Criticality and network structure drive emergent oscillations in a stochastic whole-brain model

Abstract

Understanding the relation between the structure of brain networks and its functions is a fundamental open question. Simple models of neural activity based on real anatomical networks have proven effective in describing features of whole-brain spontaneous activity when tuned at their critical point. In this work, we show that indeed structural networks are a crucial ingredient in the emergence of synchronized oscillations in a whole-brain stochastic model at criticality. We study such model in the mean-field limit, providing an analytical understanding of the associated first-order phase transition, arising from the presence of a bistable region in the parameters space. Then, we derive the power spectrum in the linear noise approximation and we show that, in the mean-field limit, no global oscillations emerge. Finally, by adding back an underlying brain network structure with homeostatic normalization, we numerically show how the bi-stability region is disrupted and concomitantly a synchronized phase with maximal dynamic range is observed. Hence, both the structure of brain networks and criticality are fundamental in driving the collective coordinated responses and maximal sensitivity of whole-brain stochastic models.

Investigating structural and functional aspects of the brain’s criticality in stroke

“…While the authors of another study simulating Ising dynamics on connectomes 47 hypothesize that structural connectivity alone cannot explain some effects observed in functional correlations of patients after severe brain injuries, no other explanation has been provided in the literature. Our explanation of the anomalous behavior of the cluster-based indicator of criticality, on the other hand, is consistent with the distinction 51 , 52 between well-known structurally-driven percolation phase transitions and dynamic transitions-for which less is known 36 . Reframed in these terms, our results would suggest that in brain strokes there is no change in whatever underlies dynamic transitions, but that there are structural changes that could disorganize the percolation-like transitions.…”

Prenatal experience with language shapes the brain

“…From a broader perspective, our findings document power law scaling of neural activity during language processing in the newborn brain. This statistical property is a hallmark of critical phenomena, and it has been suggested that criticality in the brain is linked to states of optimal information transmission and storage (49)(50)(51)(52). The newborn brain may thus already be in an optimal state for the efficient processing of speech and language, underpinning human infants' unexpected language learning abilities.…”

Does Adding of Neurons to the Network Layer Lead to Increased Transmission Efficiency?

“…Schwalger et coauthors proposed a system of equations for several interacting populations at the mesoscopic scale, starting from a microscopic model of randomly connected generalized integrateand-fire neuron models for networks varying between 50-2000 neurons [5]. In turn, in [6], it was shown that structural networks are a crucial component of the stochastic brain model on the mesoscopic scale. A metric called multiscale relevance (MSR) was proposed in [7] to capture the dynamic variability of the activity of single neurons across different scales.…”