Psychiatry

CBTp, antipsychotics, and their combination in first-episode psychosis show similar monotherapy effects, with advantages for combined treatment

In pooled individual-participant data from two similar RCTs (n=136), CBTp alone and antipsychotics alone produced comparable 6‑month PANSS improvements. Combined CBTp plus antipsychotics outperformed antipsychotics alone on PANSS total at six months, with a moderate mean difference favoring combination therapy. User-defined recovery scores favored combined treatment over either monotherapy, while the two monotherapies did not differ from each other. Serious adverse events were mainly psychiatric hospitalizations, with no clear safety signal separating CBTp, antipsychotics, or their combination. Findings suggest CBTp is a viable alternative to antipsychotic monotherapy short term, but combining both may optimize early symptom and recovery outcomes. 1

Psilocybin shows large symptom reductions in depression and anxiety across controlled studies

This PRISMA-based systematic review and meta-analysis synthesized psilocybin therapy trials for anxiety, major depressive disorder, and mood disorders. Across included studies, psilocybin produced large symptom reductions in anxiety and MDD (SMD around −1.44) and mood disorders (SMD around −1.48). Benefits often followed a single psilocybin-assisted session and were sustained beyond the acute dosing period in many trials. The authors attribute therapeutic effects to 5‑HT2A–mediated neuroplasticity and disruption of entrenched maladaptive cognitive patterns. Despite large effect sizes and robustness to publication bias analyses, implementation barriers, regulatory issues, and population-specific tailoring remain major limitations. 2

Twelve-year follow-up highlights limited full recovery despite high remission in first-episode psychosis

This 12‑year follow-up of an RCT compared extended three-year early intervention service with step-down to standard care after two years. Among 106 participants assessed, symptomatic remission reached roughly three-quarters, but full recovery occurred in only a small minority. Functional recovery rates were low and emerged as the main rate-limiting factor for clinical and full recovery. Premorbid schizoid–schizotypal traits and poorer baseline functioning predicted worse functional and clinical outcomes. Extended early intervention was retained in prediction models for symptomatic remission and full recovery, but attrition limits causal interpretation. 3

Generative AI mental health chatbots yield small average symptom improvements with wide uncertainty

This systematic review identified 26 quantitative studies of generative or hybrid AI mental health chatbots, mostly from non‑WEIRD settings. A meta-analysis of 14 RCTs showed a small but statistically significant overall benefit on negative mental health outcomes (ES≈0.30). The 95% prediction interval was wide and crossed zero, indicating substantial between‑trial heterogeneity and variable real‑world impact. Evidence in young children and older adults was sparse, limiting generalizability to these age groups. Findings support cautious use of generative AI chatbots as low-intensity adjuncts rather than standalone treatments for significant psychopathology. 4