Paranoia and Bias in Social Belief Updating in Clinical High Risk for Psychosis

Project Summary Individuals at clinical high risk for psychosis (CHR) experience attenuated positive symptoms (e.g., delusions) and may later develop psychosis spectrum disorders (e.g., schizophrenia).1 Paranoia is a common and impairing psychotic symptom, also present in CHR individuals.2–5 Emerging mechanistic research using

reinforcement learning task paradigms suggest that individuals with paranoia hold rigid prior beliefs about their environments being highly unpredictable and others being untrustworthy, which is in turn linked to reduced learning from actual environmental feedback that may challenge these beliefs.6–8 While these explanations

hold promise for explaining a complicated phenomenon, they have not been tested in CHR individuals and a number of questions remain. For instance, there is debate within the literature as to whether general vs. social learning processes are most relevant to paranoia; however, existing tasks poorly examine this issue.9,10

Furthermore, previous research has focused mainly on one reinforcement learning paradigm (e.g., learning volatile probabilities) and paradigms measuring other learning processes relevant to delusions (e.g., inappropriate learning about blocked cues) may also be related to paranoia.11 Finally, the connection between

reinforcement learning processes and symptom progression and functional outcomes (e.g., worsening) has not been examined. The proposed study will examine these gaps in the literature using samples of CHR individuals (N = 50), clinical controls (N = 50), and healthy controls (N = 50), with half being followed for one

year. Using this data, the study aims to develop and validate two distinct socially-salient reinforcement learning tasks (Aim 1), compare these two distinct learning paradigms in terms of relevance to paranoia (Aim 2), examine the value of these two social learning tasks for predicting symptom progression (e.g., worsening

paranoia) and social functioning decline, above and beyond non-social versions of these tasks (Aim 3). These aims and this project will make use of a larger, ongoing NIMH-funded project,12 which will provide data on 50 CHR individuals, as well as 50 clinical controls and 50 healthy controls. Training Plan: to complete this project

and build an independent research career the applicant will (a) pursue in-depth knowledge of psychosis risk, development, delusion formation, social cognition, and learning theory, (b) learn to develop and validate dynamic behavioral tasks, and (c) develop longitudinal data analysis skills. This training will be accomplished

through regular meetings with a team of expert scholars, coursework, workshops, and attending conferences. Training Environment: The training will take place at Northwestern University, which offers an excellent and collaborative training environment, which includes both the sponsor and the project’s data analysis consultant.

The environment, alongside the proposed study and training plan, will position the applicant to launch a unique program of research and advance our understanding of social processes in psychopathology.