The Role of Prior Uncertainty on the Neural Response to Semantically Predicted and Unpredicted Events
No Thumbnail Available
Human perception of the world and one’s own body is intrinsically uncertain. Two types of uncertainty are commonly distinguished: uncertainty in the sensory processing stream (likelihood uncertainty) and uncertainty regarding the causes of sensations (prior uncertainty). Although studies are beginning to investigate the role of prior knowledge on perception, action, and cognition, it is currently unclear to what extent neural responses to predicted and unpredicted events are weighted by the amount of uncertainty in the prior that is used to form these predictions. In the current study we devised a numerical inference task that allowed us to independently manipulate predictions and prior uncertainty on a semantic level without establishing artificial probabilistic stimulus-stimulus relationships. Contrary to our expectations, we found no brain areas where neural responses to predicted and unpredicted events are differentially weighted by prior uncertainty. However, we did find increased activity in the right visual cortex in response to semantically unpredicted events and a marginally significant increase in activation in the right supramarginal gyrus when uncertainty was low compared to when it was high. The implications of these findings are discussed with regard to the co-localization of brain regions encoding uncertainty and those responsible for computing prediction errors. Moreover, the results suggest that early sensory cortices respond to unpredicted events even when statistical relationships were kept constant and predictions had to be derived semantically.
Faculteit der Sociale Wetenschappen