Long-term scalp epileptic EEG quantification with GMA dynamics

摘要

The paper concerns the problem of automatic seizure detection based on scalp EEG and proposes to employ the generalized measure of association (GMA) to quantify the statistical dependencies and infer the dynamical interactions of brain regions with the focus area. The experimental results with clinical recordings show that the estimated GMA values changes dramatically before and during epileptic seizures reflecting the dynamic coupling and decoupling between brain regions, which can be an useful measure to quantify epileptic EEG signals.