Reliability and predictive validity of trunk and lower extremity kinematics during a jump-landing task using OpenCap markerless motion capture system

Abstract

Markerless motion capture systems offer promising alternatives to traditional marker-based systems for biomechanical assessment, but their measurement properties require thorough validation before widespread adoption. This study assessed the reliability of trunk and lower extremity joint angles and predictive validity of OpenCap, a markerless motion capture system, during a jump-landing task. Thirty-three healthy individuals performed a total of 15 jump-landing trials under two sessions: 5 self-selected (natural) landings during the first session and another 5 natural and 5 verbally cued stiff landings during the second session. Trunk and lower extremity joint angles were captured from two simultaneously operated OpenCap systems. Intersystem and test–retest reliability was assessed by computing intraclass correlation coefficients (ICC) and minimal detectable changes (MDC). Predictive validity was evaluated using area under the receiver operator characteristic (AUC) from a mixed effect logistic regression model to determine whether OpenCap-derived joint angles could accurately identify verbally cued landing conditions (natural and cued stiff). OpenCap demonstrated moderate to excellent intersystem reliability (Range: ICC2,1: 0.79–1.00) and test–retest reliability (Ranges: ICC2,k: 0.70–0.97; MDC: 1.89–11.62˚) across all joint angles. Additionally, OpenCap successfully distinguished between natural and cued stiff landings (AUC = 0.92; 95% CI: 0.90, 0.94]). By offering an accessible and efficient alternative to traditional marker-based motion capture systems, OpenCap has the potential to enhance movement assessment, injury screening, and rehabilitation monitoring in both clinical and field settings.