Wrist Photoplethysmography Predicts Dietary Information

TL;DR

This study demonstrates that wearable wrist PPG signals can predict dietary information, enabling passive monitoring of meal content and intake with significant accuracy improvements, thus opening new possibilities for dietary assessment.

Contribution

The paper introduces a novel approach using PPG signals and language models to predict meal descriptions, showing PPG's potential for passive dietary monitoring.

Findings

PPG predicts meal content with nontrivial accuracy.

Predictability decreases as PPGs are farther from meal times.

PPG improves dietary task performance, increasing AUC by 11%.

Abstract

Whether wearable photoplethysmography (PPG) contains dietary information remains unknown. We trained a language model on 1.1M meals to predict meal descriptions from PPG, aligning PPG to text. PPG nontrivially predicts meal content; predictability decreases for PPGs farther from meals. This transfers to dietary tasks: PPG increases AUC by 11% for intake and satiety across held-out and independent cohorts, with gains robust to text degradation. Wearable PPG may enable passive dietary monitoring.