Self-Adjustable Galvanic Skin Response Sensor for Physiological Monitoring

• Published in: IEEE sensors journal Vol. 23; no. 3; p. 1
• Main Authors: Miranda, Jose A., Paez-Montoro, Alba, Lopez-Ongil, Celia, Paton, Susana
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.02.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Affective computing; Biomedical monitoring; Correlation coefficients; Current measurement; Electrodes; Error analysis; Feature extraction; Galvanic skin response; Monitoring; Nervous system; Physiological effects; physiological signals; Physiology; self-adjustable sensor; Sensitivity; Sensors; signal processing; Skin; Wearable technology
• ISSN: 1530-437X; 1558-1748
• DOI: 10.1109/JSEN.2022.3233439

Continuous physiological monitoring integrated within current wearable devices is a hot topic nowadays. Despite that, measuring physiological variables is still challenging due to intrinsic and extrinsic personal factors. This results in the need for smart, adjustable, and personalised sensing devices. Among the different physiological signals that can be measured, changes in skin conductance are extensively used in Affective Computing research. This measurement presents an unequivocal relationship with the sympathetic branch of the autonomous nervous system, which relates to emotional reactions. However, there is a lack of self-adjustable skin conductance sensors. This paper presents a novel skin conductance analog front-end to deal with individual physiological dynamics. We compared the performance of our system with other skin conductance sensor circuits presented in the literature. Moreover, the experimental results in this work come from the data collected in 47 women volunteers and are compared with the measurements with a reference physiological sensing system commonly used for validations. Finally, our system presents a competitive averaged sensitivity for a typical human EDA range compared to the current state-of-the-art solutions. Experimental results show that the system achieves an averaged sensitivity of up to 0.33 nS for a range 0 μS to 40 μS, with a relative error below 1% for fixed-resistance measurements, and a 0.94 median correlation coefficient when directly comparing with the reference sensor. Moreover, we extracted and analysed different physiological features in data from both sensors and obtained comparable results. Finally, this research is intended to boost the design and development of subject-independent and self-adjustable wearable sensors.