Timing of gait events affects whole trajectory analyses: A statistical parametric mapping sensitivity analysis of lower limb biomechanics

• Published in: Journal of biomechanics Vol. 119; p. 110329
• Main Authors: Honert, Eric C., Pataky, Todd C.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 15.04.2021; Elsevier Limited
• Subjects: Algorithms; Biomechanics; Continuity (mathematics); Curve registration; Gait; Gait analysis; Hypotheses; Injury analysis; Injury prevention; Kinematics; Kinetics; Mapping; Parameter sensitivity; Power; Principal component analysis; Principal components analysis; Registration; Sensitivity analysis; Statistics; Time continuous analysis; Trajectory analysis; Walking; United States > US; Time continuous analysis; Gait analysis; Curve registration; Principal component analysis
• ISSN: 0021-9290; 1873-2380; 1873-2380
• DOI: 10.1016/j.jbiomech.2021.110329

Time continuous analyses, such as statistical parametric mapping (SPM), have been increasingly used in biomechanics research to determine differences between populations, interventions and methodologies. Currently, it is not known how sensitive time-continuous analyses are to timing variability that occur in gait data. We evaluated this sensitivity by examining the frequency of significant SPM outcomes between two walking speeds when lower limb kinematics and kinetics were segmented and aligned based on 40 repeatable gait events. These events, defined in the supplementary material, include a commonly used event like foot contact and other events that have been previously demonstrated to be repeatable. Repeatable gait events were determined from joint and segment kinematics, joint kinetics as well as ground reaction forces. We examined the frequency of statistical outcomes for a single subject with different numbers of strides analyzed and for a cohort of 10 subjects. Our findings demonstrate that gait interventions, such as changes in walking speed, can induce temporal shifts that affect time-continuous outcomes for both cohort- and subject-level analyses. As both timing and magnitude are important in gait data, researchers are encouraged to perform additional analyses to understand how both of these variables affect time-continuous analysis outcomes. Finally, we demonstrate that multiple SPM tests can be performed to determine if statistical outcomes are due to temporal shifting or differences in magnitude. It is important to understand how both timing and magnitude of biomechanical data influences time continuous analyses as these analyses inform injury prevention, device development and basic understanding of biomechanics.