Seizure warning algorithm based on optimization and nonlinear dynamics

摘要: There is growing evidence that temporal lobe seizures are preceded by a preictal transition, characterized gradual dynamical change from asymptomatic interictal state to seizure. We herein report the first prospective analysis of online automated algorithm for detecting transition in ongoing EEG signals. Such, constitutes seizure warning system. The estimates STLmax, measure order or disorder signal, signals recorded individual electrode sites. optimization techniques were employed select critical brain sites exhibit epileptic seizures. Specifically, quadratically constrained quadratic 0-1 programming problem formulated identify was tested continuous, long-term recordings obtained 5 patients with epilepsy. For patient, we use half train parameter settings, which evaluated ROC (Receiver Operating Characteristic) curve analysis. With best setting, applied all cases predicted an average 91.7% false prediction rate 0.196 per hour. These results indicate it may be possible develop devices diagnostic and therapeutic purposes.