Background: Impaired language processing is a core symptom of psychosis and is critically involved in shaping positive (e.g., auditory hallucinations) and negative symptoms (e.g., social withdrawal). Hierarchical Bayesian accounts of psychosis suggest an imbalance in the weighting of prior beliefs and incoming sensory information to be an underlying mechanism of symptom development. Thus, the goal of this study was (1) to investigate the use of semantic prior beliefs relative to sensory information during predictive language processing as a trait marker for early psychotic-like symptoms and (2) to explore the underlying neurophysiological signatures of potential alterations.

Methods: 37 participants assessed for psychotic-like symptoms completed a predictive language paradigm while their brain activity was recorded with EEG. In the task, participants listened to sentences of varying predictability, with the sentence-final words being degraded in clarity to different degrees. They were asked to report the word they perceived to assess conditioned hallucinations. To estimate the relative strength of semantic prior beliefs over the sensory input, we applied computational modelling and investigated the neurophysiological underpinnings with EEG.

Results and conclusion: Computational modelling revealed an overweighting of prior beliefs relative to sensory information with increasing psychotic-like symptoms. On a neurophysiological level, we found that this was linked to an altered N400 effect. Interestingly, the overweighting of semantic prior beliefs was also associated with experiencing more conditioned hallucinations. In conclusion, the present findings may provide an initial mechanistic explanation of how suboptimal predictive language processing contributes to the formation of hallucinations.