Acute effects of lysergic acid diethylamide (LSD) on resting brain function

Lysergic acid diethylamide (LSD) is a potent hallucinogen that produces profound alterations in perception, emotion and cognition even at moderate doses. Historical psychiatric research in the 1950s and 1960s explored LSD both as a treatment adjunct in psychotherapy and as a pharmacological model for psychosis; research largely halted after prohibition but has seen a modern revival. A key unresolved question motivating renewed work is how LSD’s characteristic subjective effects map onto neuronal activity, particularly patterns of functional connectivity measured with functional magnetic resonance imaging (fMRI). Felix and colleagues frame the present paper as a focused review of recent fMRI studies in healthy volunteers that examined resting-state functional connectivity following moderate LSD dosing (oral 100 µg or intravenous 75 µg). The authors restrict their scope to resting-state analyses (no task fMRI) and concentrate on measures of functional connectivity—that is, correlations of activity between brain regions—reporting and interpreting findings from three clinical trials (sample sizes reported in the extracted text ranged from 15 to 24 subjects).

This paper is a narrative review that summarises results from three clinical fMRI trials of LSD in healthy participants. All included studies administered moderate doses expected to produce typical subjective effects (100 µg orally or 75 µg intravenously) and acquired resting-state fMRI while participants were not engaged in tasks. Sample sizes in the examined trials were relatively small (range 15–24), and all subjects were healthy volunteers; additional sample details are referenced as being reported in a table in the original text but are not reproduced here. The reviewed analyses examined complementary connectivity metrics. One set of analyses probed within-network connectivity (synchrony among regions forming canonical resting-state networks) and between-network connectivity (correlations across different networks). Another approach used global functional connectivity (GFC), a data-driven measure defined as the mean correlation of each region with all other regions; the reviewers note that preprocessing choices such as global signal regression materially affected some GFC results. Comparisons are made across the three LSD studies and against related psilocybin fMRI studies. The authors also consider methodological confounds relevant to pharmacological fMRI, including head motion, physiological changes (heart rate, blood pressure, temperature) and potential effects on neurovascular coupling. Where specific analytic choices influenced outcomes, the review describes those differences: for example, rerunning an analysis without global signal regression yielded substantial changes in regional connectivity results in one study. The extracted text indicates that figures and tables in the original paper summarise network-level and region-level findings, but those items are not reproduced in this extraction.

Across the reviewed resting-state fMRI studies, a recurring pattern was decreased connectivity within several canonical resting-state networks alongside increased connectivity between networks. Carhart‑Harris and colleagues reported reductions in within-network synchrony for multiple networks and concurrent increases in cross-network connectivity; the initial study also found a significant correlation between reduced default mode network (DMN) connectivity and the subjective experience labelled "ego dissolution." Felix and colleagues attempted replication and confirmed the broad pattern of decreased within-network connectivity and increased between-network connectivity, but found inconsistencies in which specific networks or network pairs were affected. The reviewers also note that similar within-network decreases were observed after administration of a selective serotonin reuptake inhibitor in another study, raising questions about specificity to hallucinogens. Using global functional connectivity, several studies reported widespread increases in mean regional connectivity across frontal, parietal and temporal cortical areas. Tagliazucchi and colleagues specifically observed increased GFC in the precuneus and thalamus. However, an analysis repeated without global signal regression shifted the pattern of effects in one dataset, producing decreased connectivity in the right insula and increased connectivity in the cerebellum, basal ganglia and thalamus, illustrating sensitivity to preprocessing. Preller and colleagues described a pattern of hyperconnectivity in somatosensory regions combined with hypoconnectivity in associative areas and speculated this imbalance could impair sensory integration. Comparisons with psilocybin fMRI work revealed both overlap and divergence. A London group found widespread between-network increases after intravenous psilocybin that broadly resembled LSD findings, and a separate psilocybin GFC analysis also reported increased connectivity in regions including the precuneus and thalamus. Nevertheless, Felix and colleagues emphasise only limited agreement overall between psilocybin and LSD studies. Finally, the reviewers report that attempts to link connectivity changes robustly to subjective effects have been inconsistent: the initial DMN–ego dissolution correlation was not replicated in Felix’s follow-up study, and other expected brain–experience associations were not found.

The authors interpret the aggregated findings as indicative that LSD acutely alters brain-wide functional connectivity, with some reproducible elements such as decreased connectivity within certain resting-state networks (notably visual networks, DMN and sensorimotor network) and increased connectivity in specific regions including the thalamus. They highlight the thalamocortical system as a plausible mechanistic locus: increased thalamic connectivity would accord with models proposing that hallucinogens disrupt thalamic gating of sensory and internal information, thereby increasing cortical information flow and subjective awareness. Felix and colleagues discuss theoretical frameworks linking acute connectivity changes to potential therapeutic effects. One model proposes that destabilisation of hub connectivity by hallucinogens induces a transient state of increased ‘‘disorder’’ that allows reorganisation of pathological connectivity patterns; new patterns may then become stabilised after the acute drug effects subside, possibly aided by anti‑inflammatory processes. The reviewers note that some experimental fMRI evidence points to lasting connectivity changes after hallucinogen exposure, which could be compatible with these hypotheses, but emphasise that causal chains remain speculative. Several key limitations are acknowledged. Small sample sizes constrain reliability and generalisability. Pharmacological fMRI faces multiple potential biases: LSD can alter head motion and systemic physiology (heart rate, blood pressure, temperature), and it may affect vascular tone or neurovascular coupling; each of these can confound connectivity measures. The extracted text notes that one study reported greater head motion under LSD, and that physiological monitoring in one investigation was limited to a single pre-scan measurement rather than continuous recording during scanning. The authors therefore call for more rigorous control and measurement of these confounds in future work. In terms of safety, the review summarises historical and contemporary evidence suggesting that, under careful screening and supervision, LSD can be administered safely in research settings. Serious adverse events such as prolonged psychosis or hallucinogen persisting perception disorder are described as rare in the literature cited; modern studies excluded individuals with personal or family histories of psychotic disorders and, according to the extracted text, did not report severe adverse psychological reactions.

Felix and colleagues conclude that acute LSD administration is associated with widespread alterations in resting-state functional connectivity, with some consistent observations including decreased within-network connectivity in several resting-state networks and increased connectivity in regions such as the thalamus. They caution that many findings require replication, that links between connectivity changes and subjective or therapeutic effects remain unproven, and that methodological confounds—especially pharmacologically induced changes in physiology and neurovascular coupling—need to be more systematically addressed in future studies.