|MONTREAL NEUROLOGICAL INSTITUTE||MEMBERS|
RESEARCH (click for a topic for more information)
The laboratory is part of the Epilepsy group at the Montreal Neurological Institute. It is involved in the study of electrophysiological manifestations of human epilepsy and other neurological disorders. Research in the lab has focused mainly on EEG analysis in epilepsy, including long-term monitoring, automatic event detection, dipolar source localization, the use of ICA to examine epileptic discharges, and analysis of high frequency EEG components (HFOs) in epilepsy. Recent research has also focused on combining EEG and fMRI to study epilepsy.
Epileptic activity is
visible in the EEG in the form of prolonged discharges accompanying clinical
seizures and interictal spikes, short events without clinical accompaniments.
In epileptic patients, and particularly in those suffering from medically intractable
seizures, it is important to record such events and to find which part of the
brain generates them. This information helps classify the type of epilepsy
and therefore administer optimal medical treatment. When medical treatment
fails, it is possible to consider the surgical removal of epileptic brain
tissue, provided this tissue can be accurately localized and provided it is
not critical for normal brain function.
The laboratory is mostly supported by grants from the Canadian Institutes of Health Research.
Source Analysis: The EEG is normally recorded from the scalp surface. Methods have been developed to find the likely intracerebral source of the distribution of potential on the scalp. We have been working on establishing the domain of validity of these methods by comparing their predictions to actual intracerebral recordings. We are also working on developing new methods of obtaining intracerebral sources using innovative approaches.
Combined EEG-fMRI recordings: We have extensively used combined EEG-fMRI methods to examine interictal events. EEG is recorded in the MRI scanner during the acquisition of functional scans. The EEG allows us to determine the timing of epileptic events and to examine the fMRI activation at that time. Recording EEG during fMRI scanning is very difficult, but we have worked out a meticulous setup to optimize EEG quality.
Multi-Modal Analysis of Interictal Events: One of the advantages of the many ways to examine epileptic events is the ability to compare these methods to each other to test for reliability and to get a better understanding of the data. We compare data from EEG, fMRI, source localization/dipole modeling, SEEG, SPECT, and other methods.
It has recently become feasible to make intracranial EEG (SEEG) recordings at a
very high sampling rate (2000Hz). We have been using this method to examine the relationship
between high frequency oscillations and the epileptogenic regions.
2000Hz SEEG: It has recently become feasible to make intracranial EEG (SEEG) recordings at a very high sampling rate (2000Hz). We have been using this method to examine the relationship between high frequency oscillations and the epileptogenic regions.
Seizure Detection and Analysis: Since epileptic seizures and epileptic spikes occur unpredictably and sometimes rarely, it is often necessary to record the patientís EEG during several days or even weeks in order to fully document the patientís epileptic abnormalities. This is a long and tedious task, which can be greatly facilitated by the automatic detection of spikes and seizures. We started many years ago to detect spikes and seizures with relatively simple pattern recognition methods. These methods usually detect a large fraction of the events, but tend to make a large number of false detections as well. We have recently improved the specificity and the sensitivity of these methods by using different signal processing approaches (i.e. wavelets) and by incorporating spatial and temporal context into the analysis.
We have also worked on removing artefacts from the EEG, particularly during epileptic seizures. An automatic method was developed, based on Independent Component Analysis (ICA) and automatic component classification.
The laboratory has been
involved in a variety of studies, mostly related to epilepsy. Examples include:
Site last updated: November 2008