Multiple acquisition frame-based motion correction for awake monkey PET imaging

Motion correction for PET brain imaging of awake non-human primates (NHP) is necessary if head fixation is not used. Previous studies have used multiple acquisition frame-based motion correction methods for human and small animal studies. Due to rapid motions, this method was extended to awake NHP imaging with the addition of attenuation correction and re-grouping of sub-frames in the algorithm to achieve comparable qualitative and quantitative image quality as in an anesthetized scan. Motion data were acquired with the Vicra system with a synchronization technique that improves the time precision between the list-mode data and the motion data. Using the motion data, the raw list-mode data were framed based on an intra-frame motion threshold (IFMT), above which a new sub-frame begins. Also, a minimal-frame duration threshold was set to eliminate undesirable short sub-frames. Each sub-frame was first reconstructed without attenuation correction, transformed to a reference position, and to this animal's MR template. A transmission image from an anesthetized scan of the same subject was resliced to the reference orientation of the current study, via reslicing to the same MR template, and then transformed to the orientation of each sub-frame. Finally, each sub-frame was re-reconstructed with attenuation correction, and re-grouped into standard time frames. Our work shows that this algorithm is able to generate comparable image quality and quantitative results as in an anesthetized study.