On Improving PPG-Based Sleep Staging: A Pilot Study

TL;DR

This study investigates methods to improve PPG-based sleep staging by combining PPG signals with auxiliary data using dual-stream cross-attention models, demonstrating significant performance gains on a large sleep dataset.

Contribution

It introduces a dual-stream cross-attention approach that effectively integrates PPG and derived modalities for enhanced sleep staging accuracy.

Findings

Significant performance improvement with combined PPG and auxiliary data.

Dual-stream cross-attention outperforms single-stream models.

Validated on the large MESA sleep dataset.

Abstract

Sleep monitoring through accessible wearable technology is crucial to improving well-being in ubiquitous computing. Although photoplethysmography(PPG) sensors are widely adopted in consumer devices, achieving consistently reliable sleep staging using PPG alone remains a non-trivial challenge. In this work, we explore multiple strategies to enhance the performance of PPG-based sleep staging. Specifically, we compare conventional single-stream model with dual-stream cross-attention strategies, based on which complementary information can be learned via PPG and PPG-derived modalities such as augmented PPG or synthetic ECG. To study the effectiveness of the aforementioned approaches in four-stage sleep monitoring task, we conducted experiments on the world's largest sleep staging dataset, i.e., the Multi-Ethnic Study of Atherosclerosis(MESA). We found that substantial performance gain can be achieved by combining PPG and its auxiliary information under the dual-stream cross-attention architecture. Source code of this project can be found at https://github.com/DavyWJW/sleep-staging-models