Mathematical co-registration of functional image data (e.g., positron emission tomography, PET) to anatomical magnetic resonance (MR) imaging data allows for objective associations between function and anatomy. Baboons are often used as non-human primate models for functional neuroimaging studies. In this work, a digital MR-based high-resolution atlas of the baboon brain was generated and evaluated for PET. The atlas was generated from six SPGR-MR datasets (centered at mid-sagittal line, AC-PC orientation) that were transformed into the space of one representative MR, averaged and resampled into PET space. The atlas was evaluated by comparing blood flow and dopamine receptor and serotonin transporter binding measures determined using regions-of-interest (ROIs) generated on each individual co-registered MR (ROI(i)) and the atlas-defined ROI template (ROI(ATLAS)). Common ROIs applied to all data included frontal cortex, temporal cortex, thalamus, caudate, putamen and cerebellum. High correlations (r(2)>0.87) were found between the ROI(i) and ROI(ATLAS) data for all radiotracers (linear regression across ROIs for each baboon). The average regression slope values ranged from 0.95 to 1.02 across radiotracers. Lastly, use of the atlas for statistical parametric mapping (SPM) of [15O]water data yielded good agreement with previous ROI(i) results. Overall, the digital MR-based atlas allowed for automatic co-registration, proved useful across a range of PET Studies, and is accessible electronically via the Internet.
Download full-text PDF |
Source |
---|---|
http://dx.doi.org/10.1016/s0361-9230(02)00810-9 | DOI Listing |
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!