Early seizure detection

For patients with medically intractable epilepsy, there have been few effec tive alternatives to resective surgery, a destructive, irreversible treatme nt, A strategy receiving increased attention is using interictal spike patt erns and continuous EEG measurements from epileptic patients to predict and ultimately control seizure activity via chemical or electrical control sys tems. This work compares results of seven linear and nonlinear methods (ana lysis of power spectra, cross-correlation, principal components, phase, wav elets, correlation integral, and mutual prediction) in detecting the earlie st dynamical changes preceding 12 intracranially-re corded seizures from 4 patients. A method of counting standard deviations was used to compare acro ss methods, and the earliest departures from thresholds determined from non -seizure EEG were compared to a neurologist's judgement. For these data, th e nonlinear methods offered no predictive advantage over the linear methods . All the methods described here were successful in detecting changes leadi ng to a seizure between one and two minutes before the first changes noted by the neurologist, although analysis of phase correlation proved the most robust. The success of phase analysis may be due in part to its complete in sensitivity to amplitude, which may provide a significant source of error.