A Novel Nonlinear Dynamic Method for Stroke Rehabilitation Effect Evaluation Using EEG.

Evaluating the effect of stroke rehabilitation based on electroencephalogram (EEG) is still a challenging problem. This paper presents a novel nonlinear dynamic complexity method for the evaluation of stroke rehabilitation effect from EEG signal. Our method calculates the nonlinearly separable degree (NLSD) of EEG signal, and then employs an indicator, called mean nonlinearly separable complexity degree (Mean_NLSD), to efficiently and accurately evaluate therapy effect of stroke patients. This paper under twelve stimuli conditions on eleven patients and eleven control subjects indicates that in general Mean_NLSD is smaller at the lesion regions and that the Mean_NLSD of the control subjects is stochastic. Compared with conventional spectral methods, such as mean power spectral density (PSD), Mean_NLSD is more sensitive and robust. Overall Mean_NLSD may offer a promising approach to facilitate the evaluation of stroke rehabilitation effect.