Psilocybin-assisted neurofeedback for the improvement of executive functions: a semi-naturalistic-lab feasibility study

Executive function (EF) deficits are a transdiagnostic feature across psychiatric disorders and substantially impair daily functioning and recovery. Fronto‑cingulo‑parietal networks, and specifically frontal‑midline (fm) theta oscillations thought to originate in midcingulate and dorsolateral prefrontal cortices, have been linked to successful EF performance; the inability to upregulate fm‑theta correlates with poorer task outcomes. Neurofeedback (NF) targeting fm‑theta can enable participants to learn self‑regulation of this neural signal and has shown medium‑sized benefits for components of EF such as working memory updating, but responsiveness to NF varies and a subgroup of “non‑responders” limits clinical utility. Impaired neuroplasticity has been proposed as one factor reducing NF learning in psychiatric populations. Enriquez‑Geppert and colleagues set out to examine the feasibility and initial effects of combining a neuroplasticity‑promoting pharmacological agent, psilocybin (delivered as repeated microdoses), with fm‑theta NF. The study aimed to test whether a brief protocol—one week of dose adjustment followed by three psilocybin‑assisted NF sessions—was practical in a semi‑naturalistic lab setting, to assess effects on EEG fm‑theta self‑regulation, on objective laboratory EF tasks, and on self‑reported EF in daily life, and to evaluate expectancy/placebo proxies and participant‑centred priority outcomes.

Participants were recruited from an online microdosing workshop and screened for exclusion criteria including personal or family history of psychosis, normal or corrected vision, and three weeks' drug abstinence before baseline. Thirty‑seven volunteers were randomised using Matlab’s rand function to an experimental group (psilocybin‑assisted NF) or a passive control group; the extracted text does not clearly report the exact number of participants assigned to each arm. The passive control group was blinded to the NF element to reduce motivational differences. The study received ethical approval and followed the Declaration of Helsinki. Data collection comprised online pre‑ and post‑assessments conducted via Zoom and three in‑lab fm‑theta NF sessions performed in a sound‑attenuated EEG laboratory. Both groups followed a three‑week schedule: week 1 baseline measurements and an adjustment microdosing week for the experimental group; week 2 the experimental group completed three NF sessions on alternating dosing days while the control group had a two‑week break; week 3 post‑assessments for both groups. Online assessments included measures of placebo/expectancy (dispositional optimism, suggestibility, microdosing expectancy, NF‑specific attitudes), objective EF tasks, and self‑reported EF in daily life and personalised priority ratings. Objective EF was probed with four computerised tasks mapping to core EF domains: a reference‑back task (working memory updating), a task‑switching paradigm (set shifting), a picture‑word interference task (conflict monitoring), and a stop‑signal task (motor inhibition). Tasks used shared stimuli and trial structure; reaction time (RT) and accuracy were recorded and pre‑post difference scores computed. The picture‑word interference (conflict) task was later excluded from analysis due to poor performance by many participants. Self‑reported EF used four BRIEF‑A subscales (Working Memory, Shift, Task Monitor, Inhibit). Participants also listed up to four personal training priorities and rated pre‑, expected, and post‑functioning (0–100 scale) and satisfaction. EEG was recorded with a 15‑electrode cap and a TMSI REFA amplifier at 256 Hz. Psilocybin exposure consisted of fresh Psilocybe mexicana truffles portioned in 1 g packs; participants self‑selected an optimal microdose in an adjustment week (advised range 0.5–2 g). For NF, participants took their chosen dose about two hours before each NF appointment. Each NF session lasted roughly 70 minutes and comprised six five‑minute NF blocks (30 minutes of NF) with 5‑minute eyes‑open resting baselines at session start and end. Real‑time processing used five fronto‑midline electrodes (Fz, FC1, FC2, FCz, Cz) with nose reference and Fp1/Fp2 for ocular monitoring. Two‑second segments (shifted every 200 ms) were FFT‑analysed to compute fm‑theta (4–8 Hz) relative to the session start baseline; feedback was delivered as a colour‑saturation square updated every 200 ms (red = increase, blue = decrease, grey = blink/no change). Participants reported strategy efficacy, motivation, commitment and perceived difficulty after each block. Statistical analyses normalised NF output across 1–40 Hz into frequency bands and used two learning indices: a between‑session index (mean relative theta across six blocks per session; RM ANOVA with SESSION 1–3) and a within‑session index (block‑level changes; RM ANOVA with BLOCK including baselines). Objective task outcomes were change scores (RT and accuracy) analysed with two‑way RM ANOVAs with GROUP (experimental, passive control) and CONDITION (task‑specific levels). Self‑report BRIEF‑A subscales used pre–post difference scores and one‑sided t‑tests for group comparisons based on directional hypotheses. The stop‑signal task used RTs (SSRT) only, given dynamic accuracy adjustment. Participants with overall/subtask accuracy below 60% were excluded; after data cleaning, 31, 35 and 33 participants remained for the reference‑back, switching and stop‑signal tasks respectively. A p‑value ≤ .05 denoted statistical significance; nonparametric robust tests were applied where assumptions were violated but parametric results were reported when nonparametric tests agreed.

Sample characteristics: Thirty‑seven participants (20 female, 16 male, 1 other) with mean age 24 years (SD = 4.3) took part. Most were students, predominantly right‑handed, not on regular medication, and 20 reported prior psychedelic use. The extracted text does not explicitly state the numbers allocated to each study arm. After the microdosing adjustment week participants’ chosen microdose averaged 1.14 g of fresh Psilocybe mexicana truffles (SD = 0.52) and was taken on average 2.2 hours (SD = 0.35) before NF sessions. Participants reported good sleep and high motivation and commitment; NF sessions were rated as moderately difficult and self‑rated strategy effectiveness was moderate. Neurofeedback learning: The between‑session learning index (mean relative theta per session relative to start baseline) produced an RM ANOVA that approached significance, F(2,34) = 3.185, p = .054, with partial η2p = .158, indicating a moderate effect size after only three NF sessions. The within‑session (block‑level) learning index showed no significant block‑wise changes after Greenhouse‑Geisser correction, F(2.457,41.765) = 1.303, p = .285, η2p = .071. Placebo and expectancy measures: No significant differences were found between experimental and passive control groups on revised life orientation (optimism), suggestibility, or NF/psilocybin‑specific expectancy questionnaires. Expectations for microdosing were described as medium, slightly higher for psilocybin‑assisted NF, and moderately high for NF alone. Objective experimental EF tasks: For the analysed tasks (reference‑back, task‑switching, stop‑signal), RM ANOVAs of pre–post RT and accuracy change scores revealed no significant main effects of GROUP or GROUP × CONDITION interactions that would indicate superior objective task improvement in the experimental group. One within‑task CONDITION effect was reported for reference‑back accuracy (better improvement for no‑update than update condition) but this was not specific to group. The picture‑word interference (conflict monitoring) task was excluded because 12 of 37 participants fell below accuracy thresholds. Self‑reported EF in daily life: On BRIEF‑A subscales, the experimental group showed significantly greater pre–post improvements than the passive control group across all four domains: Working Memory (t(35) = 2.69, p = 0.0054, d = 0.885), Shifting (t(35) = 2.08, p = 0.0226, d = 0.683), Monitoring (t(35) = 2.63, p = 0.0063, d = 0.865) and Inhibition (t(35) = 2.63, p = 0.0067, d = 0.857). The Working Memory score in the experimental group moved into a range interpreted as normalised functioning, whereas the passive control group remained above deficit threshold. Participant priorities and functional ability: Participants’ self‑selected priority domains included Cognition (n = 17), Presence (n = 18), Mood (n = 13), Connection (n = 5) and Other (n = 9). Significant pre–post improvements within the experimental group were found for Cognition (t(16) = 3.895, p = .001, d = .876), Presence (t(17) = 3.566, p = .002, d = 1.052) and Mood (t(12) = 2.520, p = .027, d = .966). Connection showed no significant change. Satisfaction ratings varied, described as medium for Cognition and Mood and good for Presence. Data exclusions and analytic notes: Participants with task accuracies below 60% were excluded from respective task analyses; after this cleaning 31, 35 and 33 participants contributed to reference‑back, switching and stop‑signal analyses respectively. The study used p ≤ .05 as the significance threshold and reported parametric results where nonparametric tests corroborated findings.

Enriquez‑Geppert and colleagues interpret their findings as demonstrating the feasibility and practical implementation of psilocybin‑assisted fm‑theta NF in a semi‑naturalistic lab setting. They highlight a promising trend toward improved fm‑theta self‑regulation across three sessions, with the between‑session learning index approaching statistical significance and a moderate effect size. However, no objective task‑based EF improvements were observed after this short protocol. By contrast, participant‑reported EF in daily life and self‑selected priority areas showed sizeable and statistically significant gains, with working memory scores moving into a normalised range for the experimental group. The authors position these results within existing literature by emphasising the potential of psychedelics to enhance neuroplasticity and thereby augment NF learning, noting prior reports of rapid plasticity changes following high doses and emerging evidence for plasticity markers after repeated microdosing. They argue that microdosing has practical advantages for NF integration because it reduces acute psychoactive effects and the need for intensive clinical monitoring, while still engaging putative neuroplastic mechanisms. The study’s dual assessment strategy (objective lab tasks and ecological self‑reports) is presented as a strength that captures both controlled performance and real‑world functional change. Key limitations acknowledged by the authors include the small number of NF sessions (three rather than typical seven or eight), absence of an active control condition that would better isolate specific treatment effects, and implicit placebo mechanisms that cannot be fully excluded despite no between‑group differences on measured expectancy proxies. They also note methodological constraints such as the exclusion of the conflict monitoring task due to poor online task performance and the semi‑naturalistic dosing procedure that, while ecologically valid, introduces variability in absolute dose across participants. Implications offered by the authors focus on further optimisation of combined pharmacological‑neuroscientific interventions: determining the optimal timing between psilocybin ingestion and NF to balance acute substance effects with the neuroplasticity window, increasing session number and sample size, and employing active control groups in future trials. If replicated, they suggest psilocybin‑assisted NF could become a transdiagnostic approach targeting core cognitive dysfunctions by transiently enhancing neuroplasticity to improve NF efficacy and sustain functional gains.