Thanks for your response. In general, is it better practice to run the first level analysis in the native space, and then resample the subjects to fsaverage space during concatenation, or to resample the subjects to fsaverage space during first level analysis?
Best, Anais ________________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu [freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Douglas N Greve [greve@nmr.mgh.harvard.edu] Sent: Tuesday, November 03, 2015 12:50 PM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] First-level analysis (native vs fsaverage)
In the native space analysis the data get smoothed in the 3d volume and then projecting to the surface. This is very different from a surface-based analysis where you project to the surface and then smooth in 2D. For one thing, 5mm FWHM in one analysis is not the same as 5mm FWHM in the other. You get much more smoothing in the volume at 5mm than you do on the surface. Why one is better than the other is hard to say. You might try cranking up the smoothing for the surface-based analysis.
doug
On 11/02/2015 04:29 PM, Rodriguez-Thompson, Anais wrote:
Hi freesurfer experts,
I have a question about how running first-level analyses in either native space or fsaverage space changes activation. My original analysis was a per-session analysis run on native space. The command used to make the analysis was
ANALYSIS 1: mkanalysis-sess -analysis SIRP_LoadRegression_Stable5_050514 -TR 2 -paradigm slopepar -event-related -runlistfile runlistfile -tpef tpef_1.5mm.txt -native -fwhm 5 -timewindow 20 -TER 2 -nconditions 9 -gammafit 2.25 1.25 -refeventdur 2
However, my lab was interested in running our analyses -per-run, so I created a new per-run analysis on the fsaverage brain using the command:
ANALYSIS 2: mkanalysis-sess -analysis SIRP_LoadRegression_Stable5_072915_lh -TR 2 -paradigm slopepar -event-related -runlistfile runlistfile -tpef tpef_1.5mm.txt -surface fsaverage lh -fwhm 5 -per-run -timewindow 20 -TER 2 -nconditions 9 -gammafit 2.25 1.25 -refevendur 2 -fsd bold
The group maps for the two looked comparable. However, I covaried those maps with performance on our task. The maps looked significantly different, and while 9 cluster survived MC correction from analysis 1 (3 LH, 6 RH), only 3 survived MC correction from analysis 2 (0 LH, 2 RH).
I thought that the difference could have been due to per-run vs. per-session MC, but I created a test analysis (per-session, fsaverage space), and this analysis looked more like the 2nd analysis than the 1st, leading me to believe that the difference in activation was due to running the first-level analysis on fsaverage space rather than native space.
The command I used for the test analysis was: mkanalysis-sess -analysis TEST_fsaverage_persession_lh -TR 2 -paradigm slopepar -surface fsaverage lh -event-related -fsd bold -runlistfile runlistfile -tpef tpef_1.5mm.txt -timewindow 20 -TER 2 -nconditions 9 -gammafit 2.25 1.25 -refeventdur 2 -fwhm 5 -per-session
Do you know why running first-level analyses on fsaverage space rather than native space would decrease activation effects, leading to fewer significant clusters after MC correction?
I've attached a pdf file with the maps in question.
Thanks, Anais
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 www.nmr.mgh.harvard.edu/facility/filedrop/index.html Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer