tag:blogger.com,1999:blog-1405968300258104460.post3569873175924681382..comments2024-03-27T01:18:24.341-04:00Comments on Andy's Brain Blog: AFNI Bootcamp: Day 2Andrew Jahnhttp://www.blogger.com/profile/16435706598096921650noreply@blogger.comBlogger4125tag:blogger.com,1999:blog-1405968300258104460.post-37163830007192420762016-10-15T13:54:25.522-04:002016-10-15T13:54:25.522-04:00Hi Josiane,
I would use 3dFWHMx on the Res.hdr fi...Hi Josiane,<br /><br />I would use 3dFWHMx on the Res.hdr file that is output after SPM's group-level analysis. For example,<br /><br />3dFWHMx -acf -mask mask.hdr ResMS.hdr<br /><br />The -acf option will give you three parameters needed for an accurate cluster correction with 3dClustSim. For example, the output of the command above might be something like the following:<br /><br />0.578615 6.37267 14.402 16.1453<br /><br />The first three numbers are the parameters, and the last number is the average smoothness in the x-, y-, and z-directions. You can then use those numbers in 3dClustSim:<br /><br />3dClustSim -mask mask.hdr -acf 0.579 6.373 14.40 -athr 0.05 -pthr 0.01<br /><br />Note that this command assumes a voxel-wise uncorrected, or "primary" threshold, or 0.01, and looks for cluster sizes that occur less than 5% of the time due to chance. I would recommend using a primary threshold of 0.001 or stricter, to be on the safe side.<br /><br /><br />Hope this helps!<br /><br />-AndyAndrew Jahnhttps://www.blogger.com/profile/16435706598096921650noreply@blogger.comtag:blogger.com,1999:blog-1405968300258104460.post-9915519021365371562016-09-27T16:16:04.057-04:002016-09-27T16:16:04.057-04:00Hi Andrew,
Thank you for your blog, it is really ...Hi Andrew,<br /><br />Thank you for your blog, it is really helpful ! I ran all of my fMRI analyses (preprocessing, 1st-level and 2nd-level) on SPM, however my reviewers suggested that I used 3dClustSim on these fMRI results. I finally get how 3dClustSim works. I took the residual smoothing needed in 3dClustSim directly from SPM with SPM.xVol.FWHM. However, someone recommended me to use 3dFWHMx and not SPM to estimate the residual smoothness. What do you think ? And also, how 3dFWHMx works ? What are the options I need to use excep -acf ? <br /><br />A Huge thanks<br /><br />Josiane BourqueJosiane Bourquenoreply@blogger.comtag:blogger.com,1999:blog-1405968300258104460.post-6117046656113510462016-09-23T11:30:45.969-04:002016-09-23T11:30:45.969-04:00Hey DGM,
I haven't done ClustSim on VBM resul...Hey DGM,<br /><br />I haven't done ClustSim on VBM results, but you should be able to get the nxyz details from the SPM.xVol.DIM field. You should also be able to see the dimensions by opening up the volume in the Image Viewer through the SPM GUI.<br /><br />And yes, those smoothness estimates seem correct. Remember that there is already some smoothness in the data before you apply any additional smoothing to it, and that it is going to be slightly different in the x-, y-, and z-directions. If you want to restrict the amount of smoothness to a specific level, then look into a tool such as AFNI's 3dBlurInMask.<br /><br /><br />Best,<br /><br />-AndyAndrew Jahnhttps://www.blogger.com/profile/16435706598096921650noreply@blogger.comtag:blogger.com,1999:blog-1405968300258104460.post-67951339865800034022016-09-08T12:50:32.029-04:002016-09-08T12:50:32.029-04:00Hi Andrew,
I would like to do a 3ClustStim on VBM ...Hi Andrew,<br />I would like to do a 3ClustStim on VBM results. However I don't really know where to find the -nxyz parameter. Do you know where I can get that? Maybe in the SPM.mat? For -fwhmxyz I used a 8 mm smoothing kernel size, loading a contrast and pushing whole brain I think it's something like 13.7 14.3 13.2 . Does it seem correct?DGMhttps://www.blogger.com/profile/06001983909435744967noreply@blogger.com