Purpose: To evaluate the effect of fMRI localization approaches and region size on the reproducibility of digit localization in the human somatosensory cortex. Materials and Methods: Vibrotactile stimulation was applied to digits 2 and 4 producing cortical activation sites relating to each digit. Thirteen subjects were scanned twice on separate occasions in a 3 Tesla scanner using a voxel size of 2 mm. Regions of activity were thresholded to different sizes varying from 50 to 1000 voxels. Three measures of position were acquired from these regions: center of gravity (COG), center co-ordinate and peak voxel. To account for registration errors, Euclidean distance between the two digits was computed. Reproducibility was determined in terms of the 95% confidence interval for individual position in X, Y, and Z and also the distance between the two digit locations. Results: Region size of 200 most significant voxels was shown to have the best reproducibility. Center co-ordinate proved to be the most precise form of localizing activity with a 95% CI of 2.1 mm, 2.6 mm, and 3.1 mm in the X, Y, and Z axes. Euclidean distance between the center coordinates of the two digit activation sites was shown to be a reliable means of overcoming registration errors with a 95% CI of 1.7 mm. Conclusion: This study shows a high level of reproducibility for fMRI localization in the somatosensory system. ; Times Cited: 0