An algorithm that can help decode brain scans to identify the occurrence and type of epilepsy

An algorithm that can help decode brain scans to identify the occurrence and type of epilepsy

To develop and train the algorithm, the researchers examined EEG data from 88 human subjects. The team then analyzed the data and classified the different wave patterns into sharp signals, spikes and slow waves

Researchers at the Indian Institute of Science (IISc) in collaboration with AIIMS Rishikesh have developed an algorithm that can help decode brain scans and identify the incidence and type of epilepsy.

Epilepsy is a neurological disorder in which the brain sends out sudden flashes of electrical signals over a short period of time, resulting in seizures, fits, and in extreme cases, death.

Based on where the irregular brain signals originate, epilepsy is classified as focal or generalized epilepsy. Focal epilepsy occurs when erratic signals are confined to a certain area in the brain. If the signals are in random places, then it is called generalized epilepsy.

EEG check

In order to determine whether a patient is epileptic, neurophysiologists need to manually review EEGs (electroencephalograms), which can pick up such erratic signals, an IISc press release explains.

Visual inspection of EEG can be tiring after a long period of time and can occasionally lead to errors, said Hardik J. Pandya, assistant professor in the Department of Electronic Systems Engineering (DESE) and corresponding author of the study published in Biomedical Signal Processing and Control.

The aim of the research is to distinguish the EEG of normal individuals from the epileptic EEG, and in addition, the developed algorithm attempts to identify seizure types. “Our work is to help neurologists perform efficient and rapid automated screening and diagnosis,” he added.

Algorithm training

To develop and train the algorithm, the researchers first examined EEG data from 88 human subjects obtained at AIIMS Rishikesh. Each subject underwent a 45-minute EEG test, divided into two parts: an initial 10-minute test while the subject was awake, which included photic stimulation and hyperventilation, followed by a 35-minute sleep period in which the subject was asked to sleep.

Next, the team analyzed this data and classified the different wave patterns into sharp signals, spikes, and slow waves. Spikes are patterns where the signal rises and falls over a very short period of time (~70 milliseconds), while sharp waves are those whose rises and falls are spread out over a slightly longer period of time (~250 milliseconds) and slow waves have a much longer duration. (~400 milliseconds), the report said.

An epileptic subject shows a different set of patterns compared to a healthy individual. Researchers have developed an algorithm to calculate the total number of spikes – Cumulative spike count – and use it as a parameter to determine whether a subject is epileptic or not (a higher value means a greater chance that the subject is epileptic).

Spikes and Curves

The algorithm also calculates the sum of the areas under spikes and sharp curves to distinguish between focal and generalized epilepsy (a higher value indicates generalized epilepsy, as opposed to focal epilepsy, which has a lower value), the report explains.

The researchers add that the study shows a way to identify absence seizures (those involving brief, sudden lapses of consciousness) using a cumulative Spike-Wave Count; in some cases these absence seizures are critical and can be fatal.

The team then ran their algorithm on a new set of EEG data from subjects whose classification (whether they had epilepsy and, if so, what type of epilepsy they had) was already known to doctors. This blind validation study successfully classified subjects accurately nearly 91% of the time.

The work is currently patent pending and the algorithm is being tested for reliability by doctors at AIIMS Rishikesh.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top
%d bloggers like this: