Background and hypothesis: Processing speed and verbal learning are linked to functioning in people at clinical high-risk for psychosis (CHR-P). However, a clear understanding of these relationships has been limited by the use of single-value assessor ratings of functioning and limited incorporation of self-report and longitudinal functional assessments. This study aims to examine the effect of processing speed and verbal learning on longitudinal assessor-rated and self-reported functioning.

Study design: Individuals at CHR-P (n = 1282) and community controls (n = 425) who participated in Accelerating Medicines Partnership-Schizophrenia were examined. Processing speed and verbal learning abilities were ascertained using the Penn Computerized Neurocognitive Battery. Assessor-rated functioning was measured using the Global Functioning Scales. Self-reported functioning was collected using Ecological Momentary Assessment. Analyses were linear models and mixed-effects models.

Study results: Both processing speed and verbal learning were associated with assessor-rated functioning at baseline and through 2 months (Ps < 0.001). Processing speed abilities predicted changes in self-reported ability to function (P = .036), motivation (P = .040), and concentration (P = .020) over time. Verbal learning abilities predicted changes in self-reported loneliness (P = .044) over time. Processing speed and verbal learning were associated with assessor-rated and self-reported functioning over and above estimated intelligence quotient and positive symptoms.

Conclusions: Processing speed and verbal learning are important predictors of short-term longitudinal functioning, and may be important targets for early recognition and mitigation of psychosis risk.