A split between reported vision and the visual task
Low doses of LSD changed how participants described their visual experience, but did not measurably change their performance on a behavioral test of visual contrast surround suppression. In the study, 20 micrograms produced reports of blurrier, more saturated and more dynamic vision, while the main contrast-suppression measure did not differ from placebo. The findings separate what participants felt they saw from what this particular task detected.
The researchers were testing several possibilities at once: that low-dose LSD might reduce or enhance surround suppression, that visual self-reports might track changes in the suppression score, and that the drug would produce acute general and visual effects.
How the visual test worked
Surround suppression was the study's measure of how visual context affects a contrast judgment. Participants completed a 100-trial task using two interleaved adaptive staircases to estimate a point of subjective equivalence, or PSE, meaning the reference contrast perceived as equivalent to the target. Because the target contrast was fixed at 0.5, the researchers used PSE minus 0.5 as the surround-suppression index.
The researchers fitted psychometric curves to the task responses, excluded fitted PSEs outside 0 to 0.5, and tested the drug conditions with a linear mixed-effects model, a method that accounts for repeated observations from the same participant.
The main contrast measure stayed put
The task produced a significant average suppression signal, but the comparison across placebo, 10 and 20 micrograms found no drug-condition effect, F(2, 36.87) = 0.47, P = .628. Study day and the interaction between study day and drug condition were also nonsignificant.
That null result survived checks for missing data. After session-quality exclusions, the primary task analysis retained 68 of 93 completed sessions from 24 participants, and missing-session rates did not differ significantly across drug conditions. When the researchers repeated the analysis using minimum and maximum imputation, they obtained the same result. A Bayesian comparison of models, one including a drug effect and one without it, produced a Bayes factor of 0.005, which the authors interpreted as very strong evidence against a drug-condition effect.
Self-reports changed, especially at 20 micrograms
Self-reports did move. On the study's visual rating scale, 20 micrograms increased any-drug-effect ratings by 17.12 millimetres versus placebo, with a 95% credible interval, the model's uncertainty range, of 11.42 to 23.19. Ratings for good drug effects rose by 11.95 millimetres, with an interval of 6.87 to 17.55, bad effects by 2.26 millimetres, with an interval of 0.64 to 4.98, and altered visual perception by 7.51 millimetres, with an interval of 4.23 to 11.54. The any-drug-effect rating also rose at 10 micrograms, by 4.93 millimetres, with an interval of 0.88 to 9.41.
On the visual-perception items, both doses shifted ratings toward blurrier perception: the change was -3.67 points at 10 micrograms, with a 95% credible interval of -6.91 to -0.88, and -7.22 points at 20 micrograms, with an interval of -11.34 to -3.72, relative to placebo. At 20 micrograms, participants also reported more saturation, by 2.63 points, and more dynamicity, by 3.65 points. Moving from 10 to 20 micrograms further increased saturation by 3.60 points and dynamicity by 2.39 points. The higher dose also raised three altered-consciousness dimensions, Oceanic Boundlessness, Visionary Restructuralization and Vigilance Reduction, compared with placebo. Visionary Restructuralization was higher at 20 than at 10 micrograms, while the direct comparison between the two doses for blurriness was not credible.
The two measures did not move together
The self-reported changes did not line up with the task score. Between the 20-microgram and placebo conditions, the correlation between reported visual effects and contrast decrement was -0.15, P = .519. The reported association was not statistically significant, so the data did not show a reliable link between the subjective and behavioral measures.
For this experiment, participants reported altered vision while the behavioral index remained unchanged, and the two measures did not correlate. The result is a divergence between self-report and task performance under the tested conditions.
The caveat: participants could often guess the dose
The experiment used a randomized, double-blind, within-subject crossover design. Participants received placebo, 10 micrograms and 20 micrograms in three separate sessions at least seven days apart, with equal randomization across six dosing sequences. The report lists 41 screened volunteers, 33 included participants, three replacements and 30 full-study completers. The analysis included 31 people who completed at least two sessions, 15 of them female.
The blinding was not fully successful. Treatment guesses were correct in 54% of sessions and 64% at the end of the study, above the 33.3% chance level. Placebo and 20 micrograms were identified above chance, while 10 micrograms was not; the lower dose was guessed correctly in 38% of session-end guesses and 50% of study-end guesses. Because participants could sometimes tell which condition they had received, expectancy-related influence on subjective reports remains possible.
The task hypotheses and planned analyses were preregistered before unblinding and group-statistic computation, but individual psychometric curves had already been inspected for data quality.
Study disclosures
No study-specific funding was declared. The authors reported that M.E.L. had consulted for Mind Medicine, Inc. and Lykos Therapeutics, while the other authors declared no conflicts. The underlying data are available on reasonable request to the corresponding author.
Paper data and sources
Original title: Low-dose lysergic acid diethylamide alters subjective visual perception but not visual contrast surround suppression.
Authors: Lucca F Jaeckel, Deborah Logvinski, Mika Kanana et al.
Journal/Repository: Neuroscience of consciousness
Status: Peer-reviewed
First online: 2026-08-20
DOI: 10.1093/nc/niag050
Original paper