Skip to main content

The dynamics of reinforcement learning in the basal ganglia


Research objectives and content
The objective is to study, analytically and numerically, the dynamical behaviour of a model of those brain regions thought to perform (reinforcement) learning. We are interested in how this system can develop memory traces and represent the expectation of reward. We aim to develop an abstract and a detailed model that closely follows the experimental data and can predict the activity patterns in the brain during learning. To test this, experiments for functional Magnetic Resonance Imaging will be developed.
Training content (objective, benefit and expected impact)
The candidate learns analytic tools from dynamical systems analysis to study a complex biological system. He will also learn numerical analysis of complex and more biologically realistic systems. The collaboration with the Institute of Psychiatry will greatly benefit the candidate's knowledge of experimental neuroscience, particularly with respect to fMRI. This combination of analysis, simulation and devising experiments for the new imaging tools, is expected to play a major role in neuroscientific research in the years to come.
Links with industry / industrial relevance (22)
The results of the project will provide input to project MAX (with British Telecom) about how autonomous agents can learn efficiently from their environment. Such autonomous agents are being developed for tasks varying from information gathering on the Internet to underground cable repair.

Funding Scheme

RGI - Research grants (individual fellowships)


King's College London
The Strand
WC2R 2LS London
United Kingdom

Participants (1)

Not available