Tracking astronaut physical activity and cardiorespiratory responses with the Bio‐Monitor sensor shirt

Abstract

Astronauts develop insulin resistance, and are at risk for cardiovascular deconditioning, during long-duration missions to the International Space Station (ISS) despite their daily exercise sessions (Hughson et al. Am J Physiol Heart Circ Physiol 310: H628–H638, 2016). Chronic unloading of the musculoskeletal and cardiovascular systems in microgravity dramatically reduces the challenge of daily activities, and the astronauts’ schedules limit them to approximately 30-min/day aerobic exercise. To understand the physical demands of spaceflight and how these change from daily life on Earth, the Vascular Aging experiment is equipping astronauts for 48-72h continuous recordings with the Canadian Space Agency's Bio-Monitor wearable sensor shirt. The Bio-Monitor (Bio-M), developed from the commercial Hexoskin® device, consists of 3-lead ECG, thoracic and abdominal respiratory bands, 3-axis accelerometer, skin temperature and SpO2 sensor placed on the forehead. Our utilisation of this equipment necessitated the development of novel processing and visualisation techniques, to better interpret and guide subsequent data analyses. Here we present initial data from astronauts wearing the BioM prior to launch and aboard the ISS, demonstrating the ability to extract useful data from BioM, using software developed ‘in-house’. Astronauts wore the Bio-M continually for 72-h except for periods of water immersion or when the device conflicted with another activity. After physical exercise, astronauts changed to a dry shirt. First, we assessed the key data-quality metrics to provide initial appraisals of acceptable recordings. Mean total recording length pre-flight (60.5 hours) was similar to that in-flight (66.5 hours), with a consistent distribution of recorded day (44% vs 45%, 6am-6pm) and night (56% vs 55%, 6pm-6am) hours (pre-flight vs in-flight respectively). For each recording, quality assessment of ECG signals was performed for individual leads, before combining signals and cross-correlating R-waves to produce reliable heart-rate timings. Mean ECG quality for individual leads, represented here as the percentage of usable signal to total recording duration, was somewhat lower in-flight (92%) when compared to pre-flight (96%), likely caused by poor skin contact or dry shirt electrodes; combining lead signals as mentioned above improved the proportion of usable data to 97% and 98% respectively. Accelerometer recordings identified a significant reduction in high-force movements over the 72-hour recordings, with just over 2.5 hours/day of high-force activity in astronauts pre-flight vs 50 minutes/day in-flight. It should be noted however that accelerometer measurements in zero-gravity are likely to be reduced, and future refinement of activity data continues. Average heart rates in-flight showed little difference when compared to pre-flight, although future analyses will compare periods of sleep, rest, and activity to further refine this comparison. We conclude that utilisation of the BioM hardware with our own analysis techniques produces high-quality data allowing for future interpretation and investigation of spaceflight-induced physiological adaptations.