Reliability and comparability of the accelerometer as a measure of muscular power

The purpose of this study was to determine the effectiveness of using accelerometry as a reliable measure of upper body muscular power and its comparability with other conventional measurement tools.

Thirty men, aged 19 to 25 yr, gave informed written consent before performing a one-repetition maximum (1RM) bench press on a linear bench press apparatus. Three trials of 60% 1RM were then performed at 1-min intervals and the entire procedure was repeated the following day. Each trial was analyzed for average power (AP), average velocity (AV), and total displacement (TD) by three instruments: a uni-axial piezoresistive accelerometer (ICS Sensors Model 3145, Milpitas, CA) mounted on the Cybex Smith Press (Owatonna, MN) apparatus, a 17-mm video camera that recorded the lift, and an infrared photocell and timer system arranged to analyze a 20-cm segment of the lift. Acceleration data collected at 60 Hz obtained a direct measurement of force and an integrated measure of velocity to calculate muscular power.

Repeated measures ANOVA and intraclass correlation coefficients indicated high trial to trial reliability (r = 0.99) for all measurement variables. Film AP was significantly greater (P < or = 0.05) than the accelerometer AP and the photocell AP in the 20-cm segment (356.6 +/- 94.8 W vs 335.5 +/- 97.7 W, and 342.0 +/- 97.2 W, respectively). Also, significant mean differences (P < or = 0.05) between the accelerometer and film measurements existed for AP (246.0 +/- 70.2 W vs 286.1 +/- 83.6 W), AV (44.4 +/- 9.2 cm x s(-1) vs 51.3 +/- 12.3 cm x s(-1)) and TD (43.2 +/- 7.9 cm vs 47.4 +/- 7.4 cm) when examined over the entire lift, but there were significant correlations between the two methods (AP, r = 0.95; AV, r = 0.98; TD, r = 0.93).

These results suggest that although minor data acquisition errors were present, accelerometers can provide a reliable and versatile means to assess muscle power.