Extracting timeseries using custom ROI on CPAC

Software Support
1

Hello everyone,

I’m using C-PAC (v1.8.7.dev1) for pre-processing rs-fMRI data. I modified the pipeline configuration file to extract the mean timeseries of custom ROIs as follows:

timeseries_extraction:
  run: On
  tse_roi_paths:
    /gpfs/scratch/martil42/ost_fix/masks/ROI_dACC_7.nii: Avg
    /gpfs/scratch/martil42/ost_fix/masks/ROI_DefaultModeNetwork.nii: Avg
    /gpfs/scratch/martil42/ost_fix/masks/ROI_Frontoparietal.nii: Avg
    /gpfs/scratch/martil42/ost_fix/masks/ROI_Limbic.nii: Avg
  connectivity_matrix:
    measure:
      - Pearson
      - Partial

However, the output creates Mean_timeseries.1D files that have only one column (#Mean_1) instead of four (one for each ROI). What am I missing?

Thanks in advance!!!

~L

2

Hi @luisagulleiro,

Thank you for reaching out! Could you clarify how many ROIs are in each atlas you are using? You should expect to get one 1D file per atlas, and a column per ROI (within those atlases).

3

Hi @tamsinrogers,

Instead of atlases, I’m using custom ROI masks (for dACC, DMN, FP, LIM), so I was expecting one 1D file for each mask, with one column per file.

4

Hi @luisagulleiro , can you confirm that you see four Mean_timeseries.1D files as your output, each with one column?

5

Thanks @tamsinrogers! I can only see two Mean_timeseries.1D files in the output (one with GSR, another without GSR), with only one column inside each file. An example for one subject:

sub-678313_ses-post_task-fix_acq-normal_run-01_atlas-ROI_space-MNI152NLin6ASym_reg-default_desc-Mean_timeseries.1D

sub-678313_ses-post_task-fix_acq-normal_run-01_atlas-ROI_space-MNI152NLin6ASym_reg-defaultNoGSR_desc-Mean_timeseries.1D

6

Hi @luisagulleiro ,

Thanks for the additional info! Can you try renaming your atlases to:

  • /gpfs/scratch/martil42/ost_fix/masks/ROI-dACC-7.nii: Avg
  • /gpfs/scratch/martil42/ost_fix/masks/ROI-DefaultModeNetwork.nii: Avg
  • /gpfs/scratch/martil42/ost_fix/masks/ROI-Frontoparietal.nii: Avg
  • /gpfs/scratch/martil42/ost_fix/masks/ROI-Limbic.nii: Avg

It’s possible that the naming of your atlas files is getting misinterpreted by the C-PAC’s BIDS parsing.
Your outputs have the atlas-ROI tag but not the rest of the atlas names - maybe the _ underscore in your atlas names is causing the names to be split, so CPAC sees it as all the same file?

Let me know if that gives you a different result!

7

Thanks a lot, @tamsinrogers! Just to be safe, I modified the names of the masks to dACC7.nii, DMN.nii, FP.nii and LIM.nii and everything work perfectly (it creates 4 Mean_timeseries.1D files with one column each). An example of the files:

sub-678301_ses-post_task-fix_acq-normal_run-01_atlas-dACC7_space-MNI152NLin6ASym_reg-default_desc-Mean_timeseries.1D

Thanks again!

8

@luisagulleiro glad to hear it!! Let me know if there are any other issues you come across.