Fig. 5.

Simulated neuronal drive associated with one neuronal population. DCM for MEG was first used to infer the neuronal parameters of CMC models. a) The ensuing neuronal parameters were used to generate condition specific neuronal responses (e.g., pre-synaptic signals). b) To generate the input for the haemodynamic model, the neuronal drive functions were convolved (or shifted in time) with the onset of each trial of the fMRI experiment. c) All condition specific neural responses were then summed to generate the neuronal drives to neurovascular coupling units. This was repeated for each neuronal population and brain region.