Computational modeling of seizure reconstruction and attenuation via model reference control.

Epilepsy is one of the most common and widely studied brain disorders in the world, but its precise triggering mechanisms are not yet fully understood. This has led to the development of mathematical models capable of describing seizures in order to study their generation, propagation, and control. Data-driven models and state-feedback controllers have been used together to artificially reconstruct and mitigate seizures, providing information that may help physicians better define the shape, frequency, and intensity of the input stimulus required for seizure attenuation. However, there are several constraints to their applicability, such as reliance on stationarity-dependent techniques, restriction to local field potentials (LFPs), and the assumption of hybrid controllers. The main goal of this article is to overcome these obstacles by adopting a new approach consisting of orthogonal functions, model reference controllers (MRCs), and adaptive gains or fixed gains, applied to electroencephalograms (EEG) and intracranial EEGs (iEEG). The results demonstrate the possibility of reconstructing seizures and achieving attenuation for both types of signals artificially, also parametrizing the level of expected attenuation based on the magnitude of input and adherence of the model to the real signature. Moreover, the concept of hybrid controllers can be further formalized, demonstrating both practical appeal and consistency with previous studies in the literature on real electrical stimulation therapies. At last, the parameters of both controllers can be tuned so that the magnitude of the stimulation input remains within 5 to 18 V, close to reference voltage levels (up to 14V) from the literature.