Calibration of the Computer Science and Applications, Inc. accelerometer

TLDR

We established accelerometer count ranges for the CSA activity monitor corresponding to commonly employed MET categories. Data from 50 adults exercising on a treadmill at three speeds were used to develop a model predicting energy expenditure from activity counts and body mass, achieving r² = 0.82 and SEE = 1.40 kcal min⁻¹. Activity counts correlated highly with steady‑state oxygen consumption (r = 0.88), yielding count ranges for light (≤1951), moderate (1952–5724), hard (5725–9498), and very hard (≥9499 cnts min⁻¹) intensities, and cross‑validation showed no significant prediction error, confirming the template for classifying activity intensity with the CSA accelerometer.

Abstract

Purpose: We established accelerometer count ranges for the Computer Science and Applications, Inc. (CSA) activity monitor corresponding to commonly employed MET categories. Methods: Data were obtained from 50 adults (25 males, 25 females) during treadmill exercise at three different speeds (4.8, 6.4, and 9.7 km·h-1). Results: Activity counts and steady-state oxygen consumption were highly correlated(r = 0.88), and count ranges corresponding to light, moderate, hard, and very hard intensity levels were ≤ 1951, 1952-5724, 5725-9498, ≥ 9499 cnts·min-1, respectively. A model to predict energy expenditure from activity counts and body mass was developed using data from a random sample of 35 subjects (r2 = 0.82, SEE = 1.40 kcal·min-1). Cross validation with data from the remaining 15 subjects revealed no significant differences between actual and predicted energy expenditure at any treadmill speed (SEE = 0.50-1.40 kcal·min-1). Conclusions: These data provide a template on which patterns of activity can be classified into intensity levels using the CSA accelerometer.