Respiratory rate (RR) monitoring during sleep is critical for assessing sleep quality and detecting disorders like sleep apnea. While traditional monitoring methods require specialized clinical environments, recent technological advances have introduced less invasive alternatives that can be categorized as wearable (body-Attached) and nearable (environment-embedded) sensors. This study compared the performance of mattress-embedded inertial measurement units (IMUs) with jugular-mounted IMUs for overnight RR monitoring in real-world conditions. Four participants underwent monitoring with both systems during sleep, with a commercial chest band for breathing waveform collection serving as reference. Data were processed using adaptive sensor selection algorithms. Results demonstrated that mattressembedded sensors achieved slightly better performance (MAE: 0.3 brpm, MAPE: 3.9%) than the jugular-mounted sensors (MAE: 0.3 brpm, MAPE: 4.5%), with better results observed in the prone position if compared to lateral and supine. BlandAltman analysis confirmed strong agreement with the reference for both systems, with the mattress-based configuration showing narrower limits of agreement. These findings challenge the conventional assumption that nearable technologies necessarily sacrifice accuracy for convenience. The performance comparison across different sleep postures and the adaptive sensor selection approach provides valuable information for optimizing non-invasive respiratory monitoring during sleep, offering a promising alternative for long-Term home-based sleep assessment.
Comparison of Mattress-Embedded and Jugular-Mounted Inertial Sensors for Respiratory Monitoring during Sleep in Real-World Settings
Romano C.;Silvestri S.;Schena E.;Massaroni C.
2025-01-01
Abstract
Respiratory rate (RR) monitoring during sleep is critical for assessing sleep quality and detecting disorders like sleep apnea. While traditional monitoring methods require specialized clinical environments, recent technological advances have introduced less invasive alternatives that can be categorized as wearable (body-Attached) and nearable (environment-embedded) sensors. This study compared the performance of mattress-embedded inertial measurement units (IMUs) with jugular-mounted IMUs for overnight RR monitoring in real-world conditions. Four participants underwent monitoring with both systems during sleep, with a commercial chest band for breathing waveform collection serving as reference. Data were processed using adaptive sensor selection algorithms. Results demonstrated that mattressembedded sensors achieved slightly better performance (MAE: 0.3 brpm, MAPE: 3.9%) than the jugular-mounted sensors (MAE: 0.3 brpm, MAPE: 4.5%), with better results observed in the prone position if compared to lateral and supine. BlandAltman analysis confirmed strong agreement with the reference for both systems, with the mattress-based configuration showing narrower limits of agreement. These findings challenge the conventional assumption that nearable technologies necessarily sacrifice accuracy for convenience. The performance comparison across different sleep postures and the adaptive sensor selection approach provides valuable information for optimizing non-invasive respiratory monitoring during sleep, offering a promising alternative for long-Term home-based sleep assessment.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


