Application of functional data analysis to explore movements: walking, running and jumping - A systematic review

Background Signals are continuously captured during the recording of motion data. Statistical analysis, however, usually uses only a few aspects of the recorded data. Functional data analysis offers the possibility to analyze the entire signal over time. Research question The review is based on the question of how functional data analysis is used in the study of lower limb movements. Methods The literature search was based on the databases EMBASE, PUBMED and OVID MEDLINE. All articles on the application of functional data analysis to motion-associated variables trajectories, ground reaction force,electromyography were included. The references were assessed independently by two reviewers. Results In total 1448 articles were found in the search. Finally, 13 articles were included in the review. All were of moderate methodological quality. The publication year of the studies ranges from 2009 to 2019. Healthy volunteers and persons with cruciate ligament injuries, knee osteoarthritis, gluteal tendinopathy, idiopathic torsional deformities, slipped capital femoral epiphysis and chronic ankle instability were examined in the studies. Movements were analyzed on basis of kinematics (3D motion analysis), ground reaction forces and electromyography. Functional Data Analysis was used in terms of landmark registration, functional principal component analysis, functional t-test and functional ANOVA. Significance Functional data analysis provides the possibility to gain detailed and in-depth insights into the analysis of motion patterns. As a result of the increase in references over the past year, the FDA is becoming more important in the analysis of continuous signals and the explorative analysis of movement data.