Electrical Activity Characteristics Of Mouse Brain Slices In A Low-Mg²⁺ Epilepsy Model

Mesial temporal lobe epilepsy?mTLE?is the most common form of epilepsy in clinical practice,and is also among the most refractory type of epilepsy.Studying the pathogenesis and network mechanisms of mTLE helps to provide more comprehensive and effective antiepileptic strategies,and is thus of great importance.mTLE is mainly associated with the pathological change and abnormal activity of hippocampus,entorhinal cortex?EC?and other medial tempo-ral lobe structures.In this study,multi-channel epileptiform discharges in the combined EC-hippocampal slices induced by low-Mg²⁺artificial cerebrospinal fluid?ACSF?were analyzed,in order to clarify the properties and initiation-propagation patterns of different epileptiform discharges.In addition,this study proposed an unsupervised clustering algorithm,and utilized network analysis techniques to characterize the dynamic evolution of brain networks with var-ious graphic indicators,which improves the understanding of the network mechanisms over different epileptic states.In the first part of this work,the spatio-temporal characteristics of epileptiform discharges including the discharge frequency,duration,amplitude and initiation-propagation pattern were calculated.Recordings were subjected to time-frequency analysis by using a local multi-scale Fourier transform method.In the second part of this work,the time-variant autoregressive model was constructed based on multi-channel epileptiform discharges,and the time-variant partial directed coherence?tvPDC?was calculated on the coefficient matrix of the model to measure the causal dynamics in the EC-hippocampal networks.In order to investigate the topological changes over different phases of epileptiform discharges,this paper proposed an unsupervised clustering algorithm to track the transitions of network states.Furthermore,graph theoretical measures including degrees,modularity and small-worldness were adopted to quantify the net-work features.The spatio-temporal analysis showed that low-Mg²⁺induced three types of epileptiform discharges in combined slices,including two types of interictal epileptiform?interictal-like?discharges which initiated in the CA3 and the dentate gyrus?DG?respectively.The other syn-chronous events were ictal epileptiform?ictal-like?discharges which displayed tonic and clonic phases.The ictal wavefront?tonic discharges?initiated in the EC,but the late clonic discharges occurred first in the DG.Moreover,the characteristics and propagation pattern of the clonic com-ponents were in accordance with those of the DG-initiated interictal-like events.The clinical antiepileptic drug valproate?VPA?suppressed the CA3-initiated interictal-like discharges,while inversely facilitated the DG-initiated interictal-like discharges.The network analysis showed that the hub nodes with high degrees in the network coincided with the epileptogenic zone in previous electrophysiological findings,and the node communities were also in agreement with the anatomical structures of the combined slices.The ictal-like discharges consisted of two dis-tinct network states.The first state was similar to a scale-free network,while the second state was close to a small-world network.Our results indicate that low-Mg²⁺ACSF induced two types of interictal-like discharges in EC-hippocampal slices,which initiated in CA3 and DG respectively,and had different gener-ation mechanisms.The high correlation of DG-initiated interictal-like discharges and another clonic discharges that also originated in DG suggests that these interictal-like discharges may contribute to the maintenance of ictal-like discharges.The tvPDC measures the relationship between different nodes in the brain network from a perspective of directed information flow,which is conducive to understanding the network mechanisms of epileptiform discharges.The state transitions of the epileptic brain networks reveal important topological features in the gen-eration and maintenance of ictal-like discharges.