Neuropediatrics 2004; 35 - V4
DOI: 10.1055/s-2004-819349

Source analysis of ictal EEG patterns in focal epilepsy

T Bast 1, G Kammerer 1, Ö Özkan 1, G Ramantani 1, M Scherg 2, S Rona 3, D Rating 1
  • 1Department of Pediatric Neurology, University Hospital of Heidelberg
  • 2Department of Neurology, University Hospital of Heidelberg
  • 3Department of Neurosurgery, University Hospital of Freiburg, Germany

Objective: To determine the significance of non-invasive source analysis of ictal EEG patterns.

Methods: 65 seizure patterns recorded by surface EEG (35–53 channels, 256Hz) in 16 patients (0.7–20 years) were included in the analysis. 8 patients had dysplastic lesions (33 seizures) and 8 non-dysplastic lesions (5 tumor, 3 gliosis, 32 seizures). 14 patients underwent epilepsy surgery (11 seizure free). The earliest seizure pattern in preoperative Video-EEG was identified visually, the dominant frequency was determined by FFT and identical sinus periods were averaged (r>0,85 for 3–5 channels). A 4-shell spherical head model was used in the BESA™ source analysis. At first, one regional source was modeled (RS1). A regional source provides information on localization of activities independent from current orientation. An application of a two-independent-regional source model (RS2) followed. Anatomical relations were analyzed by 3D-T1-MRI superposition (BrainVoyager®).

Results: Source localization corresponded to the lesional hemisphere in 62 (RS1), i.e. 57 seizures (RS2). Localization within the lesional lobe was found in 77% (RS1), i.e. 82% (RS2) with a tendency to higher concordance in dysplastic lesions. The MRI-superposition showed a significantly higher number of source localization within or near (<2cm) the lesion for seizures by dysplastic compared to non-dysplastic lesion: RS1: 22/33 vs. 9/32 and RS2: 29/33 vs. 17/32 (p<0,001).

Conclusion: Source analysis of ictal EEG patterns is a promising method, though still inadequately valid for the clinical routine. An identification of the lesional lobe was possible in just 70–82%. In dysplasia a higher rate of source localization within or near by the lesion was observed, which corresponds well with an intrinsic epileptogenicity in this type of lesion.

Keywords: epilepsy; EEG; dysplasia