Wieger Scheurer

The general focus of my research pertains to uncovering the prediction-generating models in the human brain during natural perception. My work capitalises on recent advancements in artificial intelligence (AI) to allow for interpretable and computationally explicit tests of our internal generative models. In other words, I study the neural information processing of our brain as a prediction machine using machine learning. For instance, I investigate the representational content of predictions by combining the hierarchical feature extraction and generative capabilities of artificial neural networks (ANNs) to quantify the inherent predictability of naturalistic stimuli at different granular levels. This eliminates the need for controlled experimental designs that involve high-level manipulation of stimulus probability, which, in turn, allows for experimental settings to be made more naturalistic. Together with extensive comparisons to neural recordings (fMRI/MEG) this constitutes a rich empirical toolbox to investigate how predictions shape our perception in "the wild". Now we know that our brain predicts, I will use this novel toolbox to further integrate our understanding of the predictive brain by addressing questions about what, when, and how the brain predicts.

Prior to joining the Predictive Brain Lab I obtained a BSc, MSc, and a research MSc at the Psychology department of the University of Amsterdam (UvA) where I specialised in Brain & Cognition and Research Methods with a particular focus on the intersection between Cognitive Neuroscience and Artificial Intelligence. 

Apart from how the brain works I am also interested in pianos, puzzles, bicycles, books, large machines, coffee, and an increasing number of birds.