Abstracts

Rationale: Depth EEG in animal models suggests that dorsal midline thalamic nuclei play an important role in initiation and spread of TLE seizures. The aim of this study was: 1. to seek hippocampal and thalamic structural changes supporting the existence of hippocampal-thalamic seizure spread in TLE-MTS. 2. to compare those changes with changes in the thalamo-hippocampal network of TLE-no. Methods: T1 weighted whole brain images and high resolution T2-weighted hippocampal images were acquired on a 4T magnet in 63 subjects (32 controls, 14 TLE-MTS, 17 TLE-no). Thalamus volumes were obtained by warping the T1 image of an atlas subject with thalamus markings onto each other subject’s T1. A region of interest VBM analysis using each subject’s spatially normalized and modulated thalamus masks was performed to identify regions of thalamic volume loss. Hippocampal subfield volumes were obtained from the high resolution T2 image using a manual marking scheme (1), FreeSurfer was used for cortical thickness measurements. Multiple regression analysis was used for group comparisons, voxel-, resp. vertex-based analyses were done in SPM2 and FreeSurfer. Results: TLE-MTS had volume losses in the ipsilateral anterior thalamic, dorsolateral, medial, lateral posterior nucleus and pulvinar and the contralateral ventral posterior nucleus. TLE-no had thalamic volume losses in the ipsilateral pulvinar and contralateral ventral posterior nucleus. TLE-MTS but not TLE-no had ipsilateral volume losses in CA1, CA2 and CA3 and dentate. ERC and Sub were smaller than in controls but not significantly so. In TLE-MTS ipsilateral cortical thinning was most prominent in medial-posterior temporal regions and in TLE-no in inferior-lateral temporal regions. Volume losses in CA1 were correlated with volume losses in the anterior thalamic nucleus and volume losses in ERC and Sub with volume losses in the anterior thalamic, medial, lateral dorsal nucleus and pulvinar. Thalamic volume losses correlated with regions of ipsilateral cortical thinning in the superior frontal region, posterior and retrosplenial cingulate, medial and lateral temporal lobe, occipital lobe, and pre- and postcentral gyrus. Conclusions: The findings of this study suggest that ipsilateral thalamic volume losses observed in TLE-MTS can be mostly explained by loss of efferent/afferent connections between CA1, ERC/Sub and thalamus, i.e. provide evidence for a hippocampal-thalamic seizure spread in TLE-MTS. In contrast, thalamic volume losses in TLE-no are consistent with loss of afferent/efferent thalamo-cortical projection from lateral, inferior and medial temporal regions, i.e., provide evidence for a cortico-thalamic seizure spread in TLE-no. A better understanding of the networks involved in these two types of TLE might not only allow a better discrimination between primarily and secondarily epileptogenic brain regions and could help guide the selection of brain regions for invasive EEG studies. Reference: 1. Neurobiol Aging. 2007 28:719-26