Functional neuroimaging is a 4-dimensional modality that requires models to analyze the temporal domain.
Generally, analysis for a single voxel or region of interest is quite rapid, but analyzing all voxels can be a time-consuming task. The philosophy used here is that the display program should provide single time-vector analysis and visualization, and that it should also load pre-analyzed results performed by stand-alone executables.
Each time model requires a specification file that defines the model (see sidebar). The first line of the file indicates the type of the model. Subsequent lines provide key-word delimited options that are specific to the model (for instance, SRTM requires a reference region, but fMRI analysis does not). Following a list of required and/or optional keywords, the keyword “runs” or “files” indicates a new section of the specification file that will begin a list of data files, each potentially with one or files that describe the model and data.
Refer to individual pages on jip-glm, jip-srtm, jip-rem, and jip-fit for more information about standalone executables and the requirements and options associated with each model. Note that each executables will produce a “hidden” file named “.display” that contains the simple line “xd -t [file name]”, so that you bring up the entire model and all associated files by simp lying typing “xd &” within the directory. Of course, you can add other options to this command line in the “.display” file.
1st-level GLM for fMRI: xd -t glm.dat
1st-level GLM for PET using SRTM: xd -t srtm.dat
2nd-level GLM applied to fMRI data: xd -t rem.dat
