Servicenavigation

Project Details

Back

Projekt Print View

Perceptual decision-making under changing prior beliefs

Applicant Professor Dr. Andreas Heinz, since 5/2022
Subject Area Human Cognitive and Systems Neuroscience
Clinical Psychiatry, Psychotherapy, Child and Adolescent Psychiatry
Term from 2018 to 2023
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 407062764
 
Final Report Year 2024

Final Report Abstract

According to the computational framework of Bayesian predictive processing, the brain uses prior beliefs to make inferences regarding the probable causes of sensory data. If these sensory data contradict prior beliefs, prediction error signals are generated that update prior beliefs. We investigated behavioural effects of prior beliefs on perceptual decision-making, used model-based fMRI to examine the underlying neural mechanisms, and investigated alterations in these mechanisms in relation to psychosis. One key result from the project is that an important source of prior beliefs in perceptual decision-making is recent perceptual history. For example, if the sensory information is noisy and perceptual uncertainty therefore high, perceptual decisions are biased by preceding perceptual decisions. We found this to be true both under perceptual uncertainty and ambiguity and in both humans and mice. In individuals with high psychosis proneness, the effect of perceptual history was reduced, pointing to a reduced weighting of perceptual priors as a mechanism underlying psychosis, as previously suggested. In addition, we also investigated the role of other sources of priors in perceptual decision making, including associative learning and context effects. We could show that learned auditory-visual associations have a stronger effect on perceptual decisions unter high compared to low perceptual uncertainty. Moreover, we uncertainty induced by unreliable feedback lead to corrupted encoding of stimulus information in early visual cortex. Finally, we could show that prior beliefs based on visual context information shaped neural signal patterns in early visual cortex even in the absence of direct visual input. In addition to prior beliefs, the project also yielded insights into the processing of perceptual prediction errors and their role in psychosis. We devised a new experimental paradigm which allowed us to manipulate the ambiguity of visual stimuli in a graded fashion and to thereby quantify the effect of prediction errors caused by sensory information that is incongruent with perceptual history. We showed that inferior frontal cortex plays a key role in the signaling of perceptual prediction errors and that disruption of its function by rTMS decreased the effect of prediction errors on perceptual decisions. Interestingly, we also found that patients diagnosed with schizophrenia, especially those reporting frequent hallucinations, were more sensitive to disambiguating sensory information in this paradigm, suggesting a stronger weighting of sensory information relative to prior beliefs and thus enhanced prediction error signaling. In sum, this project helped to elucidate predictive processing mechanisms in perceptual decision making in their alterations in psychosis. The findings are a promising starting point for intervention studies, e.g. using transcranial magnetic stimulation to manipulate prediction error signaling as a potential treatment strategy in patients diagnosed with schizophrenia.

Publications

 
 

Additional Information

© 2026 DFG Disclaimer / Copyright / Privacy Policy

Textvergrößerung und Kontrastanpassung