Rationale: Numerous dichotomous clinical features have been studied to identify the core, unique neurobiology of focal TLE (e.g., pathology subtype, seizure subtype, seizure onset zone, surgical outcome, pre-/post-surgical brain organization). To date, no study has assessed all these clinical “conditions” of TLE in a unified, comparative framework, testing for commonalities or differences across “conditions”. In this project, we sought to determine the regional and whole brain functional connectome and structural features that are shared (or distinct) across these “conditions”.


Methods: Sample was comprised of 54 ATL-treated focal TLE patients (right=28; left=26) and matched healthy participants (HP, n= 32) with presurgical/baseline and 1 year postsurgical/follow-up multi-modality MRI data. After flipping the data from right-sided TLE patients to the left side, we calculated 8 resting-state fMRI measures (ALFF, ReHo, integration and recruitment of dynamic community structure, modal controllability from control theory, and degree centrality, local efficiency and participation coefficient from graph theory), along with gray matter volume for 116 brain regions. We then tested (independent/paired t-tests, FDR corrected) for the differential effects of these measures on key binary characteristics of the TLE condition (TLE vs. HP organization, TLE post- vs. pre-surgical organization, seizure outcome [poor vs. good], pathology [hippocampus sclerosis vs. non-HS], seizure subtype [focal vs. focal-to-bilateral tonic-clonic seizures], and seizure onset zone lateralization [left vs. right]).


Results: After t-test analysis of the 54 sets of statistics, we found that across “conditions” the thalamus played a prominent role in TLE. The ipsilateral thalamus was the brain region with the largest absolute mean Hedge’s g effect size and displayed the largest number of “condition” discriminations (number significant at p< 0.05, uncorrected). In terms of specific brain functional organization measures, thalamic local functional harmony (ReHo) characterized pre-surgical organization. Thalamic spontaneous activity (ALFF) and gray matter volume characterized post-surgical organization. Particularly noteworthy was increased ipsilateral thalamic ALFF, but decreased contralateral thalamic ALFF, postoperatively. Hubness (degree centrality), control (modal controllability) and ALFF were reliable predictors of poor seizure outcome (AUC = 0.79 in XGboost classifier, five-fold validation).


Conclusions: Our unified, comparative framework, testing for functional/structural commonalities or differences across TLE “conditions” revealed the importance of the thalamus across numerous “conditions”. Drivers of the functional status of the thalamic hemispheres differed before and after TLE surgery. Seizure outcomes after ATL were best predicted by thalamic hubs and regions controlling “hard-to-reach” brain states. Overall, the data clearly showed that each functional/structural feature plays a different role in each clinical context (“condition”), but in terms of regions, beyond even the temporal lobe, the thalamus has most consistent impact across “conditions”.


Funding:

Joseph Tracy (PI), NIH/NINDS, R01 NS112816-01