Publication:

A qualitative and quantitative comparison of sleep tracking devices

In the last decade, a growing interest in personal health tracking devices has been reported worldwide. Wearable devices provide the possibility of measuring different physiological parameters such as e.g., heart rate, heart rate variability, blood oxygen levels and body temperature. These parameters could help address the problem of accurate sleep tracking in non laboratory scenarios. The goal of this project is to provide a qualitative and quantitative comparison of data collected at night by two wearable devices usable in everyday life settings. We investigated the performance of one commercial (Oura ring 3) and one research-grade device (Empatica E4) used for tracking sleep, with the specific focus on monitoring heart-rate data. We conducted a pilot-study with 5 participants. Participants wore both devices for 4 weeks (28 nights), on which physiological and self-reported data was collected. After that, a data analysis phase was carried out where statistical analysis was performed. We compared the data from both devices using two approaches. First, correlation results were (r = 0.614±0.20) and (r = 0.686±0.19) for Pearson and Spearman correlation coefficients, respectively. Second, bias analysis results were on average (md = −2.306±7.66). Although correlation and bias analysis results vary depending on the participant, some participants showed highly correlated data with a low mean difference, which may indicate a good agreement between devices. However, using inferential statistics we found that both devices have different medians for all participants. These results show a statically significant difference between measurements, and high variability between participants. Although the sample size of the study was small, future research involving these two devices may take these findings into account.