Abstract:
During slow-wave sleep, the brain is in a self-organized regime in which slow oscillations (SOs) between up- and down-states propagate across the cortex. Here we address how SOs emerge locally and can recruit large parts of the brain. A whole-brain model that produces bistable states and adaptation-induced SOs is fit to fMRI and EEG data using a multi-objective evolutionary algorithm. Optimal fits are found at critical values of the adaptation strength where the model produces a wide range of local and global SOs with realistic statistics. Local oscillations are more frequent, last shorter, and have a lower amplitude. Global oscillations spread as waves of silence across the brain. The well-known directionality of propagation from anterior to posterior regions is explained by heterogeneities of the connectome. Our results demonstrate the applicability of whole-brain models beyond the resting-state and their potential for elucidating the spatiotemporal patterns of brain oscillations in general.