Dynamic Autoregressive Neuromagnetic Causality Imaging (DANCI)

This presentation provides a demonstration of how Granger causality (GC) can be applied to MEG data to visualize dynamic functional connectivity and causality between cortical regions on a millisecond time scale. GC is derived from autoregressive models and provides directionality information. We apply the GC technique to dynamic statistical parameter map source space to demonstrate that the dynamics of neural networks can visualized during a perceptual task. The results from this demonstration coincide with models of speech perception and suggest that Dynamic Autoregressive Neuromagnetic Causality Imaging (DANCI) can be used to investigate and verify theoretical neural network models of brain function. Key-Words: Magnetoencephalography; Neural connectivity; Autoregressive modeling; Granger causality; Large-scale neural networks; Neural network dynamics; Cognitive neuroscience; Neuroimaging.