Paradigm files are already distributed with Bert's data, but you normally would have to create your own, using any text editor. The paradigm files simply describe what was going on during your study experiment with respect to time of presentation and stimulus type. The following will explain how Bert's paradigm files were created. The methodology of the experiment is described in detail in References/Wagner2001 (Note that Bert's data were not part of Wagner et al's study, neither was that study processed with FsFast. Russ Poldrack of the MartinosCenter generously reran the experiment using a subject who agreed to provide data, specifically for this tutorial.)

The experiment involved an associative memory task. The subject is shown one of four stimulus types. Each stimulus has five words in the following pattern:

                                     targetword
                              testword1      testword3
                              testword2      testword4

The subject is asked to choose which test word is most related to the target word.

By design, the target word is either highly related or loosely related to one of the test words.

Sometimes there are four test words presented, and sometimes there are two test words presented along with two X's.

This makes four conditions:

1. H2 - highly related, two words/two X's (used in run 006)
2. H4 - highly related, four words (used in run 007)
3. L2 - loosely related, two words/two X's (used in run 008)
4. L4 - loosely related, four words (used in run 009)

This number coding is used in the paradigm files to identify which stimulus was presented when. The contrast (L2+L4) - (H2-H4) should show activity in the left frontal region.

Based on your study description, you could create a paradigm file for each functional run, using any text editor. For this tutorial, the paradigm files are already provided in the subdirectory called parfiles:

    - sem_assoc-s001-r001.dat (for functional run 007)
    - sem_assoc-s001-r002.dat (for functional run 008)
    - sem_assoc-s001-r003.dat (for functional run 009)

IMPORTANT: For any subject whose functional data you process, a copy of each run's paradigm files should be placed into the directory containing that functional run's bshort files. All paradigm files for all runs averaged together should have the same name, even though their contents may differ. In Bert's case, we will call the copied paradigm files "sem_assoc.par".

To copy Bert's paradigm files into the three functional run directories, cd to bold/ and type:

cp parfiles/sem_assoc-s001-r001.dat 007/sem_assoc.par
cp parfiles/sem_assoc-s001-r002.dat 008/sem_assoc.par
cp parfiles/sem_assoc-s001-r003.dat 009/sem_assoc.par