# Multimodal Physiological Monitoring Using Novel Wearable Sensors: A Pilot Study on Nocturnal Glucose Dynamics and Meal-Related Cardiovascular Responses

## TL;DR

This pilot study explores using wearable sensors to monitor glucose and cardiovascular changes during sleep and after meals, showing potential for real-time health tracking.

## Contribution

The study introduces a novel multimodal wearable sensing approach to capture nocturnal glucose dynamics and meal-related physiological responses.

## Key findings

- Nocturnal interstitial fluid glucose levels decreased significantly in the latter half of sleep in four out of five participants.
- Heart rate increased significantly during meal consumption compared to pre-meal periods, with a large effect size.
- Skin temperature and heart rate showed a weak and non-significant correlation during meal-related phases.

## Abstract

This pilot study investigated multimodal physiological monitoring using minimally invasive and wearable sensors across two experimental settings. Experiment 1 involved five healthy adults (1 female) who simultaneously wore an interstitial fluid glucose (ISFG) sensor and a ring-type wearable device during sleep (00:00–06:00). Time-series analyses revealed that ISFG levels decreased during sleep in four of the five participants. ISFG values were significantly lower in the latter half of the sleep period compared with the first half (0–3 h vs. 3–6 h, p = 0.01, d = 2.056). Four participants also exhibited a mild reduction in SpO2 between 03:00–04:00. These results suggest that nocturnal ISFG decline may be associated with subtle oxygen-saturation dynamics. Experiment 2 examined whether wearable sensors can detect physiological changes across meal-related phases. Nine male participants were monitored for heart rate (HR) and skin temperature during three periods: pre-meal (Phase 1: 09:00–09:30), during meal consumption (Phase 2: 12:30–13:00), and post-meal (Phase 3: 13:00–13:30). A paired comparison demonstrated a significant difference in median HR between Phase 1 and Phase 2 (p = 0.029, d = 0.812), indicating a large effect size. In contrast, HR–temperature correlation was weak and not statistically significant (Pearson r = 0.067, p = 0.298). Together, these findings demonstrate that multimodal wearable sensing can capture both nocturnal glucose fluctuations and meal-induced cardiovascular changes. This integrative approach may support real-time physiological risk assessment and future development of remote healthcare applications.