New AI Algorithm Detects Rare Epileptic Seizures from EEG Data

By HospiMedica International staff writers
Posted on 07 Jun 2024

Image: The new AI system accurately detects epileptic seizure types (Photo courtesy of 123RF)

Over 65 million people around the globe are affected by epilepsy, a neurological disorder that impacts the nervous system and causes seizures. Statistically, one in 26 individuals will experience epilepsy during their lifetime, and each year, 1 out of 1000 people with epilepsy die from unexpected deaths. Early detection is crucial for effective epilepsy treatment. Machine learning techniques have been employed to detect and classify seizures from electroencephalography (EEG) signals, which are captured using electrodes on the brain, identifying patterns too complex for human analysis alone. However, these systems have faced challenges in detecting rare forms of epileptic seizures due to their reliance on large data sets to learn patterns and make predictions, resulting in inadequate performance when encountering less common seizures. Researchers have now developed an advanced AI system capable of accurately detecting various types of epileptic seizures, thereby enhancing the diagnosis of rare and complex cases, even in young children.

The AI system, created by computer science researchers at the University of Southern California (Los Angeles, CA, USA), enhances the diagnosis of rare and complex epilepsy cases by analyzing brain interactions. This new system integrates multiple sources of information typically overlooked by AI systems in epilepsy detection, such as the positions of EEG electrodes and the brain regions they monitor. By doing so, the AI can identify patterns or features that signal an impending seizure. This approach enables the system to produce accurate results with minimal data, even for rare seizure types that have limited examples in the training data.

For example, in the case of atonic seizures—a rare type of seizure often affecting children and causing sudden loss of muscle control and collapse—the system focuses on spatial relationships in brain regions. It prioritizes brain areas involved in muscle control, such as the motor cortex, basal ganglia, cerebellum, and brainstem, to detect activity patterns indicative of atonic seizures. The researchers aim to supplement doctors' expertise in diagnosing difficult cases rather than replace them. They view this AI technology as a significant advancement in clinical neurology, with the potential to be integrated into wearable sensors that can relay information to a smartphone in the future.

“Brain seizures happen very suddenly, and so detecting seizures earlier really could save lives. The system could prompt an alert if it detects any irregularities in the brain waves. This would open up incredible opportunities for diagnosis and treatment of epilepsy,” said Cyrus Shahabi, a computer science, electrical engineering, and spatial sciences professor.

Related Links:
University of Southern California