Abstracts

Rationale: Distinction of specific BOLD correlates of two close EEG events (as spike and slow wave, SW) is not possible in EEG-fMRI because of low temporal resolution. As shown in trial-by-trial evoked potential studies however, the BOLD signal can be correlated to the time and amplitude of EEG events. This could improve the specificity of the BOLD response, eliminating changes occurring at the time of the event but not directly linked to it. Even if the spike and the SW are close in time, these components are characterized by their own amplitude fluctuations. Our hypothesis is that the BOLD response to spikes and SW should be a function of their amplitude. Methods: We selected 5 focal epilepsy patients with spike and SW complexes isolated in time from each other. The beginning of each complex was marked as well as the beginning and peak for each SW on the channel with maximum amplitude (figure, A). A standard analysis was performed using the timing of each spike and SW complexes (figure, B). To specifically analyze the SW BOLD correlates, a first analysis was performed with one regressor coding for the time and amplitude of each SW (figure, C top). A second analysis was combining two regressors, one for the SW timing and one for the fluctuation in SW amplitude (amplitude of each SW normalized by average amplitude, figure, C bottom)). Regressors were convolved with 5 models of the hemodynamic response (peak 3, 5, 7, 9 s, Glover) and included in the general linear model for each analysis. The results obtained for SW were compared to the standard analysis. Results: On average, for each patient we analyzed 67 spikes and SW complexes (26-210). The first analysis was very similar to the standard analysis except for one patient with only a deactivation in the surround of the focus. The second analysis showed specific correlation between the SW amplitude and the BOLD in surrounding and distant areas from the focus with involvement of the contralateral side. Fluctuations of the SW amplitude were always associated with a negative correlation with the BOLD except in one case. Conclusions: This is the first method allowing the discrimination of the specific BOLD correlates of the SW in spike and SW complexes. We showed that the SW amplitude fluctuation is associated with the extent of the BOLD response and possibly the propagation of the SW, the amplitude reflecting the spatial extent of neuronal synchrony. This study also argues for a global effect of the SW on brain activity since SW amplitude is also correlated with responses in distant areas (contralateral side). Moreover this specific effect of the SW is associated with a negative BOLD correlation. It strengthens previous findings suggesting that the SW reflects the membrane hyperpolarisation following the spike, an inhibitory process.