NU Sci Magazine

Predicting the unpredictable: The next frontier in managing epilepsy

November 12, 2025

By

Iba Baig

HealthNeuroscienceIssue 65

Imagine the brain occasionally sending out a massive, overwhelming electrical storm. For approximately 50 million people living with epilepsy worldwide, this is an unfortunate reality. Epilepsy is a neurological disorder in which a seizure can strike at any time in any place, with triggers including flashing lights, loud noises, bright sunlight, and even humidity. Although the constant uncertainty is debilitating, scientists are coming up with ways to predict these storms. If there were a silent signal, a biological message from the brain, that could give a five-minute or five-day warning, those with epilepsy would be able to take back control in their lives.

Predicting seizures using biomarkers could revolutionize the management of epilepsy by providing techniques that don’t just rely on unpredictable safeguards like medication or invasive surgeries.

Predicting seizures using biomarkers could revolutionize the management of epilepsy by providing techniques that don’t just rely on unpredictable safeguards like medication or invasive surgeries.

These signals are known as biomarkers: quantifiable biological footprints indicating health. For example, a fever is a biomarker for an infection, and high LDL cholesterol is a biomarker for heart disease risk. For epilepsy, researchers are searching for predictive biomarkers beyond diagnosis or treatment. Currently, doctors use EEGs on patients to find hidden patterns of brain signals that act like a countdown to a seizure. Outside of the hospital, scientists from Zealand University Hospital are developing wearable seizure detection devices, like a wearable EEG. Using wearable, offline multi-modal seizure detection systems, scientists can collect brain signals that can provide long-term detection of focal (brain area-specific) seizures. Then, by using a machine learning predictive algorithm, it can alert the wearer of an oncoming seizure, allowing them to take measures such as seizure rescue medications.

If there were a silent signal, a biological message from the brain, that could give a five-minute or five-day warning, those with epilepsy would be able to take back control in their lives

If there were a silent signal, a biological message from the brain, that could give a five-minute or five-day warning, those with epilepsy would be able to take back control in their lives

Predicting seizures using biomarkers could revolutionize the management of epilepsy by providing techniques that don’t just rely on unpredictable safeguards like medication or invasive surgeries. In addition to technological innovations, researchers at Universiti Malaya identified metabolic biomarkers and pathways linked to drug-resistant epilepsy. A group of scientists identified several metabolites, including amino acids (glycine, glutamate, isoleucine), organic acids (lactate), and glucose. This suggests a future where a simple blood test detecting key biomarkers could provide a faster and more objective answer.

Currently, epilepsy diagnosis is a slow and lengthy process, often relying on witness accounts of seizures and multiple hospital stays. Even then, proper treatment is not guaranteed, with many patients developing intolerance to medications over time. This calls for the development of a reliable warning system, allowing someone to safely pause and move before the onset of a seizure. Many who suffer from unexpected and sudden seizures endure physical injuries, fractures, traumatic brain injuries, and minor trauma that can lead to long-term complications of health and epilepsy symptoms . Biomarker development would alleviate these challenges.The journey to decode the brain and body’s silent signals is a technical challenge to restore autonomy. Biomarker research to create predictive power of wearable devices and metabolic testing has the potential to shift reactive care to proactive prediction. After all, the best treatment lies in prevention.

Sources

Built and maintained by the NU Sci web team. 2026.