• Eliane Kobayashi
  • Y. Aghakhani
  • Andrew Bagshaw
  • Anna Jansen
  • Cg Bear
  • C. Hawco
  • Frederick Andermann
  • E Andermann
  • Jean Gotman
  • François Dubeau
  • American Academy Of Neurology (Editor)
BACKGROUND: PMG is a widespread cortical malformation, but patients often have partial seizures and focal spikes. EEG-fMRI may allow better localization of the generator of the epileptic activity.

DESIGN/METHODS: Six PMG patients (3 men, mean age 35 years) underwent a two-hour continuous EEG-fMRI study (1.5 T Siemens scanner). EEG data were acquired using 21 MRI compatible electrodes and EMR32 amplifier. We used FEMR software to filter and mark the EEG spikes, and fMRI statistical analysis was performed using the timing of the spikes as events. We used a standard haemodynamic response function (HRF) which peaks at 5.4 seincs, plus four other HRFs with a time to peak of 3, 5, 7 and 9 seincs. For each spike localization we determined the presence of activation defined as five or more contiguous voxels with t>3.0 or one single voxel with a t>4.74 (corrected p value of 0.01). The same absolute parameters but with negative t values were considered to define significant deactivation. Anatomic localization for each type of response was determined by co-registration using in house software.

RESULTS: Four patients had perisylvian (3 bilateral, 1 unilateral), one bilateral frontal, and one bilateral parietooccipital PMG. Spikes during scan ranged from 8 to 204 events, and only one patient had no fMRI response. Two patients showed both activations and deactivations, and three only activations. fMRI responses were within the boundaries of the malformation, except in a patient with unilateral PMG. Maximum activation was always concordant with epileptic activity. In three patients with unilateral focus, activation was ipsilateral, more widespread in one, and more posteriorly in the other two. Two patients (one with uni- and one with bilateral perisylvian PMG) with diffuse and bilateral epileptic activity, fMRI activation was also bilateral and maximum concordant with the predominant scalp EEG activity. Deactivation was only found in these two patients and was contralateral to the PMG in one and bilateral in the patient with bilateral PMG, but mostly outside the lesion.

CONCLUSIONS: This study demonstrated for the first time that fMRI can measure metabolic or blood flow changes related to epileptic activity generated by polymicrogyric cortex. In our PMG patients the localization and extension of the activation correlated well with the source generator of spikes. The deactivation tended to occur in areas outside the lesion. EEG-fMRI appears to be a new and powerful non-invasive method to study the relationship between malformed cortex and epileptiform activity.
Original languageEnglish
Pages (from-to)177-178
Number of pages2
Publication statusPublished - 2004

    Research areas

  • Epilepsy, Polymicrogyric cortex, EEG-fMRI

ID: 2338536