Review Papers Against Their Own Results

 Quick Explanation Copied

 Long Explanation

What the Evidence Shows

In head-fixed Sst-IRES-Cre;Ai32 mice performing a go/no-go change-detection task, 1 mg/kg psilocybin (given 24 h after saline, within-subject) collapsed perceptual sensitivity (AUC) to chance in the post-injection epoch, with 11 saline sessions from 9 mice vs 6 psilocybin sessions from 6 mice Change-detection AUC collapse under psilocybin. Across 11,012 recorded units from six Neuropixels probes, psilocybin induced a visually evoked 4-Hz oscillation affecting 58.9% of SST, 45.7% of FS, and 35.2% of RS neurons (vs ~6-10% under saline), strongest in VISp 4-Hz oscillation proportions by cell type.

Crucially, psilocybin selectively enhanced onset responses of change-encoding neurons to repeated (no-change) images — SST: 0.88±0.18 → 2.52±0.32 (P=0.0060) — while image-identity-encoding neurons were unchanged (hierarchical bootstrap, P>0.05), and receptive-field spatial tuning was preserved (all uncorrected p>0.05) Selective change-encoding enhancement. PCA showed psilocybin shifted no-change population trajectories toward genuine-change states (RS: P=0.0064; SST: P=0.044 vs saline) Population trajectory shift toward change states.

Critical Blind Spots

The authors themselves flag that a direct causal link between these neural 'surprise' signatures and the behavioral deficit is not established, and reduced task attempts (without increased false alarms) are not fully accounted for Authors. BGPT infers additional weaknesses: fixed saline-then-psilocybin session order without randomization, neural analyses dropping the one 'resilient' animal, only n=6 neural sessions, and a single acute dose window. A human EEG reanalysis found psilocybin modulated early sensory precision with limited higher-level prediction-error effects, contrasting with the mouse 'surprise' framing Human MMN psilocybin EEG reanalysis — cross-species convergence is suggestive, not confirmatory. Falsification: if psilocybin recruits change-encoding ensembles in animals with intact behavior, the causal interpretation weakens.

Code (github.com/RobertoDF/psycode) and data (DANDI:001417) are public, supporting reproducibility. Christof Koch declared a financial interest in a consciousness-assessment company; other conflicts unstated. What would change these conclusions: a randomized-order, dose-response replication with the resilient animal retained, and causal perturbation (e.g., SST optogenetic 4-Hz driving) tested against behavior.

 BGPT Paper Review

Study Novelty

First cell-type-resolved (SST/FS/RS) Neuropixels evidence that psilocybin recruits change-encoding ensembles during expected stimuli; prior human work (MMN reanalysis) tested predictive coding but not at this circuit granularity.

Scientific Quality

Strong methods (Kilosort4, optotagging, public code/data) but fixed drug-order design, small neural sample (n=6 sessions), exclusion of the resilient animal, single dose, and acknowledged unestablished causal link to behavior lower quality.

Study Generality

Findings are specific to a trained visual change-detection task in SST-ChR2 mice at one acute dose; generality to other stimuli, doses, species, or psychedelic classes is untested.

Study Usefulness

Public data/code and a clear circuit-level mechanism (SST-driven 4-Hz surprise signal) provide a testable target for predictive-coding theories of psychedelic action, though behavioral causality remains open.

Study Reproducibility

Full analysis code on GitHub and dataset on DANDI (001417) with documented pipelines (Kilosort4, optotagging); only the specialized mouse training and surgical expertise limit replication.

Explanatory Depth

Mechanistic at circuit level (cell-type-specific oscillation, ensemble recruitment, trajectory alignment) but the causal chain to perception is correlational, and the origin of the 4-Hz rhythm is unexplained.