Researchers and physicians who work in the field of neurology, neuroscience, specifically epilepsy.The most frequent pathologic finding in temporal lobe epilepsy (TLE) is hippocampal sclerosis (HS).^1,2 TLE with HS has also evolved towards the view that this syndrome affects widespread brain networks and in terms of lateralization, left and right TLE seem to show a different pattern of network disease. These patients with hippocampal sclerosis (HS) also benefit from medial temporal lobectomy. Since the hippocampus is involved in many cognitive tasks, we hypothesized that resting state network (RSN) alterations would occur in these patients with network disease following medial temporal lobectomy.

IRB was obtained for this study and all the participants gave signed consent form.

Subjects: We studied 20 right-handed patients (11 left HS, 9 right HS) with diagnosis of isolated unilateral HS on basis of clinical, electrophysiological and MRI findings. These patients had medial temporal lobectomy. They were categorized into left HS (F/M: 4/7; 28.6±5.56) and right HS (F/M: 4/5; 29±5.86 years).

Image Acquisition: Imaging of the brain was performed on a 3T MR scanner (Magnetom, Trio TIM system, Siemens, Germany) equipped with a 32-channel phase-array head coil. Resting-state (rs) fMRI imaging applied T2* weighted gradient echo spiral pulse sequence (TR/TE: 2000/35 msec, FA 75°, FOV: 230 mm, matrix: 64 x 64, in-plane spatial resolution of 3.6 mm) while the subjects kept their eyes closed without a specific concentration. Structural 3D T1-weighted high resolution (magnetization prepared rapid gradient echo-MPRAGE) (TR/TE: 1900/3.4 msec; FA: 90; FOV: 256mm; matrix: 224x256; distance factor: %50) sequence was also obtained.

All patients had pre- and post-operative neurocognitive tests, and imaging with the same protocol on the same scanner. Post-operative studies were performed at 1-year follow-up.

Data Processing and Analysis

Preprocessing of the rs-fMRI data: The rs-fMRI scans were preprocessed using SPM8 ³ . Preprocessing of the rs-fMRI data included realignment, slice-timing correction, co-registration and normalization (to Montreal Neurological Institute (MNI) template), and spatial smoothing with an 6 mm³ isotropic Gaussian kernel.

Independent components analysis (ICA): Melodic ICA version 3.14 (Beckmann and Smith, 2004) ⁴ with a multisession temporal concatenation tool was used to perform groupICA analysis. Preprocessed data were used as input for ICA. 30 spatiotemporal component were yielded by using a dimensionality estimation using the Laplace approximation to the Bayesian evidence of the model. We used 17 popular Resting State Network (RSN)s (Yeo et al., 2011) ⁵ to compare the spatial map of each ICA component. We used “fslcc” (tool of FSL) to do the reference network correlations (Pearson's r > .207) . This procedure was repeated for each group. For pre- and post- operative comparisons within left and right HS groups, 10 and 8 reference network correlated components were extracted respectively (Fig.1a: left HS, 1b: right HS).

Dual Regression: To investigate spatial and intensity differences in those RSNs as a function of within-group difference between pre- and post- operative studies of left and right HS, we used FSL dual regression technique that allows for voxel-wise comparisons of rs- fMRI (Filippini et al., 2009; Littow et al., 2010; Veer et al., 2010; Abou Elseoud et al., 2011). ^6-9

Dual regression results showed alterations in default-mode network (DMN) and fronto-parietal network (FPN) in both right and left HS groups with temporal lobectomy. Post-operatively, left HS group showed decreased activation in posterior cingulate cortex (PCC) (Fig. 2a) with an increased activation in anteror cingulate cortex (ACC) (Fig. 2b) in DMN (p_corrected = 0.029). In FPN, subtle reduction in activation of right sub-gyral frontal (Fig. 3a) with marked increased activation in right superior frontal gyrus (Fig. 3b) (p_corrected = 0.001) . In comparison of pre- and post-operative imaging of right HS group, there was significantly reduced activation in medial prefrontal cortex (p_corrected = 0.023) without increased activation in DMN (Fig. 4). On the contrary, right inferior frontal gyrus showed increased activation (p_corrected = 0.023) (Fig. 5) with no significant reduction in activation in FPN in this group.Following temporal lobectomy, left and right HS patients showed different patterns of BOLD alterations on rs-fmri obtained at one-year follow-up. Significant decreased and increased activations occurred in DMN and FPN. A pre-operative extent of tissue damage or dominancy of the epileptic hemisphere may be responsible for the different patterns of adaptation/change of brain networks after temporal lobectomy.