Use of short interscan interval [15O]water positron emission tomography (PET) studies reduces the overall study duration and may allow an increased number of scans for single-subject analysis of unique cases (e.g., stroke). The purpose of this study was to examine how subtraction of residual radioactivity from the previous injection (corrected scan) compared to nonsubtraction (uncorrected scan) in a PET short interscan interval (6 minutes) study affects single-subject and group data analysis using a motor activation task. Two currently widely used analytic strategies, Worsley's method and the SPM technique, were applied. Excellent agreement between activation maps obtained from corrected and uncorrected data sets was obtained both in single-subject analyses performed on data sets from the six normal subjects and three stroke (subcortical infarct) patients, and in group analysis (six normal subjects) within a particular statistical method. The corrected and uncorrected data were very similar in the (1) number of activated brain regions; (2) size of clusters of activated brain voxels; (3) Talairach coordinates of the activated region; and (4) t or Z value of the peak intensity for every significantly activated motor brain structure (both for large activations such as in motor cortex and small activations such as in putamen and thalamus). [15O]Water PET data obtained with a short interscan interval (6 minutes) produce similar results whether or not the background is subtracted. Thus, if injection dose and timing are constant, one can achieve the advantage of a short interscan interval without the added complexity of correcting for background radioactivity.