Analyzing the Data

To describe the steps in fMRI data analysis, we will consider a hypothetical session performed on subject entered on the scanner as study-subj001 on Oct. 19, 2003 at eleven o'clock in the morning.

Note: Please take note that we now require all scanned participants to have an anonymous codified ID which we'll use to enter on the scanner console. We will not use participant's real names, birthdays or any other identifying information.

Let's say that this session consisted of:

• a localizer scan
• a T1-weighted anatomic scan
• two  functional scanning runs with 120 time frames and 25 slices each; block design:
  
  
  1. three frames rest, three frames hot stimulus,
  2. three frames rest, three frames warm stimulus, repeated 10 times

Finding and Transferring the Data

• study_subj001_20031019_111740
  

  The significance of the different fields in the file name is as follows:

  1. {study_subj001} is the subject's name.
     
     
  2. {_20031019_111740} is the date and time of the session.
     
     

• study_subj001_20031019_111740_1_mri.mnc.gz
  
  1. {_1_mri} the '1' indicates that this is the first scan performed (it is usually the localizer. You generally don't need to keep this scan, but it is transferred anyway.
     
     
  2. {.mnc} indicates that this is a MINC file
     
     
  3. {.gz} indicates that the file has been compressed using the `gzip' program.
     
     
  Here is a link to a sample image of what a localizer image looks like. Note the lack of fine anatomical detail.
  
  
• study_subj001_20031019_111740_2_mri.mnc.gz

  The field {_2_mri} indicates that this is the second scan of the MRI session, which is usually the anatomical scan, the T1-weighted anatomical scan. We will use this scan to interpret the functional data in an anatomical context. The file contains all of the slices of this 3D cross-sectional data set (usually 160-180 sagittal slices). The following is an axial slice of a 1mm x 1mm anatomical image. Note the increase in resolution and detail compared with the localizer.

• study_subj001_20031019_111740_3_mri.mnc.gz

  The next scan was the first functional run. These files contain multiple volumetric scans - one for each time point in a functional scanning run. These are the files which will be processed to produce activation maps. Here is a sample image of a slice of a functional image set. Note how blurry the image appears. This particular image has been processed with the motion correction and blurring algorithm.

You may have many MINC images in your transfer directory and it may be difficult to determine which imaging sequence the files correspond to. Rest assured that the protocol information is always described in the MINC header. You can run the following command for a quick listing of the protocol name for each of your image files:

~mferre/bin/lsminc <input_files>

Finding the analysis software

You can access the fmristat software directly in MATLAB if you are logged into a BIC machine. It is not necessary to add any path to your MATLAB environment - simply type 'matlab' at the command line to start the program.

You will only need a basic understanding of Matlab to run fmristat. There are many online guides and tutorials for Matlab. The official Mathworks documentation site may be a good starting point.

Setting up a Temporary Working Directory

mkdir /data/scratch/scratch1/yourname

mkdir /data/scratch/scratch1/yourname/study_subj001

cd /data/scratch/scratch1/yourname/study_subj001

cp /data/transfer/minc/study_subj001_20031019_111740 .

Note: The /data/scratch directories are erased on a weekly and monthly basis. Please contact the BIC network administrators for more information.

Preprocessing the Data

During dynamic data acquisition the subject may execute slight, transient head movements due to breathing, swallowing, or whatever functional task the subject was asked to perform. Their head position may also drift over a longer time scale. Such motion can cause false intensity changes in the dynamic data which can either obscure or mimic true activation induced signal changes. We can correct the data, if the movements are less than 1 mm or 1 degree, by registering all image frames in each dynamic file to one single target in that run. Frames with movements larger than 1mm or 1 degree may have to be excluded from the analysis - more information on this topic in the subsequent section on fmrilm.

Preprocessing is performed using the program fmr_preprocess, which also low-pass filters, or blurs, the dynamic data (blurring is the default but it can be disabled). Low-pass filtering increases the signal-to-noise ratio of the data, increases the tolerance of the subsequent analysis steps to residual motion in the scans, and minimizes resampling artifacts.

We would preprocess each dynamic run one at a time, by going to the directory /data/scratch/scratch1/yourname/subj001_study/dynamic and typing the command 'fmr_preprocess' on the Unix command line, followed by the list of options you want to process and the file name. You can write for example:

fmr_preprocess  -fwhm 6 -target 3 study_subj001_20031019_111740_3_mri.mnc.gz

This will blur your image with a 6 mm FWHM Gaussian filter and align all the frames to the third frame in the first run. The new files that are created have the suffix  `_MC'  (for Motion Corrected).
For example:

• study_subj001_20031019_111740_3_mri_MC.mnc
• study_subj001_20031019_111740_4_mri_MC.mnc

Note that the output files are not compressed, so they don't have the .gz extension. Make sure you have enough space in the directory where the motion-corrected images will be written.

All 120 time frames in both scanning runs will be registered by default to the third frame in the respective run. Each time frame will also, by default, be blurred with a 6 mm FWHM Gaussian kernel. If you would like use a different blurring value, you just have to change the -fwhm parameter in the call to fmr_preprocess. If you would like to register your frames to a different target frame, simply specify the target number in the -target option.