Purchase this article with an account.
A. Miki, C.J. Liu, G.T. Liu, J.C. Haselgrove; Effects of Voxel Size and Post-processing on Detection of Lateral Geniculate Nucleus Activation in Functional Magnetic Resonance Imaging at 1.5 Tesla . Invest. Ophthalmol. Vis. Sci. 2003;44(13):2723.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose. The optimization of data acquisition and analysis is crucial in functional magnetic resonance imaging (fMRI). We investigated the effect of voxel size and post-processing on the detection of activation of lateral geniculate nucleus (LGN) with fMRI at 1.5 Tesla. Methods. Ten normal volunteers (5 male and 5 female, mean age 21.1 years) were studied. Imaging was performed with a clinical 1.5-T MRI system (Sonata, Siemens). One hundred image sets of 28 functional images were acquired using a T2*-weighted echo-planar image sequence (3 second interscan interval, matrix=64x64, field of view (FOV) =200, 300, 400, or 500 mm) with the slice thickness of 5mm. All subjects were scanned four times with the 4 different FOVs (and therefore with 4 different voxel sizes) in a counterbalanced order across subjects. Five scans of visual stimulation of both eyes with light-proof binocular goggles flashing at 8Hz alternated with 5 scans of darkness. Data analysis was performed with IDL and SPM99. The functional images of each subject were corrected for the delay associated with slice acquisition timing and realigned to the first volume. Then the images were spatially normalized and smoothed. T-statistics were calculated for each voxel for the contrast of the visual stimulation condition versus the dark condition. The statistical analysis was performed with or without global normalization. The statistical threshold was set at the uncorrected P value of 0.001. Results. When statistical analysis was performed without the global normalization, activation of LGN was detected on both hemispheres in all subjects. The activation was not found for 300 and 500 FOVs in one subject and for 400 FOV in one hemisphere of another subject. However, when the global normalization was used, activation of LGN was found only in 6 subjects, in which cases the activation could be most frequently detected at the smallest voxel size. Conclusions. Our results suggest that the voxel size and post-processing may affect the detectability of LGN activation in fMRI.
This PDF is available to Subscribers Only