Hi,
My lab encountered an error when trying to run ICA-AROMA using DPABISurf (MATLAB V2024a; DPABI V8.2_240510; DPABISurf V3.2).
For your reference, I’ve copied the error code below.
“””
Error using DPABISurf_run (line 1360)
Index exceeds the number of array elements. Index must not
exceed 0.
“””
In the GUI, we selected both ICA-AROMA and ICA-AROMA Noise Regression. However, no ICA-AROMA-related files were created, even though the fMRIPrep pipeline had completed. We think the error arose because the fMRIPrep (V23.2.2) version in the docker no longer runs ICA-AROMA (https://fmriprep.org/en/stable/changes.html).
It was previously suggested to run fMRIPrep twice, once using an older version to get ICA-AROMA time-courses, and once again using a newer version (https://neurostars.org/t/using-ica-aroma-after-preprocessing/27442/13). We also saw that we might be able to run ICA-AROMA post fMRIPrep (i.e., fMRIPost-AROMA) (https://github.com/nipreps/fmripost-aroma) if fMRIPrep is run with MNI152NLin6Asym:res-2 derivatives (https://fmripost-aroma.readthedocs.io/stable/usage.html).
We were wondering if you had any other insight on why this may have happened. Furthermore, what steps would you recommend we take to resolve this problem?
Additionally, when we skip the ICA-AROMA regression step, we encounter a problem where the _res-2 automask files cannot be read. It seems the script does not expect the "_res-2" suffix in the filename, causing the process to fail. We considered breaking the pipeline to manually modify the automask filenames before they are read. However, we would like to know if there is a more efficient way to address this filename issue?
Thank you for your help!
Best,
Hsiang-Yuan Lin