Night Owl: A Quantitative Sleep Journal


Timothy Phan
Tuan Le
Ashanti Hughes
Ngoc Ngan Nguyen

Dr. D. Pompili
Night Owl.png
Night Owl (1).png

Introduction: The goal of the project is to incorporate multiple sensors on the body to assess and classify on the fly the state of the user, and extract real-time sleep analytics for assessing sleep quality. An application of this project is to provide objective and quantitative measures to assist healthcare providers in tracking the progress of their patients in terms of sleep quality and physiological variability during sleep.

Motivation: Obstructive Sleep Apnea (OSA) syndrome and Delayed Sleep Phase Disorder (DSPD) are common disorders and are major health concerns due to comorbidities with a number of cardiovascular diseases such as hypertension and cardiac dysrhythmia. Predictors of sleep abnormalities include overnight level of respiratory sinus arrhythmia, sleep onset latency, and sleep efficiency, which can all help screen for any early symptoms and keep track for better sleep-time management.

Design: The main process of the project is using Fourier analysis to decompose skin resistance into tonic and phasic components, and assess Respiratory Sinus Arrhythmia (RSA) from heart-rate variability. The collected data is then used to train the computer for on-the-fly pattern recognition and classification of the arousal state of the user.

In the results, the sleep screening report is consisting of sleep onset latency, sleep efficiency, and average overnight respiratory sinus arrhythmia. The data acquisition portion is a group of sensors worn at the wrist of the user and interfaced with a wireless node, Shimmer, running on the TinyOS platform. Three sensors are used to measure the skin resistance of the user, heart rate, and blood oxygen saturation level, as well as movements during a night of sleep.

Conclusion: By using skin-resistance variability, the classifier can detect the sleep onset of the subject. Furthermore, the data collected can be used to classify the sleep latency, sleep efficiency, and average overnight respiratory sinus arrhythmia. This information can be used as a tool for the user to pre-determine any early symptom of OSA and/or DSPD thus avoiding expensive and time-consuming tests at the clinics.