-10629-7 ref-CR45 </sup and meditation<sup 66 https://www.nature.com/articles/s42003-026-10629-7 ref-CR66 </sup —and offer a complementary, more holistic perspective on which neural markers most reliably distinguish drug states. Training procedures were identical for both single-domain and multi-domain models. Model hyperparameters were optimized via nested StratifiedGroupKFold cross-validation outer K=4, inner K=3 with grid search, where the inner loop handled hyperparameter selection, keeping outer test folds unseen and yielding an unbiased performance estimate. To assess significance, we constructed surrogate null distributions for both accuracy and Feature Importances by permuting the target labels 1000 times. For each permutation, the full StratifiedGroupKFold training procedure was repeated, and surrogate accuracies and surrogate Feature Importances were obtained by averaging across folds. Empirical values were then compared against their respective null distributions to obtain permutation-based p-values, which quantify the probability that feature-label associations arise under the assumption of independence. Feature Importances surviving p < 0.05 thus reflect permutation-corrected, multiple-comparison-controlled indices of discriminability between LSD and PLA. In addition to reporting accuracy and feature importances in the main text, we also report the AUC-ROC plots in the supplementary Supplementary Fig. S8 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary https://www.nature.com/articles/s42003-026-10629-7 MOESM2 linked to this article. Results In what follows, we report four complementary sets of findings aligned with our three research questions. First, we characterize how LSD induces region-specific shifts in alpha and beta peak frequencies Q1 . Second, we isolate genuine changes in oscillatory power after accounting for these peak displacements, allowing us to dissociate true dampening from apparent reductions driven by frequency shifts Q2 . Third, to contextualize these oscillatory changes at the systems level, we examine how these neurophysiological effects are organized across large-scale functional systems using meta-analytic functional decoding. Finally, we assess how these neurophysiological effects are modulated by music, addressing the interaction between pharmacological and contextual influences Q3 . LSD induces genuine attenuation of alpha and beta oscillations We examined the main effect of LSD on oscillatory power by parameterizing cortical power spectra and isolating the periodic component Fig. 2 https://www.nature.com/articles/s42003-026-10629-7 Fig2 . At the whole-brain level, power was first averaged across all cortical regions. In the total spectra Fig. 2A https://www.nature.com/articles/s42003-026-10629-7 Fig2 , top , LSD was associated with a broad reduction in low-frequency power and a relative increase at higher frequencies; however, these differences did not survive cluster-based multiple-comparison correction p < 0.05 . After removing the aperiodic 1/ f component Fig. 2A https://www.nature.com/articles/s42003-026-10629-7 Fig2 , second row , a clearer pattern emerged: LSD robustly weakened oscillatory activity from theta through beta bands up to 30 Hz , while mid-gamma power 60-90 Hz showed a significant increase p < 0.05; cluster-corrected . These region-averaged spectra summarize the global spectral effects of LSD; we next examined their spatial distribution across cortical regions. Using spectral parametrization to probe the presence of genuine oscillatory components, we observed that oscillatory activity is most consistently expressed across cortical regions in the alpha and beta bands, with reliable oscillatory structure also present in the high-gamma range see Supplementary Fig. S3 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . Accordingly, subsequent analyses focus on these frequency bands. Permutation-based statistical contrasts LSD vs. PLA; 50,000 permutations revealed spatially specific effects across the cortex Fig. 2 https://www.nature.com/articles/s42003-026-10629-7 Fig2 B-D , with negative t -values blue indicating power suppression and positive values red indicating power increases. In the total power spectra Fig. 2B–D https://www.nature.com/articles/s42003-026-10629-7 Fig2 , first row , LSD predominantly reduced alpha and beta power while producing focal increases in high-gamma activity. To isolate genuine oscillatory effects, we focused on power at oscillatory peaks Peak Power , rather than power within canonical frequency bands, as LSD induced systematic shifts in alpha and beta peak frequencies. Peak-based contrasts permutation-corrected, p < 0.05; Fig. 2B, C https://www.nature.com/articles/s42003-026-10629-7 Fig2 , second row revealed frequency-specific spatial patterns. Alpha Peak Power reductions were most prominent in temporal regions, whereas beta Peak Power showed widespread suppression across the cortex. In the high-gamma range, oscillatory power was reduced primarily in posterior regions. Across all three bands, analyses no longer showed significant effects in the posterior cingulate cortex, which had appeared prominently in total power contrasts, suggesting that earlier effects were driven by changes in the aperiodic component. Analyses of band-limited power without accounting for peak shifts yielded more spatially extensive suppression Supplementary Fig. S4 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 , indicating that ignoring peak-frequency shifts inflates estimates of drug effects. To complement inferential statistics, we applied supervised machine learning to assess the discriminability of LSD and PLA states. Employing machine learning here aims to improve the robustness of our findings and provide more interpretability to inferential statistics. Random Forest classifiers were trained using regional spectral power features alpha and beta Peak Power; high-gamma band-limited power to predict drug condition. Classifiers achieved robust decoding performance accuracy 68%; Supp. figure S8 for AUC-ROC plots , indicating a good to strong degree of separability between LSD and placebo states. Feature importance analysis revealed that the regions contributing most strongly to classification closely matched those identified by permutation-based t -tests Fig. 2B–D https://www.nature.com/articles/s42003-026-10629-7 Fig2 , third row . The statistical significance of feature importances was assessed against null distributions generated via label permutation 1000 iterations , with regions exceeding the 95th percentile deemed significant p < 0.05 . The convergence between ML-derived importance maps and classical statistical results across frequency bands supports the robustness and spatial specificity of LSD-induced oscillatory changes. LSD robustly shifts peaks, flattens slope and increases fractality Figure 3 https://www.nature.com/articles/s42003-026-10629-7 Fig3 summarizes the main effects of LSD on additional spectral and temporal features. Figure 3 https://www.nature.com/articles/s42003-026-10629-7 Fig3 A shows paired permutation-based statistical contrasts LSD vs. PLA; p < 0.05, corrected; 50,000 permutations , and Fig. 3 https://www.nature.com/articles/s42003-026-10629-7 Fig3 B shows feature-importance maps derived from Random Forest classifiers trained separately on each feature to decode drug condition. Statistical significance of feature importances was assessed using label permutation N = 1 000, p < 0.05 . Consistent with the global spectra Fig. 2 https://www.nature.com/articles/s42003-026-10629-7 Fig2 A , LSD induced systematic shifts in oscillatory peak frequencies at the regional level. With the aim of providing clarity on interpretation and on visualization, the first two rows of Fig. 3 https://www.nature.com/articles/s42003-026-10629-7 Fig3 A display the grand-averaged peak shift for alpha and beta bands. This is defined as the difference between the frequency of the LSD peak and the PLA peak in the given band. We showed only the regions that were significant p < 0.05, permutation-corrected when we compared the LSD peaks to the PLA ones. On average, alpha peaks were shifted upward by approximately 2 Hz and beta peaks by approximately 5 Hz permutation-corrected, p < 0.05 . Random Forest classifiers trained on peak-frequency features achieved high decoding accuracy accuracy 82%; Supplementary Fig.S 8 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 for AUC-ROC plots , and the corresponding importance maps closely resembled the spatial patterns observed in the statistical contrasts. LSD also produced pronounced changes in the aperiodic component of the spectrum. Both the 1/ f exponent slope and offset showed significant, brain-wide effects Fig. 3A https://www.nature.com/articles/s42003-026-10629-7 Fig3 , third and fourth rows , characterized by slope flattening and reduced offset. The strongest effects were localized to posterior regions, including the posterior cingulate cortex and visual cortices, as identified by the highest decile of effect magnitudes p < 10<sup −3</sup , magenta . The ML decoding analysis revealed that feature importance generally matched the regions with the strongest drug effect. Temporal complexity measures revealed a complementary pattern. Both LZC and HFD increased across the cortex under LSD Fig. 3A https://www.nature.com/articles/s42003-026-10629-7 Fig3 , final two rows , with maximal effects again observed in posterior regions p < 10<sup −5</sup , yellow . Decoding models trained on LZC and HFD achieved accuracies 76% Supp. figure S8 for AUC-ROC plots , and their feature-importance maps overlapped with regions showing the strongest statistical effects. Finally, panel Fig. 3C https://www.nature.com/articles/s42003-026-10629-7 Fig3 extends these analyses using a multidomain Random Forest model that jointly incorporates all temporal and spectral features across cortical regions. Permutation-corrected feature importances revealed that complexity metrics, aperiodic parameters, and spectral power features contributed most strongly to decoding LSD versus PLA. This multidomain ML analysis confirms that LSD-related neural signatures are distributed across multiple feature domains, complementing the single-domain statistical results in Fig. 3A, B https://www.nature.com/articles/s42003-026-10629-7 Fig3 . Functional decoding reveals differentiated effects across cognitive and affective domains To relate LSD-induced electrophysiological changes to cognitive function, we performed functional decoding of each neural feature using NiMARE meta-analytic topic maps Fig. 4 https://www.nature.com/articles/s42003-026-10629-7 Fig4 . Unthresholded LSD-PLA statistical maps were binned into deciles, and each decile was compared against NiMARE topic activation maps to identify which cognitive and affective systems are most affected. Each row of the heatmaps corresponds to a feature family: oscillatory Peak Power A, B , peak-frequency shifts C, D , aperiodic parameters E, F , and temporal complexity measures G, H . The ordering of topic terms reflects their weighted average across deciles see Section , and axis directionality is indicated in the figure. The strongest attenuation of alpha Peak Power occurred in auditory, language, and emotion-related systems. Motor and attention regions showed the weakest attenuation, with memory retrieval and imagery networks exhibiting intermediate effects. Domains showing the largest alpha power reductions also exhibited the minimal alpha peak-frequency shifts, while domains with relatively preserved alpha Peak Power motor, attention showed larger peak shifts. Consistent with this dissociation, no significant relationship was observed between the magnitude of alpha peak-power reduction and peak-frequency shift Spearman’s ρ , p 0.05 , indicating that these effects follow distinct functional hierarchies rather than reflecting a simple inverse relationship. Beta Peak-Power attenuation displayed a comparable organization, with the strongest reductions in auditory, emotional, and language domains and weaker effects in visual, attentional, motor, and imagery systems. In contrast to the alpha band, beta peak-frequency shifts were largest in domains showing strong beta power attenuation e.g., auditory and language systems , whereas domains with relatively preserved beta power exhibited smaller frequency shifts. Aperiodic parameters exhibited a complementary system-level organization. Flattening of 1/ f slope and the corresponding changes in offset were most pronounced in primary sensory areas visual and auditory , language-related domains, and higher-order imagery networks, while motor systems showed comparatively minimal effects. Changes in temporal complexity mirrored this pattern: increases in HFD and LZC were largest in mental imagery, visual, auditory, and memory-related domains, and smallest in motor systems. Across domains, systems showing the strongest aperiodic reconfiguration also exhibited the largest increases in temporal complexity. Exploratory analysis of LSD × Music interaction We investigated the interaction between LSD and music using a 2 × 2 factorial design. While no regions survived whole-brain FDR correction, we observed nominally significant interaction effects p<0.05, uncorrected . These trends point to a negative interaction, where the presence of music appears to attenuate the typical neural signature of LSD Fig. 5 https://www.nature.com/articles/s42003-026-10629-7 Fig5 . Specifically, we observed a steepening of the aperiodic slope and increased broadband power—effects that run counter to the spectral flattening typically induced by psychedelics. Similarly, signal complexity metrics LZC and HFD indicated that the combination of LSD and music resulted in reduced complexity and fractality relative to LSD in silence, suggesting that auditory stimuli may constrain the effects of the drug. Discussion Psychedelics inducing desynchronization in alpha and beta is a hallmark effect that numerous studies have reported about<sup 12 ref-CR12 , 13 ref-CR13 , 14 https://www.nature.com/articles/s42003-026-10629-7 ref-CR14 </sup . A limited number of studies have decomposed the power spectrum in its periodic and aperiodic components to show that psychedelics do indeed weaken the oscillatory activity<sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 , 22 https://www.nature.com/articles/s42003-026-10629-7 ref-CR22 </sup . Evidence for LSD-related increases in alpha and beta frequencies has been noted in the PSDs across few studies<sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 , 22 https://www.nature.com/articles/s42003-026-10629-7 ref-CR22 </sup , but these effects have not yet been systematically quantified or spatially characterized. If LSD speeds up alpha and beta activity, interpreting the weakening of alpha and beta oscillations as desynchronization risks misinterpretation, as systematic shifts without significant power weakening within the canonical frequency range will masquerade as desynchronization. By focusing on the pure oscillatory mode, we showed robust shifts in alpha and beta across brain regions. By analyzing changes in alpha and beta power defined using oscillatory peaks while accounting for peak shifts, we have evidence suggesting that LSD induces a genuine weakening of alpha and beta power. This study overall shows LSD induces both weakening of power and at the same time increases alpha and beta peak frequencies. In addition, we investigated the changes in spectral and temporal features as a result of the interaction between LSD and music, building on the previous studies<sup 41 https://www.nature.com/articles/s42003-026-10629-7 ref-CR41 </sup . Our analyses did not identify a statistically significant interaction after correction; however, nominal trends in the exponent, offset, HFD, and LZC point to a relative attenuation of these neural effects during music listening. LSD elicits localized reductions in alpha and beta oscillation amplitudes Current neuropharmacological consensus holds that the hallucinogenic effects of psychedelics are primarily mediated by serotonergic signaling see refs. <sup 9 https://www.nature.com/articles/s42003-026-10629-7 ref-CR9 , 67 https://www.nature.com/articles/s42003-026-10629-7 ref-CR67 </sup for review , acting predominantly—though not exclusively—through 5-HT2 A receptors<sup 8 https://www.nature.com/articles/s42003-026-10629-7 ref-CR8 </sup . This view is supported by studies showing that pretreatment with the 5-HT2 A antagonist Ketanserin blocks the subjective and neural effects of these drugs<sup 68 https://www.nature.com/articles/s42003-026-10629-7 ref-CR68 </sup . Although the effects of psychedelics on neural circuitry are still being investigated, converging evidence suggests that the 5-HT2 A agonism induces an asynchronous mode of glutamate release<sup 69 https://www.nature.com/articles/s42003-026-10629-7 ref-CR69 , 70 https://www.nature.com/articles/s42003-026-10629-7 ref-CR70 </sup . This mechanism appears to disrupt the temporal synchrony of neural populations, leading to widespread dysregulation of network dynamics. This receptor-mediated disruption is consistent with the regionally selective broadband desynchronization observed in our data, particularly in cortical regions with high 5-HT2 A receptor density. Such broadband desynchronization is among the most consistently reported electrophysiological signatures of various psychedelics<sup 14 https://www.nature.com/articles/s42003-026-10629-7 ref-CR14 , 71 https://www.nature.com/articles/s42003-026-10629-7 ref-CR71 </sup . Across serotonergic psychedelics, attenuation of alpha and beta oscillatory power is a consistent finding, reported for N,N-DMT <sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 , 17 https://www.nature.com/articles/s42003-026-10629-7 ref-CR17 , 72 https://www.nature.com/articles/s42003-026-10629-7 ref-CR72 </sup , 5-MeO-DMT <sup 73 https://www.nature.com/articles/s42003-026-10629-7 ref-CR73 </sup , and psilocybin <sup 14 https://www.nature.com/articles/s42003-026-10629-7 ref-CR14 </sup . In N,N-DMT studies, increases in low-frequency activity delta/theta are not consistently observed and depend on analytical approaches, such as the separation of oscillatory and aperiodic components <sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 </sup . Consistent with this literature, LSD induces genuine attenuation of oscillatory power. However, we also observe robust increases in alpha and beta peak frequencies, indicating spectral acceleration rather than the replacement of faster rhythms by slower ones. These effects are partly dissociable in space, suggesting distinct but co-occurring mechanisms of cortical reorganization. By comparison, ketamine shows a different electrophysiological profile, with increased gamma activity and reduced lower-frequency coordination <sup 74 https://www.nature.com/articles/s42003-026-10629-7 ref-CR74 </sup . Overall, while reduced oscillatory power may be a shared feature across altered states, LSD is distinguished by a systematic upward shift in alpha and beta peak frequencies. Cross-drug comparisons should nevertheless be interpreted cautiously, given differences in modality and analytical methods. Importantly, electrophysiological recordings reflect a mixture of rhythmic synchronized and arrhythmic background activity rather than purely oscillatory processes<sup 75 https://www.nature.com/articles/s42003-026-10629-7 ref-CR75 </sup . The latter typically follows a scale-free, 1/ f -like distribution<sup 76 https://www.nature.com/articles/s42003-026-10629-7 ref-CR76 </sup . Analyzing raw spectra without separating periodic and aperiodic components risks misinterpretation, because apparent power reductions can arise either from genuine oscillatory desynchronization or from changes in the arrhythmic background that lower total power without altering synchrony. Therefore, distinguishing true oscillatory desynchronization requires explicit parameterization of the spectrum<sup 20 https://www.nature.com/articles/s42003-026-10629-7 ref-CR20 </sup . Recent analytical frameworks have been developed to parameterize and disentangle oscillatory peaks from the aperiodic background<sup 20 https://www.nature.com/articles/s42003-026-10629-7 ref-CR20 , 77 https://www.nature.com/articles/s42003-026-10629-7 ref-CR77 </sup . Studies adopting this approach reported an LSD-induced weakening of oscillatory activity<sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 , 22 https://www.nature.com/articles/s42003-026-10629-7 ref-CR22 </sup . Applying this approach, we show that LSD-induced oscillatory weakening is neither uniform across frequencies nor spatially homogeneous. At the global level Fig. 2A https://www.nature.com/articles/s42003-026-10629-7 Fig2 , spectra replicate the well-established pattern of low-frequency desynchronization both before and after correction for the aperiodic component, attesting to the robustness of our preprocessing pipeline. In raw non-parametrized spectra at the regional level, however, relatively few regions exhibit significant LSD-PLA differences. Spectral parameterization reveals that the true oscillatory components are present primarily in alpha and beta and additionally in high-gamma Supplementary Fig. S3 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . Accounting for LSD-induced shifts in oscillatory peak frequencies is critical for interpreting changes in power. Although peak frequency shifts under LSD have been noted previously<sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 , 22 https://www.nature.com/articles/s42003-026-10629-7 ref-CR22 </sup , their implications for power analyses have not been examined in detail. Analyzing power within canonical frequency ranges may mischaracterize oscillatory amplitude when peaks shift; we therefore propose measuring power directly at oscillatory peaks "Peak Power” . This peak-based analysis Fig. 2B, C https://www.nature.com/articles/s42003-026-10629-7 Fig2 , second row revealed clear, spatially extended reductions in alpha and beta oscillatory amplitude under LSD. In high-gamma, raw spectra suggested increased power under LSD, whereas parameterized spectra revealed focal power suppression, particularly in posterior regions Fig. 2D https://www.nature.com/articles/s42003-026-10629-7 Fig2 , first and second row . This discrepancy is consistent with a flattening of the aperiodic slope under LSD, which shifts raw spectral power toward higher frequencies without reflecting increased oscillatory activity. A striking illustration of this dissociation was observed in the posterior cingulate cortex PCC , a core hub of the default mode network with disproportionately high 5-HT2 A receptor expression<sup 11 https://www.nature.com/articles/s42003-026-10629-7 ref-CR11 </sup . In raw spectra, the PCC exhibited robust LSD-PLA differences across frequency ranges. These differences disappeared after removal of the aperiodic component, indicating that they were driven primarily by changes in the 1/ f background rather than by oscillatory power per se. Across the cortex, LSD consistently flattened the aperiodic slope, with the strongest effect p < 10<sup −5</sup localized to the PCC. This finding underscores that apparent power reductions in key hub regions may reflect shifts in background spectral properties, rather than true oscillatory desynchronization. In this sense, spectral parametrization does more than merely correct for confounds; it also facilitates the interpretation of psychedelic effects in physiologically meaningful terms. One possible interpretation of this flattening is that it reflects changes in the temporal coordination of neural population activity. Prior theoretical, computational, and empirical work suggests that the aperiodic exponent is sensitive to excitation-inhibition balance and population synchrony, with flatter spectra potentially reflecting more irregular, less synchronized population activity <sup 21 https://www.nature.com/articles/s42003-026-10629-7 ref-CR21 , 78 https://www.nature.com/articles/s42003-026-10629-7 ref-CR78 , 79 https://www.nature.com/articles/s42003-026-10629-7 ref-CR79 </sup . In the context of psychedelics, this interpretation is compatible with evidence that 5-HT2A receptor activation modulates cortical excitability and glutamatergic signaling, including asynchronous glutamate release in prefrontal pyramidal neurons <sup 14 https://www.nature.com/articles/s42003-026-10629-7 ref-CR14 , 69 https://www.nature.com/articles/s42003-026-10629-7 ref-CR69 </sup . However, the precise physiological basis of LSD-induced aperiodic changes in human MEG remains to be established. To further illustrate the consequences of ignoring peak frequency shifts, we also examined band-limited power within canonical frequency ranges after the removal of the aperiodic component Supplementary Fig. S4 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 . This conventional approach yielded more spatially extensive apparent desynchronization, particularly in alpha, compared with peak-based analyses. Such differences indicate that fixed-band analyses can inflate the apparent spatial extent of power changes when peak frequencies shift, reinforcing the need for caution when interpreting band-limited power under LSD. The frequency specificity of LSD-induced power suppression aligns with known neurophysiological principles. Low-frequency oscillations reflect large-scale, spatially distributed synchronization, whereas higher-frequency rhythms tend to arise from more localized circuit interactions<sup 80 https://www.nature.com/articles/s42003-026-10629-7 ref-CR80 , 81 https://www.nature.com/articles/s42003-026-10629-7 ref-CR81 </sup . A drug that perturbs global coordination, as serotonergic psychedelics do through 5-HT2 A -mediated disruption of recurrent processing, would be expected to preferentially weaken large-scale, lower-frequency rhythms. In this context, the topography and frequency specificity of oscillatory suppression observed here are consistent with mechanistic accounts of psychedelic action on cortical hierarchy. Given the close association of these rhythms with top-down signaling<sup 82 https://www.nature.com/articles/s42003-026-10629-7 ref-CR82 </sup , their suppression under LSD is consistent with the REBUS Relaxed Belief under Psychedelics framework<sup 83 https://www.nature.com/articles/s42003-026-10629-7 ref-CR83 </sup . Complementing these inferential results, machine-learning analyses provided convergent evidence for the robustness of low-frequency oscillatory suppression. Models trained on oscillatory power features achieved high discriminative accuracy between LSD and PLA, with feature importance maps highlighting posterior cortical regions that closely overlapped with those identified by classical statistics. This convergence across independent analytical approaches supports the reliability of the observed effects. Finally, high-gamma oscillatory activity exhibited a distinct, more focal pattern. Unlike alpha-beta suppression, the effects in the gamma band are not consistently reported in the literature, in part because the gamma activity is highly susceptible to muscle artifacts and therefore often excluded from the analysis. To our knowledge, this study provides the first evidence isolating the oscillatory gamma response to psychedelics. Spectral parametrization Supplementary Fig. S3 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 confirmed a considerable amount of oscillatory activity in high-gamma. We found that LSD produced a focal decrease in high-gamma, primarily in the posterior regions, including the occipital cortex. Previous studies have reported both increase<sup 84 https://www.nature.com/articles/s42003-026-10629-7 ref-CR84 </sup and a decrease in gamma activity under psychedelics<sup 14 https://www.nature.com/articles/s42003-026-10629-7 ref-CR14 </sup , but those observations were based on raw non-parametrized spectra i.e., with 1/ f component , which confounds interpretations. In contrast, our results reflect strictly oscillatory power. Notably, classifiers trained exclusively on high-gamma oscillatory power achieved robust decoding performance and exhibited spatially coherent feature-importance profiles. If these effects were driven predominantly by electromyographic contamination, one would expect diffuse or inconsistent spatial patterns, which were not observed. While caution remains warranted given the susceptibility of high-frequency activity to artifacts, the convergence of parameterized spectral and decoding results suggests that the observed high-gamma effects likely reflect condition-related neural activity. LSD increases alpha and beta peak frequencies Alpha oscillations are among the most extensively studied rhythms in human electrophysiology, dating back to Hans Berger’s earliest EEG recordings<sup 85 https://www.nature.com/articles/s42003-026-10629-7 ref-CR85 </sup . Peak alpha frequency PAF has been shown to track attentional state, arousal, and broader cognitive performance<sup 86 https://www.nature.com/articles/s42003-026-10629-7 ref-CR86 </sup . Conceptually, the alpha rhythm has been described as a “fulcrum” of global brain state transitions, reflecting the natural resonant frequency of the resting brain and the stability of cortical-thalamic loops<sup 87 https://www.nature.com/articles/s42003-026-10629-7 ref-CR87 </sup . Because alpha oscillations play a central role in sensory processing, thalamocortical communication, and the regulation of inhibition and excitation balance, even modest shifts in their peak frequency carry mechanistic significance, as they imply changes in the basic operating frequency of cortical networks. By isolating oscillatory components via spectral parameterization, we show that LSD induces consistent upward shifts in alpha and beta peak frequencies across widespread cortical regions Figs. 2 https://www.nature.com/articles/s42003-026-10629-7 Fig2 and 3 https://www.nature.com/articles/s42003-026-10629-7 Fig3 . This effect differs from those of anesthesia, which reliably slows alpha rhythms, a pattern associated with reduced cortical excitability and strengthened top-down constraints<sup 88 https://www.nature.com/articles/s42003-026-10629-7 ref-CR88 </sup . Psychedelics, by contrast, generally relax high-level constraints while increasing cortical excitability, and the observed speeding of alpha and beta peak frequencies is consistent with this mechanistic dissociation. Notably, comparable peak-frequency modulation has not been robustly observed for psilocybin or DMT, which predominantly flattens oscillatory power without clear peak modulation<sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 , 89 https://www.nature.com/articles/s42003-026-10629-7 ref-CR89 </sup . Thus, under LSD, reductions in alpha and beta power co-occur with a shift of these rhythms toward faster operating frequencies. What neuropharmacological features might account for this effect? Several features distinguish LSD from other serotonergic psychedelics. First, LSD exhibits markedly higher affinity for 5-HT2 A receptors than psilocybin or DMT<sup 8 https://www.nature.com/articles/s42003-026-10629-7 ref-CR8 </sup . Second, crystallography and pharmacology studies reveal that LSD “shuts the lid” on 5-HT2A/2B receptors, giving an exceptionally long residence time, more than twice that of psilocybin<sup 90 https://www.nature.com/articles/s42003-026-10629-7 ref-CR90 , 91 https://www.nature.com/articles/s42003-026-10629-7 ref-CR91 </sup . Furthermore, LSD induces stronger increases in thalamocortical connectivity compared with other serotonergic drugs<sup 92 https://www.nature.com/articles/s42003-026-10629-7 ref-CR92 </sup . Given that the alpha-band dynamics depend on factors including cortico-thalamo-cortical connectivity and intracortical circuits mechanisms<sup 93 https://www.nature.com/articles/s42003-026-10629-7 ref-CR93 </sup , one plausible, though necessarily speculative, interpretation is that LSD’s unique pharmacokinetic and circuit-level effects simultaneously a disinhibit alpha and beta oscillatory generators, b alter local excitation/inhibition balance in ways that retune these rhythms to faster frequencies, and c amplify thalamocortical drive, shifting the natural resonant frequency of the system. Although robust peak-frequency shifts were observed only under LSD, cross-psychedelic studies with larger cohorts, as well as studies elucidating neuropharmacological similarities and distinctions, are both necessary to better establish the specificity of this effect. Thalamocortical connections serve as a central relaying architecture linking sensory input with cortical hierarchies. LSD-induced increases in thalamocortical connectivity have been interpreted as reduced sensory gating and enhanced bottom-up flow, consistent with alterations in functional hierarchy<sup 92 https://www.nature.com/articles/s42003-026-10629-7 ref-CR92 </sup . Within the REBUS framework, which integrates the entropic brain hypothesis<sup 53 https://www.nature.com/articles/s42003-026-10629-7 ref-CR53 </sup and the free-energy principle<sup 94 https://www.nature.com/articles/s42003-026-10629-7 ref-CR94 </sup , psychedelics transiently relax high-level top-down priors, weakening top-down predictions while facilitating bottom-up information flow i.e., less sensory gating . In this context, the combined weakening and speeding up of alpha and beta rhythms may reflect a hierarchical rebalance: a weakening of power indicates a loosening of top-down constraints, while peak-frequency acceleration might indicate heightened thalamic drive bottom-up flow and increased cortical excitability. Together, these effects align with the hierarchical reconfiguration proposed by the REBUS framework<sup 83 https://www.nature.com/articles/s42003-026-10629-7 ref-CR83 </sup . LSD exerts system-selective effects Although LSD induces both oscillatory power reductions and peak-frequency increases across the cortex, the relative expression of these effects differs markedly across functional systems. To characterize this system specificity, we assessed the functional impact profiles of each effect using meta-analytic decoding based on NiMARE. This approach leverages large-scale neuroimaging databases to map regional neurophysiological changes onto functional domains, allowing effect magnitudes to be quantified across cognitive systems. Using this framework, we show that the two principal oscillatory effects of LSD—attenuation and acceleration of alpha-beta rhythms—follow distinct functional hierarchies across the cortex. LSD-induced suppression of alpha and beta power was strongest in sensory, language, and emotion-related systems, with comparatively minimal effects in motor networks. In particular, auditory, visual, and language systems exhibited pronounced alpha suppression. This pattern is consistent with the high expression of 5-HT2 A receptors in sensory and multimodal association cortices<sup 11 https://www.nature.com/articles/s42003-026-10629-7 ref-CR11 </sup and aligns with prior electrophysiological work showing widespread psychedelic-induced desynchronization, especially within default mode network hubs, such as the posterior cingulate cortex<sup 14 https://www.nature.com/articles/s42003-026-10629-7 ref-CR14 </sup . Emotion-related networks similarly showed strong alpha-beta attenuation, consistent with evidence that LSD modulates limbic and affective circuits and alters emotional processing<sup 95 https://www.nature.com/articles/s42003-026-10629-7 ref-CR95 , 96 https://www.nature.com/articles/s42003-026-10629-7 ref-CR96 </sup . In contrast, motor systems remained largely preserved, exhibiting minimal oscillatory suppression. This pattern corroborates the recent connectivity findings that LSD “flattens” the brain’s functional hierarchy by reducing the differentiation between transmodal e.g., default mode, frontoparietal and unimodal primary sensory cortices<sup 97 https://www.nature.com/articles/s42003-026-10629-7 ref-CR97 </sup . Our results provide electrophysiological support for this account: the largest power weakening in transmodal networks language/emotional hubs and in unimodal regions primary sensory areas reflects a substantial reconfiguration of the normal hierarchical order of oscillatory activity under LSD. In contrast to power suppression, oscillatory peak-frequency modulation followed a distinct system-level organization. Regions exhibiting the strongest alpha power attenuation—such as auditory and language networks—showed minimal shifts in alpha peak frequency, whereas areas with more preserved alpha power, including motor and attentional systems, exhibited larger frequency increases. In the beta band, this relationship was reversed: systems showing the strongest suppression of beta power also exhibited the largest beta peak-frequency shifts. These dissociations could indicate two neurophysiological processes –dampening of oscillatory power in some circuits and acceleration of rhythmic activity in others – organized along distinct functional hierarchies. However, our results comparing the association between dampening of oscillatory power and peak frequency shifts revealed no statistical association region-averaged Spearman’s ρ = −0.04 and −0.12 for alpha and beta, respectively, p 0.05 uncorrected . These results show no evidence of a strong association between the two LSD-induced effects, and thus do not support a common spatial covariation. This interpretation is consistent with prior work: Carhart-Harris et. al<sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 </sup noted that LSD not only reduced alpha power but was also associated with a slight increase in the dominant alpha rhythm frequency. Although direct literature on frequency shifts remains limited, our observations reinforce the view that LSD modulates not just the power but also the timing of neural oscillations. Further studies are needed to disentangle these co-occurring effects to shed light on their underlying neuropharmacological mechanisms. Beyond oscillatory components, LSD also produced pronounced, system-selective changes in the aperiodic background of the MEG spectrum. We observed robust flattening of the 1/ f slope, most prominently in visual, auditory, language, and imagery-related systems. A flatter 1/ f slope reflects reduced dominance of low-frequency power and a shift toward more broadband, irregular dynamics<sup 79 https://www.nature.com/articles/s42003-026-10629-7 ref-CR79 </sup . In sensory cortices, where 5-HT2 A receptor density is high, such flattening likely reflects disrupted inhibitory constraints and increased excitability. Notably, higher-order language and imagery networks exhibited comparable slope flattening, whereas motor regions showed minimal changes, reinforcing the graded and system-selective nature of LSD’s effects Fig. 4 https://www.nature.com/articles/s42003-026-10629-7 Fig4 . These aperiodic changes were closely mirrored by increases in neural signal complexity, as indexed by Higuchi fractal dimension and Lempel-Ziv complexity. Regions with the strongest 1/ f flattening visual, imagery areas, auditory and language cortices showed the largest increases in neural diversity, whereas motor regions remained relatively stable. This inverse relationship between spectral slope and signal complexity has been reported previously<sup 98 https://www.nature.com/articles/s42003-026-10629-7 ref-CR98 </sup and suggests a shared mechanism whereby broadband, irregular activity gives rise to richer dynamical repertoires. These findings are consistent with the entropic brain hypothesis<sup 53 https://www.nature.com/articles/s42003-026-10629-7 ref-CR53 </sup and align with prior reports of elevated signal diversity under psychedelics<sup 13 https://www.nature.com/articles/s42003-026-10629-7 ref-CR13 , 42 https://www.nature.com/articles/s42003-026-10629-7 ref-CR42 , 99 https://www.nature.com/articles/s42003-026-10629-7 ref-CR99 </sup . Given that Lempel-Ziv complexity is sensitive to the regularity of oscillatory activity, such that more rhythmic signals e.g., alpha tend to yield lower complexity <sup 100 https://www.nature.com/articles/s42003-026-10629-7 ref-CR100 </sup , we performed complementary band-specific analyses. These revealed increased complexity in beta and low-gamma bands, with no significant changes in alpha, indicating that the observed LZC effects are not driven by alpha-band dynamics. Rather, they reflect a frequency-specific increase in neural signal diversity, consistent with a reorganization of cortical dynamics beyond simple power changes. Together, LSD reorganizes cortical dynamics through convergent mechanisms that are differentially expressed across cortical functional hierarchies. Alpha-beta power suppression, peak-frequency modulation, aperiodic flattening, and increased signal complexity co-localize preferentially within sensory, language, emotional, and imagery systems, while motor networks remain comparatively preserved. This system-selective reorganization provides electrophysiological support for hierarchical and entropic models of psychedelic action, including REBUS, by showing that LSD relaxes structured top-down constraints while expanding the dynamical repertoire of perceptual and cognitive circuits. Why might music not be interacting with LSD to potentiate effects One of our central questions was whether music potentiates the neurophysiological effects of LSD. Across all spectral and temporal features examined, the interaction between music and LSD did not survive correction for multiple comparisons and can therefore only be considered as a nominally significant effect. This absence of a robust interaction is noteworthy, given that music is widely regarded as a potent emotional and physiological stimulus<sup 25 https://www.nature.com/articles/s42003-026-10629-7 ref-CR25 </sup , has been intertwined with psychedelic ritual practices for millennia<sup 7 https://www.nature.com/articles/s42003-026-10629-7 ref-CR7 </sup , and is increasingly incorporated into psychedelic-assisted therapy to enhance emotional depth and meaning-making<sup 30 https://www.nature.com/articles/s42003-026-10629-7 ref-CR30 , 101 https://www.nature.com/articles/s42003-026-10629-7 ref-CR101 </sup . Although the interaction effects did not survive correction for multiple comparisons, the resulting maps showed a qualitatively consistent pattern indicative of attenuation rather than potentiation. Specifically, LSD x music interaction was associated with increased exponent and offset, alongside reduced LZC and HFD, indicative of steeper aperiodic slopes, higher broadband power, and reduced temporal complexity relative to the NoMusic condition. Prior work provides a useful framework for interpreting this pattern. Mediano and Rosas<sup 41 https://www.nature.com/articles/s42003-026-10629-7 ref-CR41 </sup , with the same dataset, showed that psychedelic-induced increases in neural signal diversity are maximal during internally oriented states e.g. eyes-closed and diminish when attention is directed toward external stimuli. Our results extend this finding by suggesting that music may act as a form of grounding agent, an external attentional anchor that stabilizes neural activity and counteracts the unconstrained dynamics characteristic of the psychedelic state. Recent work analyzing various tasks during psilocybin sessions demonstrates similar reductions in neural and subjective intensity when attention is externally directed<sup 4 https://www.nature.com/articles/s42003-026-10629-7 ref-CR4 </sup . Together, these findings raise the possibility that music does not amplify the core neural effects of LSD, but may instead attenuate them by imposing structure and redirecting attention outward. Nevertheless, given the absence of interaction effects surviving correction and the fact that the experimental design was not optimized to isolate subtle state-context interactions, these interpretations should be viewed as tentative. Moreover, detecting interaction effects typically requires substantially greater statistical power than main effects <sup 102 https://www.nature.com/articles/s42003-026-10629-7 ref-CR102 </sup , and the present study was not specifically powered for this purpose. Accordingly, this finding should be considered exploratory and in need of replication in larger samples. In exploratory analyses linking LSD-induced neural changes to subjective experience Supplementary Fig. S6 https://www.nature.com/articles/s42003-026-10629-7 MOESM1 , region-averaged reductions in alpha and beta Peak Power were associated with increased Simple Imagery, whereas flatter spectral slopes, lower offset, reduced beta Peak Power, and lower LZC predicted stronger Complex Imagery. In contrast, LSD-induced neural changes did not significantly predict Intensity, Emotional Arousal, or Positive Mood. Ego Dissolution was selectively associated with beta peak-frequency shifts and LZC. Notably, only Emotional Arousal showed a significant association with the LSD × music interaction: smaller beta peak shifts during music listening were linked to higher emotional arousal. Although these associations are exploratory p < 0.05, uncorrected , they converge with prior reports<sup 41 https://www.nature.com/articles/s42003-026-10629-7 ref-CR41 </sup , suggesting that engagement with structured external stimuli can dampen core psychedelic neural signatures even when subjective experience remains emotionally salient. Limitations and perspectives Several limitations of the present work warrant consideration and can help guide future research. First, because we frequently observed multiple oscillatory peaks Center Frequencies within the alpha band, we defined the peak as the frequency with maximal amplitude in the 1/f-removed spectrum. Our analysis, looking at the Center Frequency, found that alpha and beta peaks were present in nearly all PSDs Subjects x Regions , and high-gamma peaks appeared frequently as well, with a slight reduction under LSD. Given that oscillations in alpha and beta rhythms were present in most PSDs, our manual peak-identification procedure produces results largely equivalent to the center-frequency approach. Nevertheless, caution is warranted when interpreting findings in the high-gamma range, where oscillatory structure is less consistent, and peak estimation is inherently more variable than in the lower frequency range. In addition, the absence of concurrent EMG recordings did not allow for direct statistical comparison of muscular activity between LSD and placebo conditions, highlighting a limitation that future data collection during LSD should address. The second limitation is regarding the nominal significant interaction between LSD and music. While this aligns with prior reports suggesting that the acute influence of psychedelics may wane when participants engage with external stimuli, it also highlights a broader limitation in how such effects are typically modeled. Most studies, including ours, implicitly assume that LSD x music interactions remain constant over time. However, music is a dynamic and richly structured stimulus, and any modulatory effects may be transient or state-dependent. Collapsing the temporal dimension may therefore obscure meaningful fluctuations. A more time-resolved approach may offer clearer insights by enabling joint tracking of neural responses and concurrent phenomenology. In addition, because the pre-assessment to ensure the preference for the genre and the music track was conducted with a separate group, the propensity of this cohort was not established. From a statistical standpoint, interaction effects are inherently more difficult to detect than main effects and typically require substantially larger sample sizes to achieve comparable statistical power <sup 102 https://www.nature.com/articles/s42003-026-10629-7 ref-CR102 </sup . This poses a particular challenge in psychedelic research, where studies are often constrained by modest sample sizes, as in the present dataset N = 17 . As a result, the current analysis may be underpowered to reliably detect subtle state-context interactions, and the observed effects should therefore be interpreted with appropriate caution. A further limitation concerns the characterization of individual differences in subjective experience. At the region level, we observed significant correlations between neurophysiological changes and phenomenology only in the Simple Effects conditions. With regards to the main effect of the drug and the interaction effect of the drug x music, global region-averaged measures did show meaningful associations. We regard these global correlations as exploratory because they were not corrected for multiple comparisons. The lack of region-level associations may be due, in part, to the timing of subjective reporting. Participants provided in-scanner ratings relative to their peak drug experience, whereas MEG acquisition occurred approximately 150 minutes after peak effect. This discrepancy likely introduces noise and reduces the sensitivity of correlations between brain phenomena and phenomenology. Conclusion In this study, we provide a comprehensive electrophysiological characterization of how LSD reorganizes human cortical dynamics, demonstrating that its acute effects arise from dissociable mechanisms that map systematically onto large-scale functional systems. By combining spectral parametrization, peak-resolved oscillatory analysis, aperiodic modeling, and functional decoding, we identify three principal axes of modulation. First, LSD induces a strong attenuation of alpha and beta power in auditory, visual, language, and emotion-related networks, regions rich in 5-HT2 A receptors, while sparing the motor cortex. Second, LSD produces a system-specific increase in alpha and beta peak frequencies that follows a distinct functional hierarchy, independent from that of power suppression. Third, LSD flattens the aperiodic 1/ f spectrum and increases temporal complexity HFD, LZC , with the largest changes in sensory and imagery domains, consistent with a shift toward a desynchronized, high-complexity cortical regime. These results address our initial questions: Q1 peak-frequency shifts are spatially structured; Q2 oscillatory power weakening, aperiodic flattening, and complexity increases are anatomically selective; and Q3 music’s modulatory influence is limited and context dependent, showing no uniform enhancement of LSD’s neural effects and, in some cases, attenuating them. Overall, LSD induces a structured reorganization of cortical dynamics across temporal and spectral properties, rather than a uniform global perturbation. Future work aimed at disentangling the neuropharmacological mechanisms underlying oscillatory power weakening and peak shifts, and establishing tighter alignment between neural activity and subjective experience, will be essential for clarifying how these LSD-specific effects arise and how they contribute to psychedelic phenomenology. Ethics statement We utilized the subset of the data collected by Carhart-Harris and colleagues<sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 </sup . All ethical regulations relevant to human research participants were followed. This original study was approved by the National Research Ethics Service Committee London-West London and was conducted in accordance with the revised declaration of Helsinki 2000 , the International Committee on Harmonization Good Clinical Practice guidelines, and the National Health Service Research Governance Framework. Imperial College, London, sponsored the research, which was conducted under a Home Office license for research with Schedule 1 drugs. Data availability Queries regarding data are to be directed to SM and RC-H. Code availability Code is available on https://github.com/venkateshness/specparam LSD https://github.com/venkateshness/specparam LSD . References 1. Nichols, D. E. Psychedelics. Pharmacol. Rev. 68 , 264–355 2016 . 2. Dos Santos, R. G., Bouso, J. C. & Hallak, J. E. Hallucinogens/psychedelics resurrected as new tools in psychiatric therapy. Braz. J. Psychiatry 43 , 119–120 2020 . 3. Geyer, M. A. & Vollenweider, F. X. Serotonin research: contributions to understanding psychoses. Trends Pharmacol. Sci. 29 , 445–453 2008 . 4. Siegel, J. S. et al. Psilocybin desynchronizes the human brain. Nature 632 , 131–138 2024 . 5. Carhart-Harris, R. L. & Goodwin, G. M. The therapeutic potential of psychedelic drugs: past, present, and future. Neuropsychopharmacology 42 , 2105–2113 2017 . 6. Schultes, R. E. Hallucinogens of plant origin: interdisciplinary studies of plants sacred in primitive cultures yield results of academic and practical interest. Science 163 , 245–254 1969 . 7. Jerotic, K., Vuust, P. & Kringelbach, M. L. Psychedelia: the interplay of music and psychedelics. Ann. N. Y. Acad. Sci. 1531 , 12–28 2024 . 8. Rickli, A., Moning, O. D., Hoener, M. C. & Liechti, M. E. Receptor interaction profiles of novel psychoactive tryptamines compared with classic hallucinogens. Eur. Neuropsychopharmacol. 26 , 1327–1337 2016 . 9. Kwan, A. C., Olson, D. E., Preller, K. H. & Roth, B. L. The neural basis of psychedelic action. Nat. Neurosci. 25 , 1407–1419 2022 . 10. Van Elk, M. & Yaden, D. B. Pharmacological, neural, and psychological mechanisms underlying psychedelics: a critical review. Neurosci. Biobehav. Rev. 140 , 104793 2022 . 11. Beliveau, V. et al. A high-resolution in vivo atlas of the human brain’s serotonin system. J. Neurosci. 37 , 120–128 2017 . 12. Carhart-Harris, R. L. et al. Neural correlates of the lsd experience revealed by multimodal neuroimaging. Proc. Natl. Acad. Sci. USA . 113 , 4853–4858 2016 . 13. Timmermann, C. et al. Neural correlates of the DMT experience assessed with multivariate EEG. Sci. Rep. 9 , 16324 2019 . 14. Muthukumaraswamy, S. D. et al. Broadband cortical desynchronization underlies the human psychedelic state. J. Neurosci. 33 , 15171–15183 2013 . 15. Carhart-Harris, R. L. et al. Neural correlates of the psychedelic state as determined by fmri studies with psilocybin. Proc. Natl. Acad. Sci. USA . 109 , 2138–2143 2012 . 16. Padawer-Curry, J. A. et al. Psychedelic 5-HT2A receptor agonism alters neurovascular coupling and differentially affects neuronal and hemodynamic measures of brain function. Nat. Neurosci. 28 , 2330–2343 2025 . 17. Timmermann, C. et al. Human brain effects of DMT assessed via EEG-fMRI. Proc. Natl. Acad. Sci. USA . 120 , e2218949120 2023 . 18. Teplan, M. et al. Fundamentals of EEG Measurement. Meas. Sci. Rev. 2 , 1–11 2002 . 19. Baillet, S. Magnetoencephalography for brain electrophysiology and imaging. Nat. Neurosci. 20 , 327–339 2017 . 20. Donoghue, T. et al. Parameterizing neural power spectra into periodic and aperiodic components. Nat. Neurosci. 23 , 1655–1665 2020 . 21. Donoghue, T., Dominguez, J. & Voytek, B. Electrophysiological frequency band ratio measures conflate periodic and aperiodic neural activity. Eneuro 7 , 2020 2020 . 22. Muthukumaraswamy, S. D. & Liley, D. T. 1/f electrophysiological spectra in resting and drug-induced states can be explained by the dynamics of multiple oscillatory relaxation processes. NeuroImage 179 , 582–595 2018 . 23. Nikolic, M., Mediano, P., Froese, T., Reydellet, D. & Palenicek, T. Psilocybin alters brain activity related to sensory and cognitive processing in a time-dependent manner. medRxiv 2024–09 2024 . 24. Peretz, I. & Zatorre, R. J. Brain organization for music processing. Annu. Rev. Psychol. 56 , 89–114 2005 . 25. Vuust, P., Heggli, O. A., Friston, K. J. & Kringelbach, M. L. Music in the brain. Nat. Rev. Neurosci. 23 , 287–305 2022 . 26. Kaelen, M. The psychological and human brain effects of music in combination with psychedelic drugs . Ph.D. thesis, Imperial College London Ph. D. dissertation. London Imperial College London, 2017 . 27. Kaelen, M. et al. Effects of lsd on music-evoked brain activity. biorxiv 153031 2017 . 28. Hartogsohn, I. Constructing drug effects: a history of set and setting. Drug Sci. Policy Law 3 , 2050324516683325 2017 . 29. Bonny, H. L. & Pahnke, W. N. The use of music in psychedelic LSD psychotherapy. J. Music Ther. 9 , 64–87 1972 . 30. Barrett, F. S., Preller, K. H. & Kaelen, M. Psychedelics and music: neuroscience and therapeutic implications. Int. Rev. Psychiatry 30 , 350–362 2018 . 31. Kaelen, M. et al. Lsd enhances the emotional response to music. Psychopharmacology 232 , 3607–3614 2015 . 32. Kaelen, M. et al. Lsd modulates music-induced imagery via changes in parahippocampal connectivity. Eur. Neuropsychopharmacol. 26 , 1099–1109 2016 . 33. Preller, K. H. et al. The fabric of meaning and subjective effects in LSD-induced states depend on serotonin 2a receptor activation. Curr. Biol. 27 , 451–457 2017 . 34. Roseman, L. et al. Emotional breakthrough and psychedelics: validation of the emotional breakthrough inventory. J. Psychopharmacol. 33 , 1076–1087 2019 . 35. Davis, A. K. et al. Development of the psychological insight questionnaire among a sample of people who have consumed psilocybin or lsd. J. Psychopharmacol. 35 , 437–446 2021 . 36. Murphy, R. et al. Therapeutic alliance and rapport modulate responses to psilocybin assisted therapy for depression. Front. Pharmacol. 12 , 788155 2022 . 37. Peill, J. M. et al. Validation of the psychological insight scale: a new scale to assess psychological insight following a psychedelic experience. J. Psychopharmacol. 36 , 31–45 2022 . 38. Lyons, T. et al. Human brain changes after first psilocybin use. Nat Commun. 17 , 3977 2026 . 39. Kugel, J., Laukkonen, R. E., Yaden, D. B., Yücel, M. & Liknaitzky, P. Insights on psychedelics: a systematic review of therapeutic effects. Neurosci. Biobehav. Rev. 173 , 106117 2025 . 40. Timmermann, C. et al. LSD modulates effective connectivity and neural adaptation mechanisms in an auditory oddball paradigm. Neuropharmacology 142 , 251–262 2018 . 41. Mediano, P. A. et al. Effects of external stimulation on psychedelic state neurodynamics. ACS Chem. Neurosci. 15 , 462–471 2024 . 42. Schartner, M. M., Carhart-Harris, R. L., Barrett, A. B., Seth, A. K. & Muthukumaraswamy, S. D. Increased spontaneous meg signal diversity for psychoactive doses of ketamine, lsd and psilocybin. Sci. Rep. 7 , 46421 2017 . 43. Pascarella, A. et al. Meditation induces shifts in neural oscillations, brain complexity, and critical dynamics: novel insights from MEG. Neurosci Conscious. 2025 , niaf047 2025 . 44. Tu, Z. et al. Accurate machine learning-based monitoring of anesthesia depth with EEG recording. Neurosci. Bull. 41 , 449–460 2024 . 45. Thölke, P. et al. Caffeine induces age-dependent increases in brain complexity and criticality during sleep. Commun. Biol. 8 , 685 2025 . 46. Bedford, P. et al. The effect of lysergic acid diethylamide LSD on whole-brain functional and effective connectivity. Neuropsychopharmacology 48 , 1175–1183 2023 . 47. Pallavicini, C. et al. Spectral signatures of serotonergic psychedelics and glutamatergic dissociatives. Neuroimage 200 , 281–291 2019 . 48. Jas, M., Engemann, D., Raimondo, F., Bekhti, Y. & Gramfort, A. Automated rejection and repair of bad trials in meg/eeg. In Proc. International Workshop on Pattern Recognition in Neuroimaging PRNI , 1–4 IEEE, 2016 . 49. Ablin, P., Cardoso, J.-F. & Gramfort, A. Faster independent component analysis by preconditioning with Hessian approximations. IEEE Trans. Signal Process. 66 , 4040–4049 2018 . 50. Dale, A. M. et al. Dynamic statistical parametric mapping: combining fMRI and MEG for high-resolution imaging of cortical activity. Neuron 26 , 55–67 2000 . 51. Glasser, M. F. et al. A multi-modal parcellation of human cerebral cortex. Nature 536 , 171–178 2016 . 52. Ziv, J. & Lempel, A. Compression of individual sequences via variable-rate coding. IEEE Trans. Inf. Theory 24 , 530–536 1978 . 53. Carhart-Harris, R. L. et al. The entropic brain: a theory of conscious states informed by neuroimaging research with psychedelic drugs. Front. Hum. Neurosci. 8 , 55875 2014 . 54. Zhang, Y., Hao, J., Zhou, C. & Chang, K. Normalized lempel-Ziv complexity and its application in bio-sequence analysis. J. Math. Chem. 46 , 1203–1212 2009 . 55. Higuchi, T. Approach to an irregular time series on the basis of the fractal theory. Phys. D: Nonlinear Phenom. 31 , 277–283 1988 . 56. Kesić, S. & Spasić, S. Z. Application of Higuchi’s fractal dimension from basic to clinical neurophysiology: a review. Computer Methods Prog. Biomed. 133 , 55–70 2016 . 57. Porcaro, C., Moaveninejad, S., D’Onofrio, V. & DiIeva, A. Fractal time series: background, estimation methods, and performances. Fractal Geometry Brain 36 , 95–137 2024 . 58. Smits, F. M. et al. Electroencephalographic fractal dimension in healthy ageing and Alzheimer’s disease. PloS one 11 , e0149587 2016 . 59. Yarkoni, T., Poldrack, R. A., Nichols, T. E., Van Essen, D. C. & Wager, T. D. Large-scale automated synthesis of human functional neuroimaging data. Nat. methods 8 , 665–670 2011 . 60. Margulies, D. S. et al. Situating the default-mode network along a principal gradient of macroscale cortical organization. Proc. Natl. Acad. Sci. USA . 113 , 12574–12579 2016 . 61. Preti, M. G. & Van De Ville, D. Decoupling of brain function from structure reveals regional behavioral specialization in humans. Nat. Commun. 10 , 4747 2019 . 62. Subramani, V., Lioi, G., Jerbi, K. & Farrugia, N. Structure–function coupling and decoupling during movie watching and resting state: novel insights bridging EEG and structural imaging. Imaging Neurosci. 3 , imag a 00448 2025 . 63. Salo, T. et al. neurostuff/nimare: 0.0. 13. Zenodo 2022 . 64. Maris, E. & Oostenveld, R. Nonparametric statistical testing of EEG and MEG data. J. Neurosci. Methods 164 , 177–190 2007 . 65. Noble, W. S. How does multiple testing correction work? Nat. Biotechnol. 27 , 1135–1137 2009 . 66. D’Andrea, A. et al. Mindfulness meditation styles differently modulate source-level meg microstate dynamics and complexity. Front. Neurosci. 18 , 1295615 2024 . 67. Carhart-Harris, R. L. & Nutt, D. J. Serotonin and brain function: a tale of two receptors. J. Psychopharmacol. 31 , 1091–1120 2017 . 68. Vollenweider, F. X., Vollenweider-Scherpenhuyzen, M. F., Bäbler, A., Vogel, H. & Hell, D. Psilocybin induces schizophrenia-like psychosis in humans via a serotonin-2 agonist action. Neuroreport 9 , 3897–3902 1998 . 69. Aghajanian, G. K. & Marek, G. J. Serotonin, via 5-hT2a receptors, increases epscs in layer V pyramidal cells of prefrontal cortex by an asynchronous mode of glutamate release. Brain Res. 825 , 161–171 1999 . 70. Carhart-Harris, R. L. How do psychedelics work? Curr. Opin. Psychiatry 32 , 16–21 2019 . 71. Le, G. H. et al. Spectral signatures of psilocybin, lysergic acid diethylamide lsd and ketamine in healthy volunteers and persons with major depressive disorder and treatment-resistant depression: a systematic review. J. Affect. Disord. 355 , 342–354 2024 . 72. Tagliazucchi, E. et al. Baseline power of theta oscillations predicts mystical-type experiences induced by DMT in a natural setting. Front. Psychiatry 12 , 720066 2021 . 73. Blackburne, G. et al. Complex slow waves in the human brain under 5-meo-DMT. Cell Rep. 44 , 116040 2025 . 74. Muthukumaraswamy, S. D. et al. Evidence that subanesthetic doses of ketamine cause sustained disruptions of NMDA and AMPA-mediated frontoparietal connectivity in humans. J. Neurosci. 35 , 11694–11706 2015 . 75. Bullock, T., McClune, M. & Enright, J. Are the electroencephalograms mainly rhythmic? Assessment of periodicity in wide-band time series. Neuroscience 121 , 233–252 2003 . 76. Allegrini, P. et al. Spontaneous brain activity as a source of ideal 1/f noise. Phys. Rev. E Stat. Nonlinear Soft Matter Phys. 80 , 061914 2009 . 77. Wen, H. & Liu, Z. Separating fractal and oscillatory components in the power spectrum of neurophysiological signal. Brain Topogr. 29 , 13–26 2016 . 78. Voytek, B. & Knight, R. T. Dynamic network communication as a unifying neural basis for cognition, development, aging, and disease. Biol. Psychiatry 77 , 1089–1097 2015 . 79. Gao, R., Peterson, E. J. & Voytek, B. Inferring synaptic excitation/inhibition balance from field potentials. Neuroimage 158 , 70–78 2017 . 80. Buzsaki, G. & Draguhn, A. Neuronal oscillations in cortical networks. Science 304 , 1926–1929 2004 . 81. Von Stein, A. & Sarnthein, J. Different frequencies for different scales of cortical integration: from local gamma to long-range alpha/theta synchronization. Int. J. Psychophysiol. 38 , 301–313 2000 . 82. Fries, P. Rhythms for cognition: communication through coherence. Neuron 88 , 220–235 2015 . 83. Carhart-Harris, R. L. & Friston, K. J. Rebus and the anarchic brain: toward a unified model of the brain action of psychedelics. Pharmacol. Rev. 71 , 316–344 2019 . 84. Schenberg, E. E. et al. Acute biphasic effects of ayahuasca. PloS ONE 10 , e0137202 2015 . 85. Mushtaq, F. et al. One hundred years of EEG for brain and behaviour research. Nat. Hum. Behav. 8 , 1437–1443 2024 . 86. Angelakis, E., Lubar, J. F., Stathopoulou, S. & Kounios, J. Peak alpha frequency: an electroencephalographic measure of cognitive preparedness. Clin. Neurophysiol. 115 , 887–897 2004 . 87. Garcia-Rill, E. et al. The 10Hz frequency: a fulcrum for transitional brain states. Transl. Brain Rhythmicity 1 , 7 2016 . 88. Hagihira, S. Changes in the electroencephalogram during anaesthesia and their physiological basis. Br. J. Anaesth. 115 , i27–i31 2015 . 89. Soltanzadeh, M. et al. Divergent effects of ketamine and psilocybin on EEG power spectral density in a mismatch negativity paradigm. Psychopharmacology 1–15 Springer, 2025 . 90. Wacker, D. et al. Crystal structure of an LSD-bound human serotonin receptor. Cell 168 , 377–389 2017 . 91. Holze, F. et al. Direct comparison of the acute effects of lysergic acid diethylamide and psilocybin in a double-blind placebo-controlled study in healthy subjects. Neuropsychopharmacology 47 , 1180–1187 2022 . 92. Müller, F. et al. Increased thalamic resting-state connectivity as a core driver of LSD-induced hallucinations. Acta Psychiatr. Scandinavica 136 , 648–657 2017 . 93. Hindriks, R. & van Putten, M. J. Thalamo-cortical mechanisms underlying changes in amplitude and frequency of human alpha oscillations. Neuroimage 70 , 150–163 2013 . 94. Friston, K. The free-energy principle: a unified brain theory? Nat. Rev. Neurosci. 11 , 127–138 2010 . 95. Mueller, F. et al. Acute effects of lsd on amygdala activity during processing of fearful stimuli in healthy subjects. Transl. Psychiatry 7 , e1084–e1084 2017 . 96. Kraehenmann, R. et al. Psilocybin-induced decrease in amygdala reactivity correlates with enhanced positive mood in healthy volunteers. Biol. Psychiatry 78 , 572–581 2015 . 97. Girn, M. et al. Serotonergic psychedelic drugs lsd and psilocybin reduce the hierarchical differentiation of unimodal and transmodal cortex. NeuroImage 256 , 119220 2022 . 98. Donoghue, T. et al. Evaluating and comparing measures of aperiodic neural activity. BioRXiv 2024 . 99. Varley, T. F., Carhart-Harris, R., Roseman, L., Menon, D. K. & Stamatakis, E. A. Serotonergic psychedelics lsd & psilocybin increase the fractal dimension of cortical brain activity in spatial and temporal domains. NeuroImage 220 , 117049 2020 . 100. Ibáñez-Molina, A. J. et al. EEG multiscale complexity in schizophrenia during picture naming. Front. Physiol. 9 , 1213 2018 . 101. Carhart-Harris, R. L. et al. Psilocybin with psychological support for treatment-resistant depression: six-month follow-up. Psychopharmacology 235 , 399–408 2018 . 102. Leon, A. C. & Heo, M. Sample sizes required to detect interactions between two binary fixed-effects in a mixed-effects linear regression model. Comput. Stat. Data Anal. 53 , 603–608 2009 . Acknowledgements We thank several colleagues who contributed to the progress of this study. We are grateful to Jordan O’Byrne for assistance with dataset organization; Kenneth Shinozuka for identifying missing data for two subjects; Arianna Caputo for sharing resources and insights on the Higuchi Fractal Dimension; and Yassine El Ouahidi for adapting the functional decoding pipeline from Neurosynth to NiMARE. We utilized a subset of the data collected by Carhart-Harris and colleagues<sup 12 https://www.nature.com/articles/s42003-026-10629-7 ref-CR12 </sup . Data collection of this study was supported by Safra Foundation and the Beckley Foundation as part of the Beckley-Imperial research program, and by supporters of the Walacea.com crowdfunding campaign. Funding K.J. is supported by funding from the Canada Research Chairs 950-232368 program and a Discovery Grant from the Natural Sciences and Engineering Research Council of Canada 2021-03426 . Ethics declarations Competing interests RC-H reports providing scientific advice for TRYP therapeutics, Osmind, Otsuka, and Red Light Holland. Others declare no competing interests. Peer review Peer review information Communications Biology thanks Santiago Castro-Zaballa, Yasir Çatalfor their contribution to the peer review of this work. Primary Handling Editor: Jasmine Pan. A peer review file is available. Additional information Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Rights and permissions Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author s and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ http://creativecommons.org/licenses/by/4.0/ . About this article Cite this article Subramani, V., Nest, T., Pascarella, A. et al. LSD reconfigures cortical dynamics through faster brain rhythms and increased fractal dimension. Commun Biol 9 , 1244 2026 . https://doi.org/10.1038/s42003-026-10629-7 - Received: - Accepted: - Published: - Version of record: - DOI: https://doi.org/10.1038/s42003-026-10629-7