An effective approach for correlating Na+ concentrations with sweating rates in wearable sweat sensing platform

Abstract

Accurately correlating the sweating rate and the concentration of biomarkers in sweat is essential in many sweat-based diagnostic applications. These two measurements are always done simultaneously in wearable sweat sensing platforms. However, concentration measurements of biomarkers are always delayed on the timeline compared with their production, whereas there is no such delay for sweating rate. Thus, a timeline mismatch exists between these two measurements. This means that the concentration vs rate correlation constructed on the basis of such measurements will deviate from the actual correlation. This study demonstrates the existence of this mismatch and explains its cause using sweat Na+ measurements. It also proposes an effective approach that applies a point-by-point compensation for the delay between Na+ measurements and the real-time sweating rates, such that the data on the repositioned concentration vs time curve correspond to exactly the same point on the timeline as their production. A vison sensor is developed to measure the sweating rate with high accuracy at a frequency of more than 0.1 Hz. Off-body and on-body measurements of sweating rate and Na+ concentration are carried out, and concentration–rate correlations are constructed using both measured and repositioned concentration curves. The least squares and random forest methods are employed to fit the constructed correlations and evaluate the reliability of the proposed approach. The use of the repositioned concentration curve gives a constructed correlation that is much closer to the actual one. This study indicates the necessity to rearrange sensor-measured biomarker concentration vs time curves when correlations of concentration with sweating rate need to be constructed and proposes a practical point-by-point data repositioning strategy for doing so. The results presented here will benefit the study of sweat biomarkers with unclear correlations with sweating rate, as well as providing a basis for the development of more reliable sweat-based diagnostic methods.