Principal component analysis as a novel approach for cardiorespiratory exercise testing evaluation

Our purpose was to apply a principal component analysis (PCA) approach to cardiorespiratory exercise to test evaluation and its sensitivity to workload accumulation.

Twenty-five healthy young adults performed a progressive and maximal cycling test, which was divided into two parts: moderate and high workload intensities, using a ventilatory threshold as a cut point. A PCA of the time series of cardiovascular and respiratory variables was performed in each part and the number of principal components (PCs), the eigenvalues of the first PC (PC₁), and the information entropy were calculated.

The number of PCs increased, the eigenvalues of PC₁ decreased (t  =  5.32; p   <  0.001; d  =  1.39) and entropy was significantly higher (Z  =  3.10; p   =  .002; d  =  1.16) at high workload intensities, compared to moderate intensities.

Results showed the sensitivity of the PCA approach to workload accumulation and corroborates its potential for improving the evaluation and interpretation of cardiorespiratory exercise testing. In particular, it points to being a good candidate to objectively detect qualitative changes or thresholds.