The article presents the development of a program capable of real-time detection of REM phases during the human sleep. For this purpose, 39 electroencephalogram (EEG) recordings from PhysioNet were used. To achieve the goal of the project authors selected following set of parameters: the average amplitude of the signal, alpha and delta power band in frequency domain and the ratio Alpha-Delta, for 30 second interval. An Artificial Intelligence (AI) has been developed with Keras and trained with those parameters for 34 patients. Finally, the AI has been tested on the last 5 patients, by simulating a true night. It reaches 62% in sensibility for REM Phase detection, and 85% in specificity. Obtained results are promising in terms of real-time REM phase detection, but the approach needs further development.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.