Salience indicates what to pay attention to and instigates learning to pay attention. The insula forms a central hub of the salience network. However, the nature of the salience signals processed by the insula remained a matter of debate. For example, motivational salience increases with both the aversiveness and appetitiveness of predicted outcomes and one would therefore expect a salience-processing region to show increased activity in both domains. While there is some evidence for common encoding of appetitive and aversive outcomes, the insula has also been shown to preferentially encode the aversive rather than the appetitive domain. Using both Pavlovian and instrumental tasks, in this talk I provide evidence that the anterior insula encodes errors in the prediction of motivationally salient outcomes (absolute prediction errors) in a subjective fashion (based on individual ratings) and similarly in the appetitive and aversive domain. By contrast, middle insula regions preferentially encode salience prediction errors in an objective fashion (based on actual probabilities). Moreover, at the time of outcome predicting cues, the anterior insula encodes salience prediction errors preferentially in the aversive domain, suggesting that it integrates salient outcomes more readily into predictions when they occur in the aversive domain. The findings are compatible with the anterior insula playing a role in subjective experience and reconcile distinct views on its domain specificity.

Bio:
Philippe Tobler is Professor in Neuroeconomics and Social Neuroscience at the Department of Economics, University of Zurich. His research explores value-based decision making and reward learning in both social and non-social contexts, including how constituents of subjective value (such as risk, effort, delay) are processed in the brain; he uses behavioural experiments, fMRI, and pharmacological manipulations.