Acute effects replicated everywhere; the 24-hour antidepressant-like ones did not.
Weekly Science Digest, 30 Oct 2026 to 5 Nov 2026. 32 papers read, 12 preprints, 3 new trials.
1. Psilocybin in mice: acute robust, persistent not
Five independent labs ran the same psilocybin experiments; the next-day effects vanished.
Acute psilocybin (2 mg/kg IP) reliably increased anxiety- and avoidance-related behaviour and reduced fear expression across all sites. At 24 hours there was no replicable reduction in anxiety- or depression-like behaviour, no facilitation of fear extinction, and no change in social preference or social reward learning.
How much weight it carries: Prospective five-site replication with ~200 mice per experiment; strong evidence about rodent assays, silent on human clinical durability. (Mice, Controlled, Preclinical, Peer reviewed)
2. Clinical readouts: one miss, one pooled win, one pilot
Psilocybin missed its primary anhedonia endpoint at week 1 but separated by week 8.
In a 20-person double-blind trial in treatment-resistant depression, the week-1 DARS difference was 11.2 (95% CI −1.0 to 23.3, p=0.073); the week-8 secondary was 14.7 (CI 2.5 to 26.8). Separately, a meta-analysis of 5 LSD RCTs (N=227) found a standardized mean difference of −0.69 on depression symptoms, and the first open-label MDMA-assisted therapy trial in veterans with comorbid PTSD and alcohol use disorder (N=10) reported 90% retention.
How much weight it carries: Main trial is a 20-person pilot whose primary endpoint was negative; the supporting meta-analysis pools 227 patients across heterogeneous LSD trials with likely unblinding. (Humans, RCT, Double-blind, Placebo-controlled, Small sample, Null result, Meta-analysis, Open-label, n = 20)
A 25,419-interaction support-line dataset is now public, and cannabis drove the most distress.
The Fireside Database covers five years of peer support calls by 220 trained volunteers, with 2,777 caller satisfaction surveys. Psilocybin was the most-reported substance at 59.9% of coded interactions; cannabis produced the highest distress rate at 47.2%. A separate interview study of 13 people with persisting symptoms — mostly HPPD and depersonalization — found all index experiences occurred in naturalistic settings and that few services knew what to do with them.
How much weight it carries: Large descriptive dataset from a self-selected help-seeking population plus 13 qualitative interviews; good for characterising harms reported, not for incidence. (Humans, Observational, Survey, n = 25,419, Peer reviewed)
Olson's group reanalysed the data and says Gi signaling does not sort hallucinogens from non-hallucinogens.
Responding to a recent paper proposing Gi activation at 5-HT2A as the discriminating feature, the reanalysis of the original source data finds Gi less predictive of hallucinogenic potential than canonical 5-HT2A effectors.
How much weight it carries: Reanalysis of another group's published data, not an independent replication; settles nothing until the original authors respond. (In vitro, Mechanism, Peer reviewed)