Echo-planar functional MR imaging of epilepsy with concurrent EEG monitoring
Patel MR, Blum A, Pearlman JD, Yousuf N, Ives JR, Saeteng S, Schomer DL, and Edelman RR(1999) Echo-planar functional MR imaging of epilepsy with concurrent EEG monitoring. AJNR Am J Neuroradiol 20:10 1916–9.
Abstract: BACKGROUND AND PURPOSE: The role of functional MR (fMR) imaging in the evaluation of patients with epilepsy has not been systematically studied. Our purpose was to identify the fMR correlates of interictal epileptiform discharges. METHODS: Twenty patients with epilepsy and frequent interictal discharges were studied with concurrent EEG monitoring on a 1.5-T echo-planar magnet to acquire blood-oxygenation-level-dependent (BOLD) images in the baseline (OFF) and immediate post-discharge (ON) states. Analysis was performed using subtraction of average ON and OFF data (method I); cross-correlation analysis between the ON and OFF states (method II); and individual spike analysis (ISA), with which signal intensity in the individual ON states was statistically analyzed using a weighted comparison with the mean and variance of the OFF states (method III). Agreement of fMR activation with EEG localization was determined. RESULTS: Eighteen of 20 patients had interictal discharges during the monitoring period. Method I yielded visually detectable sites of BOLD signal differences in only one patient. Method II resulted in two patients with sites of BOLD activation. Method III, ISA, resulted in regions of increased BOLD signal corresponding to the EEG focus in nine of 10 patients. CONCLUSION: fMR studies can often reveal sites of increased BOLD signal that correspond to sites of interictal EEG discharge activity. Because of variable intensity changes associated with discharge activity, ISA resulted in increased sensitivity.
Keywords:
Context
- The blood oxygenation level-dependent (BOLD) signal was measured in 20 patients with epilepsy, with comparison of baseline status and status immediately after interictal discharge. Simple methods like subtraction were ineffective at identifying the origin of seizures, but the manipulation of multiplying each voxel’s value by (baseline mean/baseline variance) permitted identification of the seizure focus in 10 cases, with correlation with the EEG-identified origin in 9 of those.