Dynamic networks in the brain inferred from the analysis of neurophysiology data

The human brain is a hierarchical complex network − always active and evolvingdynamically. Its functional organization can be studied by investigating dynamicchanges in brain network interactions. In this thesis, I investigate dynamic changesin large-scale neuronal network to address the organization of the brain. In two complementaryconditions, resting-state and action-perception, functional connectivitypatterns in electrophysiological data are investigated to identify principles of networkdynamics.I first develop a time-frequency method to measure the temporal evolution of neuronalsynchronization. The second study uses this method to detect the dynamicsof neuronal synchronization in resting-state electroencephalography. I show sevenrobust networks with distinct topographic organizations over multiple frequencies,which build up and decay over much slower timescale, capturing the low-dimensionallinear subspaces in which resting-state activities unfold.The third study explores the role of neural dynamics in human motor control. I showthat corticomuscular coherence in the beta-band is replaced by alpha- and gammabandcoherence during a transition between sensorimotor states. Analyzing the phasespectra of the dual-band coherence reveals a↵erent and e↵erent corticospinal interactions,interpreting parallel streams of corticospinal processing in parsing predictionerrors and generating new predictions.Thereafter, I turn attention towards the identification of patterns of neuronal synchronizationin cortical source-level. The problems of volume conduction and thereference electrode ensure that this is a non-trivial endeavor. The fourth study exploresthe organization of the spatial patterns of neuronal communication in restingstateEEG source reconstruction. I show significant patterns of spatial functionalconnectivity using envelope correlation and phase synchronization, providing complementaryinsights into the timescales of neuronal dynamics. In final chapter, I showfurther approaches that can be built on those presently obtained to further elucidatethe orthogonal linear subspaces of synchronous networks at distinct frequencies.In sum, the modulation of synchronous networks at multiple frequencies suggests thebrain undergoes sequential metastable states: There arise spontaneously at rest andare reshaped by task conditions. The combined studies demonstrate a few principlesof spatiotemporally reorganization of brain networks synchronization, mirroring thecomplexity of brain activity