Preliminary report on the effects of a low dose of LSD on resting-state amygdala functional connectivity

Depressive disorders are highly prevalent and current antidepressant treatments, particularly selective serotonin reuptake inhibitors (SSRIs), have limitations including delayed onset, side effects, and variable efficacy. Anecdotal reports and surveys have popularised the practice of “microdosing” LSD — repeated administration of very low doses (typically ~10–15 µg) purported to improve mood and cognition — but controlled empirical data on its behavioural or neural effects are limited. Higher doses of psychedelics alter thalamo-cortical and limbic circuits and have been linked to changes in amygdala reactivity; however, the receptor-level and network effects of very low LSD doses remain poorly understood. This study, led by Bershad and colleagues, set out to examine the acute neural effects of a single low (13 µg) dose of LSD versus placebo in healthy young adults using resting-state fMRI and arterial spin labelling (ASL) to measure cerebral blood flow (CBF). The investigators focused on seed-based connectivity of the right amygdala and bilateral thalamus and explored relationships between connectivity changes and mood assessed immediately after scanning. The 13 µg dose was selected from prior dose-ranging work as producing minimal subjective effects and matches doses reported by microdosing communities, making it relevant to real-world practice and to hypotheses about potential antidepressant mechanisms.

The study employed a within-subject, double-blind, randomised design in which 20 healthy volunteers (10 women) aged 18–35 completed two sessions separated by at least 7 days, receiving 13 µg LSD sublingually or placebo (water) in counterbalanced order. Participants were screened medically and psychiatrically; inclusion required right-handedness, English fluency, minimum education, body mass index 18–32, and at least one prior psychedelic exposure. Standard abstinence rules for drugs, alcohol and cannabis applied before each session. An orientation session familiarised participants with procedures. Each experimental visit lasted about 5 hours and began with baseline subjective and physiological measures. Drug or placebo was administered sublingually 30 minutes after arrival; subjective and cardiovascular measures were repeated at 60, 115, 180 and 240 minutes. Resting-state BOLD and ASL scans were acquired approximately 90 minutes after administration (expected peak effect) and lasted roughly 60 minutes. Blood pressure, heart rate and tympanic temperature were monitored throughout. End-of-session questionnaires included the Drug Effects Questionnaire (DEQ), the Addiction Research Center Inventory (ARCI), the Positive and Negative Affect Schedule (PANAS) and the 5-Dimensions of Altered States of Consciousness (5D-ASC), plus an item asking participants to identify the substance they believed they had received. Imaging used a 3T scanner with BOLD EPI for resting-state functional connectivity and pseudo-continuous ASL for perfusion. Preprocessing of BOLD data was performed in CONN and included realignment, slice-timing correction, co-registration, segmentation, normalization to MNI space, outlier detection (ART) and smoothing (8 mm FWHM). Denoising regressed motion parameters and tissue signals, scrubbed outlier frames (global signal threshold z>3; composite motion >0.5 mm); participants with >20% flagged frames would have been excluded, though all passed. Seed-to-voxel connectivity maps were computed using the right amygdala and bilateral thalamus (ROIs from the Harvard–Oxford atlas) with the whole brain as target. The primary voxelwise contrast was LSD > placebo, with cluster-size correction thresholded at P < 0.05 FDR after initial voxel threshold P < 0.001 uncorrected. ASL data were processed with FSL BASIL to generate CBF maps, globally normalised and analysed with paired t-tests within seed regions and gray-matter maps, using family-wise error (FWE) correction at P < 0.05. To relate connectivity changes to mood, the first eigenvariate of significant clusters was extracted and correlated with change scores (LSD minus placebo) on PANAS measured directly after scanning.

Twenty participants completed the study; most were in their mid-twenties with some prior drug experience. Subjective and physiological effects of 13 µg LSD were modest. The dose did not significantly alter 5D-ASC ratings. On placebo sessions 16 participants correctly identified placebo; on LSD sessions 8 believed they had placebo and the remainder reported a variety of other substances or uncertainty. Women reported earlier subjective onset than men (median onset 60 min vs 120 min). Peak change-from-baseline scores were analysed with repeated-measures ANOVA. Thalamic seed-based connectivity analyses showed increased thalamus–cerebellar connectivity under LSD in two cerebellar clusters (peak voxels: x=+26, y=-70, z=-32, k=226; and x=-20, y=-72, z=-32, k=189; P < 0.05 FDR). No significant thalamo-cortical or other cortical/subcortical thalamic connectivity changes were observed. These thalamic–cerebellar increases did not correlate with mood changes (all p > 0.22). Amygdala seed-based analyses revealed both increases and decreases in connectivity under LSD relative to placebo. Connectivity increases were found between the right amygdala and right angular gyrus, right middle frontal gyrus, and left cerebellum (cluster-level P < 0.05 FDR). Decreases in amygdala connectivity occurred with the left and right postcentral gyrus and the superior temporal gyrus. In an exploratory analysis, change in amygdala–middle frontal gyrus connectivity correlated positively with change in PANAS positive affect (LSD minus placebo; r = 0.49, p < 0.03). This correlation remained after controlling for blood pressure. No significant correlations were found for negative mood or for connectivity changes in other regions. ASL analyses included 18 participants (two excluded for incomplete or poor-quality data) and found no significant differences in cerebral blood flow between LSD and placebo within the seed regions or across gray matter (FWE-corrected).

Bershad and colleagues interpret their findings as evidence that a single very low, “microdose” of LSD (13 µg) alters resting-state functional connectivity in limbic and cerebellar networks despite producing negligible subjective effects and no detectable changes in cerebral blood flow. Specifically, the drug increased amygdala connectivity with right angular gyrus, right middle frontal gyrus and cerebellum, and decreased connectivity with somatosensory (postcentral) and superior temporal regions. The observed positive correlation between increased amygdala–middle frontal connectivity and increases in positive affect after LSD led the investigators to propose that modulation of cortico-limbic circuits might underlie mood-related effects, analogous in some respects to proposed mechanisms of SSRIs (acute dampening of amygdala reactivity and enhancement of cortico-limbic connectivity). The authors situate these results alongside prior work showing reduced amygdala reactivity and altered amygdala connectivity with higher doses of LSD and psilocybin, noting that detecting related connectivity changes at a dose that does not produce an altered state of consciousness is noteworthy. They also contrast the lack of thalamo-cortical effects here with prior findings at higher psychedelic doses, suggesting that microdoses may spare pathways implicated in the psychedelic experience and instead preferentially affect emotional networks. The paper acknowledges several limitations: the sample comprised healthy volunteers (so findings may not generalise to clinical populations), only a single acute administration was studied while real-world microdosing typically involves repeated dosing, behavioural tasks probing emotion or cognition were not employed, and effects were subtle which complicates attribution of mood–connectivity associations specifically to the drug. The authors conclude that further research — notably studies with repeated dosing and clinical samples — is needed to determine whether low-dose LSD can produce antidepressant effects and whether the observed connectivity changes are mechanistically related to mood improvement.