Abstracts

Rationale:
The existence of an epileptogenic region in the dominant temporal lobe (TL) often results in the reorganization of language-relevant systems in the brain (Tracy et al., Cortex 2009; 45:1178-89). The presumptive goal of such reorganization is to compensate for the dysfunctional, e.g., epileptogenic TL. These patterns likely vary depending on the language components required by the task, baseline task competency, and the unique features of the underlying pathology. To identify potentially unique patterns of task-adaptive regional communication, we tested for whole brain neural functional connectivity (FC) alterations in left TL epilepsy (LTLE) during a complex receptive language task.
Method:
Seventeen LTLE patients and 16 age/education matched healthy controls (HCs) underwent an  fMRI scan in a 3T Philips MRI scanner while performing a sentence completion task (SC) requiring them to covertly generate a word that best completes the meaning of a visually presented sentence.  A control condition involved passively viewed random letters demarcated in word length arrays. To ensure adequate task engagement and competency, patients with a verbal IQ over 80 were enrolled.  The LTLE patients did not differ from HCs on neuropsychological measures of phonemic and semantic fluency. fMRI data was preprocessed using fMRIPrep 20.0.1 (Esteban O, et al. Nat Methods, 2019; 16: 111-16). Second-level analyses focused on the group differences in activation (LTLE vs. HC) emerging from the SC versus control contrast using SPM12 (http://www.fil.ion.ucl.ac.uk/spm/software/spm12). Using the CONN toolbox, seed regions (bilaterally) comprised of established language regions of the brain  (inferior frontal pars opercularis (IFOper), inferior frontal pars triangularis (IFTrian), middle temporal gyrus (MTG), other language network and cerebellar ROIs) were chosen to identify whole brain task-based FC effects that might provide the unique computational support for implementing the SC in LTLE relative to HCs. Independent Component Analysis (ICA) was used to identify other functional networks and subnetworks, and group differences (LTLE vs HC) in the strength of their association with language regions was assessed.
Results:
No significant cluster-level activation differences between the groups were found for SC task. Both groups showed SC activation in left MTG, left IFOper, left IFTrian, supplementary motor area (bilaterally), and right cerebellum (Figure 1a, b). In contrast, increased FC during SC was found in LTLE compared to HCs, involving the left IFOper (seed) and vermis 6 of the cerebellum, and vermis 6 (seed) and intraparietal sulcus of the dorsal attention network and left IFTrian (Figure 2a).  ICA-identified network clusters revealed increased-interhemispheric FC between language and salience network clusters in LTLE group relative to HCs (Figure 2b).
Conclusion:
These findings suggest that abnormal, but compensatory recruitment of the dorsal attention and salience networks was present in our LTLE patients during SC. These recruitments provide a window into the unique patterns of reorganization and functional plasticity responses that occur in LTLE in order to maintain adaptive language functioning.
Funding:
:JIT, NIH/NINDS, R01 NS112816-01