Low doses of LSD reduce broadband oscillatory power and modulate event-related potentials in healthy adults

Research on classical psychedelics has expanded in the past decade, with clinical trials investigating relatively high doses of compounds such as psilocybin and LSD across disorders including depression, obsessive-compulsive disorder, addictions and anorexia. Parallel to these controlled trials, anecdotal and naturalistic reports of "microdosing"—regular ingestion of very low doses of LSD (around 10–15 μg)—have proliferated, with users reporting mood and cognitive benefits. Controlled laboratory work to date suggests single low doses can increase vigour, reduce attentional lapses and alter time perception, while producing few robust effects on mood or cognition; however, the neural effects of these very low doses remain poorly characterised. This study set out to characterise the acute neural signature of low doses of LSD in healthy adults, using electroencephalography (EEG). Specifically, Murray and colleagues tested whether single sublingual doses of 13 μg and 26 μg LSD, compared with placebo, would alter resting-state oscillatory power over scalp regions corresponding to default mode network (DMN) hubs and modulate event-related potentials (ERPs) evoked during a visual oddball task. The investigators also measured subjective drug effects, mood and cardiovascular responses, and explored relationships between EEG changes and subjective measures.

The study used a within-subject, double-blind design. Twenty-two healthy right-handed adults (8 women), aged 18–35, completed three 5-hour sessions separated by at least 7 days. In randomized order participants received sublingual placebo, 13 μg LSD or 26 μg LSD; the drug was prepared as a tartaric acid solution and administered in a 0.5 mL volume held under the tongue for 60 s. Screening excluded individuals with psychosis, severe PTSD or panic disorder, recent substance use disorder (past year, except nicotine), pregnancy, regular medication (except birth control) or adverse reactions to psychedelics. Participants had to have at least one prior use of a classical psychedelic or MDMA. Sessions included baseline and hourly measures of subjective state and cardiovascular function for 4 hours after dosing. EEG was recorded during the expected peak effect (120–180 min post-dose) and lasted about 60 min. Resting-state EEG comprised 5 min eyes closed followed by 5 min eyes open (fixation cross). The primary resting outcomes were oscillatory power in delta (1–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz) and gamma (30–80 Hz) bands from scalp electrodes chosen to approximate DMN hubs (medial prefrontal cortex, posterior cingulate cortex, left and right temporoparietal cortices). The event-related paradigm was an emotional faces oddball task presenting frequent happy faces and infrequent neutral or angry faces (3:1 ratio, 320 stimuli total). Participants indicated happy versus non-happy faces; incorrect trials were discarded. Primary ERP outcomes were amplitudes and latencies of P100, N170 and P300 components at predefined electrodes (Pz for P300; PO10 for N170 and P100). Subjective measures included the Drug Effects Questionnaire (DEQ), the Addiction Research Center Inventory (ARCI), the Profile of Mood States (POMS) and the 5 Dimensions of Altered States of Consciousness (5D-ASC); cardiovascular measures (heart rate, systolic and diastolic blood pressure) were recorded hourly. EEG acquisition used a 128-channel BioSemi ActiveTwo system; preprocessing included filtering, manual removal of gross artifacts, independent component analysis to remove eye/muscle/EKG artifacts, re-referencing to the average and blinding of analysts to drug condition. Oscillatory power was estimated by fast Fourier transform and analysed with repeated measures analysis of variance (dose as within-subject factor) across regions and frequency bands. ERP data were epoched −500 to 1000 ms, peak amplitudes and latencies extracted and analysed with repeated measures ANOVA.

Twenty-two participants completed the protocol. Subjective and cardiovascular responses peaked 2–3 hours after dosing. LSD produced dose-dependent increases on DEQ items (Feel Drug, Feel High, Like Drug, Want More) and on POMS subscales for elation, anxiety and positive mood. On the ARCI, LSD increased scores on LSD-like, amphetamine-like, euphoric and energy/intellectual efficiency subscales. On the 5D-ASC, only the Blissful State subscale increased significantly at 26 μg (dose F1,21 = 5.91, p = 0.024; 26 μg vs placebo p = 0.001). Cardiovascular effects included increased heart rate (dose F1,21 = 5.11, p = 0.035) and higher systolic (dose F1,21 = 8.16, p = 0.009; 26 μg vs placebo p = 0.036) and diastolic blood pressure (dose F1,21 = 4.72, p = 0.041). Resting-state EEG showed dose-dependent reductions in broadband oscillatory power across the four DMN-related scalp regions and across frequency bands (dose F1,420 = 10.39, p = 0.001) during eyes closed rest. A regional analysis indicated the posterior cingulate cortex exhibited the strongest reduction (dose F1,105 = 15.71, p < 0.001); the left temporoparietal cortex also showed significant reductions (dose F1,105 = 8.61, p = 0.004), while the right temporoparietal effect was marginal (dose × band F1,105 = 3.76, p = 0.055). Similar broadband decreases were present during eyes open rest (dose F1,420 = 5.78, p = 0.017), with significant effects in the posterior cingulate (dose F1,105 = 5.96, p = 0.016) and medial prefrontal cortex (dose F1,105 = 8.62, p = 0.004). In the oddball task, LSD reduced error rates dose-dependently (dose F1,42 = 6.39, p = 0.015). ERP analyses showed dose-dependent attenuation of P300 amplitude (dose F1,42 = 7.52, p = 0.009) and N170 amplitude (dose F1,42 = 4.87, p = 0.033), while P100 amplitude was not significantly changed; however, P100 latency increased with dose (dose F1,42 = 11.13, p = 0.002). No significant dose × face valence interactions were detected for amplitudes, latencies or error rates. Topographic maps indicated overall attenuation of the electropositive P300 and electronegative N170 components across the scalp. The investigators report that EEG measures (resting-state or ERPs) did not correlate with subjective or behavioural responses, noting that limited statistical power may have contributed to this null finding. End-of-session identification data showed most participants correctly identified placebo sessions as placebo and 26 μg sessions as a hallucinogen, whereas 13 μg sessions yielded mixed identifications.

Murray and colleagues interpret their findings as evidence that very low doses of LSD produce measurable neurophysiological effects that resemble those reported at higher doses. The principal result was a dose-dependent reduction in broadband oscillatory power over scalp regions corresponding to DMN hubs, consistent with cortical desynchronization reported after moderate or higher doses of classical psychedelics and with MEG/EEG studies of psilocybin and ayahuasca. The authors link reduced oscillatory power, particularly at lower frequencies, to increased cortical excitability and propose that 5-HT2a receptor-mediated glutamatergic excitation could underlie these changes. They also highlight that low doses attenuated ERP amplitudes (P300 and N170) and prolonged early perceptual processing (P100 latency), suggesting effects on both perceptual and cognitive stages of face processing, despite improved behavioural accuracy on the oddball task. The investigators situate their results within broader pharmaco-EEG literature, noting parallels between the effects of low-dose LSD and those of stimulants or subanesthetic ketamine, which also reduce low-frequency oscillations, and contrasting these with sedative drugs that increase low-frequency power. Subjective effects in this sample were modest and included increases in positive mood, elation and some stimulant-like effects; the authors acknowledge that dopaminergic mechanisms might contribute to some subjective responses but note that higher-dose subjective effects are blocked by 5-HT2a antagonists in other work, leaving the role of dopamine at these low doses uncertain. Several limitations are acknowledged: EEG was analysed at the scalp without source localisation or intracranial measures, limiting inferences about the precise neural generators; the sample size constrained power to detect correlations between neural measures and subjective effects; and unblinding (participants correctly identifying 26 μg sessions) may have influenced subjective reports. The authors recommend future studies to reconcile EEG and fMRI connectivity findings, to examine repeated dosing as in microdosing regimens, to test clinical populations, and to explore combinations of low doses with psychotherapeutic support.

The study concludes that single low doses of LSD (13 and 26 μg) produce subtle positive subjective effects and robust neurophysiological changes, including broadband cortical desynchronization and reduced ERP amplitudes, resembling effects observed with higher psychedelic doses. Murray and colleagues suggest these neural effects raise the possibility that very low doses might yield behavioural or therapeutic benefits without inducing full psychedelic experiences, and they call for further research in clinical samples, with repeated dosing, and in combination with psychotherapy to evaluate potential therapeutic applications.