Functional magnetic resonance imaging based on a non-BOLD-contrast mechanism, which we have termed "SEEP" (Signal Enhancement by Extravascular water Protons), has previously been demonstrated. Here the reproducibility of areas of activity identified with both SEEP and BOLD contrast is assessed in duplicate experiments in healthy volunteers, with relatively high resolution (1.6 mm) image data at 1.5 T. These areas of activity are equally well localized to voxels containing primarily gray matter with the two contrast mechanisms. As in previous studies, areas of SEEP activity are observed to be immediately adjacent to areas of BOLD activity, with very little overlap. The response functions were estimated for both SEEP and BOLD contrast, and are observed to be distinct. The peak SEEP response is observed to lag the BOLD response by approximately 1 s and to decay more slowly with no poststimulus undershoot. Average BOLD signal changes (GE-EPI, TE=50 ms) were observed to be 3.4+/-1.2% (mean+/-S.D.), whereas SEEP signal changes (SE-EPI, TE=23 ms) were 1.9+/-0.5%, consistent with previous studies carried out at 0.2 and 3 T. These observations provide further support for the existence of a non-BOLD-contrast mechanism for fMRI, based on changes in extravascular spin density.