Note: this applies to FreeSurfer version 4.5
The paradigm file is a simple text file that describes which stimulus was present when. You will need a separate paradigm file for each run/series in the visit. It contains at least 2 columns and may contain as many as 4. The first column is the onset time of the stimulus relative to the acquisition of the first STORED image in the series (i.e., not including discarded acquisitions or dummy scans). The second column is a numeric ID that codes the condition/event type that was presented at that time. The (optional) third column is the stimulus duration (redundant with column 1). The (optional) fourth column is a name of the condition/event type (redundant with column 2). An example of a paradigm file is given below:
0.0000 2 3.000 Neutral-Long
3.0000 3 3.000 Fearful-Short
6.0000 0 1.000 NULL
7.0000 4 3.000 Fearful-Long
10.0000 0 2.000 NULL
12.0000 1 3.000 Neutral-Short
15.0000 0 3.000 NULL
18.0000 2 3.000 Neutral-Long
There are 5 event types, 4 task-related and 1 fixation (NULL). The Null/Fixation condition is ALWAYS 0. The rest of the conditions must be numbered sequentially starting at 1. These condition codes will be used later in fsfast when constructing contrasts. FSFAST only pays attention to the first 2 columns; the rest are for making them easier to read. The onset time does not need to be a multiple of the TR or linked in any way (e.g., the onset time could be something like 11.340). There are a few cases where you might need to re-code the paradigm file to indicate how long the stimulus lasts, but these are rare.
Create a file for each functional run/series and copy each one into the same directory as your functional data. E.g., if you have four functional runs, then you would create four paradigm files (note that they can all have the same content). When you copy the paradigm files to the functional data, you must give it the same name in each directory. E.g., if the functional data is stored in targetdir/bold/002, 003, 004, 005 and you have four paradigm files emotion1.par, emotion2.par, emotion3.par, emotion4.par, then:
cp emotion1.par argetdir/bold/002/emotion.par
cp emotion2.par argetdir/bold/003/emotion.par
cp emotion3.par argetdir/bold/004/emotion.par
cp emotion4.par argetdir/bold/005/emotion.par
Note that the paradigm file is always called "emotion.par" regardless of the run. There is no conflict because they are in different directories.