Authors :
Presenting Author: Liliana Martinez, B.S 24' – University of Louisville
Elizabeth Gatti, n/a – Vanderbilt University; Lucas Sainburg, B.S – Vanderbilt University; Behnaz Akbarian, M.S – Vanderbilt University; Andrew Janson, PhD – Vanderbilt University; Dingjie Su, B.S – Vanderbilt University; Benoit Dawant, PhD – Vanderbilt University; Dario Englot, M.D – Vanderbilt University; Victoria Morgan, PhD – Vanderbilt University
Rationale:
In temporal lobe epilepsy (TLE) seizures often originate in the mesial temporal structures, including the hippocampus. However, these focal seizures affect widespread brain networks. The thalamus is highly connected to the hippocampus and plays a role in the propagation and reduction of seizures through neurostimulation in TLE. Therefore, characterization of thalamic networks in TLE is clinically significant. Functional connectivity (FC) between brain regions in the seizure network can be determined with functional MRI (fMRI). In this study, we subdivided the thalamus into six bilateral subregions and investigated the effects of TLE on the thalamus subregions’ functional connections to the cortical network.
Methods:
This study included 106 healthy controls (CON), 25 left TLE patients (LTLE), and 45 right TLE patients (RTLE). A T1-weighted MRI scan (1x1x1mm
3) and functional MRI (TR=2 sec, 300 time points, 3x3x3mm
3) at rest was acquired for all subjects. The thalamus was segmented into 23 subregions
1 on the T1-weighted MRI, and then regrouped into six larger subregions (Fig.1B). The cortex was divided into 115 regions using the MultiAtlas
2 algorithm (Fig. 1A).
The fMRI data were corrected for physiological motion, slice timing and motion using SPM12. The average time series was then computed for each region of interest. FC was computed as the partial correlation between each region of the thalamus and each cortical region. This resulted in a 12x127 correlation matrix of the thalamic regions to all regions for each subject. The correlations were then normalized using the Fisher Z transform.
An unpaired t-test was used to compare correlation FC for CONs v. LTLE and CONs v. RTLE. The significance was determined after Bonferroni correction for the number of region pairs (0.05/1512 = 3.3x10-5). A second p-value threshold was set at 0.001. Statistical analysis and all processing were done utilizing MATLAB 2021a.
Results:
The FC in the LTLE patients was greater than controls between the right lateral, ventral, and pulvinar to regions in the temporal, somatosensory and subcortical lobes. The lower threshold revealed many additional increases in FC from multiple thalamic regions to the same lobes (Fig. 2).
Conclusions:
Both right and left TLE patients showed significant thalamic FC increases to regions in the temporal, occipital, and subcortical lobes, with the right TLE changes being more widespread involving most regions of the thalamus. This suggests changes may not be occurring in individual thalamic regions, but rather in cortical regions. These findings will contribute to understanding the effects TLE has on functions controlled by the thalamus.
References
1 Liu et al.,
Magn Res Imaging 2020; 65, 114-128
2 Asman et al.
Med Image Anal. 2013;17(2):194-208
Funding: NIH R01 NS110130, NIH R01 NS108445, NIH R00 NS097618, AES BRIDGE Internship