MR Imaging of Brain Generators in Human Epilepsy

New Brain Scans for Epilepsy. Epilepsy is the commonest serious neurological condition in the UK. Although medication can help, it is ineffective in 80,000 people. We propose to use a new kind of brain scan to find the brain area that is responsible for the seizures (epileptic focus) and that could be surgically removed.

The new brain scan technique, called EEG-fMRI, combines the recording of brain waves (electroencephalogram or EEG) with functional magnetic resonance imaging (fMRI). EEG-fMRI has already given promising results but we do not know how useful it is in localising the focus. We will first refine the technique to make it as sensitive and specific as possible and then evaluate its performance by comparing it with the best available data. We propose to develop EEG-fMRI on the most advanced scanners to maximise its sensitivity. We also want to compare the EEG-fMRI measurements with clinical data that is collected by placing electrodes inside the patient’s skull – a complicated and relatively risky procedure, but which, at present, gives the best possible localization of the epileptic focus. We hope to show that EEG-fMRI, a non-invasive and totally painless technique, can give the same kind of information as the invasive technique.

Technical Summary

Thirty percent of epilepsy patients are intractable to drug treatment, representing approximately one hundred thousand people in the UK. Many patients with intractable epilepsy are currently deemed unsuitable for curative surgery due to the difficulty in identifying the focus. Invasive monitoring is currently the gold standard but disadvantages of this approach include the risk of complication and costs. EEG-correlated fMRI (?EEG-fMRI?) is a non-invasive technique unique in its ability to provide activation maps of the hemodynamic (blood oxygen level-dependent or BOLD) correlates of specific epileptiform discharges. Despite limited sensitivity, concordance with non-invasive electroclinical findings on a lobar basis is frequent. Up to now, validation has relied on comparison with electroclinical findings, which is unsatisfactory as it does not allow to assess the technique?s true added clinical value. The demonstration that fMRI can enhance or even replace invasive monitoring would be a great advance and could significantly increase the proportion of patients in whom surgery is an option. This will require validation based on invasive EEG monitoring and postsurgical outcome findings in a representative group of patients.

Significant improvements in sensitivity will be obtained by implementing EEG-fMRI at high field (4.7T). Furthermore, we will investigate a novel fMRI contrast mechanism that directly reflects neuronal activity. To address scalp EEG?s intrinsic limitations, and in particular its low sensitivity and specificity, we will apply data-driven fMRI analysis techniques to reveal spatio-temporal patterns reflecting epileptogenicity. These developments will greatly benefit from the availability of invasive monitoring data.

fMRI activation patterns will be compared the epileptogenic substrate identified invasively. To further investigate the relationship between epileptogenicity and fMRI, discharges recorded within the brain and fMRI signals will be recorded simultaneously, requiring implementation of intracranial EEG-fMRI. In addition to the obvious interest in improving our understanding of the imaging correlates of epileptogenicity and epileptogenic networks, this presents us with a unique opportunity to improve multimodal forward models of EEG and fMRI and our understanding of fMRI in humans. Furthermore the role of EEG-fMRI in guiding invasive monitoring will be investigated.

This programme of research will result in a major improvement in the fMRI?s sensitivity to epileptiform activity, and provide vital information on EEG-fMRI?s clinical potential. It will also represent a major leap forward in our understanding of the relationship between neuronal activity and fMRI on the other.