Sven Wientjes

Cognitive Scientist

Currently I am working as a PhD student at the Ghent University in Belgium, in the lab of Clay Holroyd. My main research interests are in hierarchical reinforcement learning, cognitive control, and episodic memory. My work centers on computational modeling of human behavior and designing novel experiments to test these models. I have technical experience with Bayesian cognitive modeling in Stan and with PyTorch. I am also active as a supervisor of student research projects and an organizer of EEG courses and seminars at Ghent University.

Selected Work

For a full list of affeliated work, see my Google Scholar page.

Humans can learn predictions of upcoming events based on experience. These predictions are believed to be stored in a "cognitive map". A simple cognitive model of such cognitive map learning is the successor representation, an algorithm derived from reinforcement learning. This cognitive map can be used to identify important central locations, which can be used as efficient "subgoals" when forming temporally extended plans. In this project, we use Bayesian cognitive modeling to test whether individual differences in learning successor representations can explain differences in the formation of subgoals during temporally extended decision making.

In this project we used machine learning methods to jointly classify neural and behavioral signatures of mind wandering in a novel random finger tapping task. The work was spearheaded by Josephine Groot and supervised by Matthias Mittner at the University of Tromso, Norway. As a master intern I worked on implementing theory-constrained hidden markov models in Rcpp and Stan.

Curriculum Vitae