My research addresses the computational neuroscience of human cognitive decision making.

State inference in uncertain environments.

Everyday we try our best to make optimal decisions, but the information we get from our environment is often uncertain and incomplete. In the machine learning, these situations are referred to as partially observable decision-making problems, in which the true (or hidden) state of the world cannot be directly observed. Instead, we can only estimate the hidden state of the world using available observations, and make decisions based on our inferred 'beliefs'. My research to date has uncovered some of the mechanisms by which uncertainty is resolved and beliefs constructed in the brain. I've been able to show how distinct regions of the prefrontal cortex support computationally precise processes during cognitive decision-making, and I currently do brain imaging (fMRI) experiments and computational modeling (using Hidden Markov Models, HMMs) to build more complete models of these processes.

Decision making with social interaction.

An especially interesting type of cognitive decision-making problem occurs in social communication, as the ability to optimise mutual interactions requires the ability to read each others minds. My research supports the contention that our strategies and ensuing behaviour are optimized recursively, and this is a key process at the heart of 'Theory of Mind'. Not only have we been able to show some of the precise neural computations performed in the 'Theory of Mind' regions of prefrontal cortex, but we have been able to identify very specific deficits in patients with autistic spectrum disorder. My current work probes some of the complexities of how Theory of Mind is achieved in the brain, and at the moment I am studying human brain (BOLD) activity during group decision making tasks, including 'hyper-scanning' two subjects as they interact (cooperate) to solve a task together in separate fMRI scanners.