When models matter: How we learn to arbitrate between model-based versus model-free control

01 November 2021 → 31 October 2025
Research Foundation - Flanders (FWO)
Research disciplines
  • Social sciences
    • Neuroimaging
    • Cognitive processes
    • Learning and behaviour
    • Research methods and experimental design
    • Statistics and data analysis
Metacontrol decision-making task switching
Project description

In cognitive psychology a common distinction is made between higher-order mental processes and more simple forms of behavior, which has led to a variety of widely popular dual system theories. However, it is often unclear what regulates these higher-order processes, and they are usually hard to define or capture in a single latent factor. Therefore, the goal of this project is to take a different approach and test the hypothesis that the regulation of different forms of control is guided by low-level associative processes, and hence can be learned in and of itself. Specifically, I will test whether people can reliably alternate between model-based and model-free control strategies depending on environmental demands, and whether respective successes of either strategy can be linked to contextual cues or selectively reinforced. Subsequently, I will use fMRI to distinguish between two candidate processes through which this regulation might be achieved, namely, forward planning or generalization. Lastly, I aim to look at individual differences in learning this regulation which have the potential to be of high transdiagnostic value in clinical practice. In sum, this project will advance a holistic understanding of cognitive control and decision-making and could inform treatments by identifying target regulatory mechanisms and their underlying neural processes.