A new algorithm can now help identify the type and occurrence of epilepsy, thanks to researchers from the Indian Institute of Science (IISc) and AIIMS Rishikesh. According to the press release, this new breakthrough is expected to play a vital role in efficient and automated screening and diagnosis of the disease.
Epilepsy is a neurological disorder in which a sudden burst of electrical signals is released from the brain in a short period of time. This can cause seizures, convulsions or even lead to death.
Sign up for your weekly dose of what’s happening in emerging technologies.
The disease is classified as “focal” or “generalized” depending on the origin of the erratic signals produced in the brain. Focal epilepsy occurs when signals are confined to a specific region, while generalized epilepsy is confined to random locations in the brain. Neurophysiologists must manually inspect EEGs (electroencephalograms), which are used to capture erratic signals and determine if a patient has epilepsy.
Source: Detect seizures and interpret EEGs, the direct algorithmic pathway, IISc
Assistant Professor, Department of Electronic Systems Engineering (DESE) and corresponding author Hardik J Pandya says that visual inspection of EEGs could become tedious or exhausting after prolonged periods, and can also lead to errors. “The research aims to differentiate the EEGs of normal subjects from epileptic EEGs. In addition, the developed algorithm tries to identify the types of seizures. Our job is to help neurologists do efficient and rapid automated screening and diagnosis,” the professor said.
According to the researchers, in their study, the team developed a new algorithm that sifts through EEG data and identifies signatures of epilepsy from electrical signal patterns. After the initial training, the algorithm was used to detect whether the subject might have the disease – based on the patterns of their respective scans – with a high degree of accuracy.
Rathin K Joshi, PhD student at DESE, says: “We hope to refine this by testing more data to account for more human EEG variabilities until we reach the point where this becomes fully translational and robust.
Currently, a patent has been filed for the work, along with doctors at AIIMS Rishikesh testing the reliability of the developed algorithm.