Generalized fMRI Analysis

The Triumph PET/CT scanner provides the capability to scan one or more small subjects while providing co-registered CT images for anatomical reference. This section describes an analysis stream for co-registering whole-brain images to standardized neuroanatomical spaces.


Analysis proceeds from study data as it is originally saved on the scanner. Scripts expect data from CT and PET to be separated into different directories. The name of the CT and PET directories are initially defined by an initialization file on the first invocation of the script, and then they can be changed. All session-specific and subject-specific data are stored on local files and retrieved by the scripts. A typical CT directory (default name: ct) will contain a series of 224 DICOM images (1 for each slice). A typical PET directory (name like "output.job2.BQML.i16”) will contain a series of DICOM files separated into slices and time points. Suppose that these 2 subdirectories on stored a directory named “fdg-study1”. Go there and invoke the analysis script, which is assumed to be an alias named “mpet” below. Output is shown below.

    linux> cd fdg-study1

    linux> mpet

    -----------------------

    settings for group

    -----------------------

    1) CT  Input Directory   : ct

    2) CT 1st file name       : CTImage_0001.dcm

    3) PET Input Directory   : output.job2.BQML.i16

    4) PET 1st file name     : S_00001

    5) Species                     : mouse

    6) counts/sec to Bq/ml   : 93.0843

    7) Image range for SUV : 25-n

    8) Spatial smoothing      : 0

   -----------------------

   actions on group

   -----------------------

    c) convert group CT

    p) convert group PET

  Input selection [h = help, q = quit, s = subject] >



Joseph B. Mandeville, Athinoula A. Martinos Center for Biomedical Imaging at MGH/MIT/Harvard