Recent publications

An early-psychosis analysis examined how parental socioeconomic status, cerebellar connectivity, verbal ability, and alogia were associated within a serial mediation model.

The observational mediation model does not establish a causal sequence; its 100-patient HCP-Early Psychosis sample overlaps the earlier cognition study and is not an independent replication.

Diffusion MRI found lower mean diffusivity along cerebellum–ventral tegmental area pathways in first-episode psychosis, with working-memory associations but no significant fractional-anisotropy differences.

The cross-sectional study included 46 participants with first-episode psychosis and 30 controls. Diffusion measures do not directly measure dopamine signaling or establish causality.

Connectivity mapped to verbal ability, working memory, and cognitive flexibility in 100 people with early psychosis; a separate randomized, sham-controlled pilot tested left Crus I/II TMS and found an initial cognitive-composite signal.

The TMS pilot included only 16 participants, eight per arm, treated for two weeks. It did not support symptom improvement; the preliminary cognitive finding does not establish an effective routine treatment.

A secondary imaging analysis of 127 people with first-episode psychosis linked connectivity in cerebellar auditory and cognitive systems to better or worse positive-symptom response over 12 weeks of antipsychotic treatment.

The treatment follow-up was prospective, but this association analysis was not an externally validated individual prediction model. Consistency across scanner and drug subgroups does not establish clinical benefit.

Cerebellar anatomy was associated with distinct cognitive dimensions in a transdiagnostic youth sample, while cerebellar volumes did not differ significantly across psychopathology categories.

The cross-sectional analysis retained 662 of 1,452 initially considered participants after eligibility, data-availability, and quality checks. It does not establish developmental change, causality, or out-of-sample prediction.

A striatal functional-imaging score replicated diagnostic discrimination but did not predict improvement during antipsychotic treatment.

Diagnostic discrimination in 101 patients and 51 controls did not translate into treatment prediction in a separate 97-person cohort. Test-retest reliability was modest, with intraclass correlations of 0.22–0.48.

A functional-connectivity model predicted 12-week positive-symptom response to antipsychotic treatment in a 49-person discovery sample and a separate 24-person validation sample.

Both samples came from the same institutional setting. Small samples, no placebo comparison, and a discovery-stage outcome-estimation dependency limit claims of transportability, drug specificity, or improved clinical decisions.

MRI phase-encoding direction affected the repeatability of functional-connectivity measures, highlighting acquisition choices that matter for reproducible brain research.

The study used 32 healthy participants and 49 healthy HCP-Early Psychosis participants. Effects depended on processing choices and scan timing; reliability alone does not establish clinical validity.

Across 17 studies, an image-based meta-analysis found no meaningful regional activation differences during negative-emotion processing between 677 people at risk for schizophrenia and 805 controls.

This null result concerns regional responses to negative emotional stimuli in at-risk populations. It neither excludes all brain differences in psychosis nor validates alternative connectivity biomarkers.