The All India Institute of Medical Sciences (AIIMS), Rishikesh, and the Indian Institute of Science (IISc) have jointly created an algorithm that, according to the researchers, can help decode brain scans to determine the presence and kind of epilepsy. A patent application has already been filed for the work, and doctors at AIIMS Rishikesh are evaluating the algorithm’s dependability, according to Bengaluru-based IISc. A neurological condition known as epilepsy causes the brain to suddenly and rapidly release large amounts of electrical signals, which can lead to seizures, fits, or in severe cases, even death. Focal epilepsy and generalized epilepsy are divided into two categories, depending on where in the brain the irregular impulses originate.
When the irregular impulses are limited to a single area of the brain, it is called focal epilepsy. It is referred to as generalised epilepsy if the signals appear randomly. Neurophysiologists must manually examine EEGs (electroencephalograms), which might record such irregular signals, to determine whether a patient is epileptic, according to a statement released by the IISc on Wednesday. According to Hardik J. Pandya, Assistant Professor at the Department of Electronic Systems Engineering (DESE) and the corresponding author of the study published in “Biomedical Signal Processing and Control,” visual inspection of EEG can become exhausting after prolonged periods of use and may occasionally result in errors. “The purpose of the study is to distinguish between normal and epileptic EEGs. The created algorithm also makes an effort to categorise the different kinds of seizures.
Our task is to aid the automated screening and diagnosis made by neurologists, he continues. According to the announcement, the team’s study presents a revolutionary algorithm that can sort through EEG data and locate epilepsy markers in the electrical signal patterns. According to the researchers, after initial training, the algorithm was highly accurate at determining whether a human individual would have epilepsy or not based on these patterns in their respective analyses.
The researchers first looked at EEG data from 88 human participants collected at AIIMS Rishikesh in order to create and train the system. Each subject performed a 45-minute EEG test that was split into two sections: a 10-minute awake test that included photic stimulation and hyperventilation, and a 35-minute sleep session where the patient was instructed to fall asleep. As a result of their analysis, the team was able to group various wave patterns into sluggish waves, sharp signals, and spikes. When compared to a healthy person, an epileptic subject exhibits a different collection of patterns.
On a fresh batch of EEG data from individuals whose classification (if they had epilepsy and, if so, what sort of epilepsy they had) was already known to the doctors, the team ran its algorithm. According to the statement, the participants were correctly identified in this blind validation research in roughly 91% of the cases. Rathin K Joshi, a PhD candidate in DESE and the study’s first author, states, “We intend to enhance this further by testing on more data to consider more variabilities of human EEGs until we reach a point where this becomes entirely translational and resilient.”
