Altered correlation of concurrently recorded EEG-fMRI connectomes in temporal lobe epilepsy

Abstract

Whole-brain functional connectivity networks (connectomes) have been characterized at different scales in humans using EEG and fMRI. Multimodal epileptic networks have also been investigated, but the relationship between EEG and fMRI defined networks on a whole-brain scale is unclear. A unified multimodal connectome description, mapping healthy and pathological networks would close this knowledge gap. Here, we characterize the spatial correlation between the EEG and fMRI connectomes in right and left temporal lobe epilepsy (rTLE/lTLE). From two centers, we acquired resting-state concurrent EEG-fMRI of 35 healthy controls and 34 TLE patients. EEG-fMRI data was projected into the Desikan brain atlas, and functional connectomes from both modalities were correlated. EEG and fMRI connectomes were moderately correlated. This correlation was increased in rTLE when compared to controls for EEG-delta/theta/alpha/beta. Conversely, multimodal correlation in lTLE was decreased in respect to controls for EEG-beta. While the alteration was global in rTLE, in lTLE it was locally linked to the default mode network. The increased multimodal correlation in rTLE and decreased correlation in lTLE suggests a modality-specific lateralized differential reorganization in TLE, which needs to be considered when comparing results from different modalities. Each modality provides distinct information, highlighting the benefit of multimodal assessment in epilepsy.

Keywords: Concurrent EEG-fMRI; Functional connectome; Multimodal data integration; Resting state; Temporal lobe epilepsy.

Figure 1.

Overview on the construction of EEG and fMRI connectomes. EEG and fMRI data were parcellated into the 68 regions of the Desikan atlas (coregistered to each subject’s individual T1) as follows: for fMRI, the BOLD signal time course was averaged over the voxels in each region for each subject. The Pearson correlation of the region averaged fMRI-BOLD time course was calculated to build a function connectivity matrix/connectome (FC_fMRI). For the EEG, the signal of each sensor was source reconstructed to the cortical surface (15,000 solution points) using the Tikhonov-regularized minimum norm. Then, the time courses of the solution points were averaged per cortical region. The corrected imaginary part of the coherency (ciCoh) of averaged EEG source signals were used to calculate FC_EEG for each subject (Figure adapted from Wirsich et al. (2021). Please refer to the Methods for a detailed description of each step).

Figure 2.

Cross-modal correlation between group-averaged FC_EEG and FC_fMRI (pooled across centers according to Wirsich et al. [2021]) using the Desikan atlas (*rTLE patients > controls Bonferroni threshold: p < 0.05/5 = 0.01, permutation test with 5,000 iterations; for all results, see Supporting Information Table S4. Black lines depict the 95% confidence interval of the bootstrapped mean FC_EEG-FC_fMRI correlation with 10,000 iterations; for variability of correlation derived from permuted group labels, see Table S17).

Figure 3.

Scatter plots of all pair-wise FC_fMRI and FC_EEG connection strengths (each point samples the FC_EEG and FC_fMRI connection strength of one region pair of the group-averaged FC). (A, left) Significant FC_EEG-FC_fMRI correlation differences in controls and rTLE patients in the θ-band (bootstrapped 95% confidence intervals with 10,000 iterations of mean FC_EEG-FC_fMRI correlation controls: +0.036/−0.039 and rTLE: +0.047/−0.058) and (A, right) controls and lTLE patients in the β-band (bootstrapped 95% confidence intervals with 10,000 iterations of mean FC_EEG-FC_fMRI correlation in controls: +0.032/−0.034 and lTLE; +0.051/−0.057). (B, left) Spatial contribution to FC_fMRI-FC_EEG-β correlation of lTLE patients (yellow circle depicts the spatial contribution of the DMN network exhibiting a trend decrease in lTLE patients as compared to healthy controls p = 0.0074, uncorrected; Supporting Information Table S8 and Table S13). (B, right) Scatter plot of FC_fMRI and FC_EEG-β connection strengths in the DMN that are significantly less correlated in lTLE patients as compared to healthy controls (bootstrapped 95% confidence intervals with 10,000 iterations of controls: +0.056/−0.067 and lTLE: +0.104/−0.129); we did not find any significant local alterations of the cross-modal relationship when comparing rTLE patients to healthy controls (see Table S4); VIS: visual; SM: somato-motor; DA: dorsal attention; VA: ventral attention; L: limbic; FP: fronto-parietal; DMN: default mode network.