- Top of page
Purpose: The thalamus plays an important role in seizure propagation in temporal lobe epilepsy (TLE). This study investigated how structural abnormalities in the focus, ipsilateral thalamus and extrafocal cortical structures relate to each other in TLE with mesiotemporal sclerosis (TLE-MTS) and without hippocampal sclerosis (TLE-no).
Methods: T1 and high-resolution T2 images were acquired on a 4T magnet in 29 controls, 15 TLE-MTS cases, and 14 TLE-no. Thalamus volumes were obtained by warping a labeled atlas onto each subject’s brain. Deformation-based morphometry was used to identify regions of thalamic volume loss and FreeSurfer for cortical thickness measurements. CA1 volumes were obtained from high-resolution T2 images. Multiple regression analysis and correlation analyses for voxel- and vertex-based analyses were performed in SPM2 and FreeSurfer.
Results: TLE-MTS had bilateral volume loss in the anterior thalamus, which was correlated with CA1 volume and cortical thinning in the mesiotemporal lobe. TLE-no had less severe volume loss in the dorsal lateral nucleus, which was correlated with thinning in the mesiotemporal region but not with extratemporal thinning.
Discussion: The findings suggest that seizure propagation from the presumed epileptogenic focus or regions close to it into the thalamus occurs in TLE-MTS and TLE-no and results in circumscribed neuronal loss in the thalamus. However, seizure spread beyond the thalamus seems not to be responsible for the extensive extratemporal cortical abnormalities in TLE.
Temporal lobe epilepsy (TLE) is the most common form of partial epilepsy. Based on imaging and histopathologic findings, two types of nonlesional TLE are distinguished: TLE with mesiotemporal lobe sclerosis (TLE-MTS) characterized by an atrophied hippocampus with signal abnormalities on magnetic resonance imaging (MRI) and severe neuronal loss on histologic examination (TLE-MTS, about 60–70%), and TLE with normal-appearing hippocampus on MRI and no or only mild hippocampal neuronal loss on histologic examination (TLE-no, about 30–40%). Depth electroencephalography (EEG) recordings (Vossler et al., 2004) and neuroimaging (Carne et al., 2007,Mueller et al., 2007a) suggest that in TLE-MTS seizures arise from a circumscribed region in the mesiotemporal region/hippocampus as opposed to TLE-no, wherein they arise from a more widespread, less well-defined region in inferolateral temporal lobe. In both conditions, seizures are not restricted to the hippocampus and/or temporal lobe, but can also spread to other brain regions and cause structural and metabolic abnormalities similar to those in the focus (Jutila et al., 2001; Moran et al., 2001; Keller et al., 2004).
A structure of particular interest in this regard is the thalamus, in which bilateral volume loss and/or uni- or bilateral hypometabolism have consistently been described in TLE-MTS (Juhasz et al., 1999; Mueller et al., 2006, 2007a;Gong et al., 2008; Labate et al., 2008; Seidenberg et al., 2008). The thalamus not only receives massive neuronal projections from temporal limbic regions but has also widespread reciprocal connections to subcortical structures and other cortical regions. Because of its unique role as gateway between brain structures, it has been suggested that the thalamus might play an active role in seizure propagation to other brain regions (Margerison & Corsellis, 1966;Lothman & Collins, 1981;Cassidy & Gale, 1998;Rosenberg et al., 2006; Guye et al., 2006; Bertram et al., 2008; Sloan & Bertram, 2008). Seizure spread beyond the thalamus could cause neuronal loss due to deafferentation or local excitotoxic effects and thus contribute to the widespread extratemporal structural abnormalities observed in TLE. To our knowledge the relationship between thalamic volume losses and structural changes in extratemporal brain regions in TLE has not been studied systematically. The overall aim of this study was to further investigate how volume losses in the epileptogenic focus (TLE-MTS, hippocampus; TLE-no, cortical: thickness of the inferior temporal gyrus), ipsilateral thalamus, and extrafocal cortical structures relate to each other using the FreeSurfer software package (version 3.05; https://surfer.nmr.mgh.harvard.edu) to assess cortical thickness and deformation-based morphometry to identify regions of thalamic volume loss. Specifically, the following hypotheses were tested: (1) In TLE-MTS, thalamic volume loss is most prominent in the anterior thalamic and lateral dorsal nucleus, that is, nuclei receiving direct projections from CA1, the only hippocampal subfield with direct connections to the thalamus (Herkenham, 1978; Cenquizca & Swanson, 2006). Volume/neuronal loss there leads to loss of thalamocortical projections and thus to cortical thinning in the corresponding temporal and extratemporal projection areas. (2) Based on the assumption that the focus in TLE-no is located in the inferolateral temporal lobe, the medial pulvinar, which receives projections from this area, will show the most prominent volume loss. The distribution of extrafocal cortical thinning corresponds to cortical projection areas of the pulvinar.