Oxygen Uptake Plateau Diagnosis Using a New Developed Segmented Regression Estimation Method for Autocorrelated Data

Area of Science:

• Physiology
• Biostatistics
• Exercise Science

Background:

• Linear regression inadequately models nonlinear oxygen uptake dynamics.
• Segmented regression offers a more satisfactory approximation for oxygen uptake.
• Existing methods for plateau identification struggle with autocorrelated data.

Purpose of the Study:

• To extend segmented regression and Wald-type tests for oxygen uptake plateau identification to autocorrelated data.
• To develop and evaluate an algorithm for estimating segmented regression models under autocorrelation.

Main Methods:

• Proposed an algorithm using generalized least squares for segmented regression with autocorrelation.
• Developed a bootstrap method for resampling Wald's statistic null distribution.
• Evaluated performance via Monte Carlo simulations and applied to real oxygen consumption data.

Main Results:

• The proposed estimator outperforms classic methods in autocorrelated scenarios, particularly with small sample sizes.
• The plateau diagnosis test demonstrated coherent empirical Type 1 Error rates and good statistical power.
• The method objectively identifies oxygen consumption plateaus based on a significance level.

Conclusions:

• The developed methods provide a robust approach for estimating parameters in segmented regression models for autocorrelated data.
• The proposed bootstrap test effectively diagnoses oxygen consumption plateaus.
• The methods show good performance in both simulated and real-world case studies.