Brain Organization into Resting State Networks Emerges at Criticality on a Model of the Human Connectome

Abstract

The relation between large-scale brain structure and function is an outstanding open problem in neuroscience. We approach this problem by studying the dynamical regime under which realistic spatio-temporal patterns of brain activity emerge from the empirically derived network of human brain neuroanatomical connections. The results show that critical dynamics unfolding on the structural connectivity of the human brain allow the recovery of many key experimental findings obtained with functional Magnetic Resonance Imaging (fMRI), such as divergence of the correlation length, anomalous scaling of correlation fluctuations, and the emergence of large-scale resting state networks.Understanding the relation between brain architecture and function is a central question in neuroscience. In that direction, important efforts over recent years have been devoted to map the large-scale structure of the human cortex, including attempts to build brain structural connectivity matrices from imaging data. An example is the connectivity matrix of the entire human brain, recently derived from fiber densities measured between a large number (500-4000) of homogeneously distributed brain regions [1]. This and related work encompasses a large collaborative project dubbed the brain "connectome" [3], whose ultimate goal is to understand in detail the architecture of whole-brain connectivity. However, "like genes, structural connections alone are powerless", thus "the connectome must be expressed in dynamic neural activity to be effective in behavior and cognition" [2]. The results presented in this Letter show that very relevant aspects of brain dynamics can be predicted from the structure provided that the underlying dynamics are critical.To guide our comparison with available experimental results, we choose to concentrate on robust findings concerning brain dynamics. Specifically, we ask how spontaneous brain dynamics at the large scale organize into the relatively few spatio-temporal patterns revealed experimentally in recent years [4]. This is important because a wide range of experiments using functional Magnetic Resonance Imaging (fMRI) have emphasized that these spatial clusters of coherent activity, termed Resting State Networks (RSN) [5], are specifically associated with neuronal systems responsible for sensory, cognitive and behavioural functions [6]. Furthermore, the pattern of correlations in these networks has been shown to change with various cognitive and pathophysiological conditions [4]. Of interest here are studies showing that the RSN activity exhibits peculiar scaling properties, resembling dynamics near the critical point of a second order phase transition [7][8][9], consistent with evidence showing that the brain at rest is near a critical point [10]. These empirical findings are in line with computational modeling results [11][12][13].Here we study whether a simple dynamical model running over the empirical structure of neuroanatomical connections [1] suffices to replicate the aforementioned fundamental features of ...

Rare regions of the susceptible-infected-susceptible model on Barabási-Albert networks

“…Studies of the CP, as well as other processes [8,9], have shown that quenched disorder in networks is relevant in the dynamical systems defined on top of them. Very recently it has been shown [16][17][18] that generic slow (power-law or logarithmic) dynamics is observable by simulating CP on networks with finite d. This observation is relevant for recent developments in dynamical processes on complex networks such as the simple model of "working memory" [19], brain dynamics [20], social networks with heterogeneous communities [21], or slow relaxation in glassy systems [22].…”

Mean-field behavior as a result of noisy local dynamics in self-organized criticality: Neuroscience implications

“…Our main motivation is due to the recent observation of neuronal avalanches and their presumed relation to SOC. In a wide range of recent experiments [14][15][16][17][18][19][20][21] , neuronal avalanches have been shown to exhibit power law behavior with mean-field exponents. Whether neural dynamics [22] is dissipative or not is still debated, but their noisy dynamics [23] is a certainty, as the post-synaptic neurons receive more or less than their fair share of the ion distributed by the pre-synaptic neuron in the ionic plasma, which permeates the space between synapses.…”

Synchronization, non-linear dynamics and low-frequency fluctuations: Analogy between spontaneous brain activity and networked single-transistor chaotic oscillators

“…19,28,29 At the same time, the fact that realistic RSNs emerge even in highly simplified simulation scenarios, for example, from networked phase (Kuramoto) oscillators or discrete excitable units, raises the intriguing possibility of recapitulating some dynamical phenomena underlying brain function in other physical systems, where direct manipulation of connectivity is possible and causal relationships between connectivity and non-linear dynamics can be explored experimentally. [8][9][10]30 In particular, it has recently been shown that singletransistor oscillators can exhibit strikingly complex activity depending on an easily tunable control parameter (DC voltage source series resistance), oscillating periodically, chaotically, or close to criticality. 31 An experimental investigation of a ring of 30 diffusively coupled such oscillators, each consisting of a bipolar junction transistor, three reactive components and a resistor, has furthermore demonstrated the spontaneous formation of multi-scale community structure as a function of coupling strength, with elements of similarity to the modular organization observed in brain networks.…”

“…[5][6][7] Based on realistic models of structural connectivity, numerical simulations predict the emergence of functional connectivity (activity synchronization between regions) in the form of discrete "resting-state" networks (RSNs), such as the "default-mode network" and "fronto-parietal network." [7][8][9][10] These networks have been detected by means of independent component analysis of blood oxygen level dependent (BOLD) time-series recorded using functional MRI during awake idleness (resting-state), whose fluctuations primarily represent spontaneous neural activity. 11,12 Graph-based analyses of BOLD signal synchronization have also confirmed high node degree of functional connectivity (representing a measure of "synchronization density") in the aforementioned cortical hub regions, in broad agreement with the underlying structural connectivity.…”