Propagation And Response Dynamics On Heterogeneous Brain Networks

The brain is a network with multi-scale features that encompasses multiple spatial scales from individual neurons and synapses to brain regions and large-scale fiber tracts.Advanced electrical activity recording and imaging techniques allow us to construct brain networks at different scales and study their topology and dynamics.There have been many studies on the heterogeneity of network structure at different scales,and personalized brain network models constructed accordingly can be used to study brain function or brain disease.Nodes in the brain network have the same dynamic parameters in most studies.However,there is increasing evidence that node heterogeneity exists widely in brain networks at different scales,which leads to the results obtained in brain network models with homogeneous nodes being less reliable.At the same time,the brain still maintains these heterogeneous properties after a long evolution,which likely means that node heterogeneity plays a critical role in improving the working efficiency of the brain.In this paper,we investigated the propagation and control of focal seizures in a large-scale whole-brain model that accounts for the heterogeneity of nodal excitability,and the effect of excitatory-inhibitory neuronal networks incorporating neuronal resting potential heterogeneity on network responses and dynamic range.These studies can contribute to the understanding of brain function and the personalized treatment of brain disorders.Correct prediction and efficient control of seizure spread are vital to the successful treatment of focal epilepsy with surgery.As a promising approach to fulfill this purpose,personalized brain network models derived from noninvasive structural data of individual patients with epilepsy have attracted much attention.Although increasing evidence suggests that regional cortical excitability exerts a substantial effect on seizure spread,the existing seizure prediction methods are mostly based on structural connection information.Therefore,accurate prediction of seizure spread in whole-brain models with heterogeneous regional excitability is crucial for seizure control and has potential applications in the clinical treatment of epilepsy.In this study,we aim to solve this problem by proposing a prediction method based on random walk on graphs,named the modified random walk with extended restart(m RWER).Simulations in the largescale whole-brain model show that heterogeneous excitability substantially influences seizure propagation patterns.In this case,while the prediction of seizure spread with structural connections was jeopardized by heterogeneous excitability,prediction with m RWER yielded better performance and was more robust against heterogeneity.Surgical control of seizures propagation can be achieved by identifying key nodes responsible for the early spread of seizures from the focal region and remove connections between them and focal node.Compared to strategies based on structural connections,virtual surgery with a strategy based on m RWER generated outcomes with a high success rate while maintaining low damage to the brain by removing fewer anatomical connections.Understanding how the brain efficiently completes various complex tasks is the main goal of studying the brain.Many studies have shown that the brain has selforganized critical features,which can optimize the brain's information processing capacity,such as the dynamic range of the network in response to external stimuli.These features are closely related to the excitatory-inhibitory balance of cortical networks.On the other hand,more and more studies have begun to focus on the heterogeneity of neurons,and it has been proved that neuronal heterogeneity can optimize the efficient coding,responsiveness,and information flow of the network.However,the effect of neuronal heterogeneity on the dynamic range of excitatoryinhibitory neuronal networks and the excitatory-inhibitory balance remains unclear.Here,we construct a network with excitatory and inhibitory neurons with heterogeneous resting potentials and find that balanced excitatory-inhibitory synaptic currents and moderate heterogeneity maximize the dynamic range of the network.In addition,the excitatory-inhibitory balance of the network is disrupted when excitatory and inhibitory neurons have different degrees of heterogeneity.The mRWER approach that introduces nodal heterogeneity can better predict and control the spread of focal seizures in a heterogeneous large-scale whole-brain network,which has potential applications in precision medicine based on personalized data.In addition,we also verified that neuronal heterogeneity can enhance the dynamic range of the network and can affect the excitation-inhibitory balance of the network,suggesting that neuronal heterogeneity can both optimize the dynamic range of brain responses to stimuli and possibly affect the brain's information processing ability by changing the excitation-inhibitory balance in the network.