Ultrahuman Ring AIR Sensor : The
Ultrahuman Ring AIR features an advanced temperature sensor designed for continuous peripheral temperature tracking. The device operates at a range of -20°C to 40°C and records data with high precision at a frequency of one reading every 5 minutes.
iButton DS1921H Sensor : The
iButton sensor is a well-established device for temperature monitoring, widely used in scientific studies for its accuracy and reliability.
5,6 It records temperature at user programmable intervals from 1 to 255 minutes and can measure temperatures across the range of +15°C to +46°C with a resolution of 0.125°C.
Off body experimental set up- Ring AIR to iButton Sensor
Experimental Setup : Pairs of Ring AIR and iButton devices were placed in polypropylene tubes to ensure uniform temperature exposure and prevent contamination, with experiments conducted across a wide range of temperatures ranging from 0 to 42 °C. Enhanced sampling was conducted at two discrete temperature ranges (20-25°C and 34-37°C) which represents the most frequent temperature ranges of user data and replicated in a controlled water bath.
Data Collection : The experiment was conducted over a continuous three-day period, with temperature measurements logged at 5-minute intervals by the Ultrahuman Ring AIR and 3-minute intervals by the iButton sensor. Both devices employed time-stamped digital recording to ensure precise synchronization and facilitated robust analytical comparisons.
On body- finger comparison- Ring AIR to iButton Sensor
Study Participants : 7 healthy adults volunteers (aged 20-50 years, nearly equal male to female ratio) from the Ultrahuman team were included, representing diverse skin tones, and body compositions.
Experimental Setup : Volunteers wore one Ultrahuman Ring AIR on their preferred finger and the iButton secured to the base of the index finger of their non-dominant hand using medical-grade adhesive tape for stable contact to skin throughout the measurement period.
Data Collection : Temperature readings were recorded over 3-5 consecutive nights to capture variations during different sleep stages and environmental conditions. Volunteers maintained their usual sleeping habits and environmental settings to simulate real-world conditions. Both devices recorded temperature data digitally with timestamps to enable accurate synchronization and analysis. Participants ensured proper placement of the iButton and functionality of the Ultrahuman Ring AIR each night to maintain data integrity and prevent loss or corruption.
Statistical Analysis:
To ensure data accuracy, consecutive data points within a 5-minute interval were selected, minimizing the impact of measurement errors and unexpected deviations. Key statistical metrics were calculated to evaluate the performance and agreement between the devices at individual data point levels. The mean difference (bias) assessed the average deviation between paired temperature measurements, while the upper and lower limits of agreement (LOA) evaluated the range within which most differences fell. Group variance (MLM) quantified variability across participant groups to account for individual differences. Subsequent to normality testing, Spearman correlation was used to measure the strength and direction of the monotonic relationship between the devices, and the p-value determined the statistical significance of the observed correlations. Data was analyzed using scikit-learn, an open source statistical and machine learning library in Python. Since this was an internal volunteer study, two separate data analysts independently analyzed the results.
