Multiscale Eigen Analysis on EEG and Heartbeat Dynamics: a Pilot study

Abstract : The continuous reciprocal interplay underlying cerebral and cardiovascular interactions has been shown to generate complex and nonlinear dynamics. To this extent, differences in induced cross-temporal dynamics are here investigated via wavelet-based multivariate multiscale analysis. Twelve features were extracted from both EEG and ECG recordings from 24 healthy subjects at rest and during a coldpressor test. The proposed multivariate analysis using the eigenstructure of the multiscale decomposition was compared with a classical multivariate analysis. Preliminary results show that differences between experimental conditions are enhanced by the application of the proposed multivariate analysis.