Stair ascent comparison of lower limb kinematics with differing time normalization techniques

Simultaneously assessing amplitude and temporal effects in biomechanical trajectories using nonlinear registration and statistical nonparametric mapping

“…3B.2). The importance of nonlinear registration for reducing variance improving homology has been extensively demonstrated in previous research (Ramsay and Li, 1998;Sadeghi et al, 2000;Weiske et al, 2021;Moudy et al, 2018), but its relevance to amplitude-phase separation has not, to our knowledge, been previously reported in the biomechanics literature.…”

“…This process is often referred to as temporal 'normalization' in the biomechanics literature (e.g. Weiske et al, 2021), but this paper uses 'registration' (Sadeghi et al, 2000) to follow the broader literature (Ramsay and Li, 1998;Srivastava et al, 2011;Tucker et al, 2013;Marron et al, 2015;Wrobel et al, 2019).…”

“…Many nonlinear registration approaches exist for trajectory data including: event landmarking (Crane et al, 2010;Moudy et al, 2018) and manual warping (Pataky et al, 2019) along with many automated, algorithmic techniques (Ramsay and Li, 1998;Sadeghi et al, 2000;Marron et al, 2015). Nonlinear registration has been employed in the biomechanics literature to demonstrate both reduced variance (Sadeghi et al, 2000;Weiske et al, 2021) and improved correlation with performance measures (Moudy et al, 2018). Nonlinear registration algorithms remain in active development in the statistics literature (e.g.…”

Evaluation of the power deficit of elderly people during stair negotiation: Which joints should be assisted at least by an exoskeleton and with what amount?

“…User-specific gait patterns vary over several timing parameters like cadence or speed, while only the power deficit in amplitude during similar event times are necessary. To highlight differences in amplitude rather than differences due to shifted event times, we normalize time by DTW with one randomly drawn group sample each as reference signal (Honert and Pataky, 2021;Weiske et al, 2021).…”

“…Traditional findings such as Novak and Brouwer ( 2011 ) are based on linear time normalization and a multitude of trigger events. However, recent findings suggest that time normalization techniques (Weiske et al, 2021 ) and trigger events (Honert and Pataky, 2021 ) modify biomechanical results greatly. To quantify the power deficit between two groups appropriately, the use of dynamic time warping (DTW) for time normalization with statistical parametric mapping (SPM) for alpha level correction is therefore useful.…”