The spatial tracking system previously developed for monitoring head movement in patients undergoing PET brain imaging was subjected to extensive bench top and human factors testing during this reporting period.These tests revealed several anomalies in system operation that were identified and corrected.Test procedures were devised by which movements in the tracker coordinate system could be translated into movement in the PET gantry coordinate system.Tracker measurements could then be used to correct PET image data for head movement. Software for performing these corrections on real PET data were written and the entire system prepared for testing in patients. Experiments have been devised in which corrected and uncorrected images in these subjects will be compared in order to quantify the efficacy of this system in removing head movement artifacts from PET brain images. Previously developed methods to minimize the distorting effects of scattered radiation on the quantitative accuracy of planar and SPECT images were successful in reducing this source of error substantially. However, additional experiments indicated that the energy spectrum fitting method originally proposed did not accurately account for the presence of small angle scatter. As a result, a perturbation correction to the original method was proposed that used the large angle scatter fraction at each image point to estimate the small angle scatter component at that point. This new method allowed quantification of the activity in test objects to within 2% of their true value compared to 20% for the old method.Without correction, such errors can easily exceed 300% depending on source configuration.
