An Explainable Machine Learning Approach to Explain the Effects of Training and Match Load on Ultra-Short-Term Heart...

Area of Science:

• Sports Science
• Exercise Physiology
• Data Science in Sports

Background:

• Optimizing athlete monitoring requires understanding the influence of training and match load on autonomic recovery.
• Heart rate variability (HRV) is a key indicator of autonomic recovery and overall physiological status.

Purpose of the Study:

• To investigate the impact of training and match load on next-day HRV.
• To utilize explainable machine learning (SHAP) to identify influential load metrics on internal physiological responses.

Main Methods:

• Monitored five semi-professional basketball players throughout a season, collecting daily HRV and load metrics.
• Employed linear mixed models to analyze HRV differences across training and non-training days.
• Developed a Gradient Boosting Machine model with SHAP analysis to interpret feature importance for HRV prediction.

Main Results:

• Next-morning HRV (LnRMSSD) was significantly lower on match days compared to training and non-training days.
• SHAP analysis identified Rate of Perceived Exertion (RPE), days since last match, minutes played, and recent training load as primary drivers of HRV changes.
• Individual player responses showed variability, emphasizing personalized physiological responses.

Conclusions:

• Match load significantly impacts autonomic recovery, leading to reduced next-day HRV.
• Explainable AI (SHAP) offers valuable insights into individualized athlete responses to training and match loads.
• Integrating subjective and objective load measures is crucial for effective athlete monitoring and training prescription.