Multimodal Time-Series Analysis of Physiological Signals Associated with Interstitial Fluid Glucose Fluctuations During Sleep

Continuous monitoring of glucose dynamics during sleep is critical for early detection of asymptomatic hypoglycemia and other metabolic risks. In this study, we aimed to characterize the physiological responses associated with interstitial fluid glucose (ISFG) fluctuations during sleep by performing multimodal time-series analyses of heart rate (HR), peripheral oxygen saturation (SpO₂), and motion signals. 5 healthy participants (1 female) simultaneously wore a minimally invasive ISFG sensor and a ring-type sensor that recorded HR, SpO₂, and actigraphy. Throughout the sleep period (00:00–06:00), ISFG levels decreased in four out of five participants. Subsequently, in participants 1, 3, 4, and 5, ISFG increased between 12:00–15:00 and then gradually declined again towards the night. Paired t-tests comparing ISFG levels between 0–3 h and 3–6 h showed a significant decrease during the later half of the night (n=5, p=0.01). Regarding SpO₂, four participants exhibited a mild decrease between 3:00–4:00 after sleep onset, while one dataset was incomplete. Our findings suggest that ISFG reduction during sleep is associated with subtle changes in oxygen saturation patterns. These multimodal signal patterns could serve as potential new indicators for detecting unrecognized nocturnal hypoglycemia. Furthermore, the integration of low-invasive continuous monitoring technologies may enhance online consultations and remote medical services by providing real-time physiological risk assessments during sleep. Future studies with larger cohorts are warranted to validate these promising applications.