Dynamic Analysis of Gray Matter and White Matter Functional Network Connectivity in Individuals with Temporal Lobe...

Accurately localizing the epileptic zone in temporal lobe epilepsy (TLE) is challenging. Emergent resting-state functional connectivity (rs-FC) analyses have demonstrated that the epileptic network of TLE likely involves both gray (GM) and white matter (WM). Clustering dynamic functional network connectivities (dFNCs) reveals time-varying dependencies of functional networks (FNs) and may provide additional insight into TLE, as interictal discharges are often transient and spread across FNs. In this study, we analyzed data from 103 participants from the Epilepsy Connectome Project (ECP), including 51 healthy controls, 34 individuals with left TLE, and 18 with right TLE. We used K-means clustering to derive FNs in GM and WM. We calculated static and dynamic FNC between these FNs, using the average correlation for static FNC and standard deviation (STD) for dynamic FNC. We then obtained dynamic FNC states by clustering windowed FNCs. We did not find any significant connections in dynamic FNC (dFNC) using STD. However, dFNC states demonstrated the association of WM FNs, altering connectivity in TLE. The dynamic state that was slightly negatively correlated with the sFNC showed significant changes in TLE, affecting not only GM but also the GM-WM and WM-WM connections. Using dFNC, we observed notable connections that were absent in both higher negative and positive FNC patterns and did not fall within the strongest 1% connections. Furthermore, a significant GM-WM connection was found in a state in which the probability of self-transition differs between TLE and HC. The state analysis (dynamic) reveals that the FCs between temporooccipital - lateral occipital cortex and Superior frontal gyrus - left frontal orbital cortex - paracingulate gyrus, Superior parietal lobule - postcentral gyrus - precentral gyrus and corpus callosum - forceps minor - superior and anterior corona radiata - right superior longitudinal fasciculus - left temporal sub gyral, and left inferior longitudinal fasciculus and hippocampus increased significantly in TLE. The FCs between right angular gyrus - right middle frontal gyrus -right middle temporal gyrus and left inferior longitudinal fasciculus, and right angular gyrus - right middle frontal gyrus -right middle temporal gyrus and left temporal sub gyral - left superior longitudinal fasciculus - left splenium of the corpus callosum decreased significantly. These increase and decrease exhibit only in a low negative connectivity state in TLE. In general, our findings suggest that including WM along with GM FNs and considering time-varying network interactions are important to understanding TLE.