European Journal of Applied Physiology

, Volume 98, Issue 6, pp 601–612

Estimating energy expenditure using accelerometers

  • Scott E. Crouter
  • James R. Churilla
  • David R. BassettJr
Original Article

DOI: 10.1007/s00421-006-0307-5

Cite this article as:
Crouter, S.E., Churilla, J.R. & Bassett, D.R. Eur J Appl Physiol (2006) 98: 601. doi:10.1007/s00421-006-0307-5

Abstract

The purpose of this study was to examine the validity of published regression equations designed to predict energy expenditure (EE) from accelerometers (Actigraph, Actical, and AMP-331) compared to indirect calorimetry, over a wide range of activities. Forty-eight participants (age: 35 ± 11.4 years) performed various activities that ranged from sedentary behaviors (lying, sitting) to vigorous exercise. The activities were split into three routines of six activities, and each participant performed at least one routine. The participants wore three devices (Actigraph, Actical, and AMP-331) and simultaneously, EE was measured with a portable metabolic system. For the Actigraph, 15 previously published equations were used to estimate EE from the accelerometer counts. For the Actical, two published equations were used to estimate EE from the accelerometer counts. For the AMP-331 we used the manufacturer’s equation to estimate EE. The Actigraph and Actical regressions tended to overestimate walking and sedentary activities and underestimate most other activities. The AMP-331 gave a close estimate of EE during walking, but overestimated sedentary/light activities and underestimated all other activities. The only equation not significantly different from actual time spent in both light and moderate physical activity was the Actigraph Freedson kcal equation. All equations significantly underestimated time spent in vigorous physical activity (P < 0.05). In conclusion, no single regression equation works well across a wide range of activities for the prediction of EE or time spent in light, moderate, and vigorous physical activity.

Keywords

Motion sensor Physical activity Oxygen consumption Accuracy 

Supplementary material

421_2006_307_MOESM1_ESM.jpg (235 kb)
Supplementary material
421_2006_307_MOESM2_ESM.jpg (229 kb)
Supplementary material
421_2006_307_MOESM3_ESM.jpg (231 kb)
Supplementary material
421_2006_307_MOESM4_ESM.jpg (231 kb)
Supplementary material
421_2006_307_MOESM5_ESM.jpg (232 kb)
Supplementary material
421_2006_307_MOESM6_ESM.jpg (223 kb)
Supplementary material
421_2006_307_MOESM7_ESM.jpg (226 kb)
Supplementary material
421_2006_307_MOESM8_ESM.jpg (225 kb)
Supplementary material
421_2006_307_MOESM9_ESM.jpg (214 kb)
Supplementary material
421_2006_307_MOESM10_ESM.jpg (226 kb)
Supplementary material
421_2006_307_MOESM11_ESM.jpg (221 kb)
Supplementary material
421_2006_307_MOESM12_ESM.jpg (213 kb)
Supplementary material
421_2006_307_MOESM13_ESM.jpg (292 kb)
Supplementary material
421_2006_307_MOESM14_ESM.pdf (566 kb)
Supplementary material

Copyright information

© Springer-Verlag 2006

Authors and Affiliations

  • Scott E. Crouter
    • 1
    • 2
  • James R. Churilla
    • 1
  • David R. BassettJr
    • 1
  1. 1.Department of Exercise, Sport, and Leisure StudiesThe University of TennesseeKnoxvilleUSA
  2. 2.Division of Nutritional SciencesCornell UniversityIthacaUSA

Personalised recommendations