Causality analysis of P300 recordings focused on the localization of active brain areas

Abstract

This paper focuses on the causal influence in the brain. These effects are analyzed by stimulating visually or aurally a patient with an external sensorial stimulus and investigating the outcome into the brain (event-related potentials) in order to detect causal influences from one neuron or neuronal population to another. This analysis is achieved by applying time varying multivariate autoregressive models to the ERP signals. The causal interaction is measured using the Granger Causality Index (GCI). The obtained results using causality can validate the analysis performed by an specialist, and they demonstrate the causal interactions between electrophysiological signals of different sites of the cortex, occurring during cognitive processes related with ERP.