Countermovement Jump Force-Time Mechanics Differentiate ACL Injury Status in Elite Alpine Ski Racers.

Area of Science:

• Biomechanics
• Sports Medicine
• Machine Learning

Background:

• Countermovement jump (CMJ) assessments evaluate neuromuscular function in alpine ski racers post-anterior cruciate ligament reconstruction (ACLR).
• Numerous CMJ force-time metrics complicate data interpretation and return-to-sport decisions.
• Machine learning (ML) offers a method to analyze complex biomechanical data and identify patterns.

Purpose of the Study:

• To apply ML classification algorithms to CMJ force-time metrics to distinguish between elite alpine ski racers with ACLR and healthy controls.
• To identify key biomechanical features predictive of ACLR status.
• To explore ML's potential in assessing neuromuscular recovery and rehabilitation progression.

Main Methods:

• Trained ML algorithms (random forest, SVM, logistic regression, naïve Bayes, k-NN) using 23 CMJ force-time features from 24 ACLR athletes and 42 controls.
• Utilized 5-fold cross-validation and an independent test dataset for evaluation.
• Analyzed CMJ data collected longitudinally from 836 testing sessions.

Main Results:

• ML models achieved high classification performance, with balanced accuracies from 0.59 to 0.88 and AUCs from 0.63 to 0.95.
• Features related to the propulsion phase of the CMJ were most critical in differentiating ACLR athletes from controls.
• The study demonstrated ML's capability to identify significant biomechanical differences.

Conclusions:

• ML classification models can aid in interpreting CMJ force-time data for athletes recovering from ACLR.
• Identifying high-information features related to injury status can support return-to-sport decision-making.
• CMJ mechanics becoming indistinguishable from controls may indicate successful neuromuscular recovery.