Abstracts

RATIONALE: Spike-triggered functional MRI (fMRI) has been used to study interictal epileptic discharges (IED) in patients with epilepsy. Recently, continuous fMRI acquisition with simultaneous EEG recording and off-line EEG processing has become available. This method is easier to implement and provides a stronger statistical analysis. We are presenting the results on a large group of patients with focal epilepsy investigated using this new technique.
METHODS: The only inclusion criterion was the presence of focal intractable epilepsy and frequent IEDs. EEG was recorded simultaneously during BOLD-fMRI acquisition. Each session lasted about 80 minutes. The IEDs were identified off-line after removal of the artifact from fMRI scanning. IED timing was used to perform statistical analysis of the fMRI data. For those patients with an activated area, the volume of activation was calculated. The study population was divided into two groups according to the presence or absence of fMRI activation. The two groups were compared in terms of their clinical and electrophysiological variables.
RESULTS: A total of 32 studies were performed on 28 patients. Nine studies were not included in the analysis: seven in which the patients had no IEDs during the scanning period and two in which some data were lost. Activation was obtained in 12 of the remaining 23 studies (52%). The average volume of activation was 5.1 cm3 (SD 7 cm3). In 11 of these 12 studies there was at least one region of activation concordant with the EEG (same lobe). In 2 studies, there were also regions of activation at a distance from the EEG focus. In one study, the activation was contralateral to the EEG focus. Patients who had bursts of IEDs were more likely to have fMRI activation than patients with isolated spikes (p[lt]0.05). The number of spikes per study did not influence the presence of activation, nor did the existence of a lesion. A set of other electrophysiological and clinical variables was also examined, and none was correlated to the presence of activation.
CONCLUSIONS: Continuous acquisition of fMRI with simultaneous EEG recording is feasible and yields regions of activation that are concordant with EEG activity. However, activation was only found in about 50% of studies and no clear explanation could be found for this result, with the exception of presence or absence of bursts of IEDs. The significance of regions of activation outside the expected EEG focus remains to be investigated. This technique will require further validation and some technical improvements prior to its routine use in pre-surgical evaluation of epileptic patients.
[Supported by: Canadian Institutes of Health Research under grand MOP-38079]