Wearable Multi-Biosignal Analysis Integrated Interface With Direct Sleep-Stage Classification

Authors:

Abstract
This paper presents a wearable multi-biosignal wireless interface for sleep analysis. It enables comfortable sleep monitoring with direct sleep-stage classification capability while conventional analytic interfaces including the Polysomnography (PSG) require complex post-processing analyses based on heavy raw data, need expert supervision for measurements, or do not provide comfortable fit for long-time wearing. The proposed multi-biosignal interface consists of electroencephalography (EEG), electromyography (EMG), and electrooculography (EOG). A readout integrated circuit (ROIC) is designed to collect three kinds of bio-potential signals through four internal readout channels, where their analog feature extraction circuits are included together to provide sleep-stage classification directly. The designed multi-biosignal sensing ROIC is fabricated in a 180-nm complementary metal–oxide–semiconductor (CMOS) process. For system-level verification, its low-power headband-style analytic device is implemented for wearable sleep monitoring, where the direct sleep-stage classification is performed based on a decision tree algorithm. It is functionally verified by comparison experiments with post-processing analysis results from the OpenBCI module, whose sleep-stage detection shows reasonable correlation of 74% for four sleep stages.
Funding Information
  • Institute of Information and Communications Technology Planning and Evaluation
  • Ministry of Science and ICT, South Korea (2018-0-00756, 2019-0-00208)
  • Samsung
  • National Research Foundation of Korea (2017M1A2A2087833)

This publication has 31 references indexed in Scilit: