Abstracts

Rationale: Medial temporal lobe epilepsy (MTLE) is a common form of adult focal epilepsy. Hippocampal sclerosis (HS) is an important diagnostic marker for MTLE and, when resected, is associated with good surgical outcome. Extra-hippocampal brain damage is often more subtle and, when detected, is associated with poorer outcome. Voxel-based morphometry (VBM) is a group-wise neuroimaging analysis tool capable of detecting subtle brain pathology and has often been used in MTLE to survey extra-hippocampal pathology. Prior MTLE VBM studies have reported pathology in a large number of inconsistently distributed areas and could not derive a focused consensus on where extra-hippocampal pathology is most common, limiting VBM's impact on MTLE research and treatment. Recently refined coordinate-based meta-analysis methods offer unprecedented analytic rigor and could identify which regions and pathways demonstrate the most consistent MTLE pathology. To this end, we applied anatomic likelihood estimation (ALE) to 22 whole-brain VBM experiments representing 562 MTLE patients with unilateral HS. We then tested whether the results of this ALE analysis were functional and anatomic constituents of the epileptogenic network using meta-analytic connectivity modeling and preliminary single-subject diffusion tractography. Methods: We adopted standard coordinate-based meta-analysis, MTLE-HS specific, and data non-redundancy inclusion filters in our paper selection. Paper retrieval included an internet search and a bibliographic search of relevant retrieved papers. Using the latest-released Eickhoff (2012) ALE method, we meta-analyzed 300 VBM foci in 3 groups: left MTLE, right MTLE, and rectified (left MTLE rectified to right + right MTLE). We created a meta-analytic connectivity model (MACM) according to the Robinson (2010) method and used ALE-reported, hand-painted tissue labels to create a diffusion tractography map in one healthy subject within 3D Slicer. Results: The three VBM ALE analyses yielded similar results: spatial convergence across studies exists in the epileptogenic hippocampus and the bilateral thalamus (voxel-level FWE p>0.05). Within the thalamus, the medial dorsal nucleus (MDN) represented the greatest consensus from each group. Confirmatory analyses of hippocampal-MDN connectivity confirmed functional (MACM) and anatomical (diffusion tractography) connectivity. Importantly, anatomical connectivity was validated in a single subject study. Conclusions: In unilateral MTLE-HS, we found consistent gray matter pathology only within the epileptogenic hippocampus and bilateral thalami. Our ALE meta-analysis yields significant dividends from VBM studies: the MDN thalamus represented the most significant thalamic pathology, or 76% of included studies—nearly equal to the epileptogenic hippocampus. We demonstrated hippocampo-MDN functional and anatomic connectivity, aligning with other studies that report MDN pathology as part of an MTLE epileptogenic network. Our observation of hippocampo-MDN anatomic connectivity in a single subject further suggests the thalamus as a clinically useful indicator in MTLE.