A 2D clustering approach investigating inter-hemispheric seizure flow by means of a Directed Transfer Function

Abstract

Determination of seizure origin is often challenging due to the rapid speed at which electrical activity propagates throughout the brain. The Directed Transfer Function (DTF) has been proposed and validated as a quantitative approach to determine the flow of seizure activity. In this work, outflow and inflow features are extracted from the DTF matrix and used as inputs to a Kmeans unsupervised clustering approach. Results demonstrate the ability of the proposed methodology in automatically identifying sources and sinks of seizure activity as well as discriminating primary from secondary generators. Such distinction could lead to more tailored surgical resections.