Premium
Correspondence between large‐scale ictal and interictal epileptic networks revealed by single photon emission computed tomography (SPECT) and electroencephalography (EEG)–functional magnetic resonance imaging (fMRI)
Author(s) -
Tousseyn Simon,
Dupont Patrick,
Goffin Karolien,
Sunaert Stefan,
Van Paesschen Wim
Publication year - 2015
Publication title -
epilepsia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.687
H-Index - 191
eISSN - 1528-1167
pISSN - 0013-9580
DOI - 10.1111/epi.12910
Subject(s) - ictal , electroencephalography , magnetic resonance imaging , ictal interictal spect analysis by spm , single photon emission computed tomography , temporal lobe , epilepsy , nuclear medicine , psychology , emission computed tomography , functional magnetic resonance imaging , correlation , neuroscience , medicine , radiology , positron emission tomography , mathematics , geometry
Summary Objective Epilepsy is increasingly recognized as a network disorder, but the spatial relationship between ictal and interictal networks is still largely unexplored. In this work, we compared hemodynamic changes related to seizures and interictal spikes on a whole brain scale. Methods Twenty‐eight patients with refractory focal epilepsy (14 temporal and 14 extratemporal lobe) underwent both subtraction ictal single photon emission computed tomography ( SPECT ) coregistered to magnetic resonance imaging ( MRI ) ( SISCOM ) and spike‐related electroencephalography ( EEG –functional MRI (f MRI ). SISCOM visualized relative perfusion changes during seizures, whereas EEG ‐f MRI mapped blood oxygen level–dependent ( BOLD ) changes related to spikes. Similarity between statistical maps of both modalities was analyzed per patient using the following two measures: (1) correlation between unthresholded statistical maps (Pearson's correlation coefficient) and (2) overlap between thresholded images (Dice coefficient). Overlap was evaluated at a regional level, for hyperperfusions and activations and for hypoperfusions and deactivations separately, using different thresholds. Nonparametric permutation tests were applied to assess statistical significance (p ≤ 0.05). Results We found significant and positive correlations between hemodynamic changes related to seizures and spikes in 27 (96%) of 28 cases (median correlation coefficient 0.29 [range −0.12 to 0.62]). In 20 (71%) of 28 cases, spatial overlap between hyperperfusion on SISCOM and activation on EEG ‐f MRI was significantly larger than expected by chance. Congruent changes were not restricted to the territory of the presumed epileptogenic zone, but could be seen at distant sites (e.g., cerebellum and basal ganglia). Overlap between ictal hypoperfusion and interictal deactivation was statistically significant in 22 (79%) of 28 patients. Despite the high rate of congruence, discrepancies were observed for both modalities. Significance We conclude that hemodynamic changes related to seizures and spikes varied spatially with the same sign and within a common network. Overlap was present in regions nearby and distant from discharge origin.