Issues selecting structual or functional images via GUI

[flostor96]

Hi,

thanks for your great toolbox! It is precisely what I've been looking for; however, I'm having issues with the selection of the structural and functional images for denoising.

I have a unique data structure where the "prepro_SID" folder contains a subfolders for each subject. That subject folder (e.g.; sub-001mr01_ses-01) contains the the original as well as processed structural (e.g.,0249rs-001-t1.nii) and functional data.
As such, I selected the parent folder "prepro_SID" in the first row, selected the folder of the first subject in the second row, and even with the most general filter, I don't get any results and the following error:

No images matched the filter ".*.nii" within "/biondata/0249RS/work/auswertung/prepro_SID/sub-001mr01_ses-01". Please check the root/subfolder and filter.

I'm not sure if I'm doing anything wrong or whether this is a bug.
My workaround by selecting the files manually only worked for the structural images, but not for the functional ones.

I'm running TMFC denoise v1.4.1 and SPM 12 v6685.

[Masharipov]

@flostor96
Hi,

Thank you for trying TMFC_denoise and reporting this issue. I've updated the TMFC_denoise.m file.

If I understand correctly, your data structure looks like this:

prepro_SID/
└── sub-001mr01-ses-01/
        ├── 0249rs-001-t1.nii          # unprocessed T1
        ├── <functional *.nii files>   # preprocessed and unprocessed
        └── GLM_subfolder_name/
                 └── SPM.mat           # first-level GLM 

Could you please confirm that the SPM.mat file is stored inside
prepro_SID/sub-001mr01-ses-01/GLM_subfolder_name/?

Can you also share:

• The full path to your first subject’s SPM.mat file
• An example filename of your unsmoothed functional images (first file for the fisrt subject)

Please try the following:

1. Download the updated file: https://github.com/IHB-IBR-department/TMFC_denoise/blob/main/TMFC_denoise.m
2. During Subject selection, select your folders: sub-001mr01-ses-01, sub-002mr01-ses-01, etc.
3. In the file filter, use *.nii (not .*.nii).
   You can also try a more specific filter such as *t1.nii for structural files.

Best,
Ruslan

[flostor96]

Hi Ruslan,

Thanks for your swift response! Unfortunately, my data structure is not yet the best, looking like this:

analysis/
└── prepro_SID/
        └── sub-001mr01-ses-01/
                ├── 0249rs-001-t1.nii          # unprocessed T1
                ├── <functional *.nii files>   # preprocessed and unprocessed
└──flevel_SID/
        └──model02/
                        └── sub-001mr01-ses-01_firstlevel/
                                 └── SPM.mat           # first-level GLM 

Could you please confirm that the SPM.mat file is stored inside prepro_SID/sub-001mr01-ses-01/GLM_subfolder_name/?

As such, this is not the case. The SPM.mat is in analysis/flevel_SID/model02/sub-001mr01-ses-01_firstlevel/

The unsmoothed functional images are named like this: wau0249rs-002_SID_epi.nii

I tried the updated file, but still no success. I was able to get the denoising running using a MATLAB script s based on your example:
I'm not sure if it was intended, but the code only worked when I adjusted line 73 from
func_paths.fname{iScan,1} = fullfile(data_path,['sub-' sprintf('%02d', iSub)], ...
to
func_paths{iSub}.fname{iScan,1} = fullfile(data_path,['sub-' sprintf('%02d', iSub)], ....

Thanks so much for the immense effort you put into this project!

Best,
Florian

[Masharipov]

@flostor96

Hi Florian,

Thanks for the detailed follow-up!

You were absolutely right about the indexing.
The correct fix is:

func_paths(iSub).fname{iScan,1}

instead of

func_paths.fname{iScan,1}

(you used {} in your message, but it should actually be ()).

I also recognized the cause of the automatic file selection issue — it occurs when subject folders in ANAT and STAT roots have slightly different names, e.g.
sub-001mr01-ses-01 vs. sub-001mr01-ses-01_firstlevel.

I’ll soon release an update to both tmfc_select_anat_GUI and tmfc_select_func_GUI to handle such prefix/suffix differences automatically.

Thanks again for catching this and for your helpful feedback!

Best,
Ruslan

[Masharipov]

Hi @flostor96,

The new release TMFC_denoise v1.4.3 (and TMFC toolbox v1.8.3) is now available — it includes major improvements in the GUI for both structural and functional file selection (tmfc_select_anat_GUI and tmfc_select_func_GUI).
These updates should fully address the issues you previously encountered with folder-name differences and automatic path detection.

Could you please try the new TMFC denoise GUI and let me know if the selection now works as expected?

When testing functional file selection, you can try several approaches (all should now work):

1. Simply enter the correct text filter (e.g. war.nii) and press Apply filter.
2. Change only the parent folder (FUNC root) and apply the filter again.
3. Change both the parent folder and the functional subfolder, then apply the filter.

Your feedback would be very helpful in confirming that the issue is fully resolved in this version.

Best,
Ruslan