Its pitch is short in-clinic time; its control arm is a low dose, not a placebo.
Weekly Science Digest, 23 Oct 2026 to 29 Oct 2026. 28 papers read, 5 preprints, 3 new trials.
1. A short-acting psychedelic gets a phase 2b readout
Intranasal mebufotenin for treatment-resistant depression reported in JAMA, with discharge time as a co-headline.
Adults with depression nonresponsive to 2-5 antidepressant trials were randomised at 38 sites to a single intranasal dose of 8 mg (n=46), 12 mg (n=73), or 0.3 mg, with 8 weeks of blinded follow-up. The logistical claim is the point: 5-MeO-DMT's experience is brief, so in-clinic monitoring is short.
How much weight it carries: Phase 2b, 38 sites, double-blind with 8-week follow-up; the control was a 0.3 mg active dose rather than inert placebo, so effects are dose contrasts. (Humans, RCT, Double-blind, Controlled, n = 119)
Open-label psilocybin after an AUD trial helped mainly the heaviest drinkers, and a sex-difference reanalysis found little.
Fifty-nine adults who completed the NYU alcohol trial got one open-label 25-40 mg/70kg dose; gains in heavy drinking days concentrated in those still drinking heavily at entry, with no serious adverse events. Separately, a Bayesian reanalysis of the 59-person psilocybin-vs-escitalopram trial found no overall sex difference, with a single signal: lower post-treatment anxiety in women on psilocybin.
How much weight it carries: Open-label extension of a phase 2 RCT with no blinded comparator, plus a post hoc Bayesian reanalysis with only 20 women; both are secondary, not confirmatory. (Humans, Open-label, No control group, Small sample, n = 59)
3. Separating the antidepressant effect from the trip
Repeated low-dose LSD sustained antidepressant-like effects in mice without psychedelic-like behaviour.
Structure-guided 5-HT2A agonists biased toward β-arrestin drove the long-lasting effect, and LSD's slow receptor dissociation let repeated low doses accumulate occupancy while avoiding the Gq calcium peak. No psychedelic-associated behavioural disruption appeared in mice or cynomolgus macaques. A separate medicinal chemistry paper mapped a druggable hydrophobic tunnel beside the 5-HT2A orthosteric site using synthesised phenethylamines.
How much weight it carries: Mouse and macaque behaviour plus receptor pharmacology; behavioural proxies for hallucinogenesis do not establish what humans would experience. (Mice, Animals, Preclinical, Mechanism, Drug design)
4. What 40 years of psychosis reports actually show
Across 52 studies, population data showed no consistent link between psychedelic use and psychosis.
Controlled trials turned up one documented case of transient delusions. Emergency and toxicology data showed higher rates, and among the 34 case reports, 58% of individuals had personal or family psychiatric history.
How much weight it carries: Systematic review of 52 heterogeneous papers, 34 of them case reports; no denominator for naturalistic risk and trial data come from pre-screened samples. (Humans, Systematic review, Observational)
