Artificial Intelligence and Computational Approaches for Epilepsy

Abstract

Studies on treatment of epilepsy have been actively conducted in multiple avenues, but there are limitations in improving its efficacy due to between-subject variability in which treatment outcomes vary from patient to patient. Accordingly, there is a growing interest in precision medicine that provides accurate diagnosis for seizure types and optimal treatment for an individual epilepsy patient. Among these approaches, computational studies making this feasible are rapidly progressing in particular and have been widely applied in epilepsy. These computational studies are being conducted in two main streams: 1) artificial intelligence-based studies implementing computational machines with specific functions, such as automatic diagnosis and prognosis prediction for an individual patient, using machine learning techniques based on large amounts of data obtained from multiple patients and 2) patient-specific modeling-based studies implementing biophysical in-silico platforms to understand pathological mechanisms and derive the optimal treatment for each patient by reproducing the brain network dynamics of the particular patient per se based on individual patient's data. These computational approaches are important as it can integrate multiple types of data acquired from patients and analysis results into a single platform. If these kinds of methods are efficiently operated, it would suggest a novel paradigm for precision medicine.

Keywords: Artificial intelligence; Epilepsy; Patient-specific modeling; Precision medicine; Seizures.

Figure 1

Overview of computational approaches for personalized medicine. Using accumulated large amount of medical data and advanced in-silico techniques, the personalized medical platform could be built to provide accurate diagnosis, prognosis prediction and treatment optimization for individual epilepsy patients. Currently, these computational studies are mainly divided into two approaches: a machine learning approach that implements a model capable of performing a specific function, such as automatic diagnosis, and a biophysical modeling approach that reproduces and simulates the patient’s brain network dynamics system itself.

Figure 2

Machine learning-based computational approach. Many recent studies have focused on implementing the computational predictive models to localize SOZs or judge epileptic brain states, such as pre-ictal and ictal onset, by employing traditional machine learning algorithms or deep learning algorithms based on scalp EEG and/or iEEG data recorded from epilepsy patients. Compared to the traditional machine learning approach, which consists of two step processes of manually extracting the features of the data and training the machine by applying the features as inputs, deep learning approach automatically figures out the discriminative features from data and learns them. SOZ, seizure onset zones; EEG, electroencephalogram; iEEG, intracranial EEG.

Figure 3

Biophysical modeling-based computational approach. Recent studies based on personalized brain network modeling approach have demonstrated the its feasibility, in which the model can reproduce seizure propagation characteristics of each patient and suggest optimal intervention method for each patient via systematic simulations in the patient-specific environments.