A direct comparison between whole-brain PET and BOLD fMRI measurements of single-subject activation response

We present the results of a direct comparison of single-subject activation using identical tasks for both functional PET and fMRI whole-brain studies. We examined the most commonly employed methods for each modality, For fMRI this is the blood oxygenation level-dependent (BOLD) contrast method with echo-planar imaging. In PET single-subject activation studies are based on the development of high sensitivity 3D imaging: of regional cerebral blood how from multiple [O-15]water injections. The identical activation paradigm of a visually cued sequential finger opposition was used for PET and fMRI. For both modalities the entire brain volume difference images were smoothe d to the same final resolution and the peak t value within the primary sens ory/motor (PSM) area was then identified. All contiguous voxels in the PSM above a predetermined threshold of statistical significance were determined . Finally, the difference-weighted centroid location was calculated for the PSM region for each modality. These studies showed a very similar pattern of activation, with the volume of activation greater in fMRI and higher lev els of statistical significance. The centroids of activation, however diffe red by 9 +/- 3 mm between the modalities, with the fMRI centroid location d orsal to that for PET. These results were stable across all processing opti ons including differing levels of image smoothing and thresholds of statist ical significance. These results are consistent with the hypothesis that dr aining veins contribute a substantial signal for fMRI activation studies an d indicate caution for the interpretation of BOLD fMRI images with activati on sites near draining veins. (C) Academic Press.