Acute LSD effects on response inhibition neural networks

Classic hallucinogens such as lysergic acid diethylamide (LSD) and psilocybin have been used as pharmacological models to probe neurobiological mechanisms underlying psychotic symptoms, particularly visual hallucinations and cognitive disturbances. Activation of 5-hydroxytryptamine2A receptors (5-HT2A R) is thought to mediate the visual hallucinations produced by these substances, and prior work shows that 5-HT2A R antagonists can attenuate both hallucinations and certain cognitive impairments. The introduction frames cognitive control, particularly inhibitory processes, as central to distinguishing internally generated imagery from external perception, and notes clinical observations that patients with visual hallucinations (for example in Parkinson's disease) often show greater prefrontal/executive impairments than those without hallucinations. Schmidt and colleagues set out to test two linked hypotheses in healthy volunteers acutely given LSD: first, that subjective LSD-induced impairments in cognitive control would be positively related to reports of visual imagery; second, that LSD would reduce activation in prefrontal and related regions supporting response inhibition (including middle/superior/inferior frontal gyri, middle temporal gyrus, pre-supplementary motor area, anterior cingulate cortex and putamen) and that such reductions would be associated with LSD-induced visual imagery. The study therefore combines behavioural, subjective and fMRI measures using a Go/No-Go paradigm to probe neural networks of inhibitory control under LSD versus placebo.

The study was a double-blind, randomised, placebo-controlled, cross-over trial approved by the relevant ethics committee and registered on ClinicalTrials.gov. Twenty-four volunteers were recruited and screened with comprehensive medical and psychiatric exclusions; occasional recreational drug use (<10 lifetime uses) was permitted if no adverse reactions had occurred. After excluding participants for excessive motion during earlier scans and during the Go/No-Go task, the final analysed sample comprised 18 healthy adults (nine men, nine women; mean age 31 ± 9 years, range 25–58). Each subject completed two sessions at least 7 days apart, receiving oral placebo or 100 µg LSD at 09:00. The fMRI session was performed approximately 2.5–3 h after dosing to capture peak subjective and pharmacological effects. Subjective effects were assessed 3 h after intake using the 5 Dimensions of Altered States of Consciousness (5D-ASC) scale; the analysis focused on the 'impaired cognition and control' factor and the 'elementary' and 'complex imagery' factors, with item examples provided for each. Plasma LSD concentrations were measured at 0, 1, 2 and 3 h post-dose using validated liquid chromatography–tandem mass spectrometry; correlations between mean 2- and 3-h plasma concentrations and subjective effects were tested with Pearson correlations and Bonferroni correction (p < 0.017). Behavioural performance on an event-related Go/No-Go task was indexed by probability of inhibition on No-Go trials, number of correct Go responses and reaction time to Go trials. The task included 160 Go, 24 No-Go and 24 oddball trials, with oddball stimuli designed to control for novelty/attentional effects; total task duration was approximately 6 minutes. fMRI data were acquired on a 3 T Siemens scanner (EPI sequence; TR 2.5 s, TE 28 ms, 38 slices, voxel resolution 3 × 3 × 3 mm) producing 160 volumes. Preprocessing in SPM12 included realignment, normalisation to MNI space, smoothing with an 8 mm kernel and motion/artefact checks; volumes with >3 mm deviation were replaced by the average of neighbouring volumes and subjects with >10% corrupted volumes were excluded. First-level contrasts compared No-Go versus oddball trials (to control for attentional novelty), convolved with the canonical haemodynamic response and including motion regressors; group-level treatment differences (LSD v. placebo) used paired t tests. Significance used cluster-level family-wise error correction p < 0.05 with a cluster-forming threshold p < 0.001 uncorrected and an extent threshold of 20 voxels. Additional SPM analyses tested whether (a) the relationship between brain activation and probability of inhibition differed by treatment, and (b) brain activation during response inhibition related to plasma concentration or subjective 5D-ASC covariates.

Subjective ratings and plasma: LSD produced large increases over placebo on the 5D-ASC measures: impaired control and cognition (t17 = -8.007, p < 0.0001), elementary imagery (t17 = -9.099, p < 0.001) and complex imagery (t17 = -7.052, p < 0.001). After LSD, impaired control/cognition correlated positively with elementary imagery (r = 0.620, p = 0.006, corrected) but not with complex imagery (r = 0.235, p = 0.347). Mean plasma LSD concentrations were reported as 1.29 ng/mL (SD 0.722) at 1 h, 1.33 ng/mL (SD 0.59) at 2 h and 1.15 ng/mL (SD 0.52) at 3 h. Plasma concentration (mean of 2 and 3 h) correlated positively with visual imagery 3 h after administration (r = 0.56, p = 0.016, corrected) and showed a trend with cognitive impairments (r = 0.461, p = 0.054). Behavioural performance: Acute LSD impaired inhibitory performance. Probability of inhibition decreased from a placebo mean of 0.79 (SD 0.025) to 0.77 (SD 0.044) under LSD (t17 = 2.19, p = 0.043). LSD also reduced the number of responses to Go trials (placebo mean 151.00, SD 12.84; LSD mean 129.33, SD 24.83; t17 = 4.23, p = 0.001) and prolonged reaction times to Go trials (placebo mean 413.90 ms, SD 27.10; LSD mean 432.64 ms, SD 12.89; t17 = -3.42, p = 0.003). The probability of inhibition under LSD was negatively correlated with subjective impaired cognition/control (r = -0.523, p = 0.026). fMRI activation: Combined across treatments, the Go/No-Go task engaged frontal, striato-thalamic and cerebellar regions. Relative to placebo, LSD significantly decreased activation during response inhibition in several regions: right middle temporal gyrus, angular gyrus and inferior frontal gyrus; right anterior cingulate cortex; left postcentral gyrus; and left cerebellum (cluster-level p < 0.05 FWE, cluster-forming p < 0.001). The relationship between right parahippocampal activation and probability of inhibition differed markedly by treatment: under placebo the correlation was strongly positive (r = 0.88, p < 0.001), while under LSD there was no relationship (r = -0.06, p = 0.8). Brain–subject relationships: Severity of subjective impaired cognition correlated negatively with activation in the right middle frontal gyrus and bilateral superior frontal gyri. Elementary imagery severity correlated negatively with activation in the right precentral gyrus and left superior frontal gyrus. No significant relationships were found between plasma LSD concentration and brain activation during response inhibition.

Schmidt and colleagues interpret these findings as evidence that 5-HT2A receptor activation by LSD disrupts inhibitory processing via reduced parahippocampal‑prefrontal engagement, and that this disruption may promote the emergence of visual imageries. The authors note three linked observations that support this account: first, subjective reports of impaired cognitive control and visual imagery were positively associated after LSD; second, LSD impaired behavioural indices of inhibition and slowed/attenuated responses to Go trials; third, LSD reduced activation in regions implicated in error detection and top-down control (anterior cingulate cortex, parahippocampus, and dorsolateral prefrontal regions including superior/middle/inferior frontal gyri). They link the reduced anterior cingulate and parahippocampal activations to impaired error processing and a failure to generate corrective top-down signals to lateral prefrontal cortex, which in turn could impede learning from errors and updating of goal-directed behaviour. The disappearance of the positive relationship between parahippocampal activation and inhibitory performance under LSD is taken to indicate that LSD prevents the adaptive recruitment of additional resources following errors. Reduced activation in the left superior frontal gyrus was specifically associated with greater subjective cognitive impairments and visual imagery, and the authors suggest this could reflect enhanced internally generated representations at the expense of external sensory processing, thereby facilitating hallucination-like imagery. The discussion situates the results relative to clinical findings in schizophrenia and Parkinson's disease, where similar prefrontal and limbic dysfunctions have been linked to hallucinations and impaired inhibition. The authors acknowledge important limitations: the modest number of No-Go trials prevented separation of neural responses to successful versus failed inhibitions and limited connectivity analyses; the findings are correlational so causality between cognitive impairment and visual imagery cannot be established; LSD's strong subjective effects make blinding difficult and expectations might have influenced self-reports; and the brief task and fast event-related design constrain some inferences. They recommend future studies employ more extensive cognitive batteries, multiple prefrontal-mediated tests and designs better suited to probe connectivity and error-specific responses. In sum, the authors conclude that acute 5-HT2A R stimulation by LSD reduces parahippocampal‑prefrontal activation during inhibition, which may contribute to the formation of LSD-induced visual imageries and offers a neuropsychopharmacological insight relevant to visual hallucinations in neuropsychiatric disorders.