DPARSFA Functional connectivity fills empty spaces with data

Submitted by vvillalta on Tue, 02/21/2012 - 10:55

Hi all! We have run a Functional Connectivity analysis (FCA) with our own spm scripts and we have also tried to run it with DPARSFA. Some of our images were acquired lacking some of the top brain slices, as you will see in the file attached. DPARSFA is giving a value to these regions lacking when doing the FCA. DPARSFA will not fill this information lacking throughout all the processing (including detrend and filter) but when showing the FCA maps and moreover it has given different values to them. Any suggestion about what it is happening? The file attached shows two pictures at top resulting from our batch FCA and two pictures below showing DPARSFA results. Please see whole brain at FCA map resulting from DPARSFA. Thanks very much for any help in advance! Best, Victoria

Re

Hi! If you look at the original BOLD images, the voxels outside brain had much smaller intensity than the voxels inside the brain. However, the values there are NOT zeros. They were also not fixed across time. Since the intensity of the voxels outside brain varies across time, then they will have a correlation with the seed time course. The correlation was not strong, but they exist. Since DPARSF/REST will not create a mask based on the original BOLD images (which SPM does), that means these week correlations will not be excluded as your own scripts with SPM.

Re

Thanks very much!! V

Forums

Discuss