Seminar 13/6/2024

Alteration in predictive language processing in different stages of psychosis

Abstract: Alterations in speech represent a core symptom of schizophrenia, the underlying mechanisms and links to positive and negative symptoms are still limited. Here, I am using the predictive coding framework to investigate how predictive language alterations link to positive symptoms. Specifically, I am interested in exploring how an imbalanced weighting of prior knowledge and sensory evidence links to hallucinations and delusions. We developed a novel predictive language task which reflects Bayesian inferential processing. Additionally, we set up a Bayesian belief updating model which allows us to extract and simulate the prior weight during language perception, informing us whether individuals rely more on prior knowledge or more on sensory evidence when listening to speech. I will present results from a healthy cohort with schizotypy scores and schizophrenia patients first in acute psychosis and then in psychotic remission. In both cohorts, I will be able to show a link between an overweighting of high-level priors and positive symptoms. Additional, I will link the results to EEG and spectroscopy data.

Bio: Franziska Knolle is a research group leader (since 2022) at the Department of Neuroradiology at the Technical University of Munich, in Germany, and former Marie Skłodowska-Curie Fellow (2020). She did her PhD at the Max-Planck-Institute of Human Cognitive and Brain Science in Leipzig in 2012. After a break from academia during which she studied Human Medicine, she conducted two postdocs at the Department of Physiology, Development and Neuroscience, and the Department of Psychiatry at the University of Cambridge from 2016-2019.

Email Website