My research focuses on developing new analysis methods to study neural information processing, with an emphasis on dynamic aspects of human brain activity.

Real-life behaviour such as conversation, sports, household tasks and so on, depend on the dynamic coordination of brain activity in multiple brain regions. Elucidating the nature of such inter-regional communication is critical for us to understand behaviour. Currently, we have a number of non-invasive human brain measurement techniques available: fMRI, MEG, EEG and NIRS, each with its advantages and disadvantages but with no single technique sufficient to comprehensively understand the brain. My research focuses on multi-modal integration analysis – developing quantitative models that combine information for multiple measurement techniques. In particular we pursue integration of MEG and fMRI data to provide high temporal and high spatial resolution. Our approach is based on constructing a generative model of the dynamics of brain activity, and inferring the model structure using Bayesian statistics.

One of our current projects is construction of sophisticated models of the network dynamics of brain activity so that we can perform realistic simulations of signals recordable by MEG, EEG, fMRI, and NIRS. In particular, we are developing methods to incorporate anatomical network information within these models. Elsewhere, we are developing new diffuse optical tomography algorithms to make NIRS measurements a more reliable tool for ambulant brain activity measurement.