Converging PET and fMRI evidence for a common area involved in human focal epilepsies

Abstract

Objectives: Experiments in animal models have identified specific subcortical anatomic circuits, which are critically involved in the pathogenesis and control of seizure activity. However, whether such anatomic substrates also exist in human epilepsy is not known.

Methods: We studied 2 separate groups of patients with focal epilepsies arising from any cortical location using either simultaneous EEG-fMRI (n = 19 patients) or [¹¹C]flumazenil PET (n = 18).

Results: Time-locked with the interictal epileptiform discharges, we found significant hemodynamic increases common to all patients near the frontal piriform cortex ipsilateral to the presumed cortical focus. GABA(A) receptor binding in the same area was reduced in patients with more frequent seizures.

Conclusions: Our findings of cerebral blood flow and GABAergic changes, irrespective of where interictal or ictal activity occurs in the cortex, suggest that this area of the human primary olfactory cortex may be an attractive new target for epilepsy therapy, including neurosurgery, electrical stimulation, and focal drug delivery.

Figure 1. EEG-fMRI group analysis

Results of a second-level random-effects group analysis of 19 patients with focal epilepsy syndromes. For visualization, consistent common activations (p < 0.001) are overlaid on axial slices of a mean T1-weighted template brain (X, Y, Z = −30, 6, −2, coordinates in Montreal Neurological Institute space). The activation within the region of interest near the presumed area tempestas was significant at p < 0.05 (family-wise error), when correcting for multiple comparisons across the search region (2,744 mm³).

Figure 2. Flumazenil PET group comparison

Regions of significantly increased flumazenil volume of distribution in 18 patients with focal epilepsy syndromes compared with those of 24 normal control subjects. The hot metal color scale displays all voxels falling below p < 0.01 for display; increasing intensity corresponds to increased significance.

Figure 3. Flumazenil PET correlational analysis

(Top row) Regions of increased flumazenil volume of distribution (VD) in 18 patients with focal epilepsy syndromes in a parametric analysis of patient data alone that showed reduced flumazenil binding with increased number of seizures per month (p < 0.05 corrected). (Bottom) Scattergraph of seizure frequency vs flumazenil volume of distribution at the voxel with a maximum z score (indicated by + in the left panel).

Figure 4. Combined EEG-fMRI/PET results

Clusters around the peak voxels for EEG-fMRI group analysis (yellow) and correlation between flumazenil binding and seizure frequency (blue) are superimposed on a T1 template. ce = capsula externa; ci = capsula interna; Cl = claustrum; CN = caudate nucleus; fPC = frontal piriform cortex; GP = globus pallidus; IC = insular cortex; oc = optic chiasm; Pu = putamen; tPC = temporal piriform cortex.