Abstracts

Rationale: A strong driving force in epilepsy research is the desire to delineate differences in spontaneous BOLD signal as measured by resting state fMRI (rsfMRI). Understanding these differences may lead to the establishment of a biomarker for disease severity, a predictor of treatment response, and/or could facilitate new drug developments. To date, multiple rsfMRI studies have been conducted. However, no study to our knowledge has examined different epileptic groups using a voxel-to-voxel rsfMRI connectivity analysis. This approach may be more robust than region-of-interest motivated analyses given the evidence for extratemporal involvement. Methods: Participants were recruited from the Scott & White Healthcare Epilepsy Center. 14 healthy controls (HCs), 12 TLE patients (7 LTLE/5 RTLE, the latter were flipped left-right prior to processing to be consistent with LTLEs), and 8 IGE patients underwent a 5-minute rsfMRI study (TR/TE: 2500/30ms; flip angle 90°, slice thickness 3mm, 38 axial slices), on a Siemen's 3T Tim Trio. A high-resolution 3D volume was also collected for anatomical considerations, and to register functional data. RsfMRI data were first preprocessed in SPM8 using standard steps including slice-time correction, realignment, reslicing, normalization, and smoothing. In preparation for functional connectivity analyses, physiological and other potentially confounding sources of noise were removed using the CompCor method. This step included regressing out the BOLD signal contributions from white matter and cerebrospinal fluid masks as well as head motion parameters. A temporal band pass filter of 0.01Hz to 0.1Hz was applied to the data. Voxel-to-voxel intrinsic global connectivity analyses were performed. Results: We found that HCs had greater intrinsic connectivity (voxel threshold p<0.001 uncorrected; extent threshold p< 0.05 FDR-corrected) compared to TLE in 4 distinct clusters encompassing the left inferior and middle temporal gyri, bilateral somatosensory association cortex, bilateral posterior cingulate cortex, and bilateral angular gyrus. Compared to IGE, HCs had greater connectivity in the right inferior and middle temporal gyri, and the right fusiform gyrus. TLEs had greater connectivity (voxel p<0.01 uncorrected; extent threshold p<0.05 FDR-corrected) throughout the visual cortex. TLEs had greater connectivity to the right premotor and primary motor cortices, as well as the right dorsolateral prefrontal cortex compared to the IGEs. IGEs had greater connectivity (voxel p<0.01 uncorrected; extent threshold p<0.05 FDR-corrected) in the left dorsolateral prefrontal cortex, premotor cortex, and anterior cingulate compared to TLEs and evidence of greater connectivity throughout the visual cortices compared to HCs. Conclusions: These results suggest that whole-brain rsfMRI connectivity analyses may yield important insights into the neuropathophysiological mechanisms of epilepsy. Our data suggest that extratemporal differences may be present in both forms of epilepsy.