Comparison of training load models for predicting heart rate variability changes

Abstract EN

The aim of this study is to compare different training load (TL) models to predict changes in heart rate variability (HRV).

TL was evaluated using the heart rate stress score (HRSS), while HRV was measured in the morning using a polar H 7 chest strap and an android application during a 2-minute period for an individual adult engaged in recreational running.

the results indicate that the REDI model exhibits the highest correlation with HRV indices.

specifically, the REDI model's lambda values of 0.1, 0.03, and 0.02 demonstrate the strongest correlations with low frequency (LF), high frequency (HF), and Ln-RMSSD-RA (P< 0.001).

respectively.

the acute and chronic load of both the EWMA and Coggan models exhibit similar correlations with lf (P< 0.05).

conversely, no linear correlation was observed between any of the models' training stress balance (TSB) and HRV.

in non-linear regression, the Chronic-EWMA showed stronger relationship with Ln-RMSSD-RA compared to other models.