My research interests are in the development of Neuroimaging Methods for fMRI, MEG and EEG analysis. The research is embodied in a mathematical framework called Probabilistic Generative Modeling in which biophysical forward models are inverted using approximate Bayesian inference. I have recently developed a family of spatio-temporal models for analysing neuroimaging data which combine information from both spatial and temporal forward models to provide better estimates of neuronal activity. I have also developed algorithms for brain connectivity analysis which allow one tomake inferences about how activity in one area affects activity in another, and how this is changed by experimental manipulation. These approaches can be combined with Bayesian model comparion to enable neural network theories of brain function to be formally compared using neuroimaging data. Current projects include the development of an EEG-fMRI fusion model and analysis of the nested oscillations that support working memory. I am a co-author of the Statistical Parametric Mapping (SPM) software package for neuroimaging data analysis.