Automated movement analysis predicts transition to non-psychotic disorders in individuals at ultra-high risk for psychosis.

Objective: Ultra-high risk (UHR) criteria were developed to identify prodromal symptoms of psychosis, but most individuals do not transition. This highlights the need to identify transition markers like movement analysis. In the first stage of our study, movement analysis differentiated UHR from controls, showing reduced movement and increased erraticism. Our aim is to verify if these variables can predict UHR outcomes after follow-up.

Methods: UHR individuals were recorded performing two speech tasks at baseline. Videos were analyzed using motion energy analysis (MEA) for head and torso movements-mean amplitude, frequency, and variability-and manually coded for gesticulation. After follow-up, 7 UHR converted to psychosis, 21 to other DSM-5 disorders (GD), and 18 did not convert (NC).

Results: The GD group showed lower torso frequency and higher variability in both regions compared to Psychosis, as well as greater variability for torso when compared with NC. No differences were found between Psychosis and NC. Gesticulation did not differ between groups.

Conclusions: Baseline movement variability distinguishes UHR transition outcomes, with higher variability seen in those converting to non-psychotic disorders. This supports the importance of movement analysis as a potential transition marker and suggests that treating UHR individuals as a single group may overlook important information.