Validity of smartphone accelerometers for assessing energy expenditure during fast running

Abstract

Recent advances in smartphone technology have facilitated the generation of mobile applications to monitor exercise performance. The aim of the present study was to investigate the validity and reliability of the triaxial accelerometers in the HTC and Samsung platforms for assessing energy expenditure (EE) during walking and fast running. Eleven endurance-trained participants (nine males and two females, mean ± s.d.: Age 35 ± 10 years; Maximal aerobic capacity 58.5 ± 7.2 ml·min-1·kg-1) completed two identical discontinuous incremental exercise tests on a treadmill. The tests consisted of walking (4, 5, and 6 km·h-1) and running (8, 10, 12, 14, 16, 18, and 20 km·h-1, or until volitional exhaustion) for 3 min at each speed, followed by 3 min of recovery for all speeds faster than 10 km·hr-1. Acceleration was recorded in all three axes of motion using a novel application on each device and expressed as vector magnitude (VM). Respiratory variables used to assess EE were measured using indirect calorimetry. The reliability of EE and acceleration measured between tests 1 and 2 were assessed using intraclass correlation coefficients (ICC). The validity of the accelerometers to assess EE was determined using Pearson's correlation coefficient. Measurements of EE and VM recorded using both devices were highly reproducible between tests 1 and 2 (ICC: EE = 0.976; HTC VM = 0.984; Samsung VM = 0.971). All accelerometer outputs rose linearly with speed during walking and running up to and including 20 km·h-1. EE was significantly correlated with both HTC VM (R = 0.98, P < 0.001) and Samsung VM (R = 0.99, P < 0.001). Data from the present study suggests that the inbuilt triaxial accelerometers in HTC and Samsung smartphone devices offer a valid and reliable method for determining the change in EE during fast running. © Springer International Publishing Switzerland 2014.