It is hard to say exactly what the difference is, but in the command below, you set the high-pass cutoff to be 128 Hz meaning that it will try to filter out everything below 128Hz, which is undoubtedly where your signal is. Probably you meant to use something like -hcp .01 or so

On 9/27/2019 11:06 AM, Zhi Li wrote:

External Email - Use Caution

Dear Freesurfer experts,

May I ask what's the difference between FSFAST and SPM in the 1st level GLM? Now I use fsfast to pre-process my task-based imaging data, then model the data with FSFAST and SPM respectively. I expected to get very similar results from FSFAST and SPM in the whole brain volume-based analysis. Although there is some overlap between FSFAST and SPM on the whole brain activation, however, there is a lot of difference. Activation in some regions based on SPM was not found in that based on FSFAST. Please see below for my mkanalysis information:

mkanalysis-sess \ -fsd Num -mni305 2 -fwhm 8 \ -event-related -paradigm Num.par -nconditions 18 \ -spmhrf 0 -TR 2 -refeventdur 16 -nskip 0 -polyfit 0 \ -hpf 128 -mcextreg \ -analysis Num.sm08.mni305 -per-run -force

Any suggestions?

In addition, I am not clear if I convert the volume data from FSFAST to SPM in a right way. I replace the mask of subcortical mask with the whole brain mask. Then I convert the pre-processed data from mni305 to mni152 and model it with SPM. However, there is a lot of loss in the prefrontal and subcortical regions. Seems like the whole brain data shrinkage in SPM (please see attached figure).

Looking forward to your kind reply.

Thanks and best,

Li Zhi

_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer