Authors :
Presenting Author: Stacy Hudgins, MS – Drexel University
Michael Sperling, MD – Department of Neurology – Thomas Jefferson University; Yolanda Kry, BS – Department of Neurology – Thomas Jefferson University; Hasan Ayaz, PhD – School of Biomedical Engineering, Science, and Health Systems – Drexel University; Joseph Tracy, PhD, PhD, ABPP/CN – Department of Neurology – Jefferson Comprehensive Epilepsy Center
Rationale: The functional connectome features unique to temporal lobe epilepsy (TLE) subtypes are not well characterized. TLE patients who suffer not just focal seizures, but also include focal-to-bilateral tonic-clonic seizures (FBTCS+) represents a more severe and intractable form of TLE and are at risk for broader network impairments. The thalamus has been implicated in generalized seizure activity, but whether the functional architecture within the thalamus distinguishes FBTCS+ from other TLE subtypes is unknown. In this study, we sought to address two questions: (1) Do thalamic regions form distinct functional networks in TLE subtypes, and (2) do thalamic regions ipsilateral versus contralateral to the ictal focus form substrates in thalamic architecture that support FBTCS+.
Methods:
A total of 166 patients with refractory unilateral TLE (left=95; right=71) comprised of two subgroups (FBTCS-=45; FBTCS+=120) and age and gender-matched healthy controls (HC=119) underwent an eight minute resting-state fMRI scan followed by a standard post-processing pipeline. Hierarchical clustering based upon ROI-to-ROI anatomical proximity and functional similarity metrics (within and between intrinsic networks) yielded functional network connectivity (FNC) sets. Using sixteen thalamic parcellations (eight left; eight right), FNC comparisons between experimental groups were conducted via MANOVA. Voxel-based Intrinsic Connectivity Contrast (ICC) was obtained. Average HC ICCs were subtracted from each patient-level TLE thalamic parcel, yielding deviations compared to normal (positive toward ictal and negative toward non-ictal), then flipped from right-to-left in RTLE patients to allow direct ictal vs. non-ictal comparisons. Degree centrality (measure of overall connectivity density, hubness) and global efficiency (regional integration) were obtained.
Results: Separate direct comparisons of TLE to HC yielded one FNC set surviving multivariate parametric GLM testing (Figure 1B; not for LTLE, nor FBTCS+ or FBTCS-). The FNC cluster showed reduced pair-wise connections between regions of the bilateral ventral anterior complex (VAia and VAip, p-FDR< 0.001). Direct comparisons of ictal and non-ictal regional ICC deviations from HC, the ventral anterior parcel was observed with that difference higher in RTLE compared to LTLE. A statistical trend was observed when comparing FBTCS+ vs. FBTCS- for ictal and non-ictal VAia. RTLE and FBTCS+ compared to HCs (separately) differed in degree centrality and global efficiency centered around the right VAia; the LTLE and FBTCS- groups did not.