Abstracts

Rationale:
Impaired naming function and verbal fluency are frequently observed in temporal lobe epilepsy (TLE), yet the neural correlates of the disrupted language and attention networks are poorly understood, especially in Chinese patients with TLE. Task-residual functional connectivity (trFC) is a way to ""emulate"" resting-state functional connectivity while gPPI access task-related functional connectivity. We aimed to describe the activation and functional connectivity of language network, and their relation to clinical performance.
Method:
56 health controls (HC) and 49 TLE patients (16 left TLE and 33 right TLE) participated in this study, which involves three language tasks: picture naming (PN), verbal fluency-Character (VFC) and verbal fluency-Pinyin (VFP). Images data were preprocessed by Statistical Parametric Mapping 12 (SPM12). Functional connectivity analysis between activated areas was performed using CONN v.19.c. Results for the main effect of activation or deactivation were shown at threshold Family-wise error (FWE)-corrected (p< 0.05) with a 10-voxel minimum cluster size extent threshold. For group comparison, p< 0.001 uncorrected was applied and FEW-corrected (p< 0.05) activations were additionally reported. The threshold for the main effect of functional connectivity was p< 0.05, FDR-corrected for cluster and connection level. The result of group comparison and correlations with clinical performance were reported at a threshold p< 0.05, FDR-corrected for connection and p< 0.05, MVPA uncorrected for cluster level.
Results:
All three tasks activated cerebellum, supplementary motor area (SMA), inferior occipital gyrus, and temporal lobe, while deactivating angular gyrus, posterior cingulate/precuneus (PCC/Pc) and mid cingulate. In addition, verbal fluency tasks also activated middle frontal gyrus and pre-supplementary motor area (pre-SMA). These areas involve in cognitive control network (CCN) and default mode network (DMN) while PCC/Pc, pre-SMA and angular gyrus are nodes intersect attention and language networks. Disrupted deactivation in the angular gyrus, posterior cingulate/precuneus, mid cingulate and temporal lobe was observed in the TLE group in all three tasks. Four seed regions (angular gyrus, pre-SMA, mid cingulate, and posterior cingulate/precuneus) were identified in areas with a significant difference between healthy controls and TLE patients. Seed regions are widely connected with task-related areas in all tasks. In LTLE group, seed regions were less connected to other regions in the gPPI analysis of PN, the trFC of VFP and both functional connectivity analysis of VFC while RTLE group only exhibit weakened task-residual connectivity in VFC (p< 0.05 FDR for connection level and p< 0.05 MVPA uncorrected for cluster level). Functional connectivity decreased with longer disease duration in trFC and gPPI analysis of PN as well as trFC of VFP. Poorer naming score correlates with greater trFC between right pre-SMA and operculum. Lower connectivity between angular gyrus and left temporal gyrus as well as greater correlation of angular gyrus and left hippocampus were associated with better verbal fluency performance.
Conclusion:
CNN and DMN participated in language tasks. Disturbed deactivation and functional connectivity in attention and language network were observed in TLE patients, especially in LTLE group. This provides a novel neural explanation of impaired naming and Chinese verbal fluency in TLE patients
Funding:
:National Natural Science Foundation of China (NSFC:81671300), Research Project of the Ministry of Science and Technology of China (2016YFC0904400), Natural Science Foundation of Hunan province (2020JJ5914), Xiangya Hospital Clinical Project (No. 2016L08).