Two recently published studies show just how important the use of computer technology and modelling have become in the study of epilepsy.
A study being carried out at Newcastle University is using a brain model to explore the cause of different epileptic seizure onset patterns. According to the study, at the onset of an epileptic seizure, differing characteristics of brain tissue surrounding the seizure’s origin site may determine which of two main patterns of brain activity will be seen. Electrical activity in the brain at the start of an epileptic seizure typically follows either a “low amplitude fast” pattern or a “high amplitude slow” pattern. Patients whose seizures follow the high amplitude slow pattern have a higher risk of continuing seizures after surgical treatment. However, the mechanisms underlying these different patterns are unclear.
To better understand the onset patterns, Yujiang Wang of Newcastle University, UK, and colleagues used a previously developed computer model that can simulate brain activity at the start of a seizure. The model output suggested that the onset pattern of a seizure may be determined not by brain tissue at the site where the seizure originates, but by characteristics of the surrounding “healthy” brain tissue.
The simulation showed that the high amplitude slow pattern occurs when surrounding brain tissue has higher excitability; that is, the brain cells have a stronger response to stimulation and can react immediately to the initiation of a seizure. Meanwhile, the low amplitude fast pattern is associated with tissue of lower excitability, which is only slowly penetrated by seizure activity.
These findings suggest why the different onset patterns are associated with different treatment outcomes. Surgical removal of seizure-triggering brain tissue may be enough to prevent seizure activity in nearby low-excitability tissue. However, high-excitability tissue may still be stimulated by alternative trigger sites after surgery, providing a possible explanation for the worse outcomes experienced by patients whose seizures follow the high amplitude slow pattern.
Next, the researchers plan to study seizure onset patterns in greater detail. “We hope to contribute towards the overall goal of associating patterns seen in seizures with an understanding of the underlying mechanism,” Wang says. “This would not only help our understanding of seizures in general, but may be useful for patient stratification in terms of treatment options.”
In a second study being carried out at Boston Children’s Hospital could enable more patients with epilepsy to benefit from surgery when medications do not help. The approach streamlines the seizure monitoring process required for surgical planning, making surgery a more feasible and less risky option for patients.
Currently, for some patients, pinpointing the diseased brain areas where their seizures originate requires invasive surgery to place grids of electrodes on the brain’s surface. This is followed by long-term electroencephalography (EEG) monitoring – typically for a week – while doctors wait for a seizure to happen. Then, patients must undergo a second brain operation to remove the diseased tissue.
The new technology, developed by Joseph Madsen, MD, Director of Epilepsy Surgery at Boston Children’s Hospital, and Eun-Hyoung Park, PhD, a computational biophysicist in the Department of Neurosurgery, could allow patients to be monitored in one short session, without the need to observe an actual seizure. Patients could then proceed directly to surgery, avoiding a second operation.
Effective use of this technology could cut the cost and risk by more than half by reducing the current two-stage procedure to one-stage, the researchers say. “We know that the diseased brain network responsible for the seizures is there all along,” says Madsen. “So rather than wait for the patient to have a seizure, we set out to find patterns of interaction between various points in the brain that might predict where seizures would eventually start.”
To identify the brain areas causing the seizures, Madsen and Park applied a special algorithm to analyze patients’ interictal EEG data – data captured between their seizures. They randomly selected 25 patients with hard-to-treat epilepsy who previously had long-term EEG monitoring at Boston Children’s, and analyzed data from the first 20 seizure-free minutes of the patients’ EEGs.
Their algorithm, known as Granger causality analysis, is based on a statistical approach developed Sir Clive Granger (for which he won the Nobel Prize in Economics in 2003). Madsen and Park adapted the Granger method, originally used for economic forecasting, to calculate the probability that activity at one brain location predicts subsequent activity at other brain locations strongly enough to be considered causative. Their analysis generated a map of the causal relations in each patient’s epileptogenic network, which Park and Madsen superimposed over images of the brain.
They then showed that the brain regions predicted to be causing seizures strongly correlated with actual causative regions on seizure EEGs – as read by ten board-certified epileptologists, usually many days later.
Madsen and Park have shown that their calculations can be done quickly enough to allow data obtained in the operating room to potentially influence surgical decision-making. They now are investigating how the Granger causality method can best augment readings of EEGs by trained neurophysiologists. “We still need to validate and refine our approach before it can be used clinically,” notes Madsen. “But we are hopeful that these advanced computer applications can help us treat more children with epilepsy – with less risk and lower cost.”
Article: Granger Causality Analysis of Interictal iEEG Predicts Seizure Focus and Ultimate Resection, Eun-Hyoung Park, PhD Joseph R. Madsen, MD, Neurosurgery, doi: 10.1093/neuros/nyx195, published 2 May 2017.
Article: Mechanisms underlying different onset patterns of focal seizures, Wang Y, Trevelyan AJ, Valentin A, Alarcon G, Taylor PN, Kaiser M, PLOS Computational Biology, doi: 10.1371/journal.pcbi.1005475, published 4 May 2017.