Spread of Epileptic Seizure Activity in Experimental and Clinical Epilepsy: The Use of Mutual Information Analysis

Abstract

The mechanism of spread of epileptic seizures re-mains an important unsolved question in the process of epileptogenesis. This question is not only of fundamental importance for understanding how focal epileptic activity propagates to neighboring and/or distant brain areas but also of practical interest. This last aspect is particularly relevant where the primary site of an epileptogenic focus must be determined, namely in those patients in whom the surgical removal of the focus is being considered. The propagation of focal epileptic activity is not easy to assess by way of visual examination of a large number of EEG records even if recorded from subdural or intracerebral electrodes. It is generally assumed that the EEG signal recorded from the focal area will present a time lead over signals recorded from distant areas. However, a main difficulty is that it is not possible to assess visually small time differences in the order of a few milliseconds which may exist between such EEG signals. Therefore, computer- based analysis methods have been devised to obtain estimates of time delays between such EEG signals. Brazier (1972,1973) carried out the first type of such an analysis. Based on the phase angle (φ₀ in degrees) at a particular frequency (f₀ in Hz) for which coherence was maximal, a time delay (Δt) was calculated according to: φ₀/360f₀. This method, however, may give ambiguous results since the phase angle at one particular frequency can be transformed into a time delay only if the phase as function of frequency fits a straight line which intersects the axes at the origin. However, this does not always occur in practice. Sometimes phase functions are encountered between EEG signals which at zero frequency intersect the phase axis at ±180°.