Habitual behavior after stress: alterations in outcome representations?

• Applicant: Professor Dr. Lars Schwabe
• Subject Area: Biological Psychology and Cognitive Neuroscience
• Term: from 2019 to 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 434153598

Instrumental learning can be controlled by a goal-directed and by a habitual process. Recent findings show that acute stress may induce a shift from goal-directed to habitual processes. Although this shift after stress is also highly relevant for clinical settings, it is still completely unknown when this shift occurs. Moreover, there is currently a controversy about whether there is a training-dependent shift from goal-directed to habit learning, irrespective of stress, in humans or not. To address these open questions, the present project will employ a novel approach: based on the idea that goal-directed and habit learning differ per definition in whether the outcome that is engendered by a response is encoded or not, the project will combine multivariate decoding analyses with electroencephalography (EEG) to assess the mental representation of action outcomes during the presentation of discriminative stimuli and during responding across an extended learning session. This approach will allow for the first time a fine-grained analysis of the dynamics of goal-directed and habitual processes across learning. In order to assess the impact of stress on these dynamics, healthy participants will undergo a standardized stress or control procedure before they perform an instrumental learning task - while EEG is recorded - during which stimulus(S)-response(R)-outcome(O) sequences are presented, thus enabling both goal-directed (S-R-O) and habit (S-R) learning. In addition to specific EEG-measures, we will include specific trials in which one outcome is devalued which allows a separation of goal-directed and habit processes at the behavioral level as well as eyetracking. It is hypothesized that there is a shift from goal-directed to habit processes with increased training and that this shift is reflected in a decrease of the (neural) representation of action outcomes. Stress is assumed to facilitate this shift, thus leading to reduced outcome representations, compared to the control condition, in particular after moderate training. In sum, this project will provide novel insights into fundamental aspects of human behavior: (when) does human responding become independent of its outcomes and how does stress affect the dynamics in the control of instrumental behavior across learning?

• DFG Programme: Research Grants
• International Connection: United Kingdom
• Co-Investigator: Professor Dr. Bernhard Staresina