Estimating Swimming Energy Expenditure using Accelerometer-based Activity Monitors

Abstract

INTRODUCTION: Traditional methods utilized to estimate energy expenditure (i.e. backward extrapolation of expired gases) during swimming are limited due to the aquatic environment. Recent investigations have demonstrated the use of miniaturized accelerometer-based activity monitors as an alternative means of estimating swimming energy expenditure utilizing sample-based regression techniques (Johnston & Stager, 2005). However, previous research has documented large inter individual variations in energy expenditure at any given swim velocity (Faulkner, 1968). PURPOSE: The purpose of this project was to first determine the relationship between accelerometer activity counts recorded from arm stroke (AS), leg kick (LK), and the sum of arm stroke and leg kick (TOT) motion during swimming and swimming energy expenditure. An additional goal was to examine the use of the accelerometer-based activity counts as a means of estimating swimming energy expenditure via individually-based regression techniques. METHODS: Actical activity counts (from arm stroke and leg kick), heart rate, and expired gases were collected from a progressive series of front crawl swim bouts performed by six collegiate competitive swimmers (3 men & 3 women, Age 20.3 ± 1.0 years) in an Endless Pool. Water velocity during the progressive swims ranged from 1.00-1.65 m•s-1. The relationship between oxygen consumption and activity counts were modeled using linear and 2nd order polynomial fits for each subject. RESULTS: Linear and quadratic regression analyses were significant (p < 0.05) for all subjects. Individualized regression analyses R2 values ranged from 0.86-0.96 and 0.98-0.99 for linear and quadratic models respectively (n = 6). There were significant correlations between AS and heart rate (ranged from 0.91-0.96; p < 0.05), LK and heart rate (ranged from 0.79-0.97; p < 0.05), and TOT and heart rate (ranged from 0.93-0.99; p < 0.05) for each subject. CONCLUSIONS: This study demonstrates that individualized regression techniques using accelerometer-based activity counts recorded from AS, LK, and TOT may provide a valid non-invasive means of estimating swimming energy in collegiate competitive swimmers.