A Single Channel EEG-Based Algorithm for Neonatal Sleep-Wake Classification

Sleep is categorized as an arrangement of modifications occurring in our body inside our brain, muscles, working its way through our eyes (occipital lobe), respiratory along with cardiac activity. It makes the human body fresh and ready for the next day. In neonates, it is essential for brain and physical development. Polysomnography is the gold standard for determining and classification of sleep stages. However, it is expensive and requires human intervention. Therefore, over the past two decades, researchers proposed multiple algorithms for automatic neonatal sleep stage classification. All the previous studies used multichannel EEG recordings for classification. Not every intensive care unit contains a multichannel EEG extraction device. For this reason, a single channel automatic neonatal sleep-wake classification algorithm, using a support vector machine, has been proposed in this paper. 3525 30-s training and testing were used to train and test the network. The proposed algorithm can reach sleep-wake classification accuracy of 77.5% with mean kappa 0.55 using single channel EEG. The results were extracted using five-fold cross-validation and the mean has been reported in this paper. Experimental results and statistical analysis show that single channel EEG can be used for neonatal sleep classification with notable accuracy.