Évaluation préchirurgicale pour l'épilepsie pédiatrique: localisation de la zone épileptogène par magnétoencéphalographie de l'activité cérébrale spontanée

 

Jeremy Moreau

CUSM-Hôpital neurologique de Montréal

 

Domaine : Neurosciences, santé mentale et toxicomanies

Programme : Formation de doctorat

Concours 2016-2017

Partenaire:

Fondation des Étoiles

Epilepsy affects over 10 million children worldwide, and about a third of those requiring therapy will fail to have their seizures controlled by medication. Brain surgery to remove tissue suspected to generate seizures can help many of these patients. Early treatment in children is crucial as it is associated with better cognitive and developmental outcomes. Many medical imaging technologies are used to locate the affected brain regions, but a relatively new technique called magnetoencephalography (MEG) shows particular promise. MEG detects the very weak magnetic fields generated by brain cells, and can locate epileptic brain tissue more accurately than other techniques in many patients.

However, one problem with MEG is that it relies on recording specific patterns of epileptic activity that don't happen very often. Keeping patients—especially children—in the MEG scanner for a long time in order to record these types of activity is impractical and expensive. Were it possible to locate the epileptic brain regions using the background brain activity present in all MEG scans, then this problem would be solved. Early trials in five adult epilepsy patients suggest that this may be possible. Nonetheless, more work is needed to see how this method compares to standard clinical MEG and whether it works for all the different kinds of epilepsy.

Moreover, some new methods based on the same basic approach could locate the epileptic brain regions more accurately. In our study, we will use MEG in the presurgical evaluation of epileptic children at the Montreal Children's Hospital. We will compare standard clinical MEG, the previously reported method, and new methods based on statistical comparisons to a large database of MEG recordings. We will investigate how all these methods compare across the different types of childhood epilepsies, and see which one predicts the best post-surgical outcome.