When a person or an animal is surprised, i.e. confronted with an event that deviates from our expectations, learning happens: expectations are adjusted to reality. We want to figure out how organisms integrate the diverse information from the environment into a model of their surroundings and then adjust their experiences, so that prediction errors no longer occur, as the model was brought into accordance with the reality. In our experiments we use a widespread theory, that is based on error correction. We are especially interested in which way errors are processed in extinction learning.
A negative prediction error dictates that an anticipated event does not occur. A positive prediction error describes an event that occurs unexpectedly. In our project we investigate how positive and negative predictions errors affect extinction learning.
Find out more on the project page >>