Ventilatory Thresholds Estimation Based on ECG-derived Respiratory Rate

Diego García1, Spyridon Kontaxis2, Adrián Hernández-Vicente3, David Hernando1, Javier Milagro1, Esther Pueyo1, Nuria Garatachea3, Raquel Bailón4, Jesus Lazaro1
1University of Zaragoza, 2Biomedical Research Networking Centre on Bioengineering, Biomaterials and Nanomedicine, 3GENUD Research Group, FCSD, IA2, University of Zaragoza, Spain, 4I3A, IIS, Universidad de Zaragoza, CIBER-BBN


The purpose of this work is to study the feasibility of estimating the first and second ventilatory thresholds (VT1 and VT2, respectively) by using electrocardiogram (ECG)-derived respiratory rate during exercise testing. The ECGs of 25 healthy volunteers during cycle ergometer exercise test with increasing workload were analyzed. Time-varying respiratory rate was estimated from an ECG-derived respiration signal obtained from QRS slopes' range method. VT1 and VT2 were estimated as the points of maximum change in respiratory rate slope using polynomial spline smoothing. Reference VT1 and VT2 were determined from the ventilatory equivalents of $O_2$ and $CO_2$. Estimation errors of -13.96 (54.84) W for VT1 and -8.06 (39.63) W for VT2 (median (interquartile range)) were obtained, suggesting that ventilatory thresholds can be estimated from solely the ECG signal.