Diminished default mode network recruitment of the hippocampus and parahippocampus in temporal lobe epilepsy

Abstract

Object: Functional neuroimaging has shown that the brain organizes into several independent networks of spontaneously coactivated regions during wakeful rest (resting state). Previous research has suggested that 1 such network, the default mode network (DMN), shows diminished recruitment of the hippocampus with temporal lobe epilepsy (TLE). This work seeks to elucidate how hippocampal recruitment into the DMN varies by hemisphere of epileptogenic focus.

Methods: The authors addressed this issue using functional MRI to assess resting-state DMN connectivity in 38 participants (23 control participants, 7 patients with TLE and left-sided epileptogenic foci, and 8 patients with TLE and right-sided foci). Independent component analysis was conducted to identify resting-state brain networks from control participants' data. The DMN was identified and deconstructed into its individual regions of interest (ROIs). The functional connectivity of these ROIs was analyzed both by hemisphere (left vs right) and by laterality to the epileptogenic focus (ipsilateral vs contralateral).

Results: This attempt to replicate previously published methods with this data set showed that patients with left-sided TLE had reduced connectivity between the posterior cingulate (PCC) and both the left (p = 0.012) and right (p < 0.002) hippocampus, while patients with right-sided TLE showed reduced connectivity between the PCC and right hippocampus (p < 0.004). After recoding ROIs by laterality, significantly diminished functional connectivity was observed between the PCC and hippocampus of both hemispheres (ipsilateral hippocampus, p < 0.001; contralateral hippocampus, p = 0.017) in patients with TLE compared with control participants. Regression analyses showed the reduced DMN recruitment of the ipsilateral hippocampus and parahippocampal gyrus (PHG) to be independent of clinical variables including hippocampal sclerosis, seizure frequency, and duration of illness. The graph theory metric of strength (or mean absolute correlation) showed significantly reduced connectivity of the ipsilateral hippocampus and ipsilateral PHG in patients with TLE compared with controls (hippocampus: p = 0.028; PHG: p = 0.021, after correction for false discovery rate). Finally, these hemispheric asymmetries in strength were observed in patients with TLE that corresponded to hemisphere of epileptogenic focus; 87% of patients with TLE had weaker ipsilateral hippocampus strength (compared with the contralateral hippocampus), and 80% of patients had weaker ipsilateral PHG strength.

Conclusions: This study demonstrated that recoding brain regions by the laterality to their epileptogenic focus increases the power of statistical approaches for finding interhemispheric differences in brain function. Using this approach, the authors showed TLE to selectively diminish connectivity of the hippocampus and parahippocampus in the hemisphere of the epileptogenic focus. This approach may prove to be a useful method for determining the seizure onset zone with TLE, and could be broadly applied to other neurological disorders with a lateralized onset.

Fig. 1

Functional MR images showing the DMN in healthy controls. Midline structures, including the medial prefrontal cortex and PCC, are observed along with bilateral hippocampal and parahippocampal regions. Independent component analysis was conducted on the healthy controls’ resting-state fMRI data to generate 20 components. The component representing the DMN was thresholded to include voxels with a significant contribution to the component time course (z ≥ 1.00 and 10 voxels) and over-laid atop the MNI ICBM452 structural brain atlas provided with the AFNI software package. Images are displayed in neurological convention (left side is on the left of the axial slice).

Fig. 2

Default mode network ROIs. The ROIs were defined as 6 mm radius spheres centered on the center of mass for each of the 10 ROI clusters generated from the independent component analysis in Fig. 1 and described in Table 4. The ROIs were used in connectivity analysis across groups. Images are displayed in neurological convention. MPFC = medial prefrontal cortex; L PAR = left inferior parietal lobule; R PAR = right inferior parietal lobule; L HPC = left hippocampus; R HPC = right hippocampus; L PHG = left PHG; R PHG = right PHG; L SFC = left superior frontal cortex; R SFC = right superior frontal cortex.

Fig. 3

Graph theory depiction of DMN connectivity in healthy controls and patients with TLE. There is less connectivity of the ipsilateral hippocampus (HPC_I) and contralateral hippocampus (HPC_C) to the rest of network. The strength of connectivity of these regions to the PCC was significantly less in the patient group than the control group. Each subject’s ROI correlation matrix was binarized into an adjacency matrix describing the presence or absence of a connection (edge) between 2 ROIs (nodes). Lines depict edges present in the adjacency matrices at different population thresholds (that is, at least 20%, 30%, and 40% of each sample). PHG_I = ipsilateral PHG; PHG_C = contralateral PHG; MPFC = medial prefrontal cortex; PAR_I = ipsilateral inferior parietal lobule; PAR_C = contralateral inferior parietal lobule; SFC_I = ipsilateral superior frontal cortex; SFC_C = contralateral superior frontal cortex.