The effect of body placement site on ActiGraph wGT3X-BT activity counts

Document Type


Publication Date



Purpose. When the ActiGraph GT3X is worn on the hip, y-axis and vector magnitude (VM) counts level off at running speeds over 10 kph. Currently, it is not known if the counts level off when the device is worn on the wrist or ankle. Thus, the primary purpose of this study was to determine if ActiGraph counts level off with increasing running speeds at the wrist and ankle. Methods. Participants (N = 20) completed ten treadmill walking and running speeds (3–20 kph) for 30 s each. An ActiGraph wGT3X-BT was worn on the right hip, both wrists, and both ankles. Acceleration data for x-, y-and z-axes and VM were converted to 5 s epochs. Repeated measures analyses of variance were used to assess differences across speeds for each axis and VM for each wear location. Pair-wise comparisons with Bonferroni adjustments were performed to determine where differences occurred. Pearson correlations were used to assess the association between counts and speed. Results. Hip y-axis and VM counts increased significantly with speeds up to 10 kph and significantly decreased at speeds beyond 14 and 16 kph, respectively. However, at the wrists and ankles, significant increases in counts were seen for y-axis and VM counts across all running speeds. Conclusion. When worn on the wrist or ankle, ActiGraph y-axis and VM counts do not level off as is seen with the hip location. Wearing the ActiGraph on the wrist or ankle results in a stronger, more linear relationship between speed and counts than wearing it on the hip, and should result in more accurate estimations of energy expenditure for running speeds above 10 kph.

Publication Title

Biomedical Physics & Engineering Express





First Page


Last Page


Digital Object Identifier (DOI)




Rights Statement

In Copyright