Automatic sleep staging: from young adults to elderly patients using multi-class support vector machine

Towards Improving Sleep Quality Using Automatic Sleep Stage Classification and Binaural Beats

Sleep disorders are extremely common in today's society and are greatly affecting the health and safety of every person suffering from one. Over the last decades, Automatic Sleep Stage Classification (ASSC) systems have been developed to assist specialists in the sleep stage scoring process and therefore in the diagnosis of sleep disorders. Binaural beats are auditory phenomena that have been shown to have a positive impact in sleep quality and mental state. This paper introduces a framework that combines an ASSC system and a binaural beats generator in real time. Our goal is to pave the way for developing systems which could reproduce specific binaural beats depending on the detected sleep stage, in order to entrain the brain into a more efficient sleep. For the ASSC stage, different classifiers were evaluated using data signals retrieved from a public sleep stage signals database, corresponding to ten subjects. The complete framework was tested using the database signals and signals from a test subject, captured and processed in real time. Our proposed framework may lead to a fully automated system to improve sleep quality without the need of medication.

A Clinically Applicable Interactive Micro and Macro-Sleep Staging Algorithm for Elderly and Patients with Neurodegeneration

Elderly and patients with neurodegenerative diseases (NDD) often complain about sleep problems and show altered sleep structure. Automated algorithms for efficient and specific sleep staging are needed. We propose a new algorithm using only one electroencephalographic and two electrooculographic channels to score wakefulness, rapid eye movement (REM) sleep and non-REM sleep in a cohort of elderly healthy controls (HC), patients with Parkinson's disease (PD), isolated REM sleep behavior disorder (iRBD), considered the prodromal stage of PD, and patients with PD and RBD (PD+RBD). The proposed method scores both standard 30-s epochs (macro-staging) and 5-s mini-epochs (micro-staging), whose evaluation may help to better understand sleep micro-structure. Moreover, the algorithm is interactive, as it labels the classified sleep epochs as either certain or uncertain, so that experts can manually review the uncertain ones. The algorithm performances were evaluated for macro-sleep staging, where it achieved overall accuracies of 0.87±0.05 in 41 HC, 0.86±0.10 in 57 PD, 0.76±0.10 in 31 iRBD and 0.77±0.10 in 30 PD+RBD patients when all 30-s epochs were considered. The accuracies increased to 0.91±0.05, 0.90±0.08, 0.85±0.09, 0.88±0.08 respectively when considering only the certain ones. The epochs labeled as uncertain were 9.95±4.15%, 11.13±7.86%, 18.39±7.38% and 18.90±8.00% in HC, PD, iRBD and PD+RBD respectively. The proposed interactive micro and macro sleep staging algorithm can be used in clinics to reduce the burden of manual sleep staging in elderly and patients with NDD.

A New Fully Automated Random-Forest Algorithm for Sleep Staging

Rapid eye movement (REM) sleep behavior disorder is considered the prodromal stage of alpha-synucleinopathies. Its diagnosis requires careful detection of REM sleep and the gold standard manual sleep staging is inconsistent and expensive. This work proposes a new automatic sleep staging model to add robust automation to such applications, using only electroencephalography (EEG) and electrooculography (EOG) recordings. The publicly available ISRUC-Sleep database was used to optimize the design of the proposed model. The model was trained and tested on subgroup-I consisting of 100 subjects with evidence of having different sleep disorders and the polysomnographic data were manually scored by two individual experts. We divided the EOG and EEG recordings in overlapping moving 33-s epochs with step of 3s and for each of them we computed several time and frequency-domain features. The features were used to train a random forest classifier that was able to label each 33-s epoch with the probabilities of being wakefulness, REM and non-REM. The mean of the probability values of ten 33-s epochs were calculated, and the sleep stage with the highest probability was chosen to classify a 30-s epoch and matched with the manual staged hypnogram. The performance of the model was tested using 20-fold cross validation scheme. When the epochs where the scorers agreed were used, the classification achieved an overall accuracy of 92.6% and a Cohen's kappa of 0.856. Future validation on RBD patients is needed, but these performances are promising as first step of development of an automated diagnosis of RBD.

An Intelligent Sleep Apnea Classification System Based on EEG Signals

Sleep Apnea is a sleep disorder which causes stop in breathing for a short duration of time that happens to human beings and animals during sleep. Electroencephalogram (EEG) plays a vital role in detecting the sleep apnea by sensing and recording the brain's activities. The EEG signal dataset is subjected to filtering by using Infinite Impulse Response Butterworth Band Pass Filter and Hilbert Huang Transform. After pre-processing, the filtered EEG signal is manipulated for sub-band separation and it is fissioned into five frequency bands such as Gamma, Beta, Alpha, Theta, and Delta. This work employs features such as energy, entropy, and variance which are computed for each frequency band obtained from the decomposed EEG signals. The selected features are imported for the classification process by using machine learning classifiers including Support Vector Machine (SVM) with Kernel Functions, K-Nearest Neighbors (KNN), and Artificial Neural Network (ANN). The performance measures such as accuracy, sensitivity, and specificity are computed and analyzed for each classifier and it is inferred that the Support Vector Machine based classification of sleep apnea produces promising results.

Louis EK, Duwell EJ, Timm PC, Sandness DJ, Boeve BF, Silber MH. Diagnostic thresholds for quantitative REM sleep phasic burst duration, phasic and tonic muscle activity, and REM atonia index in REM sleep behavior disorder with and without comorbid obstructive sleep apnea

OBJECTIVES

We aimed to determine whether phasic burst duration and conventional REM sleep without atonia (RSWA) methods could accurately diagnose REM sleep behavior disorder (RBD) patients with comorbid OSA.

DESIGN

We visually analyzed RSWA phasic burst durations, phasic, "any," and tonic muscle activity by 3-s mini-epochs, phasic activity by 30-s (AASM rules) epochs, and conducted automated REM atonia index (RAI) analysis. Group RSWA metrics were analyzed and regression models fit, with receiver operating characteristic (ROC) curves determining the best diagnostic cutoff thresholds for RBD. Both split-night and full-night polysomnographic studies were analyzed.

SETTING

N/A.

PARTICIPANTS

Parkinson disease (PD)-RBD (n = 20) and matched controls with (n = 20) and without (n = 20) OSA.

INTERVENTIONS

N/A.

MEASUREMENTS AND RESULTS

All mean RSWA phasic burst durations and muscle activities were higher in PD-RBD patients than controls (P < 0.0001), and RSWA associations with PD-RBD remained significant when adjusting for age, gender, and REM AHI (P < 0.0001). RSWA muscle activity (phasic, "any") cutoffs for 3-s mini-epoch scorings were submentalis (SM) (15.5%, 21.6%), anterior tibialis (AT) (30.2%, 30.2%), and combined SM/AT (37.9%, 43.4%). Diagnostic cutoffs for 30-s epochs (AASM criteria) were SM 2.8%, AT 11.3%, and combined SM/AT 34.7%. Tonic muscle activity cutoff of 1.2% was 100% sensitive and specific, while RAI (SM) cutoff was 0.88. Phasic muscle burst duration cutoffs were: SM (0.65) and AT (0.79) seconds. Combining phasic burst durations with RSWA muscle activity improved sensitivity and specificity of RBD diagnosis.

CONCLUSIONS

This study provides evidence for REM sleep without atonia diagnostic thresholds applicable in Parkinson disease-REM sleep behavior disorder (PD-RBD) patient populations with comorbid OSA that may be useful toward distinguishing PD-RBD in typical outpatient populations.