Authors :
Presenting Author: Sam S Javidi, PhD – Thomas Jefferson University
Ankeeta A, PhD – Thomas Jefferson University; Xiaosong He, PhD – University of Science and Technology of China; Qirui Zhang, MD – Thomas Jefferson University; Salvatore Citro, MD – Thomas Jefferson University; Yolanda kry, B.S – Thomas Jefferson University; Michael Sperling, MD – Thomas Jefferson University; Joseph Tracy, PhD – Thomas Jefferson University
Rationale:
Focal to bilateral tonic-clonic seizures (FBTCS+) is one of the most challenging temporal lobe epilepsy (TLE) subtypes in terms of disease severity and treatment response. Current studies have implicated subcortical-cortical connectivity, and commonly take a regional/nodal approach that is agnostic to abnormalities related to changes in the white matter (WM) fiber pathways that may carry epileptogenic signaling between regions. We take this pathway approach by interrogating the integrity of WM sources indexed by diffusion-tensor-derived WM streamlines (STR). First, we test for FBTCS+ and FBTCS- (focal only TLE) differences in whole brain network-level graph measures. Second, we focus on individual STRs and examine their association with clinical variables that may be associated with FBTCS+ risk.
Methods:
We acquired T1-weighted and diffusion-weighted HARDI images on 22 FBTCS-, 43 FBTCS+ and 105 matched healthy participants (HP), processed with QSIprep and DSIstudio to obtain STR connectome matrices (AAL atlas, 90 regions of interest). We re-oriented the AAL STR matrices of the right TLE group to match the ictal (.i) and non-ictal (.ni) tracts of the left TLE group. Subsequently, we transformed the connectivity matrix into a 4005x4005 STR connectome matrix and calculated graph metrics (degree, betweenness, eigenvalue centrality, clustering coefficient). We performed ANOVA on individual STR with permutation testing to account for type I error and imbalanced sample sizes. We also reported associations between graph metrics and clinical variables.
Results:
Figure1 displays the STR that exhibit significant differences between our experimental groups. Across all graph measures, FBTCS+ displayed interhemispheric structural connection (SC) alterations compared to HP. These interhemispheric SC showed mostly involved STR decreases. TLE showed a large number of abnormally increased STRs involving AMYG.i SC with the frontal and cingulate regions (n.b., more significant in FBTCS-). In contrast, the AMYG.ni showed abnormally decreased SCs to a separate set of frontal regions (n.b., more significant in FBTCS+).
Table1 showed that in FBTCS+, age and illness duration had limited associations with graph metrics. Increased seizure frequency was associated with higher SC involving in the AMYG.i SC to SMA.i, SFGmed.i, REC.i, DCG.i, and CUN.ni in FBTCS+. Lastly, the more severe the generalized seizure activity, the less non-ictal SMA-PUT SC.Conclusions:
Our pathway approach yielded important information not just about regions involved in abnormality but the broader structural circuit expressing that abnormality. Two key findings appeared. TLE demonstrated contrasting SC abnormalities for the ictal and non-ictal amygdala, consistently involving frontal pathways, suggesting even in the absence of FBTCS amygdala-frontal pathways are selectively affected. FBTCS+ findings showed reliable interhemispheric abnormalities pointing to the specific SCs that may support the unique seizure propagation features of this group.Funding:
Joseph Tracy (PI), NIH/NINDS, R01 NS112816-01