DCNet: A Self-supervised EEG Classification Framework for Improving Cognitive Computing-enabled Smart Healthcare.

Abstract

Cognitive computing explores brain mechanisms and develops brain-like computing models for cognitive processes. EEG measures brain activity, and EEG classification identifies patterns using machine learning algorithms. Combining EEG classification with cognitive computing offers insights into cognitive processes, brainmachine interfaces, and cognitive state monitoring. We propose (DreamCatcher Network) DCNet, a self-supervised learning method for diagnosing sleep disorders using EEG. DCNet autonomously learns comprehensive representations through contrast learning, reducing annotation time. The training process involves feature learning, classification, time-series contrast learning, and data enhancement. Experimental results on the Sleep-EDF dataset achieved 81.28% average accuracy. Validation on the HAR dataset showed model efficiency and scalability, with 92.51% accuracy on the test set. DCNet has the potential to revolutionize sleep disorder diagnosis and enhance the development of cognitive computing-enabled smart healthcare systems.