Machine Learning-Based Classification of Alertness Levels in Elite Shooting Athletes Using Heart Rate Variability.

Area of Science:

• Sports Science
• Psychophysiology
• Machine Learning in Sports

Background:

• Elite athletes require sustained attention during competition.
• Monitoring alertness is crucial for performance and safety.
• Heart Rate Variability (HRV) offers a non-invasive window into physiological stress.

Purpose of the Study:

• To develop a predictive model for alertness in elite shooting athletes.
• To analyze Heart Rate Variability (HRV) dynamics under simulated competitive stress.
• To identify key HRV predictors of alertness using machine learning.

Main Methods:

• 83 national-level shooting athletes underwent a 60-minute Psychomotor Vigilance Task (PVT).
• Continuous HRV data were recorded and analyzed for key features.
• Machine learning algorithms (SVM, RF, XGBoost, AdaBoost) built binary classification models for alertness.

Main Results:

• The AdaBoost model achieved the highest performance (Accuracy: 0.75, F1-score: 0.73, AUC: 0.77).
• Very Low Frequency percentage (VLF%) was the most significant predictor of alertness.
• Elevated VLF% correlated with decreased alertness levels.

Conclusions:

• A binary classification model integrating HRV indices (VLF%) and AdaBoost effectively distinguishes alertness levels in shooting athletes.
• This provides a validated, non-invasive tool for objective psychophysiological monitoring.
• The model offers actionable insights for pre-competition readiness assessment in sports training.