Enhanced visual contrast suppression during peak psilocybin effects: Psychophysical results from a pilot randomized controlled trial

Abstract

In visual perception, an effect known as surround suppression occurs wherein the apparent contrast of a center stimulus is reduced when it is presented within a higher-contrast surrounding stimulus. Many key aspects of visual perception involve surround suppression, yet the neuromodulatory processes involved remain unclear. Psilocybin is a serotonergic psychedelic compound known for its robust effects on visual perception, particularly texture, color, object, and motion perception. We asked whether surround suppression is altered under peak effects of psilocybin. Using a contrast-matching task with different center-surround stimulus configurations, we measured surround suppression after 25 mg of psilocybin compared with placebo (100 mg niacin). Data on harms were collected, and no serious adverse events were reported. After taking psilocybin, participants (n = 6) reported stronger surround suppression of perceived contrast compared to placebo. Furthermore, we found that the intensity of subjective psychedelic visuals induced by psilocybin correlated positively with the magnitude of surround suppression. We note the potential relevance of our findings for the field of psychiatry, given that studies have demonstrated weakened visual surround suppression in both major depressive disorder and schizophrenia. Our findings are thus relevant to understanding the visual effects of psilocybin, and the potential mechanisms of visual disruption in mental health disorders.

Introduction

In visual perception, the appearance of a given object is influenced by neighboring objects (Schwartz, Hsu, & Dayan, 2007). These interactions, known as spatial context effects, produce many well-known visual illusions (Ebbinghaus, 1902; Fraser, 1908; Gibson & Radner, 1937). Surround suppression of apparent contrast—when the perceived contrast of a target stimulus is reduced (suppressed) by the presence of higher-contrast surrounding stimuli—has been robustly demonstrated using psychophysics (Cannon & Fullenkamp, 1991; Chubb, Sperling, & Solomon, 1989; Ejima & Takahashi, 1985; Olzak & Laurinen, 1999; Petrov & McKee, 2006; Schallmo & Murray, 2016; Snowden & Hammett, 1998; Xing & Heeger, 2000; Xing & Heeger, 2001; Yu, Klein, & Levi, 2001). However, the neurochemical basis of this phenomenon is incompletely understood. It remains unclear which neuromodulator systems are critically involved in surround suppression. One approach to investigate the role of neuromodulators in visual perception is to use drugs as pharmacological probes in combination with visual tasks. Previous inquiries into surround suppression have probed GABAergic pathways with compounds such as ethanol (Read et al., 2015) and lorazepam (Schallmo et al., 2018); cholinergic pathways with donepezil (Gratton et al., 2017; Kosovicheva, Sheremata, Rokem, Landau, & Silver, 2012); adenosine pathways with caffeine (Nguyen et al., 2018); dopaminergic pathways with bromocriptine (Gratton et al., 2017); and noradrenergic pathways with guanfacine (Gratton et al., 2017). 

The psychedelic drug psilocybin (4-phosphoryloxy-N,N-dimethyltryptamine) is a naturally-occurring tryptamine compound found in at least 144 different species of mushrooms worldwide (Guzmán, 2005; Guzmán, Allen, & Gartz, 1998). Psilocybin, which is rapidly metabolized into psilocin in the body, is a serotonin 2A (5-HT_2A) receptor agonist that can alter perception, emotion, cognition, and sense of self (Hirschfeld & Schmidt, 2021; Holze et al., 2022; Studerus, Gamma, & Vollenweider, 2010). Visual effects (“psychedelic visuals”) are the most common and prominent dose-dependent subjective effects of psilocybin (Aday, Wood, Bloesch, & Davoli, 2021; Bayne & Carter, 2018; Carbonaro, Johnson, Hurwitz, & Griffiths, 2018; Hirschfeld & Schmidt, 2021; Holze et al., 2022; Studerus et al., 2010; Vollenweider & Preller, 2020). Relatively few studies have investigated psilocybin using visual psychophysical methods (Aday et al., 2021). Psilocybin appears to impact only select visual processes, while others are left intact [e.g., impairing global but not local motion detection; Carter et al. (2004); Kometer, Cahn, Andel, Carter, & Vollenweider (2011); Carbonaro et al. (2018); Aday et al. (2021)]. The phenomenology of psilocybin's visual effects can include alterations in the appearance of textures, the shapes/sizes of objects, motion, and colors—the visual field can take on a kind of fluidity, exhibiting animated dynamics described as pulsing, glowing, breathing, shifting, and dancing (Swanson, 2018). Given this visual phenomenology, it is plausible that center-surround interactions are in some way impacted by psilocybin, as surround suppression plays a critical role in certain visual processes (Angelucci et al., 2017; Nurminen & Angelucci, 2014), including visual saliency and pop-out (Knierim & van Essen, 1992), perception of object boundaries (Nothdurft, Gallant, & Van Essen, 2000), perceptual constancies (Allman, Miezin, & McGuinness, 1985), figure-ground segmentation (Lamme, 1995; Supèr, Romeo, & Keil, 2010), contour integration (Field, Hayes, & Hess, 1993; Hess & Field, 1999; Kapadia, Ito, Gilbert, & Westheimer, 1995; Polat, Mizobe, Pettet, Kasamatsu, & Norcia, 1998), and motion detection (Jones, Grieve, Wang, & Sillito, 2001). Michaiel, Parker, & Niell (2019) found reduced surround suppression in mice after administration of the psychedelic drug DOI. Relatedly, Azimi et al. (2020) provide evidence for the role of 5-HT_2A in mouse visual gain control. To our knowledge, no previous study has measured visual contrast suppression under psilocybin or any other serotonergic psychedelic in humans. The present study aims to address this gap in knowledge. 

Psilocybin-assisted therapy has rapid and enduring antidepressant effects in patients with major depressive disorder (MDD) (Carhart-Harris, Bolstridge, et al., 2018; Davis et al., 2021; Goodwin et al., 2022; Griffiths et al., 2016; Gukasyan et al., 2022; Ross et al., 2016; von Rotz et al., 2023). Visual surround suppression has been found to be weaker in patients with MDD (Golomb et al., 2009; Salmela et al., 2021). The neurochemical bases of weakened surround suppression in MDD are currently unknown. The role of serotonin function in MDD is widely acknowledged but not well understood (Carhart-Harris & Nutt, 2017; Cowen & Browning, 2015). Examining the impact of psilocybin on surround suppression could thus address multiple knowledge gaps in our understanding of the links between MDD, serotonin, and visual surround suppression. 

Despite its therapeutic promise, psilocybin's subjective effects have paradoxically (Osmond & Smythies, 1952) been characterized as “a schizophrenia-like psychosis” (Vollenweider et al., 1998, p. 3897). This claim has remained controversial for nearly a century (Swanson, 2018) and has been called into question by recent studies (Leptourgos et al., 2020). Multiple investigations have found weakened visual surround suppression in people with schizophrenia compared with healthy controls (Dakin, Carlin, & Hemsley, 2005; Linares et al., 2020; Pokorny, Lano, Schallmo, Olman, & Sponheim, 2019; Schallmo, Sponheim, & Olman, 2015; Serrano-Pedraza et al., 2014; Tadin et al., 2006; Tibber et al., 2013; Yang et al., 2012; Yoon et al., 2009). Understanding the impact of psilocybin on surround suppression could thus illuminate the debates about the similarities, or lack thereof, between psychedelics and psychosis, as well as the possible role of serotonergic neuromodulation in schizophrenia. 

The present study aimed to test a psychophysical contrast-matching task to examine suppression of perceived contrast using different center-surround stimulus configurations under peak effects of a high dose of psilocybin in healthy human participants, compared to placebo control. The data reported here are results from a small pilot feasibility study in advance of a larger trial. 

Data on harms were collected, and no serious adverse events were reported. We hypothesized that psilocybin would yield weaker visual surround suppression, like what is observed among individuals with psychotic disorders. Surprisingly, our findings instead suggest that psilocybin may selectively enhance surround suppression during visual contrast perception. 

Methods

This triple-blind, randomized, placebo-controlled clinical trial was conducted at the University of Minnesota in Minneapolis, Minnesota. Experimental protocols were approved by the Institutional Review Board of the University of Minnesota (STUDY00009765/SITE00000856/Pro00045074). All experiments were performed in accordance with approved guidelines and regulations, including approval from the United States Food and Drug Administration and Drug Enforcement Administration (clinicaltrials.gov study no. NCT04424225). 

Participants

Adults aged 25 to 38 years with no current or previous mental health diagnosis, not currently using any prescription medications, with at least one reported previous experience taking a moderate to high dose of psilocybin, and without histories of psychotic disorder were eligible to participate. Six healthy participants with normal or corrected-to-normal vision (three male and three female, mean age 33 years; see Supplemental Table S1 for full demographic information) were included in this study. This sample size was chosen to demonstrate feasibility for our pilot study and was informed by results from previous studies of surround suppression in healthy adults (e.g., Schallmo & Murray, 2016). All participants provided written informed consent before participation and were compensated at a rate of $20 to $30 per hour. No participants withdrew from the study due to harms. 

Crossover design

Participants completed two drug dosing sessions—one day with active drug (25 mg synthetic psilocybin); one day with placebo (100 mg niacin)—two weeks apart to achieve drug washout. Study drug and placebo control were provided by Usona Institute (Madison, WI, USA) through their drug supply program. Doses of psilocybin and placebo were administered in identical opaque gelatin capsules in the setting of a comfortable, decorated hospital room, with study visits lasting approximately nine hours. Participants were randomized into groups where group A received psilocybin on dosing day 1 and placebo on dosing day 2, whereas group B received the drugs in the reverse order. Participants, experimenters, and all study staff were blind to which drug was administered on any given experimental session. Participants were informed that they would be receiving 25 mg psilocybin during one session and niacin (placebo) during the other–with a 50% chance of getting psilocybin on the first versus the second session–and that the drug administration was double-blind. After both dosing sessions were complete, participants were asked to guess which drug was given at each of their dosing sessions. See Supplemental Methods for additional details on our blinding procedure. 

Apparatus

Visual stimuli were presented on a Dell P2319H 23-in LED-backlit monitor (1920 × 1080 pixels, 60 Hz refresh rate) at a viewing distance of 60 cm. The monitor was calibrated using a spectrophotometer. Mean luminance was 111.2 cd/m². Stimuli were generated and presented using PsychoPy (Peirce et al., 2019) version 2021.2.3. 

Stimuli

Stimuli consisted of annulus sinusoidal luminance modulation gratings presented on a mean gray background at 3° eccentricity to the left and right of fixation. The central fixation mark was constructed according to Thaler, Schütz, Goodale, & Gegenfurtner (2013) (shown in Figure 1). The contrast of one grating (target) was fixed at 50%, whereas the contrast of the other grating (reference) varied across trials using an adaptive staircase (detailed in section Paradigm). Both gratings had an outer diameter of 2° and a spatial frequency of 1.1 cycles/° and always shared the same orientation, which varied from trial-to-trial in counterbalanced randomized instances of 0°, 45°, 90°, or 135° orientation. In a subset of stimulus conditions (see Supplemental Table S2 and Figure 1), the target grating was presented within a larger (4° outer diameter) annular grating, referred to as the surround. The contrast of the surround was fixed at 100% and its spatial frequency was the same as the target grating. Surround orientation was either 0° (parallel) or 90° (orthogonal) relative to the target. A 0.5° mean luminance “gap” was placed between the target grating and the surround grating to prevent brightness induction effects (Schallmo et al., 2015; see Yu et al., 2001), as well as to help the observers attend only to the central grating in making their judgments (Xing & Heeger, 2001). The reference grating was always presented with no surround. 

Figure 1.

 

Samples of stimuli used in the experiment. (A) An example of a trial from the No-Surround staircases. (B) An example of a catch trial, in which the reference stimulus appeared with 80% Michelson contrast. (C–D) Example trials from the Orthogonal Surround staircases. (E–F) Example trials from the Parallel Surround staircases.

Figure 2.

 

Surround suppression under psilocybin (blue) and placebo (orange), shown as the difference between observers’ PSEs and the veridical contrast of the target stimulus (PSE − 50.0%) for the No Surround (NS), Orthogonal Surround (OS), and Parallel Surround (PS) stimulus conditions. (A) Suppression for individual observers in each stimulus condition. (B) Values across participants. Error bars show the standard error of the mean (SEM) calculated within subjects, according to the method of Morey (2008).

There were three experimental conditions, defined by the surround configuration: No Surround (NS), Orthogonal Surround (OS), and Parallel Surround (PS). See Figure 1 and Supplemental Table S2 for details. 

Paradigm

Participants completed the following visual tasks after a 3-hour wash-in period post drug administration. We chose the 3-hour time point because it roughly corresponds to the latter half of peak plasma concentration (Brown et al., 2017; Holze et al., 2022) and subjective effects (Carbonaro et al., 2018; Holze et al., 2022) of psilocybin, allowing participants to acclimate to peak subjective effects before starting the task. Participants did not perform other experimental tasks prior to the contrast-matching task. During the approximately three hours in between taking the drug and completing the task, participants were encouraged to relax and rest on a bed with eyeshades. Most participants simply relaxed and listened to music or conversed with study staff during this period. 

For our primary outcome measure, we used a two-alternative forced-choice visual psychophysics task designed to measure the point of subjective equality (PSE) at which the contrast of the target and reference gratings appeared equivalent to the observer. On each trial, participants were asked to report which circular grating appeared higher in contrast (target or reference, ignoring the surround). Each trial began with a fixation mark presented alone for 500 ms, after which the grating stimuli appeared on either side of fixation. Each trial ended when the participant made a key press to indicate which stimulus was higher contrast. Stimulus duration and response time were not limited. 

Two staircases per condition (target left and target right; 40 trials each) were presented in a randomized, intermixed order. The adaptive staircase (implemented using PsychoPy's (Peirce et al., 2019) default StairHandler() class configured with a one-up, one-down rule) adjusted the contrast of the reference grating on every trial in order to converge on the contrast level at which the observer reported that the reference grating was higher contrast 50% of the time (i.e., the PSE). The initial contrast of the reference grating for each staircase was 50% (the same as the target grating). 

To ensure task comprehension, subjects verbally confirmed their understanding of the task instructions and completed seven practice trials (one sample trial from each staircase plus one catch trial) before beginning the main experiment. In addition, 40 catch trials (large contrast difference, low difficulty) were included to assess off-task performance. Reference contrast in the catch trials was fixed at 80% (i.e., 30% higher than the target on all catch trials, no staircase was used). Catch trial response accuracy was analyzed to ensure that the participants understood and performed the task correctly. Total task duration was approximately 30 minutes. 

Subjective drug effects

For our secondary outcome measure, the Altered States of Consciousness questionnaire (5D-ASC) scale (Dittrich, 1998; Studerus et al., 2010) was used to assess the overall effects of drug on various elements of subjective experience. The 5D-ASC is a well-validated 94-item self-rating visual analog scale that assesses five “dimensions” of altered states of consciousness: (1) Oceanic Boundlessness, (2) Anxious Ego Dissolution, (3) Visionary Restructuralization, (4) Auditory Alterations and (5) Vigilance Reduction (Dittrich, 1998). Studerus et al. (2010) used confirmatory factor analysis to identify 11 subscales of 5D-ASC (11D-ASC): (1) Experience of Unity, (2) Spiritual Experience, (3) Blissful State, (4) Insightfulness, (5) Disembodiment, (6) Impaired Control and Cognition, (7) Anxiety, (8) Complex Imagery, (9) Elementary Imagery, (10) Audio-Visual Synesthesia, and (11) Changed Meaning of Percepts. Both the 5D-ASC and the 11D-ASC subscales are routinely used to assess subjective effects in studies that use psychedelic drugs (Carbonaro et al., 2018; Hirschfeld & Schmidt, 2021; Holze et al., 2022; Holze et al., 2020; Holze et al., 2019; Liechti, Dolder, & Schmid, 2017; Preller et al., 2017; Studerus et al., 2010). Importantly, the 5D-ASC detects the unique mind- and perception-altering effects of serotonergic psychedelic drug effects in particular, as it is sensitive enough to distinguish these from other psychoactive effects such as those produced by D-amphetamine or MDMA (Holze et al., 2020). Of particular interest to the present study is the 5D-ASC Visionary Restructuralization dimension and its 11D-ASC subscales (8) Complex Imagery, (9) Elementary Imagery, (10) Audio-Visual Synesthesia, and (11) Changed Meaning of Percepts, which are comprised of questions that assess alterations to visual experience. All participants in this study completed all 94 items of the 5D-ASC questionnaire to retrospectively rate drug effects three or four days after their psilocybin and placebo dosing sessions. 

Statistical analyses

The point of subjective equality (PSE) for each observer was calculated independently for each of the six staircases by fitting a Logistic function to the response data. We obtained the PSE threshold using the data.FitLogistic() and data.functionFromStaircase() functions provided by PsychoPy (Peirce & MacAskill, 2018). Guess rate (the proportion of responses considered guesses unrelated to stimulus levels) and lapse rate (the proportion of responses considered unintentional or erroneous and unrelated to stimulus levels) were both set to 4%, consistent with the literature on 2AFC designs and our prior work in human clinical populations (Kingdom & Prins, 2010; Schallmo et al. 2015; Schallmo et al. 2020). Surround suppression illusion strength was quantified by subtracting the veridical contrast of the target stimulus grating (50%) from the PSE value. Data from placebo and psilocybin sessions were analyzed separately, and then compared within participants. 

We performed two complementary analyses to examine differences in surround suppression between placebo and psilocybin sessions. The first was performed in a double-blinded manner; participants and study staff were blinded to which compound was received during each dosing session (i.e., psilocybin vs. placebo), and statistical analyses were performed after unblinding. In this first analysis, differences in PSE values between psilocybin and placebo were compared in a repeated measures analysis of variance (ANOVA) using drug and stimulus conditions as within-subjects factors. 5D-ASC ratings between psilocybin and placebo were compared in a repeated measures ANOVA using drug and ASC subscale/dimension as within-subjects factors. Both ANOVAs were followed-up with post hoc pairwise t-test comparisons. Effect size is reported as n_G² (for ANOVAs) and Bayes factor (for t-tests). 

Correlations between PSE and 5D-ASC ratings were calculated using the rmcorr technique (Bakdash & Marusich, 2017; Bland & Altman, 1995a; Bland & Altman, 1995b) and followed up with pairwise biweight midcorrelation tests (Langfelder & Horvath, 2012) using the difference (psilocybin minus placebo) for threshold and rating score pairs for each participant. We chose these methods because they do not require first averaging the data and avoid violating independence assumptions, making them ideal (and more sensitive) for paired repeated-measures data (Bakdash & Marusich, 2017). 

The first set of statistical analyses were performed using the Pingouin package for Python (Vallat, 2018). The second set of analyses was performed for the triple blind by an independent biostatistician who was blinded to the hypotheses and the identity of the two compounds (i.e., psilocybin vs. placebo). PSE values for psilocybin versus placebo sessions were compared using a linear mixed-effect model, assuming homogeneous variance across surround conditions and dosing sessions. This model was selected as the best fitting over alternative variance structures. Catch trials and associations with subjective experience ratings (e.g., 5D-ASC) were not examined in our second analysis. These analyses were conducted using R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria). The issue of double versus triple blinding is discussed further in the Limitations section of the Discussion. 

Results

To examine the impact of psilocybin on surround suppression of perceived contrast (primary outcome), we asked participants (n = 6) to compare target gratings (fixed contrast, with or without surrounding stimuli) to reference gratings (variable contrast) presented on opposite sides of a central fixation point, and report which appeared higher contrast. We measured PSEs in observers under psilocybin and placebo in three stimulus conditions (no surround [NS], orthogonal surround [OS], and parallel surround [PS]; see Methods). In addition, we measured the subjective effects of psilocybin and placebo using the 5D-ASC psychometric instrument. The following subsections provide results from the different comparisons and correlations across these data points. 

Effect of drug on PSE

In our first analysis (double-blind, see Methods), repeated measures ANOVA performed on the PSE values revealed a main effect of drug that reached significance (F_1,5 = 7.294, p = 0.043, n_G² = 0.128). The main effect of stimulus condition reached significance, consistent with surround suppression (F_2,10 = 8.822, p = 0.006, n_G² = 0.381). The interaction effect of drug × condition did not reach significance; however, a notable trend effect was observed (F_2,10 = 3.0, p = 0.095, n_G² = 0.056), such that PSE values appeared lower for psilocybin vs. placebo in the OS and PS conditions. Supplemental Table S3 shows all ANOVA results. 

A Shapiro-Wilk test did not show evidence of non-normality in the placebo (W = 0.983, p = 0.848) or the psilocybin (W = 0.971, p = 0.459) PSE data. 

Subsequent post hoc pairwise t-test (two-sided) comparisons, in which the PSEs from all stimulus conditions were compared between psilocybin and placebo confirmed the significant ANOVA main effect of drug on PSE (t₅ = 2.701, p = 0.043, Bayes factor = 2.342). Further t-test comparisons were performed using PSEs from each stimulus condition. For the No Surround (NS) condition, drug did not have a significant effect on PSE, suggesting that psilocybin did not significantly impact general contrast perception when stimuli were presented without a surround (t₅ = 0.665, p = 0.535, Bayes factor = 0.446). Post hoc comparisons revealed a trend effect in the OS condition (t₅ = 2.467, p = 0.057, Bayes factor = 1.905) and a trend effect in the Parallel Surround (PS) condition (t₅ = 2.268, p = 0.073, Bayes factor = 1.594). 

Our second analysis (triple-blind, see Methods) provided partial support for the findings in our first analysis. Specifically, a linear mixed-effects model showed a marginally significant effect of psilocybin on PSEs (mean = 0.045, SE = 0.025, t₅₉ = 1.80, p = 0.077). Follow-up contrasts showed that this was driven by marginally lower PSE values for psilocybin vs. placebo in the PS condition (mean = 0.075, SE = 0.043, t₅₉ = 1.73, p = 0.088), whereas differences in the NS condition (mean = 0.006, SE = 0.043, t₅₉ = 0.145, p = 0.9) and OS condition (mean = 0.053, SE = 0.043, t₅₉ = 1.24, p = 0.2) were not significant. 

Mean PSE data from psilocybin compared with placebo sessions suggest that psilocybin enhanced surround suppression. Mean PSE values were more negative under psilocybin (i.e., the suppression of perceived contrast was stronger, observers perceived contrast less veridically, compared with placebo) for the OS and PS, but not for the NS, stimulus conditions. Figure 2 displays the magnitude of illusory suppression of perceived contrast (defined as the PSE minus the target contrast of 50.0% Michelson contrast). 

Figure 3.

 

Within-subjects effects of the drug were calculated by subtracting the PSE measured under placebo from the PSE measured under psilocybin for each stimulus condition. (A) The effect of psilocybin on PSE for each individual observer. (B) The mean effect for each stimulus condition. Error bars represent standard error of the mean (SEM).

We quantified the effect of psilocybin on PSE for each stimulus condition as the psilocybin PSE minus the placebo PSE. The mean effect of psilocybin was a shift in the PSE of −7.48% contrast in the parallel surround (PS) condition and −5.34% contrast in the orthogonal surround (OS) condition. By comparison, when stimuli were presented with no surrounds (NS), the difference in the PSE from placebo to psilocybin was −0.63%. Figure 3 summarizes our results in terms of the effect of psilocybin on PSE (drug − placebo) for each stimulus condition. These results suggest that psilocybin enhanced the magnitude of visual surround suppression whereas general contrast discrimination remained unaffected by the drug. 

Catch trials

Each observer completed 40 ‘catch’ trials intermixed with all the other trails under both placebo and psilocybin (see Methods, Figure 1F). Catch trials—contrast discrimination probes with very low difficulty—were used to measure whether the observer was performing with sufficient attention, understanding of the task, and ability to judge visual contrast at a basic level. Catch trial accuracy was high in both placebo (mean = 99.17%, SD = 1.29%; Figure 4) and drug sessions (mean = 97.92%, SD = 1.88%; paired t-test (two-sided), t₅ = 2.24, p = 0.08; unblinded analysis)—suggesting that psilocybin did not significantly interfere with the participants’ ability to perform the task correctly. 

Figure 4.

 

Catch trial scores. Values represent the percentage of correct responses to 40 catch trials. (A) Catch trial scores for each individual observer. (B) Mean catch trial scores. Error bars represent standard error of the mean (SEM).

Subjective effects of drug

Psilocybin produced robust effects on subjective experience, as indicated by a significant main effect of drug (F_1,5 = 20.109, p = 0.006, n_G² = 0.518; unblinded analysis) in a repeated measures (drug × subscale) ANOVA on 5D-ASC score. There was also a significant main effect of subscale (F_4,20 = 14.051, p < 0.001, n_G² = 0.364) and a significant drug × subscale interaction (F_4,20 = 8.484, p < 0.001, n_G² = 0.342). 

Subjective drug effects and PSE

Next, we asked whether the magnitude of surround suppression correlated with the intensity of the psychedelic effects reported by the participants (secondary outcome). To assess correlation between rating scale scores and PSEs, we used repeated measures correlation (rmcorr), “a statistical technique for determining the common within-individual association for paired measures assessed on two or more occasions for multiple individuals” (Bakdash & Marusich, 2017; Bland & Altman, 1995a; Bland & Altman, 1995b).¹ We performed rmcorr analyses by pairing each participant's PSE thresholds with their 5D-ASC ratings of subjective effects from each dosing session. Rmcorr allowed us to assess the links between surround suppression and particular subjective drug effects in a drug-agnostic fashion, regardless of which drug (psilocybin or placebo) induced the effects. 

On the 5D-ASC scale, rmcorr revealed statistically significant inverse correlations between PSEs and 5D-ASC scores for the data points (OS, Visionary Restructuralization, r₅ = −0.856, p = 0.014 and PS, Visionary Restructuralization, r₅ = −0.836, p = 0.019; unblinded analysis; Figure 5)—higher rating scores (more intense subjective drug effect) were associated with greater surround suppression (lower PSEs). At the level of 5D-ASC dimensions, PSE did not significantly correlate with any other type of subjective effects, nor with the overall Total Score. Post-hoc t-test results for 5D-ASC rating scores are shown in Supplemental Table S5. 

Figure 5.

 

Rmcorr plot showing each individual's 5D-ASC “Visionary Restructuralization” dimension scores from placebo and psilocybin as a function of their PSE thresholds in each stimulus condition. The lines show the rmcorr slope (1 slope calculated across all participants) as it intercepts with each participant's placebo (triangle marker) and psilocybin (circle marker) sessions. Rmcorr plots for 11D-ASC dimensions are shown in Supplemental Figures 1 and 2.

We performed an additional rmcorr analysis using the 11D-ASC scoring method (Studerus et al., 2010), which includes eleven lower-level subscales comprised of items from three of the five dimensions of the 5D-ASC scale (see Methods). This un-blinded analysis revealed significant inverse correlations at the data points PS, Audio-Visual Synesthesia, r₅ = −0.942, p = 0.001, OS, Elementary Imagery, r₅ = −0.824, p = 0.023, OS, Audio-Visual Synesthesia, r₅ = −0.822, p = 0.023, PS, Elementary Imagery, r₅ = −0.802, p = 0.03, PS, Complex Imagery, r₅ = −0.784, p = 0.037, PS, Blissful State, r₅ = −0.772, p = 0.042, OS, Complex Imagery, r₅ = −0.77, p = 0.043, OS, Changed Meaning of Percepts, r₅ = −0.759, p = 0.048, and PS, Changed Meaning of Percepts, r₅ = −0.757, p = 0.049. The NS condition, by comparison, did not yield significant correlations in any of the 5D-ASC subscales (Visionary Restructuralization, r₅ = −0.297, p = 0.518; Oceanic Boundlessness, r₅ = −0.229, p = 0.621) or the 11D-ASC subscales (Complex Imagery, r₅ = −0.454, p = 0.307, Elementary Imagery, r₅ = −0.332, p = 0.467; Audio-Visual Synesthesia, r₅ = −0.299, p = 0.514; Changed Meaning of Percepts, r₅ = −0.106, p = 0.821; Blissful State, r₅ = −0.353, p = 0.437), as expected. However, we note that our ability to detect any possible relationships between self-reported psychedelic effects and PSE values in the NS condition was limited by lower variance (compared to the PS and OS conditions). Taken together, these results indicate that, only for these specific 11D-ASC subscales, higher rating scores (more intense subjective drug effect) were associated with greater surround suppression (lower PSEs in the OS or PS, but not the NS, stimulus conditions). Results from all rmcorr correlations are shown in Supplemental Table S4. Rmcorr plots for 11D-ASC dimensions are shown in Supplemental Figures S1 and S2. 

For each correlation that reached significance in the rmcorr analysis, subsequent post hoc pairwise correlations were performed using robust biweight midcorrelation (Langfelder & Horvath, 2012) on the (psilocybin minus placebo) difference values for each data point (Supplemental Tables S6 and S7). 

All reported adverse events are listed in Table 1. Note that none of these were considered serious adverse events. 

Table 1.

 

Adverse events reported over the duration of the study.

 

Discussion

We examined the extent to which the psychedelic 5-HT2A agonist psilocybin impacted surround suppression using a contrast discrimination task in healthy observers. The lack of serious adverse events reported (Table 1) demonstrates that this kind of study–in which participants complete psychophysical tasks under acute effects of a psychedelic drug–can be done safely. Our results, although preliminary and limited by the small sample size of our pilot study, suggest that psilocybin enhanced visual surround suppression. Furthermore, we found that the magnitude of visual surround suppression correlated with intensity of subjective psychedelic visuals. 

Psilocybin, surround suppression, and visual context processing

We found that surrounding stimuli had a greater influence on the perceived contrast of a center stimulus under psilocybin compared with placebo (niacin). During peak effects of psilocybin, the strength of the illusion was greater (observers perceived the apparent contrast of the center stimulus less veridically). We did not find strong evidence to suggest that this effect depends on relative orientation; when the surround had a parallel (collinear) orientation to the center, the effect of psilocybin on suppression was similar to when the surround orientation was orthogonal (see Figure 3). 

Importantly, psilocybin reduced PSEs only on the trials that presented a surround—contrast discrimination remained largely unaffected by psilocybin in the NS condition. This suggests that psilocybin may have a particular impact on the processing of visual context. Moreover, on catch trials—where the target grating (50% contrast) was presented within a surround, but the reference grating was fixed at a value that was easily discriminated (80% contrast)—psilocybin did not significantly impair task performance (see Figure 4). 

We did not observe a significant difference in the effect of psilocybin between the parallel and orthogonal surround conditions. Contrast surround suppression is thought to depend on both orientation selective and orientation insensitive mechanisms (Webb, Dhruv, Solomon, Tailby, & Lennie, 2005; Cai, Zhou, & Chen, 2008; Schallmo & Murray, 2016; Schallmo, Kale, & Murray, 2019). The relative sensitivity of these two mechanisms to different pharmacological manipulations is an area of active research. Nguyen et al. (2018) found equivalent effects of caffeine (an adenosine receptor antagonist) in reducing surround suppression for parallel and orthogonal surrounds (vs. placebo). In contrast, Kosovicheva and colleagues (2012) found weaker surround suppression for parallel but not orthogonal surrounds under donepezil (an acetylcholinesterase inhibitor) versus placebo. Our results may suggest psilocybin may enhance surround suppression in an orientation-insensitive fashion. Alternatively, we may not have had sufficient statistical power to detect a difference in the effect of psilocybin between our parallel and orthogonal surround conditions, owing to the small sample size. 

Taken together, our findings suggest that psilocybin may specifically impact processing of contextual stimuli in the visual system. This interpretation is consistent with previous findings from Carter et al. (2004), who used a psychophysical motion processing experiment conducted under peak effects of psilocybin and concluded that psilocybin impairs “high-level” but not “low-level” motion perception. However, an alternative interpretation of the impaired high-level motion processing in Carter et al. (2004) is that psilocybin amplified the processing of contextual information in the visual motion stimuli. This is supported by reports from the subjects stating that under psilocybin the task seemed harder as the randomly moving dots became harder to ignore, which led the authors to consider whether the effect of psilocybin on the task was due to reduced inhibition of the irrelevant contextual signals (Carter et al., 2004). This finding appears consistent with our results showing increased influence of visual context, as well as with observations of how psilocybin impacts other domains of mental function (Carhart-Harris, Roseman, et al., 2018). Interestingly, recent findings (Azimi et al., 2020) suggest that serotonin systems might play a role in divisive normalization (Heeger, 1992), a canonical neural computation linked to a wide range of contextual functions in perception and cognition, including surround suppression (Carandini & Heeger, 2012; Schallmo et al., 2018). 

Surround suppression and psychedelic visual phenomenology

Scores in the Visionary Restructuralization dimension correlated significantly with magnitude of surround suppression in the PS and OS conditions. We note that this correlation was found exclusively for the Visionary Restructuralization dimension. Interestingly, the Oceanic Boundlessness dimension had the highest mean scores for psilocybin, as well as the largest difference in score between psilocybin and placebo, yet surround suppression did not significantly correlate with this dimension. Similarly, the 5D-ASC Total Score was not predictive of the magnitude of surround suppression. Taken together, these findings may indicate that the subjective effects measured by the Visionary Restructuralization dimension arise concomitantly with visual surround suppression (Figure 5). 

Depression, surround suppression, and psilocybin

Psilocybin-assisted psychotherapy has recently shown promise in the treatment of MDD (Carhart-Harris, Bolstridge, et al., 2018; Davis et al., 2021; Griffiths et al., 2016; Gukasyan et al., 2022; Ross et al., 2016; von Rotz et al., 2023). However, the exact mechanisms are unclear, and debate is currently centered around the question of which subjective psychedelic effects (if any) are critically involved in positive clinical outcomes—some strongly emphasize the critical role of mystical-type experiences (Griffiths et al., 2016; Ross et al., 2016) whereas others question their importance (Letheby, 2021). Meanwhile, some researchers and corporations hope to discover novel drugs that could deliver psilocybin's antidepressant effect without producing hallucinogenic effects (van den Berg, Magaraggia, Schreiber, Hillhouse, & Porter, 2022; McClure-Begley & Roth, 2022). The role of psychedelic visuals has received little attention or has been downplayed (Letheby, 2021; Swanson, 2024). 

During major depressive episodes, previous studies have shown that patients with MDD exhibit weakened surround suppression (Golomb et al., 2009; Salmela et al., 2021). A possible relationship between weakened surround suppression and anhedonia in depression versus strengthened surround suppression and subjective intensification of visual qualities under psilocybin may be an interesting topic for future study. 

Psychosis versus psilocybin effects

Numerous studies have found weakened surround suppression in people with schizophrenia compared to healthy controls (Dakin et al., 2005; Linares et al., 2020; Pokorny et al., 2019; Schallmo et al., 2015; Serrano-Pedraza et al., 2014; Tadin et al., 2006; Tibber et al., 2013; Yang et al., 2012; Yoon et al., 2009). Our results suggest that psilocybin has the opposite effect on surround suppression (enhancement). Psilocybin's visual effects have previously been cited as points of similarity with early stages of acute psychosis (Vollenweider et al., 1998, p. 3897). Our results appear to be at odds with this and other long-held assumptions that psychedelic drug effects “mimic” psychosis in schizophrenia; assumptions that have been used to justify the “model psychoses” theories that began in the late 1800s and to classify the drugs as psychotomimetic (Swanson, 2018). In line with conclusions from recent comparisons (Leptourgos et al., 2020), our findings suggest caution is warranted when comparing psilocybin's visual phenomenology to that of schizophrenic psychosis. 

Neural information processing and psychedelic effects

Our results highlight a central ambiguity in recent attempts that use ‘predictive processing’ (Swanson, 2016; Swanson, 2018) to explain psychedelic effects (Carhart-Harris & Friston, 2019; Corlett, Frith, & Fletcher, 2009; Letheby & Gerrans, 2017; Pink-Hashkes, Rooij, & Kwisthout, 2017; Stoliker, Egan, & Razi, 2022). The leading proposals argue that psychedelic effects stem from a “decomposition” (Pink-Hashkes et al., 2017) or “relaxing” (Carhart-Harris & Friston, 2019; Carhart-Harris & Nutt, 2010) of top-down feedback signaling, which “liberates bottom-up signaling” (Carhart-Harris & Friston, 2019). However, we found that the surround suppression illusion, which is thought to depend at least in part on top-down feedback in visual cortex (Angelucci & Bressloff, 2006; Nassi, Lomber, & Born, 2013; Nurminen, Merlin, Bijanzadeh, Federer, & Angelucci, 2018), was strengthened under peak effects of psilocybin. In summary, our results suggest that further inquiry into the behavior of perceptual systems under psychedelics is warranted, as well as a closer theoretical examination of how psychedelics might impact the brain's balance of bottom-up (feedforward) signals against top-down (feedback) modulation. 

Limitations

The results presented here are preliminary findings from the first six (n = 6) participants in our pilot study to demonstrate feasibility of measuring surround suppression under psilocybin. While statistically significant, we recognize that not all participants showed dramatic differences in surround suppression in the psilocybin condition. We provided individual participant data for transparency in this regard. Furthermore, because this is a pilot study with a small sample size, we did not correct for multiple comparisons (i.e., across the correlations) which limits the conclusions that can be drawn from the results presented here. Because catch trial accuracy was slightly lower in the psilocybin condition, we recognize that our findings may be confounded by the effect of psilocybin on response accuracy, task attention, or other areas of task performance unrelated to a specific effect on surround suppression mechanisms. 

Maintaining an effective blind in psychedelic drug investigations is notoriously complicated (Burke & Blumberger, 2021; Muthukumaraswamy, Forsyth, & Lumley, 2021; Olson, Suissa-Rocheleau, Lifshitz, Raz, & Veissière, 2020), and our study was no exception. We attempted to minimize placebo effects using niacin as an “active” placebo; nonetheless, it is likely that unblinding occurred during study visits for some experimenters and participants alike. We take seriously the possibility that placebo/nocebo and expectancy could influence PSEs. This possibility is somewhat mitigated by our within-subjects crossover design, in which half of our participants received psilocybin first, and the other half placebo first (see Methods). 

Our inclusion criteria excluded drug-naive subjects, which may have increased the likelihood of expectancy effects in both the psilocybin and placebo conditions. Expectation effects are important to consider in any study with human subjects, especially involving psychedelic drugs. Undoubtedly participants in this study had some expectations regarding the effects of the drug, particularly on their vision since they were informed that visual measurements were being taken. We did not measure participants’ expectations before the experiment, and we did not ask participants to guess which drug was given after the first dosing session but before the second. After the second session, 100% of participants correctly unblinded the experimental condition when asked to guess the session in which they received psilocybin. However, we believe that expectation effects on our results are likely mitigated by the fact that the participants were unaware of how the experiment was measuring surround suppression. It is unclear how expectation would bias their responses to the psychophysical task in one direction or the other, even if they knew for certain which drug they had been given. 

We note that our first set of statistical analyses were performed after unblinding (i.e., double but not triple-blind), which raises the possibility of bias being introduced in our analyses. This is somewhat mitigated by the fact that our second analysis (performed by a blinded statistician, i.e., triple blind) showed results that tended to agree with those of the first analysis, at least in direction of effects, if not in the degree of statistical significance. Moreover, results from our first analyses show a pattern that is the opposite of the hypothesized effect where we anticipated stronger, rather than weaker surround suppression following psilocybin. 

The dose of psilocybin was fixed at 25 mg (oral) for every participant regardless of body mass. Thus, it is possible that dose-response differences occurred between participants, as the intensity of psilocybin's visual effects can be dose-dependent (Carter et al., 2005; Hasler, Grimberg, Benz, Huber, & Vollenweider, 2004; Holze et al., 2022). To mitigate this, future studies could dose psilocybin using a body-mass-relative (mg/kg) ratio. 

Our choice of niacin (vitamin B3) as an “active placebo” follows the common practice of using niacin as placebo in studies of psilocybin, but it is possible that niacin might also have had unanticipated effects on visual perception. One way to address this would be to compare results under psilocybin to a baseline (pre-dosing) measurement. Although we collected baseline (pre-dosing) data, we chose to exclude these data from the current study due to data quality issues in several participants. 

There have been case reports of lasting visual changes after consumption of psychedelic drugs, a phenomenon known as hallucinogen persisting perceptual disorder (HPPD) (Martinotti et al., 2018). For participants who received psilocybin first, results from the placebo session could thus have been confounded by enduring visual changes from their psilocybin session two weeks prior. However, such effects may not have had a dramatic impact on our results for the following reasons. First, all participants had prior experience with psilocybin, so if any putative visual changes from psilocybin last indefinitely, they may already have been present for all sessions. Second, if enduring visual changes did carry into the placebo session for participants who received psilocybin first, we would expect this would reduce the difference in the effects of drugs that we measured between the two sessions. 

The role of specific g-protein coupled receptors (e.g., the 5-HT2A receptor) in causing our findings is unknown, because psilocybin has affinity for a variety of serotonergic and dopaminergic receptors (Ray, 2010), and we did not employ techniques to isolate pharmacological action. However, previous studies that paired the selective 5-HT2A antagonist ketanserin with psilocybin indicate that the visual effects of psilocybin generally depend on 5-HT2A activation (Kometer, Schmidt, Jäncke, & Vollenweider, 2013; Vollenweider et al., 1998), with some exceptions (Carter et al., 2007). Future studies could use selective 5-HT2A antagonist ketanserin to probe the specific role of 5-HT2A in psilocybin-induced alterations of surround suppression, and its relation to subjective drug effects. 

Stimuli were presented for an unlimited time (i.e., until participants made their selection with a keypress). This raises the possibility that the effects of psilocybin on surround suppression that we observed were caused by fixation instability rather than a direct effect of the drug on neural mechanisms of surround suppression. This could be due to some unknown effect of the drug on fixation, or simply that participants may be more likely to dwell longer on task trials under psilocybin, thus increasing the likelihood of drifting fixation. Because we did not record participant response times, we could not compare trial duration between drug sessions. 

We administered the 5D-ASC scale more than 24 hours after each dosing session. Ideally, participants would have completed this questionnaire the same day as each dosing session or immediately the following morning. 

Conclusions

Our results suggest that psilocybin enhanced visual surround suppression in a contrast-matching psychophysics task. Furthermore, the magnitude of surround suppression correlated significantly with self-rated intensity of subjective visual effects, but not with overall psychedelic alterations in consciousness. Interestingly, subjective effects measured by the Audio-Visual Synesthesia subscale had the strongest correlation with magnitude of surround suppression. Our results suggest that (1) surround suppression is not weakened and may instead be enhanced under peak effects of psilocybin, (2) serotonergic neuromodulation may play an important role in surround suppression, (3) enhancements in surround suppression appear to correlate with the intensity of subjective visual effects across individuals, (4) psilocybin may shift visual context processing in the opposite direction of that found in patients experiencing major depressive episodes, (5) the visual effects of psychedelics may be distinct from those underlying abnormal visual perception in schizophrenia, and (6) we may have reason to question the theoretical notion that psychedelic drugs weaken top-down feedback. 

Acknowledgments

The authors thank their research participants for contribution to the study, and the Clinical Research Support Center (CRSC) at the University of Minnesota for study development, initiation, and ongoing regulatory monitoring. 

Supported by the University of Minnesota Foundation (Nielson, PI), and the Heffter Research Institute (Nielson, PI). Supported by the National Institutes of Health's National Center for Advancing Translational Sciences, grant UL1 TR002494. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health's National Center for Advancing Translational Sciences. 

Author contributions: LS and MPS designed the experiment. JLN, SJ, and LS carried out the experiments. KC and RV helped with participant safety monitoring. LS and MDE analyzed the data. LS and MPS wrote the majority of the manuscript, with editorial contributions by JLN, SJ, KRC, and RV. 

Commercial relationships: none. 

Corresponding author: Link Ray Swanson. 

Email: [e-mail]. 

Address: Center for Cognitive Sciences, University of Minnesota, MN 55455, Minneapolis. 

Footnotes

1  “Rmcorr accounts for non-independence among observations using analysis of covariance (ANCOVA) to statistically adjust for inter-individual variability. By removing measured variance between-participants, rmcorr provides the best linear fit for each participant using parallel regression lines (the same slope) with varying intercepts” (Bakdash & Marusich, 2017).