The Predictive Coding Account of Psychosis

Area of Science:

• Computational neuroscience
• Neurobiology
• Psychiatry

Background:

• Increasing interest in neurocomputational mechanisms of psychosis.
• Predictive coding and Bayesian inference as a successful approach.
• Psychosis linked to aberrant predictive coding and altered precision of prior beliefs.

Purpose of the Study:

• Review current evidence for aberrant predictive coding in psychosis.
• Discuss challenges for the canonical predictive coding account.
• Propose a nuanced framework for understanding psychosis.

Main Methods:

• Review of existing literature on predictive coding and psychosis.
• Analysis of evidence for altered precision in prior beliefs and sensory data.
• Exploration of hierarchical processing and active inference in psychosis models.

Main Results:

• Conflicting findings exist regarding prior belief strength in psychosis.
• Hallucinations and delusions may involve distinct predictive coding alterations.
• Nuanced predictive coding models accounting for modality-specific deficits and hierarchy are needed.

Conclusions:

• A more complex predictive coding framework is necessary to explain psychosis.
• Considering hierarchical organization and active inference may reconcile conflicting findings.
• This nuanced approach can better address the diverse manifestations of psychosis.