A3: Representation of aversive cues predicting appetitive outcomes

In everyday life, there are many instances in which an aversive event predicts a positive outcome. Yet, how the prospect of a reward linked to a negatively valenced stimulus can change the responding to and representation of the aversive stimulus is unknown. This project aims to establish a classical conditioning paradigm in which a negatively valenced conditioned stimulus (CS) differentially predicts a positively valenced US. This paradigm in combination with computational modelling will allow us to investigate learning related effects at the autonomic (EDA) and neurobiological level (fMRI). With respect to model based fMRI, we will focus on multivariate methods such as representational similarity analysis5 to characterize dynamically changing representations of the CS as a function of learning.

We expect that the pairing with a reward will lead to a decrease of the unpleasantness rating and EDA response to the paired CS. With respect to multivariate neural representation of each CS, we expect them to differ initially in the insula and mPFC, but over time the representations of the aversive CSs in these regions should get more similar to the pattern evoked by the rewarding US. As previous studies have suggested a role of endogenous opioids in signaling hedonic aspects of reward, we expect that the pattern changes described above can be blocked by Naloxone.