Answers are grounded in this paper and cite their sources.
This cross-sectional study investigated the test-retest reliability, measurement precision, and criterion validity of evaluating the McGill core muscle endurance test (right and left side-bridge tests) using an AI (MediaPipe) computer vision application compared to a traditional standardized stopwatch among basketball players. Thirty healthy Thai basketball players (mean age 22.9 ± 2.62 years) performed right and left side-bridge tests across two trials. Performance failure times were independently evaluated by two experts using both a stopwatch and the AI (MediaPipe) tool. Statistical evaluations included Intraclass Correlation Coefficients (ICC) for reliability, Pearson correlation coefficients for criterion validity against mean expert ratings, and Bland-Altman plots to assess method agreement, alongside Standard Error of Measurement (SEM) and Minimal Detectable Change (MDC95) thresholds. Both experts showed good reliability (ICC > 0.75) for both stopwatch and AI methods across both sides. The stopwatch method achieved lower SEM values (Right: 1.12–1.52 s; Left: 1.31–1.39 s) and smaller MDC95 change thresholds (3.09–4.23 s) than the AI system (SEM: 1.39–1.63 s; MDC95: 3.86–4.52 s), demonstrating superior measurement precision and sensitivity to true performance change. Criterion validity analyses revealed non-significant linear associations and low correlation coefficients between test failure times and mean expert ratings. Bland-Altman plots demonstrated normal distributions of differences, acceptable limits of agreement, and mostly minimal proportional bias across trials. This study can be concluded that AI (MediaPipe) application shows reliable feasibility for automated tracking, the traditional stopwatch currently retains higher measurement precision, consistency, and sensitivity for detecting true clinical changes in core endurance.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: