Neuropediatrics 2005; 36 - V10
DOI: 10.1055/s-2005-867969

Comparison of interictal EEG and MEG I: MEG detects a higher ratio of focal interictal epileptic discharges in sleep

G Ramantani 1, R Boor 2, R Paetau 3, D Rating 1, T Bast 1
  • 1Universitätskinderklinik Heidelberg, Kinderneurologie, Heidelberg
  • 2Universitätskinderklinik Kiel, Kinderneurologie, Raisdorf
  • 3LTL. Helsinki University of Technology, Brain Research Unit, Helsinki

Introduction: This study aims to explore the effect of background activity during physiologic sleep, on visual MEG and EEG spike detection in simultaneous interictal sleep recordings. Because of MEG being indifferent to radial activity, i.e. sleep changes, a higher ratio of epileptic discharges (ED) unique to MEG compared to EEG was presumed in the case of overlapping sleep activity.

Patients and methods: We examined 14 children and adolescents aged 3.5–17 years with localization-related epilepsy. 122-channel whole-head MEGs and 33 channel EEGs were recorded simultaneously for 20–40min. Segments of artefacts or bilateral polyspikes (>2 ED/200 ms) were excluded from the evaluation. The EEG and MEG data was split into 28 data segments, which were then blinded and independently reviewed by 4 experienced neurophysiologists for the presence of ED. EEG and MEG data was matched and ED detected by ≥3 investigators were identified as unequivocal spikes and classified in three categories: 1. MEG>EEG (difference ≥3 raters), 2. EEG>MEG (ditto) and 3. EEG=MEG (≥3 raters each). The presence of simultaneous sleep changes (spindles, vertex waves, K-complexes, slow wave sleep) was visually determined for every single ED-segment (±500 ms).

Results: Out of 4704 detected ED, 1387 spikes fulfilled the above criteria; 689 spikes unique to MEG (54% within EEG sleep changes) and 136 spikes unique to EEG (21% sleep). 562 Spikes were identified by ≥3 raters in both EEG and MEG (42% sleep). A significantly higher number of unique MEG spikes were associated with vertex waves or other sleep changes (including spindles, K-complexes and delta sleep) compared to both EEG unique spikes and unequivocal EEG/MEG spikes.

Conclusion: A higher total number of spikes was detected by MEG compared to EEG. Beside factors such as focus localization and source orientation, overlapping sleep changes contribute to the differences observed. In sleep, the advantage of MEG in enhancing tangential over radial sources may result in higher spike detection rates for the combination of the two modalities.