Elastic Shape Analysis of Movement Data

TL;DR

This paper introduces a shape-based analysis method for biomechanical movement curves, demonstrating that analyzing full curves provides stronger associations with osteoarthritis severity than traditional discrete summaries.

Contribution

It presents a novel shape analysis approach for biomechanical curves and quantitatively shows its advantages over conventional discrete summaries in OA research.

Findings

Full movement curve analysis yields stronger OA associations.

Shape-based representation captures more biomechanical variation.

Method improves understanding of biomechanics in OA.

Abstract

Osteoarthritis (OA) is a highly prevalent degenerative joint disease, and the knee is the most commonly affected joint. Biomechanical factors, particularly forces exerted during walking, are often measured in modern studies of knee joint injury and OA, and understanding the relationship among biomechanics, clinical profiles, and OA has high clinical relevance. Biomechanical forces are typically represented as curves over time, but a standard practice in biomechanics research is to summarize these curves by a small number of discrete values (or landmarks). The objective of this work is to demonstrate the added value of analyzing full movement curves over conventional discrete summaries. We developed a shape-based representation of variation in full biomechanical curve data from the Intensive Diet and Exercise for Arthritis (IDEA) study (Messier et al., 2009, 2013), and demonstrated through nested model comparisons that our approach, compared to conventional discrete summaries, yields stronger associations with OA severity and OA-related clinical traits. Notably, our work is among the first to quantitatively evaluate the added value of analyzing full movement curves over conventional discrete summaries.