Wearable ring bioelectrical impedance analyzer for estimation and monitoring of body fat

Abstract

The World Health Organization has declared obesity as one of the fastest growing global epidemics. People suffering from obesity are at an increased risk of serious health conditions including stroke, diabetes, high blood pressure, and other chronic diseases caused by excess fat in the body. To study and control this obesity epidemic, precise measurement of the body fat mass is necessary. Bioelectrical impedance analysis (BIA) is a common yet effective method for body composition, especially fat mass, measurement in which the electrical impedance of a human body is measured. Here we report a novel wireless BIA device, in the form of a wearable ring, for estimating the body fat as well as a smartphone application for users to manage and track their health data. The wearable analyzer was compared with an off-the-shelf body fat monitor in terms of performance for 40 healthy volunteers. In addition, the small, dry electrodes used in the ring analyzer are compared with traditional gel-based electrodes. Although the dry electrodes impose a larger skin-electrode contact impedance (5.18 ± 0.655 kΩ) than gel-based electrodes (933.37 ± 40.06 Ω), the designed system provides critical compensation for the higher electrode impedance and the parasitic effects for the dry electrodes. With the compensation implemented, these two types of electrodes are highly correlated with correlation coefficients r = 0.96 for resistance and r = 0.93 for reactance. Body fat estimation equations proposed by four different groups are analyzed and compared using the results from the wearable ring device. Our ring analyzer and the reference monitor demonstrate a linear regression with a strong correlation (r = 0.90) for estimating fat mass. Furthermore, the Bland-Altman plot shows that only 2 out of 40 subjects lie outside the 95% limit of agreement. All these findings show that the ring BIA system is highly portable, easy to use, and with comparable performance to large reference analyzers.