Correction for continuous motion in small animal PET

In small animal PET imaging experiments, animals are generally required to be anaesthetized to avoid motion artifacts. However, anaesthesia can alter biochemical pathways within the brain, thus affecting the physiological parameters under investigation. The ability to image conscious animals would overcome this problem andopen up the possibility of entirely new investigational paradigms.We have previously reported a motion-correction approach for small animal PET imaging that employs motion tracking and line of response (LOR) rebinning, and successfully demonstrated its use in phantom scans with step wise motion. In this paper we investigate an improvedsynchronization method in which TTL signals output by the motion tracker are sent to the microPET gate input to trigger the insertion of gate marks in the list mode stream that indicate the times of motion tracker measurements. The method is tested in separate microPET scans of a phantom and an anaesthetized rat which were moved continuously during data acquisition. In both cases, the motion-corrected images corresponded well with the motion-free images.We also tested the effect of pose measurement rate and synchronization error on motion correction accuracy by down-sampling and temporally misaligning list mode and motion data in a phantom study. Motion correction errors were relatively large at frequencies below -10Hz and fell rapidly to a roughly constant level above 20Hz. Motion correction errors also increased rapidly with increasingsynchronization error. In practice the acceptable limits of sampling rate and synchronization error will depend on the velocity of the motion. Using the synchronization technique presented here, and an adequate pose sampling rate, it was possible to correct for continuous motion similar to that we expect to be exhibited by conscious rats during microPET imaging experiments.