Abstracts

Rationale: Epilepsy is a functional disorder in which both interictal and ictal activity is known to propagate widely throughout the brain. Many studies have demonstrated subtle changes in functional connectivity at the group level, particularly in temporal lobe epilepsy. Fewer studies have reported on alterations in connectivity in individual patients. Here, we seek to compare functional connectivity in epileptogenic regions in individual patients to that observed in a control population of healthy volunteers (HVs).  Methods: Multi-echo resting state fMRI and 3D MPRAGE images were acquired for 13 patients (8 male, age 17-56) with medically refractory epilepsy and 18 HVs (14 male, age 13-47) at the National Institutes of Health Clinical Center. Resting state echoes were despiked, slice-time corrected, blip-corrected, and aligned to the MPRAGE. Echoes were combined using AFNI's optimal combination method and blurred to 4mm. Then, motion parameters, 5 components accounting for most variance in ventricles and white matter, and bandpass regressors (0-0.08 Hz) were regressed from the timeseries. The fMRI timeseries for each patient was extracted and reconstructed on the surface. A region of interest (ROI) was created for each patient's seizure onset zone (SOZ) using clinically-identified SOZ electrodes or a hand-drawn lesion mask in the case of focal cortical dysplasia (FCD). For each subject, an average timeseries within each ROI was calculated, and the correlation with this average time series was calculated for each surface node. The individual HV correlation maps were combined to generate an average connectivity map for each ROI. For each patient, a cosine similarity score was calculated to this average connectivity map for their ROI. This similarity score was then compared to the score observed in each HV for the same ROI. Patients were categorized as dissimilar to the group if their z-scored cosine similarity was more than two standard deviations from the mean. Results: 13 patients were studied with interpretable resting state fMRI results and localizable epileptogenic regions based on MRI or ECoG results. 7 were non-lesional on MRI (pathologically all with microdysgenesis, 1 also with MTS), 3 had FCD (2 histologically proven), 1 with bilateral periventricular nodular heterotopia and microdysgenesis on pathology, 1 with possible bilateral MTS on imaging and MTS and microdysgenesis on pathology, and 1 non-lesional patient with unknown pathology as he did not undergo surgical resection. ROI connectivity maps in 6/13 patients were more dissimilar to the average map than any of the 18 HVs (3/6 with microdysgenesis, 1/3 with MTS and microdysgenesis, 1/3 with FCD, and 1/1 with unknown pathology). Of the 6 patients with temporal lobe epilepsy, 2/6 patients were found to have abnormal SOZ connectivity maps. For those with parietal lobe epilepsy, 2 of 4 patients had abnormal maps, and for those with frontal lobe epilepsy, 2 of 3 were found to be abnormal. Conclusions: In comparison with healthy volunteers, 6/13 patients with drug-resistant focal epilepsy were found to exhibit abnormal resting state fMRI functional connectivity in epileptogenic regions of interest, including in 5/7 non-lesional patients. Our findings support the idea that epileptogenic regions can have abnormal functional connectivity that is detectable in single subjects, even in the absence of visible lesions. Funding: This work was supported by the NIH Intramural Research Program.