When data arrives on rumba7 from the console you need to first scp that data from rumba7 to psychology, and then scp the data to ravana

Then go into the directory where the data is located and run dicomsort, which is located on ravana in /usr/local/bin/dicomsort 
Dicomsort will sort the MRI data into the particular series numbers.  For example 8241 is the first sequence run, 8242 is the next sequence, etc.

Then you need to concatenate the data into a 4d image using dicomcat.
/usr/bin/dicomcat directory series 
example: dicomcat 8241 8241
This will produce both a *.hdr file and a *.img file.

Then you should run a proportional threshold on the mprage hdr file (or whatever anatomical file you choose to use.)  This will take a specified percentage of voxels that have the highest values and set them to the highest voxel value in the lower percentage. 
The usage here is: 
proportionalThreshold -i infile -o outfile -p|--proportion proportion 
proportion is the proportion of voxels that are NOT clipped.
For example, if I want to remove the top 1% of voxels, I would use proportionalThreshold -i mprage.hdr -o mprage_prop.hdr  -p .99

Then you'll need to BET the mprage. 
--Open FSL.  
--Click on BET Brain Extraction.  
--Under input image, choose your 4d mprage.hdr file that you want to be bet'd. You should choose the mprage that has already been prop'd so you'll end up with an mprage that has been prop'd and bet'd.  
You can read up on the BET options on fsl's website, but the default settings here usually work pretty well.
The BOLD data will be bet'd when you run the FEAT analysis.

Now, you're ready to run an FSL analysis on your data.
--Open FSL.
--Click on FEAT fMRI Analysis.  This will open a FEAT window.
--First choose whether you want to run a First-Level Analysis or a Higher Level Analysis.  
--Then click on Select 4D data and choose the BOLD.hdr data file that you want to run your analysis on.  
The total volumes will be set automatically from the data file that you select, but **the TR must be set manually** 

It's also a good idea to set the output directory.
If you leave this area blank, fsl will keep the same file name as the input file with a .feat extension.
If you run more than one analysis, feat will name the additional files with a + at the end.  Each additional file will get an additional +.  This can get extremely confusing, especially if you are running many analyses trying different things.  It's best to manually set the output file name being as descriptive as you can.

--Under the Pre-stats, Stats and Post-stats tabs, you can choose which settings and thresholds to apply to your data.
You can read more about these settings at fsl's website at http://www.fmrib.ox.ac.uk/fsl/feat5/index.html

--Under the Registration tab, click the box for Main Structural Image. 
In this window, choose the mprage_prop_bet.hdr file that you created earlier.
The Standard space file should already be set.  
For the degrees of freedom to use during the registration, I feel that its best to get the best registration possible using the least degrees of freedom.
You may need to play with this setting a bit.
For our analyses, 3 DOF in the Main Structural Image and 6 DOF in the Standard space tend to work best.  

When you've set all of you parameters, click go.  Featwatcher should open and you can watch the progress of your analysis.  When it's finished, the location of the output data file will be displayed.  Open a mozilla window and explore your results.