PREDICTION OF STROKE USING EEG SIGNALS

A stroke is a disruption of blood supply to the brain that leads disability. Rehabilitation can reduce damage and other complications also recover disability due stroke. One instrument be used in post-stroke patients monitoring Electroencephalogram (EEG). EEG signals are recorded from several channel pairs. So apart maintaining signal sequences, handling time connectivity between channels essential. Therefore, sequence need handled simultaneously on two dimensions. The vertical dimension was multi-channel dimension, horizontal sequences each concerning time. proposed system uses 2D-Convolutional Neural Networks method identify based signals. First, wavelet filter obtain frequency range 1–13 Hz represent patients. identification results one three levels stroke, i.e., No Stroke Minor Stroke.