APA
Click to copy
Echeverría, S., Caulier-Cisterna, R., Vergara-Quezada, J., Contreras–Briceño, F., & Fuentealba, P. (2025). Estimation of Heart Rate and Respiratory Rate from NIRS Signals in Athletes using Empirical Mode Decomposition and Continuous Wavelet Transform. CHILEAN Conference on Electrical, Electronics Engineering, Information and Communication Technologies.
Chicago/Turabian
Click to copy
Echeverría, Selim, R. Caulier-Cisterna, Jorge Vergara-Quezada, Felipe Contreras–Briceño, and Patricio Fuentealba. “Estimation of Heart Rate and Respiratory Rate from NIRS Signals in Athletes Using Empirical Mode Decomposition and Continuous Wavelet Transform.” CHILEAN Conference on Electrical, Electronics Engineering, Information and Communication Technologies (2025).
MLA
Click to copy
Echeverría, Selim, et al. “Estimation of Heart Rate and Respiratory Rate from NIRS Signals in Athletes Using Empirical Mode Decomposition and Continuous Wavelet Transform.” CHILEAN Conference on Electrical, Electronics Engineering, Information and Communication Technologies, 2025.
BibTeX Click to copy
@article{selim2025a,
title = {Estimation of Heart Rate and Respiratory Rate from NIRS Signals in Athletes using Empirical Mode Decomposition and Continuous Wavelet Transform},
year = {2025},
journal = {CHILEAN Conference on Electrical, Electronics Engineering, Information and Communication Technologies},
author = {Echeverría, Selim and Caulier-Cisterna, R. and Vergara-Quezada, Jorge and Contreras–Briceño, Felipe and Fuentealba, Patricio}
}
This study explores the estimation of heart rate (HR) and respiratory rate (RR) from near-infrared spectroscopy (NIRS) signals in elite athletes during incremental exercise. Signals from the prefrontal cortex and m. Intercostales regions were preprocessed using empirical mode decomposition (EMD) to isolate intrinsic mode functions (IMFs) associated with cardiac and respiratory oscillations. These IMFs were analyzed using continuous wavelet transform (CWT) to obtain time-resolved frequency estimates of HR and RR. Performance was evaluated per subject using mean, median, standard deviation (STD), and interquartile range (IQR) of the absolute error, along with Pearson's correlation. For HR, the mean absolute error was 7.71 cycles/min, with a median of 5.35 cycles/min, STD of 9.05 cycles/min, and IQR of 2.72 cycles/min. For RR, the mean absolute error was 3.38 cycles/min, with a median of 2.97 cycles/min, STD of 2.25 cycles/min, and IQR of 1.71 cycles/min. Considering typical physiological ranges during exercise (HR: 120–180 bpm; RR: 30–60 breaths/min), these errors are within acceptable limits for practical applications. Correlation values exceeded 0.8 in 91.5% of subjects for HR but only 51.7% for RR, indicating that while RR errors were numerically smaller due to its narrower physiological range, its consistency across subjects was lower. Despite these limitations (particularly regarding respiratory estimation, motion artifacts, and sampling frequency), the results support the feasibility of EMD+CWT as a non-invasive and adaptive framework for physiological monitoring in sports settings.