Article Impact Level: HIGH Data Quality: STRONG Summary of Nature Sensors https://doi.org/10.1038/s44460-026-00044-0 Dr. Xiaodong Wu et al.
Points
- National University of Singapore researchers created a single-channel wearable patch named X-Sig that merges electrical and mechanical physiological signals at the sensor level.
- The device vertically stacks a conductive polyurethane electrode with an ultrathin piezoelectric film to collect four distinct body signals through one output channel.
- Continuous blood pressure estimations achieved high accuracy with mean differences of 0.27 mmHg systolic and 0.33 mmHg diastolic compared to standard cuff monitors.
- Forearm gesture classification accuracy reached 96.4% across 10 movements, outperforming single-modality electromyography at 72.1% and force myography at 82.9%.
- Biocompatibility testing showed zero skin irritation following 24 hours of continuous wear, while electrode materials were fully recycled using water and ethanol.
Summary
This study evaluated the performance of a cross-modal epidermal sensor, named X-Sig, designed to simplify continuous physiological monitoring by fusing biopotential and biomechanical signals into a single composite output. Current wearable devices rely on separate sensors and circuitries for each signal modality, increasing hardware footprint and power consumption. Researchers from the National University of Singapore developed a vertically stacked architecture combining a conductive, adhesive polyurethane electrode layer for electrical sensing with an ultrathin piezoelectric film for mechanical sensing, enabling in-sensor signal fusion through one channel.
The device simultaneously acquires electrocardiography (ECG), electromyography (EMG), radial pulse pressure waves, and force myography (FMG) without skin irritation observed over 24 hours of continuous wear. Placed on the wrist, the sensor extracts heart rate and pulse arrival time to estimate blood pressure continuously. Machine-learning algorithms yielded mean differences of 0.27 mmHg for systolic blood pressure and 0.33 mmHg for diastolic blood pressure compared with a standard cuff-based monitor, meeting Class A accuracy standards under IEEE guidelines.
When tested on the forearm for hand gesture classification, the fused EMG and FMG signal achieved 96.4% classification accuracy across 10 distinct gestures using 70 training samples per gesture. In comparison, single-modality systems achieved accuracies of 72.1% for EMG alone and 82.9% for FMG alone. These results demonstrate that in-sensor signal fusion provides a highly accurate, bandwidth-efficient platform for non-invasive cardiovascular monitoring, gesture recognition, and rehabilitation without requiring complex multi-channel hardware.
Link to the article: https://www.nature.com/articles/s44460-026-00044-0
References
Wu, X., Zhu, C., Zheng, L., Song, Y., Wang, J., Zhang, X., Yao, Z., Han, Y., Wang, Z., Jiang, Z., Liu, Z., & Liu, Y. (2026). A cross-modal epidermal sensor enables single-channel fusion of biopotential and biomechanical signals. Nature Sensors, 1(4), 315–327. https://doi.org/10.1038/s44460-026-00044-0
