3. PET simulation

This step involves the generation of the response of the PET system to each of the identified tissue components within the brain model (or Regional Spread Functions). command: 
pvc_petsim [options]  labels_inPET.mnc  realPET.mnc  simuPET.mnc  atlas.mnc
pvc_petsim -help for more details where:
  • labels_inPET.mnc is the segmented MRI volume in PET native space (output from step 2).
  • realPET.mnc is the real dynamic PET image data set
  • simuPET.mnc and atlas.mnc are respectively the simulated regional spread functions (RSF's), and the corresponding ROI templates. For now, one option is actually not really one: -label label_file where label_file is a tracer-dependent transformation file that indicates which of the 90+ regions identified in the ICBM atlas need to be merged into functionnally distinct tissue components. An exemple of such file is given below in the case of the tracer [18F]fluoro-L-dopa:

    12 12 white
    10 10 gray
    21 21 caudate
    23 23 putamen



    This table may contain a restrictred number of tissue components as shown in this exemple, or can be a translation label file that lists all existing labels in Talairach space, with their corresponding value in the brain phantom (i.e., labeled MRI in PET native space). You can see another exemple of such translation file in /usr/local/mni/data/PVCcorrect/standard_labels.dat You should follow this exemple to create your own tracer-dependent label translation file. N.B. Non-active tissues such as bone or CSF should NOT be included in the label file. Only radio-active tissue components should be included, as non-active tissues will only introduce additional noise to corrected time-activity curves. Only the 2nd collumn is relevent to the program and has to correspond to the actual tissue code in the segmented volume (labels_inPET.mnc).

    <olivier@bic.mni.mcgill.ca>
    last updated: Dec. 3, 1998