Inferring multi-scale neural mechanisms with brain network modelling
Neuroimaging gives us an incomplete picture of whole-brain activity. We have extended the ‘The Virtual Brain’ connectome-based modeling platform to integrate multimodal data with biophysical models in order to support neurophysiological inference. Simulated cell populations are linked with white-matter connectivity estimates and driven by individual electroencephalography (EEG) derived electric source activity. The models are fit and cross-validated to predict subjects’ individual whole-brain resting-state functional magnetic resonance imaging (fMRI) time series. Evaluation of the simulated firing and synaptic activity underlying fMRI predictions enables the formulation of novel hypotheses that explain several empirical observations from electrophysiology, EEG and fMRI. Thus unifying biologically realistic models mechanistically links up various empirical observations. In my presentation I will emphasize the integrative role for computational modeling to complement empirical studies.