Abstracts

Rationale: Combined functional magnetic resonance imaging (fMRI) and EEG recordings allow to map brain areas involved during generation and propagation of epileptic activity. However, EEG-fMRI is a challenging technique (EEG recording inside the scanner, removing of scanner and pulse artifacts in the EEG, complex statistical fMRI analyses) and is often characterized by an insufficient sensitivity. Dynamic imaging of coherent sources (DICS) is a technique which only involves the EEG: DICS allows to study neuronal interactions by imaging power and coherence estimates using a spatial filter (Gross et al. PNAS 98(2):694-9, Muthuraman et al. IEEE, 2008:4716-9). The aim of this study was to apply DICS to different types of generalized epileptiform discharges, namely absences and photoparoxysmal responses (PPR) and compare it to the EEG-fMRI results obtained from the same data sets. Methods: Artifact corrected 32 channel EEG recordings (EEG-fMRI data sets 3T MRI) from 10 patients with absences and 7 patients with PPR during the scanning were selected; 6 patients with absences and 6 patients with PPR have been described in previous EEG-fMRI studies (Moeller et al. Epilepsia 2008; 49(9):1510-9, Moeller et al. Neuroimage 2009;48(4):682-95.). For all absences and PPRs the normalised parametric power maps in the 2-5 Hz frequency range were computed to identify the area in the brain with the strongest power. This area was defined as a reference region. The coherence between the reference region and the entire brain was computed for every patient using DICS. In a second step grand averages were computed for a) all absences and b) all PPRs. Results: Both analyses of single absences and grand average of all absences detected the source area of strongest oscillatory activity in the 2-5 Hz frequency range in the medial frontal cortex. Coherent sources with this reference region were found bilaterally in the frontal and parietal cortex, in the thalamus and in the cerebellum (figure 1). The cortical sources and the source in the thalamus were concordant with the results of the EEG-fMRI group analysis (Moeller et al. Epilepsia 2008). For PPRs the source area of strongest oscillatory activity in the 2-5 Hz frequency range was found bilaterally in the visual cortex for single subjects and in the grand average analysis. Coherent sources with these reference regions were found bilaterally adjacent to the intraparietal cortex, in the fronto-polar cortex and in the premotor cortex. No sources in the thalamus were detected (figure 2). Sources adjacent to the intraparietal sulcus and in the premotor cortex corresponded well with results from the EEG-fMRI analysis (Moeller et al. Neuroimage 2009) Conclusions: The EEG alone provides information about networks associated with absences and PPRs which is similar to EEG-fMRI results: DICS is able to find coherent sources in the 2-5 Hz range in areas detected by EEG-fMRI. Although both EEG patterns are generalized, only absences show a thalamic involvement, while PPR seems to be a cortical phenomenon.