Reliability and validity of the 3DNX accelerometer during mechanical and human treadmill exercise testing

Objective: This study aimed to assess the validity and reliability of the 3DNX triaxial accelerometer in a laboratory and mechanical setting. Methods: Experiment 1: A total of 10 units were fixed to a Multi-axis shaker table. A schedule comprising a range of accelerations was repeated on two occasions along each of the three measurement axes. Experiment 2: A total of 11 recreationally active individuals completed a treadmill protocol (4-20 km hr(-1)) on two occasions. Two 3DNX accelerometers were secured to the lower back, logging data every 5 s. Oxygen uptake was measured using the Douglas Bag technique. Reliability inter-and intra-units were assessed using a combination of limits of agreement (LoA), coefficient of variation (CV) and intra-class correlation coefficient (ICC). Validity was assessed using simple linear regression. All data are expressed in counts per 5 s (counts 5 s(-1)). Results: Experiment 1: CVintra ranged from 0.0 to 8.9% in all axes. The absolute bias +/- 95% LoA values were all below four counts 5 s(-1). CVinter did not rise above 4.5%. ICCs were 1.0 for all stages in all axes. The relationship between 3DNX counts and acceleration yielded an R-2 value of 0.99 and a standard error of the estimate (SEE) of six counts 5 s(-1). Experiment 2: CVinter ranged from 7.7 to 16.0% for trial 1 and from 7.7 to 16.2% for trial 2. ICCs between units ranged from 0.95 to 1.00 for trial 1 and from 0.90 to 0.99 for trial 2. Mean ratio bias +/- 95% LoA values for intra-unit and inter-unit reliability were -0.7% (P > 0.05) +/- 12.4% and 0.4% (P > 0.05) +/- 5.6%, respectively. The relationship between. (V) over dotO(2) and 3DNX counts for walking and running was linear (R-2 = 0.65, SEE = 1.42 ml kg(-1) min(-1); R-2 = 0.62, SEE = 3.63 ml kg(-1) min(-1)). Conclusion: The 3DNX accelerometer is a reliable and valid device for measuring acceleration in a mechanical setting and during human treadmill exercise. International Journal of Obesity (2011) 35, S88-S97; doi:10.1038/ijo.2011.39