Abstract

I will present some research of my group on predictive modeling and outline some links to selective attention. Selective attention is in the service of coping with the overwhelming informational offer available for our information processing system. At all levels of the system we have to prioritize part of the information at the expense of disregarding or even actively inhibiting other parts. Research on selective attention is one of the hottest topics in experimental psychology and cognitive neuroscience since many decades. Predictive processing is another hot topic that also has a long tradition, but became increasingly popular in the last decade only. Predictions inform our system about future states of the world. I will present recent experimental work from my group that shows that strong sensory predictions for a sound can modulate brain responses to that sound and may even elicit sensory responses when the sound is absent. This supports the idea that the brain uses sensory templates as a neural substrate for prediction. I will also argue that involuntary selective attention is a (late) consequence of a mismatch detection during predictive processing. Its psychophysical and behavioral consequences can be regarded as error signals. Even voluntary attention can be viewed from the predictive processing perspective; for example, attention can guide the establishment of predictive models or it can possibly set a bias towards one of several competing predictive models.