 |
 |
Hi, The images you showed last time don't quite make sense with what you are saying here. The one called "fslmaths.png" does not look like it has been spatially smoothed to me. In fact the other image looks like it is spatially smoothed by a lot more than 3mm. So I think this is worth investigating, to check if you do have the same spatial smoothing applied. The fslmaths command looks fine and seems to have the same bptf argument, so that's fine. If you are not adding the mean image back into your result then that would also produce an obvious visual difference. Also, the grand mean scaling (intensity normalisation) looks like a potential difference between your non-GUI processing and what the GUI is doing. I think that if you want to replicate all of these processes (and intensity normalisation is always applied and cannot be turned off), and not just the temporal filtering, then you can probably get things to work by doing the intensity normalisation, the filtering and then adding back the mean image. I hope this helps. All the best, Mark > On 22 Aug 2017, at 16:10, Usman Sheikh <[log in to unmask]> wrote: > > Hi Mark, > > Thanks for your prompt and detailed response. As far as spatial filtering is concerned, in both of these cases, I have applied spatial smoothing with FWHM of 3mm, and that I have done during ICA Aroma (temporal filtering turned OFF). For temporal filtering, for one of them, I used Feat gui (with all other options including spatial filtering turned OFF and a cutoff period of 50s) and for the other, I used fslmaths command (with sigma = 50 / 2 *TR). I was expecting to see the same filtered-output (in fslview) in both of these cases. > > "Are you sure that you are selecting the correct settings for what you want?" If what you are asking is, if I am setting everything correctly in the Feat gui, then you can see a demo design fsf file that I am attaching with this email. It is HPF with everything else turned OFF. > > Also, I tried to search for a similar command in feat report logs, it looks it does this multistage filtering with a few intermediary files like prefiltered_func_data. > > /opt/fsl/fsl-5.0.9/fsl/bin/fslmaths prefiltered_func_data -mas mask prefiltered_func_data_thresh > /opt/fsl/fsl-5.0.9/fsl/bin/fslmaths prefiltered_func_data_thresh -mul 1.02946522499 prefiltered_func_data_intnorm > /opt/fsl/fsl-5.0.9/fsl/bin/fslmaths prefiltered_func_data_intnorm -Tmean tempMean > /opt/fsl/fsl-5.0.9/fsl/bin/fslmaths prefiltered_func_data_intnorm -bptf 29.4117647059 -1 -add tempMean prefiltered_func_data_tempfilt > > Thanks, > Usman. > <run1_hpf_fsllist.fsf>