Understanding dynamic state changes in temporal lobe epilepsy.

The authors review their work in applying nonlinear dynamics to predict onset of seizures in patients with medically refractory temporal lobe epilepsy. The underlying mathematical methodology is presented in some detail. To illustrate their approach, they present an extensive discussion of the analysis of preictal data from two seizures of one patient, and from one disease-free subject. They find similar behavior in some nonlinear measures across seizures, which suggests the possibility of forming a robust method of seizure prediction. However, despite clinical and electrographic preictal and ictal similarity, they have also found marked heterogeneity in other nonlinear measures of preictal activity across seizures arising out of stage 2 nonrapid eye movement sleep. The underlying basis for this variation remains uncertain and needs to be the subject of further intense study to gain a better understanding of the dynamic basis of epilepsy. The origin of these heterogeneities may or may not be related to the much larger differences in nonlinear measures between patients and disease-free subjects. To understand these differences, the authors think it is crucial to pay close attention to potentially confounding factors such as behavioral and other state changes, and to study and report in detail the ways in which relevant nonlinear measures behave in the presence of such changes, independent of seizure onset.