Abstracts

Rationale: Since epilepsy is a complex network disorder, it is pertinent to understand the impact of changes within deep gray structures. The thalamus is abundantly connected to both cortical and subcortical structures, and in epilepsy thalamic sclerosis has been observed (Mueller et al., 2010, Spencer, 2002). Among people with drug resistant epilepsy (DRE), some thalamic subnuclei have been clinically validated as surgical targets for neurostimulation. Selection of the specific target depends on epilepsy subtype, determined by patient’s seizure onset zone (SOZ). Therefore, modalities that enhance visualization of targets, such as ultra-high field 7 tesla (7T) MRI, may enhance understanding of their involvement in seizure activity or correlations with different SOZ. Here, 7T MRI was used to more accurately segment subregions of the thalamus and volumetrically compare voxel-wise thalamic nuclei of epilepsy patients and controls.


Methods: Forty-three people with epilepsy (23 females/19 males, mean age 32.9 years) were age and sex matched with 43 healthy controls (HC) (33.4 years) without current or lifetime history of seizure or epilepsy. DRE patients (N=30) were classified by the following criteria: seizure activity could not be controlled with at least two anti-epileptic drugs and confirmed through post-scan chart review. All subjects underwent a 7T MRI that included a isotropic T1-weighted sequence (TA: 7:26 mins; TR: 6000ms; TE: 5.1ms; FA1/FA2: 5/4; Matrix: 282×146; voxel: 0.8×0.8×0.8 mm3; FOV: 255×183mm2; BW: 130 Hz/pixel; iPAT: 3,TI1/TI2: 1050/3000ms). T1 volumes were post-processed using FreeSurfer7.2 (Billot, B., et. al, 2020), including motion correction, intensity normalization, automatic segmentation, and parcellation processing steps. Following reconstruction, each thalamus was segmented into 6 subnuclei: anterior, lateral, ventral, intralaminar, medial, and posterior. All data were tested for normality. Volumetric differences between patients and HC were assessed using Welch two sample t-test for normal measures, and Mann Whitney U test was used for non-parametric data. To compare DRE and non-DRE patients, binomial linear regression was performed controlling for age and sex. The relationship between disease duration and subnuclei volume was assessed using Pearson correlation. Significant results were corrected for multiple comparisons using false discovery rate (FDR).


Results: There were significant reductions in left anterior, ventral, intralaminar, and posterior, and right ventral and posterior nuclei volume observed in epilepsy compared to HC (Table 1). DRE patients had significantly smaller left posterior subnuclei compared to non-DRE (p = 0.04), this result did not survive FDR correction. There was a significant relationship between disease duration and left and right ventral, intralaminar, and posterior subnuclei volume (Figure 1).


Conclusions: This work confirms that UHF MRI is a valuable tool to leverage for volumetric differences within thalamic subnuclei in epilepsy. These volumetric changes, particularly those correlating with disease duration, warrant further investigation along with sclerosis asymmetry, SOZ, and their relevance in neurostimulation.


Funding: NIH-R00NS070821