AFNI Soup to Nuts: How to Analyze Data with AFNI from Start to Finish

There is no single “correct” way to analyze FMRI data. The methods one chooses should be mostly decided before acquiring any data. They depend on what questions one wants to answer, and the resulting experimental design that was chosen.
- However, there are some typical processing steps that are widely used. These steps are introduced and discussed in this handout.
- The sample study used for this hands-on is a real study, although the variable names have been slightly modified, and the data resolution has been lowered.
afni_proc.py
- The data processing script discussed in this handout was generated by a program in AFNI called afni_proc.py.
- Specifically, afni_proc.py is a python script that can generate a single-subject data analysis script by asking the user to provide information regarding their study, such as input datasets and stimulus files that will be used. The program also asks for more specific information, such as the number of TRs to be removed (if any), the EPI volume that will be used to align the remaining volumes, and additional information necessary for the regression or deconvolution analysis that will follow.

There is no single “correct” way to analyze FMRI data. The methods one chooses should be mostly decided before acquiring any data. They depend on what questions one wants to answer, and the resulting experimental design that was chosen.