Local Lyapunov exponents detect epileptic zones in spike-less interictal MEG recordings

Objective: In terms of dynamical system theory the rapid alteration of elec tromagnetic brain signal properties observed with transition from intericta l into ictal epileptic activity implies an alteration between at least two dynamical states (attractors). We explored whether such a multistability is reflected also in the dynamical characteristics of the interictal signal. Methods: A combined method consisting of structural MRI, multichannel magne toencephalography (MEG) and the non-linear dynamics was applied for the det ection of subthreshold interictal activity in temporal lobe epilepsy. Emplo ying multichannel MEG we calculated local Lyapunov exponents (ILE) as a mea sure of the probability of state transitions in patients with mesiotemporal and extra-mesiotemporal focal epilepsy. Results: The spatial distribution of the variability of ILE was found to be correlated to the side of the epileptic focus, thus, providing a reliable estimate of its localization. Conclusion: Our results demonstrate that interictal epileptic activity is c haracterized by an increase of the number of transitions between different attractors. (C) 2001 Elsevier Science Ireland Ltd. All rights reserved.