Abstracts

Rationale: In simultaneous EEG-fMRI, the most common blood oxygenation level dependent (BOLD) changes resulting from scalp EEG spikes are positive responses caused by the excessive synchronized neuronal activity of the spike. Surprisingly,negative responses are sometimes seen. Recent studies describe positive BOLD signal changes even before focal spikes seen on EEG. Some of these early positive BOLD responses are followed by later negative responses in the spike field. We aimed to systematically study if the negative BOLD responses sometimes seen in the focus point to preceding positive changes. Methods: 92 patients with focal epilepsy who underwent EEG-fMRI at 3 T (TR=1.75s, 5mm voxels, up to 2 hours of scanning) were retrospectively studied. After gradient and ballistocardiogram artifact removal, interictal spikes were marked on the EEG. The data had previously been analyzed in the general linear model framework using hemodynamic response functions (HRFs) peaking from 3 to 9 seconds after the spike time. The statistical t-maps were thresholded at |t|>3.1 (p<0.001) with a minimum cluster extent of 4 voxels to control for family-wise errors (p<0.05 corrected). In the current study, we only included patients with a focal negative BOLD response in the region of the spike field, whereas deactivations in default mode areas were not considered. Patients with prolonged bursts of interictal discharges were excluded due to overlap between consecutive HRFs. Patients who met the inclusion criteria were re-analyzed using models with HRFs peaking from -9 to +9 seconds around the spike to investigate possible earlier BOLD responses. Results: 7 patients met the inclusion criteria (2 lesional temporal lobe epilepsy, 1 non-lesional temporal lobe epilepsy, 4 lesional frontal lobe epilepsy), showing negative BOLD responses in the spike field in the standard analysis. None had positive BOLD responses immediately adjacent to the areas of deactivation that had been noted. When looking for early BOLD signal changes, 2 patients (28%) had focal prespike positive BOLD responses in the regions that had shown deactivation in the standard analysis. These early activations were seen at the combined maps of -5 to -9 seconds. No preceding deactivations were detected. Conclusions: This study confirms that in a certain proportion of patients with focal epilepsy in whom the standard analysis reveals focal negative responses, earlier positive BOLD responses can be detected. One may assume that in these patients the negative BOLD response reflects the undershoot of the earlier positive response. Early changes may result from increased neuronal activity in the spike field prior to the spike seen on the scalp EEG, although no evidence for such changes has been uncovered. However, the pre-spike positive BOLD explains focal deactivations only in some patients. The cause of negative BOLD signal changes in the focus as a result of an epileptic event remains unexplained in the majority of the small proportion of patients in whom it occurs.