Computational Algorithms for Predicting Psychosis

(CNTRACS Trial)

In Study 3, all tasks and models from Aim 1 will be evaluated in a large community sample of online participants recruited through Prolific.

Aim 3 will establish 1) requirements for testing computational algorithms in population-based samples, 2) whether task length is optimal, and 3) whether computational parameters help predict psychopathology (hallucinations and delusions in particular) in the general population, and how these predictors may be impacted by demographic variables, such as race, ethnicity, gender, age and other social determinants of health.

Participant demographics will be matched to the U.S. Census 2023 American Community Survey (ACS) data release (updated as available). MPI MacDonald has successfully recruited large online samples for over a decade.

Specifically, researchers will recruit two samples of n=2,000 to preregister hypotheses and then n=3,000 to independently confirm those hypotheses. Multiple methods will affirm the highest quality human self-report and behavioral data for online samples.

Researchers use Prolific, which is known for high quality research participation, maintaining low levels of bots through strategies such as requiring a photo ID and a waiting period. Second, inclusion criteria will require an approval rating of 90-100% from previous researchers. Participants will complete three sessions, each limited to \~60 min to avoid burnout.