Disturbance of patterns in EEG spatial correlations

In the study of epileptic seizure or epileptic attack, a strategy receiving increased attention is the use of nonlinear methods in detecting the earliest dynamical changes preceding seizures. The methods usually consider continuous EEG measurements from epileptic patients to predict and ultimately control seizures. As part of the inquiry into the structure of the dynamics of the brain activity we investigate changes amongst the EEG signals being recorded at different locations on the scalp. Patterns emerging from the correlation coefficients between the EEG channels seem to be disturbed with the approach of a crisis. Results show that those patterns are often disturbed 10 to 15 minutes before the beginning of crises, helping to detect the earliest dynamical changes preceding seizures.