Cerebral blood flow and fMRI BOLD auditory language activation in temporal lobe epilepsy

Abstract

Blood oxygen level-dependent (BOLD) functional magnetic resonance imaging (fMRI), an important research and clinical tool, depends on relatively greater transient increases in (regional cerebral blood flow) rCBF than cerebral metabolic rate for oxygen during neural activity. We investigated whether reduced resting rCBF in patients with temporal lobe epilepsy affects BOLD signal during fMRI language mapping. We used [(15)O] water positron emission tomography (PET) to measure rCBF, and 3 Tesla echo planar imaging (EPI) BOLD fMRI with an auditory description decision task in 33 patients with temporal lobe epilepsy (16 men; mean age 33.6 ± standard deviation [SD] 10.6 years; epilepsy onset 14.8 ± 10.6 years; mean duration 18.8 ± 13.2 years; 23 left focus, 10 right focus). Anatomic regions drawn on structural MRI, based on the Wake Forest Pick Atlas, included Wernicke's area (WA), inferior frontal gyrus (IFG), middle frontal gyrus (MFG), and hippocampus (HC). Laterality indices (LIs), and asymmetry indices (AIs), were calculated on coregistered fMRI and PET. Twelve patients had mesial temporal sclerosis (seven on the left), two patients had a tumor or malformation of cortical development (both left), one patient a right temporal cyst, and 18 patients had normal MRI (14 left). Decreasing relative left WA CBF correlated with decreased left IFG voxel activation and decreasing left IFG LI. However, CBF WA AI was not related to left WA voxel activation itself or WA LI. There was a weak positive correlation between absolute CBF and fMRI activation in left IFG, right IFG, and left WA. Patients with normal and abnormal MRI did not differ in fMRI activation or rCBF AIs. Reduced WA rCBF is associated with reduced fMRI activation in IFG but not WA itself, suggesting distributed network effects, but not impairment of underlying BOLD response. Hypoperfusion in TLE does not affect fMRI clinical value.