How does the choice of reference frame impact the distribution of WBAM components around different anatomical axes?

Introduction: Accurate estimation of WBAM in daily activities requires a locally-defined reference frame aligned with anatomical axes, especially during activities with changing body direction [1,2]. Ho et al. [3] highlighted the significant influence of different local frames on spatial gait parameters during turning gaits. Our previous research suggested smoothing the mediolateral (ML) oscillations of local reference frames using a low-pass filter, but a detailed analysis of their distinctions and impact on WBAM distribution across anatomical axes is lacking. Additionally, it is uncertain whether low-pass filtering alters the WBAM distribution.

Research Question: How do the differences among the global and various non-low-pass filtered (Non-LP) and low-pass filtered (LP) local reference frames influence the WBAM distribution around different anatomical axes?

Methods: Ten healthy participants (four females and six males) were enrolled in the experiments. Participants were asked to perform several tasks at a self-selected speed, including walking along a straight line (StrW), slalom walking (SlaW), Zig-zag walking (ZigW), and walking with 180-degree turn (TurW) tasks. All tasks were repeated 3 times. A VICON optical motion capture system was used to collect the kinematics (100 Hz). Among local reference frames, we chose the pelvis-oriented ψ_P and vCoM-oriented ψ_vCoM local reference frames, and we introduced an average angular velocity-oriented ψ_Aω local reference frame (Figure (a)). We compared the frame angle differences among the global and various Non-LP and LP local reference frames as shown in (Figure (b)), and the Anteroposterior (AP) and ML WBAM differences expressed in these different frames. Linear mixed models were used to test the effect of the reference frame on the distribution of AP and ML WBAM components.

Results: Our results found that the choice of reference frames significantly affected the AP and ML WBAM distributions in all tasks before filtering. The low-pass filter decreased the frame angle differences between the global and all local reference frames in the StrW task, as well as among local reference frames in all tasks. All LP local reference frames showed no significant differences in both AP and ML WBAM during the StrW task compared to the global frame, with almost no significant difference observed among various LP local reference frames.

Discussion: Considering the ease of calculation, ψ_P and ψ_vCoM were more practical compared to ψ_Aω since they could be determined by a reduced optical marker set or IMUs in future applications. Non-LP ψ_vCoM showed a large frame angle difference compared to Non-LP ψ_P due to the opposite ML oscillation of vCoM, leading to significant AP and ML WBAM differences between them. Yet, applying the low-pass filter to ψ_P and ψ_vCoM effectively eliminated their frame angle and WBAM differences. Therefore, the LP ψ_P and ψ_vCoM were both more recommended in daily-life conditions for interpreting WBAM in different anatomical axes.