The 5-HT2A/1A Agonist Psilocybin Disrupts Modal Object Completion Associated with Visual Hallucinations

Kometer and colleagues frame the study within converging evidence that serotonergic receptors, particularly 5-HT2A and 5-HT1A, are highly expressed in visual cortical areas and have been implicated in visual disturbances and hallucinations in conditions such as Parkinson's disease and schizophrenia. Previous work shows that the indoleamine psilocybin produces visual distortions and hallucinations that phenomenologically resemble some clinical visual disturbances and that these effects are primarily mediated via 5-HT2A receptors. Despite this, the neural mechanisms by which 5-HT2A/1A receptor activation alters visual perception and gives rise to hallucinations remain unclear. To address this gap, the investigators used high-density electroencephalographic (EEG) recordings together with topographic mapping and source estimation to measure how psilocybin affects the spatiotemporal dynamics of modal object completion in healthy volunteers. Specifically, they tested the effects of two psilocybin dose conditions (reported in the extracted text as 125 and 250 g/kg) versus placebo during a Kanizsa figure task that elicits modal completion, and related electrophysiological changes to self-reported visual hallucination intensity assessed with the 5D-ASC scale. The aim was to identify early electrophysiological markers and cortical sources linking 5-HT2A/1A receptor activity to altered object completion and visual hallucinations.

Seventeen healthy, right-handed volunteers (eight males, nine females; mean age 28.8 ± 3.5 years) were recruited and screened to exclude current or past psychiatric disorders, substance dependence, and certain personality risk factors. Twelve participants reported prior experience with psilocybin or other hallucinogens. The study received ethics approval and participants provided written informed consent. A within-subjects, randomized crossover design was used: each participant completed three experimental sessions separated by at least two weeks and received placebo (lactose) and two graded doses of psilocybin administered in identical gelatin capsules. The extracted text reports doses as "low-dose" and "high-dose" (LD, HD) with numeric values given as 125 and 250 g/kg; the units in the extraction are unclear. Testing for the Kanizsa experiment began 120 minutes after drug ingestion. During the task participants viewed Kanizsa figures (illusory triangle induced by three ‘‘pacmen’’ inducers) and matched non-Kanizsa control stimuli while maintaining central fixation. Stimuli were presented at 1 m viewing distance, subtended 3.7° visual angle, and were shown in four orientations to reduce habituation. Each session comprised 176 trials (88 Kanizsa, 88 non-Kanizsa) presented pseudorandomly with a 3-sec stimulus-onset asynchrony in two blocks; stimuli remained on screen for up to 300 msec after a button press or 2000 msec maximum. Participants made speeded button-press judgements indicating whether they perceived an illusory triangle; response finger was counterbalanced across subjects. Subjective effects were measured with the Altered States of Consciousness (5D-ASC) questionnaire, with particular focus on the Visionary Restructuralization (VR) factor for visual alterations and the Auditory Alterations (AA) factor. EEG was recorded with a BioSemi ActiveTwo system using 64 scalp electrodes, horizontal and vertical electrooculograms, bandpass filtered at 0.01–67.0 Hz and digitized at 256 Hz. Offline EEG preprocessing re-referenced signals to the average reference, bandpass filtered at 1–40 Hz, and segmented into epochs from −150 to +600 msec relative to stimulus onset. Trials with artifacts exceeding ±100 μV were rejected; mean numbers of accepted epochs per condition are reported (around 79–85). Analyses included waveform quantification of P1 and N170 components at selected electrode pairs, topographic analysis using spatial k-means clustering to identify dominant scalp maps across time, and source estimation with standardized low-resolution electromagnetic tomography (sLORETA) over time windows defined by topographic stability. Statistical analyses used repeated-measures ANOVAs (dose, hemiscalp, electrode, stimulus as within-subject factors) with Bonferroni-corrected post hoc tests. Correlations between psilocybin-induced source-density changes and VR/AA scores were computed voxel-wise and assessed with nonparametric permutation testing to correct for multiple comparisons.

Psychometric effects: Repeated-measures ANOVA on 5D-ASC scores showed a strong main effect of dose [F(2,32)=36.596, p<.00001, η²=.70], a main effect of factor [F(4,64)=12.924, p<.00001, η²=.45], and a dose × factor interaction [F(8,128)=7.09, p<.00001, η²=.31]. On the Visionary Restructuralization factor (VR), both low-dose and high-dose psilocybin produced significant increases versus placebo (post hoc p<.00001); VR did not differ significantly between low and high dose (p=.178). Auditory Alterations increased only for high dose versus placebo (post hoc p<.05). Behavioural performance: Reaction times increased dose-dependently under psilocybin [F(2,32)=18.841, p<.00001, η²=.54], and responses were faster for Kanizsa than non-Kanizsa stimuli [F(1,16)=36.037, p<.0001, η²=.69]. Error rates rose slightly with dose [F(2,32)=3.3291, p<.05, η²=.17] but remained very low across conditions (placebo 0.88%, LD 1.18%, HD 1.92%), indicating task performance was preserved. Electrophysiology overview: Both stimulus types elicited prominent visual evoked potentials, including P100 (P1) and N170 components. Topographic k-means clustering across the 600-msec poststimulus period identified nine predominant scalp topographies explaining 96.74% of the variance. Three early stable periods were common across conditions and defined the measurement windows: 0–86 msec, 90–144 msec (P1 window), and 148–223 msec (N170 window). P1 (90–144 msec): Waveform analysis revealed a main effect of electrode [F(4,64)=33.42, p<.00001, η²=.68] and an electrode × dose interaction [F(8,128)=3.43, p<.01, η²=.18]. Post hoc tests showed that P1 amplitude at occipital electrodes O1/O2 was significantly increased relative to placebo for both low-dose (p<.05) and high-dose (p<.0001) psilocybin, with no reliable effects at other electrode sites. Source estimation localized P1 activity to V2 and lateral occipital complex (LOC) bilaterally across conditions; in the high-dose condition the increase extended to right V1. P1 source increases did not correlate with reported visual or auditory hallucinations (all p>.17 for visual, p>.98 for auditory). N170 (148–223 msec): Psilocybin produced a dose-dependent decrease in N170 amplitude [F(2,32)=17.170, p<.00001, η²=.52], with post hoc contrasts showing significant differences between each dose condition. A dose × hemiscalp interaction [F(2,32)=5.30, p<.05, η²=.25] and inspection of waveforms indicated a stronger reduction over the right hemiscalp. Consistent with prior findings, N170 magnitude was greater for Kanizsa versus non-Kanizsa stimuli [F(1,16)=35.863, p<.0001, η²=.70], but the Kanizsa–non-Kanizsa difference diminished dose-dependently under psilocybin. Topographic analyses indicated that the effects were primarily strength modulations rather than changes in scalp map topography. sLORETA source estimates placed N170-related activity in LOC and V2 bilaterally, with stronger right-lateralized LOC/V2 activation for Kanizsa stimuli. Psilocybin reduced current source density within LOC, V2, and fusiform gyrus across both stimulus types. Importantly, reductions in source density over right-lateralized LOC, V2, and posterior parietal areas during the N170 timeframe correlated positively with the increased intensity of visual hallucinations (VR scores); no significant correlations were found with auditory-alteration scores (all p>.64 for auditory).

Kometer and colleagues highlight three principal findings. First, psilocybin produced differential modulation of early visual components: a dose-dependent decrease of the N170 (148–223 msec) alongside a modest increase of the earlier P1 (90–144 msec) over occipital sites. Second, the N170 reduction was more pronounced for Kanizsa figures than for non-Kanizsa controls, implying an effect on processes underlying modal object completion. Third, the magnitude of psilocybin-induced reductions in right-lateralized extrastriate (LOC, V2) and posterior parietal activation during the N170 window correlated with self-reported visual hallucination intensity. The authors interpret the findings to implicate 5-HT2A/1A receptor activity in object completion processes. Because the N170 enhancement for Kanizsa figures is thought to reflect boundary completion and region-based segmentation, the preferential attenuation of the N170 for Kanizsa stimuli suggests psilocybin disrupts mechanisms critical for interpolating object boundaries. By contrast, the slight P1 enhancement may relate to an increase in perceived brightness commonly reported after indoleamine hallucinogens; P1 is known to be sensitive to stimulus brightness and other low-level features, and the P1 source increases localized to V2/LOC and, at high dose, right V1. Kometer and colleagues propose that decreased stimulus-evoked activation in modality-specific cortical areas during the N170 time window may characterise hallucinations more generally. They suggest a model in which internally generated visual content (hallucinations) competes with external sensory processing for neuronal resources, resulting in reduced evoked responses to external stimuli. The observed correlation between reduced posterior parietal activation and hallucinatory severity is taken to indicate impaired allocation of attentional resources to external stimuli under psilocybin. The authors relate these mechanisms to clinical observations in Parkinson's disease and schizophrenia, noting parallels such as increased 5-HT2A receptor density in patients with visual hallucinations and prior reports of reduced extrastriate activation and diminished N170 in some patient groups. Finally, the discussion notes pharmacological caveats: psilocybin is a mixed 5-HT2A/1A agonist and has downstream effects that could involve dopaminergic systems. Nevertheless, existing evidence (5-HT2A knockout mice and blockade of effects by the 5-HT2A antagonist ketanserin) supports a primary role for 5-HT2A receptors in psilocybin's psychotomimetic visual effects. The authors acknowledge that 5-HT1A receptors are also widespread in visual cortex and may contribute, and they call for further studies to disentangle the specific receptor contributions and mechanisms underlying the spatiotemporal dynamics of object completion and hallucinations.