Causal symbolic information transfer for the assessment of cardiovascular and cerebral interactions through EEG microstates occurences

Knowledge of functional brain–heart interplay is essential owing to the strong physiological and clinical reciprocal implications that exist between central and peripheral nervous systems. Indeed, many neurological, psychiatric, and cardiovascular disorders may be related to central autonomic network (CAN) dysfunctions. Hence, it is evident that a proper methodology to quantify the functional brain–heart interaction might be crucial for the objective diagnosis and treatment monitoring of CAN-related diseases. The aim of this thesis is to estimate the directional interplay between the brain and the heart through joint symbolic analysis (JSA) and information measures in a group of subjects performing a mental arithmetic task (MAT). EEG microstates analysis was employed to extract a sequence of quasi-stable topographies representing the scalp-related electric field of each subject over time. Each microstate, reflecting the activation of different neural networks, was treated as a non-ordered symbol in the successive JSA. By contrast, the heart activity was represented by R-peaks intervals series, HRV high frequency [0.15-0.40 Hz] and low frequency [0.04-0.15 Hz] instantaneous powers, which were extracted by integration of Smoothed Pseudo Wigner-Ville Distribution (SPWVD) of the re-sampled interpolated HRV sequence. Heart-related signals were symbolized according to intra-series Max-Min symbolization. After that, four classes of variations were defined in order to obtain a common symbolic alphabet for both series. Information correlates (i.e. Transfer Entropy, Joint Entropy, Kullback-Leibler divergence) were extracted from the data and their statistical significance between resting state and MAT conditions were assessed through Wilcoxon signed rank test