Functional epileptic network in left mesial temporal lobe epilepsy detected using resting fMRI

Abstract

The purpose of this study was to determine transient functional signal activity in a small, homogeneous group of left temporal lobe epilepsy (TLE) patients, without the use of EEG; and to use one of these activated regions to identify a possible epileptogenic network across the whole brain in this group. Resting functional MRI scanning was performed on five left TLE patients who underwent selective amygdalohippocampectomy resulting in seizure control and 10 healthy control subjects. Activation maps of functional signal peaks were calculated using a data-driven analysis, 2dTCA, across the group of patients. In addition to the expected region of activation in the left anterior hippocampus, the results of the 2dTCA analysis revealed activity in the bilateral insular cortex and default-mode network which are not commonly reported using fMRI, but are supported by other electrical and functional changes. The region of activation corresponding to the anterior hippocampal region of resection (presumably the epileptogenic region) was used as a seed region for fMRI functional connectivity analysis. This revealed increased negative connectivity in the patients as compared to controls across a network including thalamic, brainstem, frontal and parietal regions consistent with theories of inhibited function in subcortical and cortical structures during ictal propagation.

Figure 1

Group localization of transient BOLD peaks across five TLE patients determined using 2dTCA (p<0.01 corrected for multiple comparisons). Four different transient BOLD peak timing profiles were determined and so four maps were created localizing this activity. Color bars represent T-values. Below each map is the plot of the 2dTCA normalized regressor from one of the five patients used to create each map.

Figure 2

Activated cluster seed region compared to region of resection in a single subject. a) Post-surgical T2-weighted MRI image showing resected region of left anterior hippocampus in a single patient. b) Activated cluster from group activation map #4 from Figure 1. Color bar represents T-values. Cross hairs are shown in the same location in each figure to illustrate agreement between resection and activated region.

Figure 3

Functional connectivity to left anterior hippocampus (arrow) across group of five left TLE patients (p<0.001 FWE cluster 10). a) Positive connectivity to seed region. b) Negative connectivity to seed region. Color bars represent T-values.

Figure 4

Functional connectivity to left anterior hippocampus (arrow) across group of ten healthy controls (p<0.001 FWE cluster 10). a) Positive connectivity to seed region. b) Negative connectivity to seed region. Color bars represent T-values.

Figure 5

Difference between connectivity to left anterior hippocampus (arrow) between group of five left TLE patients and ten healthy controls patients (p<0.005 uncorrected, cluster 10). Map shows regions where connectivity to seed is increased in controls over TLE patients due to increased negative connectivity in TLE patients (see discussion). No other changes were detected at this significance level. Color bars represent T-values.

Figure 6

Negative BOLD signal changes temporally correlated with positive BOLD changes in the left anterior hippocampus in left TLE patients (Figure 1, map #4) (p<0.01 corrected for multiple comparisons). Color bars represent T-values.