Dynamic characteristics of oxygen consumption.

Area of Science:

• Cardiorespiratory Physiology
• Exercise Science
• Biomedical Engineering
• Nonparametric Modeling

Background:

• Oxygen uptake (VO2) is a key indicator of cardiorespiratory response to exercise.
• Traditional VO2 modeling often uses simplified first-order systems due to limited data quality.
• Existing methods struggle with inadequate stimulation and low signal-to-noise ratios in VO2 measurements.

Purpose of the Study:

• To propose a novel nonparametric kernel-based method for dynamic VO2 modeling.
• To achieve more robust and accurate estimations of cardiorespiratory responses during exercise.
• To overcome limitations of traditional methods in modeling VO2 dynamics.

Main Methods:

• Conducted treadmill exercises with monotonous stimulation on 20 healthy participants.
• Measured and recorded VO2 using a portable gas analyzer (COSMED).
• Applied a kernel-based estimation method for nonparametric VO2 modeling, utilizing elastic net regularization and a stable spline kernel.

Main Results:

• The proposed kernel-based method demonstrated superior performance in VO2 modeling.
• The stable spline (SS) kernel yielded the best results among tested kernels.
• The SS kernel-based method significantly outperformed the prediction error method (PEM) in accuracy.

Conclusions:

• The nonparametric modeling method effectively estimates the impulse response of the VO2-Speed system.
• The identified average nonparametric model accurately predicts VO2 response during treadmill exercise.
• This approach offers a more robust tool for assessing cardiorespiratory fitness.