Comparison Of Wrist-worn Accelerometer Output Based On Handedness

Purpose: The purpose of this study was two-fold; to compare accelerations between dominant and non-dominant wrists during structured activity bouts; and to determine whether there are differences in wrist accelerations (dominant and non-dominant) between age groups during structured activity bouts.

Methods: Participants aged 23 to 79 y performed a block of ten activities. Activities were performed for seven minutes with at least a three-minute break between activities. During each activity, participants wore ActiGraph GT3X monitors on their dominant and non-dominant wrist. Data were summed into 60-second epochs. T-tests were used to compare the x, y, z, and vector magnitude values between wrists. An analysis of variance was used to determine differences between ages for vector magnitude wrist acceleration.

Results: Overall, the vector magnitude of 17 of the 27 activities performed was significantly different between the two wrist placements. Specifically, nine, 14, and 14 of the 27 activities were significantly different between wrists in the x-, y-, and z-axes, respectively. Excluding walking, 73% of the activities performed (housework, yardwork, sedentary, leisure activities) were significantly different between wrists. These activities ranged from dominant hand activities (i.e. playing cards, washing windows) to those that require use of both hands (i.e. croquet, washing dishes). ANOVA results revealed only one (lawn bowling) of the 27 activities differed significantly across age.

Conclusion: There are significant differences between accelerations recorded between dominant and the non-dominant wrists and this difference remains consistent with age. This study provides evidence that the acceleration values using vector magnitude, x, y, or z accelerations between wrists cannot be compared, thus future research should standardize monitor placement.

Supported by NIH grant R01-HL091019-04