AI-powered Detector Improves Epilepsy Diagnosis and Treatment in Children

A new artificial intelligence (AI) tool is showing promising results in detecting subtle brain abnormalities that cause drug-resistant epilepsy in children, perhaps leading to faster diagnoses and more effective surgical interventions.Developed by researchers from The Royal Children’s Hospital, the University of Melbourne, Florey Institute of Neuroscience and Mental Health, Harvard Medical School, and the Austin Hospital, the detector identifies cortical dysplasias – often developing in utero – which are a common cause of epilepsy unresponsive to medication.

Currently,cortical dysplasias are challenging to identify using standard MRI scans,with less than half being detected on a child’s initial scan. the AI tool focuses on identifying these lesions, which can be “impossible for traditional MRI techniques to identify,” according to Dr. emily Macdonald-Laurs. This delay in diagnosis can lead to prolonged seizures and increased risk of learning difficulties,including intellectual disability.

In a recent test group of 17 children, 12 underwent surgery guided by the AI-aided scans, and a remarkable 11 are now seizure-free. Approximately one in 200 children experiences epilepsy, and this new technology offers hope for those with drug-resistant forms linked to cortical dysplasias.

The impact of this technology is powerfully illustrated by the case of Royal, a 5-year-old boy who suffered increasingly frequent seizures, escalating to as many as 19 in a two-hour period. Despite antiseizure medication, his seizures persisted, and a standard MRI failed to pinpoint the cause. Following scans aided by the detector, a cortical dysplasia was located and successfully removed at the royal Children’s Hospital (RCH). His mother, Gurjinder, reports that Royal has gone from multiple seizures daily to being completely seizure-free, regaining his appetite and returning to his normal personality.

“Without the assistance of the detector, it would have taken so much longer to achieve a diagnosis and Royal’s health would have continued to deteriorate,” Gurjinder stated.

Dr. Macdonald-Laurs and her team are now seeking additional funding to expand testing of the detector to paediatric hospitals across Australia. The research, published in Epilepsia (Macdonald‐Laurs, E., et al. (2025). Automated detection of bottom‐of‐sulcus dysplasia on magnetic resonance imaging-positron emission tomography in patients with drug‐resistant focal epilepsy. epilepsia. doi.org/10.1111/epi.18628), represents a meaningful step forward in the diagnosis and treatment of epilepsy in children.

AI Tool MindGlide Revolutionizes Multiple Sclerosis Treatment Assessment

University College London (UCL) researchers have unveiled MindGlide, an artificial intelligence (AI) tool poised to transform the evaluation of multiple sclerosis (MS) treatments. This innovation promises enhanced precision and efficiency in monitoring disease progression and treatment effectiveness.

AI Revolutionizes MS Treatment Monitoring

Image: ©Pornpak Khunatorn iStock

Understanding MindGlide’s Capabilities

MindGlide leverages AI to analyze brain images, specifically MRI scans, detecting subtle changes indicative of MS that were previously challenging to quantify.This capability allows physicians to gain a more comprehensive understanding of MS progression and the efficacy of various therapeutic interventions.

Enhanced Detection: Identifies areas of brain damage, plaques, and early signs of brain shrinkage.
Rapid Analysis: Processes images in just 5 to 10 seconds, significantly reducing analysis time.
Improved accuracy: Outperforms existing AI tools in detecting brain plaques.

The challenge of Multiple Sclerosis

multiple sclerosis is an autoimmune disorder where the immune system mistakenly attacks the brain and spinal cord. This assault leads to a spectrum of physical, mental, and emotional impairments. Early and accurate monitoring is crucial for effective management.

Key Facts About MS in the UK

Approximately 130,000 people live with MS.
The annual cost to the National Health Service (NHS) exceeds £2.9 billion.
Monitoring relies heavily on MRI scans to observe brain changes.

Traditional monitoring methods often require specialized MRI scans to capture the detailed data necessary for assessing MS progression. However,thes specialized scans are not routinely performed,creating a gap in comprehensive patient care.

MindGlide: Bridging the Gap in MRI Analysis

MindGlide addresses the limitations of conventional MRI analysis by employing AI to analyze standard MRI scans commonly used in hospitals. This approach unlocks valuable insights from existing scans, many of which were previously underutilized due to image quality constraints.

The AI model is trained on extensive datasets, enabling it to detect subtle brain changes that traditional methods might overlook. This capability is particularly crucial for identifying early signs of disease progression and assessing treatment effectiveness.

According to a study published in Nature Communications, researchers tested mindglide using over 14,000 images from more than 1,000 MS patients. The results were compelling:

MindGlide outperformed two other AI tools for similar tasks. It was found to be 60% better than one tool,SAMSEG,and 20% better than another,WMH-SynthSeg when it came to detecting brain plaques,a key indicator of disease activity in MS patients.

These plaques, or lesions, are critical indicators of MS progression and play a vital role in evaluating treatment effectiveness.

Unlocking Insights from Existing Brain Scans

MindGlide’s ability to identify lesions on the outer layers of the brain and detect changes in deeper brain areas provides a more comprehensive view of the disease. The tool’s consistent performance across short-term and long-term scans makes it invaluable for both initial assessments and ongoing monitoring.

Furthermore, the technology has validated previous high-quality research regarding the effectiveness of various MS treatments, demonstrating its potential to refine and personalize treatment plans.

The potential to unlock vital insights from a large number of existing brain scans is a key breakthrough. Many of these images were previously underutilized because they were not high-quality enough for expert analysis.

With MindGlide, these images can now be analyzed effectively, allowing doctors and researchers to understand better MS and how different treatments impact the disease. This coudl lead to faster and more accurate diagnoses and better treatment plans personalized to individual patients.

Future Directions and Expansion

While MindGlide represents a significant advancement in MS research and treatment monitoring, its current implementation is limited to brain scans. Spinal cord imaging, which is also critical for assessing MS, is not yet included in the tool’s capabilities.

researchers plan to expand its scope in future studies to offer a more comprehensive assessment of the entire neural system, including the spinal cord. This expansion will further enhance MindGlide’s utility in diagnosing and managing MS.

Medical Imaging

AI Detects Tiny Brain Lesions for Faster Epilepsy Treatment in Children