Cognitive dysfunction in a psychotropic medication-naïve, clinical high-risk sample from the...

1) to assess generalizability of neurocognitive deficits reported in previous Western clinical high-risk (CHR) for psychosis studies in a prodromal program in Shanghai, China; and 2) to investigate neurocognition in CHR subjects in relation to a broader range of clinical outcomes (e.g. remission) than presence or absence of psychosis.

Baseline neurocognitive assessments of CHR (n = 217) and healthy control (HC; n = 133) subjects were compared based on 1-year follow-up clinical status using MANOVA. CHR subjects were first divided into 'converter' (CHR-C; n = 41) and 'non-converter' (CHR-NC; n = 155) to psychosis groups and compared to HC and to each other. CHR subjects were then divided into 'remission' (i.e. achieved remission; n = 102), 'symptomatic' (persistent positive symptoms in the absence of conversion; n = 37) and 'poor-outcome' (converted and symptomatic subjects who did not respond to treatment; n = 57) groups.

CHR neurocognitive performance was broadly impaired compared to HC; CHR-C subjects showed lower performance in processing speed and visual learning than CHR-NC. CHRs with poor clinical outcomes showed lower performance on most MCCB tasks compared to HC, particularly in learning and processing speed, as clinical outcome worsened from remission to symptomatic to poor outcome groups.

Level and pattern of baseline neurocognitive weaknesses in SHARP CHR subjects were similar to those in NAPLS-2. Outcome stratification into remission, symptomatic and poor groups was associated with increasing cognitive deficits in learning and processing speed. These findings support cross-cultural generalizability and advance understanding of CHR neurocognitive heterogeneity associated with 1-year clinical outcomes.