A Deep Learning Strategy for Automatic Sleep Staging Based on Two-Channel EEG Headband Data

Abstract

Sleep disturbances are common in Alzheimer’s disease and other neurodegenerative disorders, and together represent a potential therapeutic target for disease modification. A major barrier for studying sleep in patients with dementia is the requirement for overnight polysomnography (PSG) to achieve formal sleep staging. This is not only costly, but also spending a night in a hospital setting is not always advisable in this patient group. As an alternative to PSG, portable electroencephalography (EEG) headbands (HB) have been developed, which reduce cost, increase patient comfort, and allow sleep recordings in a person’s home environment. However, naïve applications of current automated sleep staging systems tend to perform inadequately with HB data, due to their relatively lower quality. Here we present a deep learning (DL) model for automated sleep staging of HB EEG data to overcome these critical limitations. The solution includes a simple band-pass filtering, a data augmentation step, and a model using convolutional (CNN) and long short-term memory (LSTM) layers. With this model, we have achieved 74% (±10%) validation accuracy on low-quality two-channel EEG headband data and 77% (±10%) on gold-standard PSG. Our results suggest that DL approaches achieve robust sleep staging of both portable and in-hospital EEG recordings, and may allow for more widespread use of ambulatory sleep assessments across clinical conditions, including neurodegenerative disorders.

Deep learning enables accurate automatic sleep staging based on ambulatory forehead EEG

“…Based on signal measured with ear-EEG electrodes, random forest classifier has been used to achieve scoring accuracies of κ = 0.45-0.65 [24] and κ = 0.73 [25]. A fabric headband with forehead electrodes developed by Cognionics Inc. (San Diego, CA, USA) has been recently used to achieve 74.0% deep learning scoring accuracy based on two-channel data [26]. Furthermore, a commercially available device, Dreem headband (Dreem Inc., Paris, France), has been used to achieve 83.5% (κ = 0.748) accuracies with a deep learning model that was trained with consensus hypnograms of five experienced scorers [4].…”

Home EEG sleep assessment shows reduced slow‐wave sleep in mild–moderate Alzheimer's disease

“…We removed data only in the corrupted channel, duplicating the valid channel in its place. We then passed the EEG into 12 deep learning models, previously described, 11 each trained on a different overlapping subset of labelled data. Majority voting was used to predict the 12 models, keeping the majority prediction for epochs in which ≥7 models agreed, labelling epochs for which the majority of the models did not agree as “unknown.”…”

“…We used a commercially available EEG headband (Cognionics Inc.) customized for two-channel recordings, as previously described. 11 The headband was also made to use disposable clip-on Kendall Meditrace Mini adhesive hydrogel foam electrodes (Cardinal Health). Detailed picture instructions on where to attach the electrodes at home were given.…”

Sleep Monitoring of Children With Nocturnal Enuresis: A Narrative Review

“…Studies led by Ma et al suggested that deep learning methods achieve robust sleep staging results of both portable and in-hospital EEG recordings. In addition, it may allow for more widespread use of ambulatory sleep assessment tests across a variety of clinical conditions, including neurodegenerative disorders ( 48 ). The diagnostic accuracy of a novel algorithm for the estimation of sleep stages and disease severity in patients with breathing-disordered sleep is based on actigraphy and respiratory inductance plethysmography ( 49 ).…”