Computational models of dynamics in brain networks underlying action selection

Deep brain stimulators for Parkinson’s disease are about to get smarter, but doctors first need a mathematical map of how healthy brains choose which action to take. Current stimulators deliver constant electrical pulses to brain regions affected by Parkinson’s, which can reduce tremor but also cause side effects because the stimulation is one-size-fits-all. Newer devices have multiple contacts that can both record brain activity and deliver targeted stimulation in response. The problem is that no one has yet worked out what “normal” patterns of neural activity look like during action selection—the process of deciding, for example, whether to reach for a cup or stand up. Without that baseline, engineers cannot program the stimulators to restore healthy dynamics. This programme will build computational models of the brain networks involved in action selection, then simulate how different patterns of multi-contact stimulation might shift abnormal activity back toward a healthy state. If successful, the work will directly inform the next generation of closed-loop deep brain stimulators—devices that adjust stimulation in real time based on what the brain is actually doing. That could mean fewer side effects and better symptom control for people with Parkinson’s, without requiring patients to think about the technology at all.