Rationale:
Seizure propagation is poorly understood. The dynamotype model of seizures describes the basic properties leading to seizure initiation and termination based upon bifurcation theory. It provides insights into how cortical masses enter and leave bursting activity without making assumptions about underlying seizure etiology. The goal of this work is to characterize recruitment behaviors across multiple dynamotypes using this model.
Methods:
We implemented a novel coupling method using the dynamotype model of seizures to investigate recruitment of a non-epileptogenic region (node 2) into a seizure (node 1). In the model, node 1 bursts autonomously. Node 2 does not burst autonomously but can be recruited into a burst by node 1. Simulations were run for a variety of node 1 dynamotypes, node 1 excitability values, node 2 resting parameters, and coupling strengths. For each simulation with recruitment, properties including recruitment delay and recruitment energy were measured. To confirm the range of node 2 dynamics present in the model were possible in the epileptic brain, we compared to clinical data collected as part of long-term intracranial video-EEG monitoring in patients with epilepsy. All patients consented to have their deidentified EEG data saved for research, and the protocol was approved by the local IRB.
Results:
Simulations indicated that seizure recruitment properties depend on both the dynamics of the seizure onset zone and the resting state dynamics of the propagation zone. Seizures that did not display baseline shifts at onset were more likely to propagate, and propagated faster, than seizures that displayed baseline shifts at onset. Seizures that exhibited increasing amplitude at onset were unlikely to propagate. Further, the model provided an explanation for why seizure dynamics may change from onset to propagation zone. We confirmed the range of recruitment behaviors present in the model are possible in the epileptic brain by comparison to clinical propagation examples.
Conclusions:
The model’s predictions provide a theoretical framework for seizure propagation, deepening our understanding of recruitment of non-epileptogenic brain regions during a seizure. In identifying dynamical factors that influence seizure spread, the results of this study are a step toward development of novel quantitative tools to describe seizure propagation.
Funding:
This work was supported by the University of Michigan SOAR program.