Diagnosing Brain Disorders of Divers in Different Frequency Bands of Brain Using Convolutional Neural Network

Electroencephalographic signal processing of divers can reveal useful information about the brain complications caused by diving that these people suffer over the years. In this paper, an optimal method based on a convolutional neural network is presented, which extracts the optimal features from the recorded signal by the network itself. This method is suitable for recognizing the abnormalities of different brain regions in different people in different bands. The obtained results show anomalies in the alpha band and in O1, O2, T7 and T8 channels. Two groups can be separated in the alpha band with accuracy, precision and recall of 94.2%, 92.9% and 91.3%, respectively.