Abstracts

Despite the fact nonlinear EEG analyses have provided valuable insights into the complex spatio-temporal dynamics of the epileptogenic process, these techniques are nevertheless limited. This is due to the fact that the underlying assumptions of low-dimensionality and nonlinearity might not be fullfilled in all cases. We here propose a more general analysis technique which is based on concepts from the theory of stochastic processes and study the ability of our method for an interictal localization of the epileptic focus in MTLE patients., Using a moving-window technique we estimated the coefficients (drift and diffusion) of the Fokker-Planck equation from long-lasting (up to several days), multichannel EEG data recorded intracranially during the interictal state from up to now 5 MTLE patients. The resulting profiles of the estimated measures were averaged over time to compare the stochastic dynamics in different brain regions., In all patients investigated so far we observed that the stochastic qualifyers are most pronounced at recording sites that coincided with the epileptic focus as determined by the presurgical workup and the complete postoperative seizure control. Moreover the observed indications for a higher degree of nonlinearity in these brain regions confirm our previuos findings that were obtained using other measures for nonlinearity., These findings indicate that a more general characterization of the EEG in terms of a nonlinear stochastic process can provide valuable information for the interictal localization of the epileptic focus in MTLE. This technique might thus contribute to an improvement of the presurgical evaluation., (Supported by Deutsche Forschungsgemeinschaft.)