Relationship of High-resolution Triaxial Accelerometry With VO2 In Highly-trained Runners During Treadmill Locomotion

Abstract

PURPOSE: The use of accelerometers as tools to quantify human movement has become increasingly prevalent in recent years. For estimation of VO2, "Activity Monitors" are typically low resolution, and higher resolution instruments may be of value. The purpose of this study was to determine the test-retest reliability of wireless high frequency, high-resolution accelerometers (HRA), and examine the relationship of HRA signals with VO2 during treadmill locomotion in highly trained runners. METHODS: Procedures were approved by the local HSRC. Seven NCAA D-I distance runners (21.4±1.7yr, 65.5±5.7 kg, 70.1±6.2 ml/kg/min) performed repeated incremental treadmill trials at 1% grade, starting at 2 kph and increasing 2 kph/2 min to volitional exhaustion. This protocol elicited a range of gaits from slow walking to maximal running in these highly trained subjects. Metabolic gasses were collected using an Oxycon Mobile (Viasys, CA). Wireless triaxial HRA (Microstrain, VT) were affixed to subject's lower back and data was streamed at 625 Hz. Raw signal from three axes was integrated to obtain resultant acceleration scalars (RES; Matlab; Mathworks, MA). Root-Mean-Square (RMS), for the last minute of each stage, for Vertical (VT), Medio-Lateral (ML), Fore-Aft (FA) and (RES) were compared to the corresponding minute average of relative VO2 (REL). Pearson's product correlation (r) and coefficient of variation (CV) were determined for VT, LT, FA and RES. Test-retest reliability was examined using Interclass correlations (ICC) and Chronbach's Alpha (SPSS, IL; α = 05). RESULTS: RMS of VT, LT, FA and RES was strongly correlated with REL (r =.87,.88,.95 and.95 respectively; p<.001). CV's were 5.7, 24.4, 24.5 and 5.2% for VT, ML, FA and RES, respectively. Both ICC's and Chronbach's alpha were.98,.97,.97 and.99 for VT, ML, FA and RES. CONCLUSIONS: Test-retest reliability of high frequency HRA displays strong validity and reliability across a wide range of walking/running speeds during treadmill locomotion in highly trained runners. Further, RMS of RES is strongly correlated with VO2, which indicates that RMS of unfiltered HRA is a valid means to determine VO2 during treadmill locomotion. These insights may allow for both clinicians and coaches to effectively monitor subjects for either rehabilitation and/or training purposes.