Determining Stages of Sleep
PATENT STATUS
Singapore provisional patent application disclosing this technology was filed in October 2016.
OVERVIEW OF TECHNOLOGY ON OFFER
An end-to-end framework for real-time automatic sleep stage classification
BRIEF DESCRIPTION
The quality and quantity of sleep bears a strong bi-directional relationship to one’s health and well-being. Sleep disorders are known to affect 30% of the population. A sleep cycle consists of the following stages: wake, stages S1, S2 and S3/4 and rapid eye movement (REM) sleep. In order to profile these sleep stages, a polysomnographic (PSG) equipment is used to record electroencephalogram (EEG), electrooculogram (EOG) and other measurements. A trained professional known as the sleep scorer annotates these recordings manually into the abovementioned sleep stages. This manual sleep scoring method is difficult, time consuming and costly. Further, the element of subjectivity results in producing only about 80 to 82% congruence between expert scorers.
The present technology developed at Professor Mike Chee’s lab at Duke-NUS is a method and system involving deep learning algorithms that use multiple neural network layers for automatically determining sleep stages. The system has a client-server architecture for efficiently transporting PSG data measured from 2 EEG and 2 EOG channels from client (or patient) to the server and receiving back the sleep data divided into 30-second sleep epochs, each epoch designated a specific sleep stage. The accuracy of this automated scoring framework is on par with expert human scorers but much faster (~5 seconds compared to 30 to 60 minutes by human scorer). The accuracy and speed of sleep scoring using this technology opens up new applications that require real-time sleep-stage dependent interventions, one such application being memory enhancement.
POTENTIAL APPLICATIONS
- Sleep stage classification using polysomnography data (EEG and EOG data)
- Diagnosis of sleep disorders
- Automatic delivery of precisely timed acoustic stimulation during slow-wave sleep for memory augmentation
KEY BENEFITS
- Rapid, real-time and automated scoring of sleep stages
- Utilizes PSG data unlike wrist-worn watches and other wearable devices currently available in the market which only measure surrogates of sleep (such as respiratory rate, motion and heart rate), and hence is more accurate.
PUBLICATIONS
Amiya Patanaik, Ju Lynn Ong, Joshua J. Gooley, Sonia Ancoli-Israel & Michael W.L. Chee. An end-to-end framework for real-time automatic sleep stage classification. (manuscript under review).
Website: http://z3score.com/
INVENTOR BIO
Prof Mike Chee:
Profile
Dr. Amiya Patanaik: (email: amiya@z3score.com, personal website: aptnk.in/profile)
CONTACT
Please email us for further enquiries: cted@duke-nus.edu.sg