A natural evolution optimization based deep learning algorithm for neurological disorder classification.

Neurological disorders are one of the significant problems of the nervous system that affect essential functions of the human brain and spinal cord. Monitoring brain activity through electroencephalography (EEG) has become an important tool in the diagnosis of brain disorders. The robust automatic classification of EEG signals is an important step toward detecting any brain disorder in its earlier stages before status deterioration. Motivated by the computation capabilities of natural evolution strategies (NES), this paper introduces an effective automatic classification approach referred to as natural evolution optimized deep learning (NEDL). The proposed classifier that is included in a signal processing chain comprises of other state-of-the-art methods in a single framework for EEG classification. Initially, the EEG signal is decomposed into a number of sub-bands by applying wavelet transform where a number of spectral and statistical features are extracted. These extracted features are examined using the approach of artificial bee colony to optimally select the best features. Finally, the selected features are processed using the proposed NEDL classifier, after which the proposed approach is evaluated using two benchmark datasets that address epilepsy disease and motor imagery. Several experiments are conducted where the proposed approach outperforms other deep learning models along with existing approaches.