![]() A core software component of such a system is a sleep scoring algorithm, which can reliably perform automatic sleep stage scoring given the patient’s EEG signals. Wearable EEG is a strong candidate for such use. In this case an affordable, portable and unobtrusive sleep monitoring system for unsupervised at-home use would be ideal. 26 There is therefore a pressing need for longitudinal sleep monitoring for both medical research and medical practice. Recent research suggests that detection of sleep/circadian disruption could be a valuable marker of vulnerability and risk in the early stages of neurodegenerative diseases, such as Alzheimer’s disease, Parkinson’s disease and multiple sclerosis, and that sleep stabilization could improve the patients’ quality of life. 20 In Table 1 we reproduce the Rechtschaffen and Kales sleep staging criteria, 22 merging the criteria for N3 and N4 into a single stage (N3). Sleep scoring is performed according to the Rechtschaffen and Kales sleep staging criteria. Then, one or more experts classify each epoch into one of the five stages (N1, N2, N3, R or W) by quantitatively and qualitatively examining the signals of the PSG in the time and frequency domains. After the PSG is recorded, it is divided into 30-s intervals, called epochs. The PSG includes an electroencephalogram (EEG), an electrooculogram (EOG), an electromyogram (EMG), and an electrocardiogram (ECG). The totality of the signals that are recorded through these sensors is called a polysomnogram (PSG). These stages are defined by electrical activity recorded from sensors placed at different parts of the body. Formerly, stage N3 (also called slow wave sleep, or SWS) was divided into two distinct stages, N3 and N4. These are rapid eye movement (stage R) sleep and 3 non-R stages, stages N1, N2 and N3. 26Īccording to the American Academy of Sleep Medicine (AASM) manual, 11 sleep is categorized into four stages. 26 Associations between disruption of normal sleep patterns and neurodegenerative diseases are well recognized. The health consequences of reduced sleep, abnormal sleep patterns or desynchronized circadian rhythms can be emotional, cognitive, or somatic. The performance of our method appears to be uncorrelated with the sleep efficiency and percentage of transitional epochs in each recording. Our method has both high overall accuracy (78%, range 75–80%), and high mean \(F_1\)-score (84%, range 82–86%) and mean accuracy across individual sleep stages (86%, range 84–88%) over all subjects. We used a single channel of EEG from this dataset, which makes our method a suitable candidate for longitudinal monitoring using wearable EEG in real-world settings. We used an openly available dataset from 20 healthy young adults for evaluation. ![]() We used class-balanced random sampling across sleep stages for each model in the ensemble to avoid skewed performance in favor of the most represented sleep stages, and addressed the problem of misclassification errors due to class imbalance while significantly improving worst-stage classification. We used ensemble learning with an ensemble of stacked sparse autoencoders for classifying the sleep stages. Our time-frequency analysis-based feature extraction is fine-tuned to capture sleep stage-specific signal features as described in the American Academy of Sleep Medicine manual that the human experts follow. We developed a machine learning methodology for automatic sleep stage scoring. ![]()
0 Comments
Leave a Reply. |