Validation of the iButton and Flex sensors for measuring skin temperature

Purpose Skin temperature is a critical component of athletes' heat strain during exercise in the heat. This study evaluated the innovative Flex sensor (BodyCap, Caen, France), allowing real-time measurements, against a data logger (iButton hygrochron DS1923, Analog Devices, Wilmington, USA), commonly used in sports science. Methods In a first study, the accuracies of iButton faces as well as the Flex sensor were tested on a constant-temperature Peltier module affixed to a brass plate. In a second study, concurrent skin temperature measurements were conducted with both sensors on eight participants cycling in hot conditions. Results Both faces of the iButton underestimated plate temperature, with the base (ME = mean error: 0.59 +/- 0.25 degrees C at 30 degrees C, 1.62 +/- 0.48 degrees C at 40 degrees C) being significantly more accurate than the top (ME: 1.14 +/- 0.26 degrees C at 30 degrees C, 3.02 +/- 0.71 degrees C at 40 degrees C, p < 0.001). The rate of temperature increase of the base was also higher than the top at 30 degrees C (1.97 +/- 0.16 degrees C/min vs 1.49 +/- 0.15 degrees C/min) and 40 degrees C (4.41 +/- 0.46 degrees C/min vs 3.14 +/- 0.16 degrees C/min, both p < 0.001). The Flex sensor underestimated plate temperature (ME: 1.31 +/- 0.32 degrees C at 30 degrees C, 3.00 +/- 0.84 degrees C at 40 degrees C), showing larger errors than the iButton base (p < 0.001). During exercise, the Flex sensor (36.91 +/- 1.19 degrees C) measured higher skin temperature compared to the iButton (36.50 +/- 1.15 degrees C), with an ME of -0.41 +/- 0.55 degrees C. Conclusion In the absence of a gold standard, for scenarios where real-time monitoring is not required, using the iButton with its base applied to the skin is therefore recommended. To enhance the Flex sensor's accuracy, implementing a calibration protocol before use is strongly suggested, enabling its full potential for real-time monitoring.