Investigating the Role of Prior Expectations in the Psychosis Continuum

Predictive processing theory describes how internal (generative) models of the world are used to infer the causes of sensory inputs. In this account, top-down predictions (priors) are weighed against bottom-up sensory information with the aim of prediction error minimisation. Dysfunctional predictive processing is prominently cited as a mechanism for psychosis and psychosis-like symptoms in the non-clinical population. Findings from single tasks with modest sample sizes are frequently used to advance a theory for a generalised altered reliance on priors in the psychosis continuum. This is problematic since studies often present conflicting results. Recent studies have suggested that paradoxical findings may be explained by the existence of a hierarchy of priors, although few studies have investigated this directly. This thesis appraises the evidence for an aberrant weighting of priors in the psychosis continuum. The first study (Chapter 2) conducted a systematic review and meta-analysis of perceptual studies to assess whether these overall present evidence for a perturbed reliance on priors in psychosis patients compared to healthy controls. Chapter 3 employed a multi-paradigm approach to measure reliance on priors in the schizotypy spectrum, followed by a factor analysis examining whether performance on different tasks is underpinned by a general factor of reliance on priors (Chapter 4). Finally, Chapter 5 analysed data from a longitudinal study examining whether patients at clinical high-risk of psychosis display an altered reliance on high- and low-level priors and whether this is associated with a subsequent transition to frank psychosis. The meta-analysis provided no evidence for a generalised predictive processing deficit in psychosis. The relationship between schizotypy and the weighting of priors varied between tasks, and this heterogeneity could not be explained by subsetting tasks according to the hierarchical level of the prior. Individual analyses of tasks in Chapter 3 also presented heterogeneous results for the relationship between schizotypy and reliance on priors. The subsequent factor analysis indicated that individuals’ reliance on priors is task-specific. Chapter 5 found no difference in the weighting of either low- or high-level priors in the at-risk group compared to controls, and preliminary findings indicated no association between task performance and subsequent transition to psychosis. Overall, these results do not support the theory that psychotic and schizotypal symptoms are associated with a generalised predictive processing deficit. Researchers should therefore exercise caution when using results from single tasks to inform hypotheses about the role of priors in illness. Further research is required to more precisely delineate categories of priors, their interactions, and their contribution to psychotic symptoms. This may have important implications for the design and interpretation of paradigms for aiding diagnosis, predicting clinical outcomes, and targeting early interventions.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2025

Author dc:creator

• Miller-Silva, Chantal

Advisor dc:contributor.advisor

• Murray, Graham

• psychosis
• schizotypy
• predictive processing
• delusions
• hallucinations
• priors

Licence

• https://www.repository.cam.ac.uk/bitstreams/2eb314aa-a9ee-4cfb-880d-28218cad910e/download
• http://purl.org/NET/rdflicense/allrightsreserved

Language dc:language

eng

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.126525

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/397371