Publications

All Publications

BallistoBud: Heart Rate Variability Monitoring using Earbud Accelerometry for Stress Assessment

Abstract

This study explores the potential of commercial earbuds for sensing physiological biomarkers like heart rate (HR), heart rate variability (HRV), and stress arousal. We collected accelerometer (IMU) and photoplethysmography (PPG) data from earbuds to estimate these biomarkers, comparing them to reference electrocardiograms (ECG) across 97 healthy participants. Despite the high accuracy of PPG-based sensing, its power consumption, additional cost, and potential for skin irritation may limit its adoption. We investigated using IMU sensors to capture ballistocardiographic (BCG) signals, which, under specific conditions, matched PPG performance. We introduced ECG-gated BCG heatmap, a novel visualization technique for accurate signal quality annotation, and trained a Random Forest model to differentiate usable from unusable BCG signals, achieving 82\% test accuracy. Filtering out unusable signals significantly reduced HR/HRV estimation error, making these estimates comparable to PPG. Our results highlight the feasibility of accurate physiological sensing with earbuds, paving the way for user-friendly wearable health technologies.

Author: Mahbubur Rahman, Mehrab Bin Morshed, Li Zhu, Jilong Kuang

Published: Association of Computing Machinery, Conference on Human Factors in Computing Systems (ACM CHI)

Date: Apr 28, 2025