search

bbfrederick / rapidtide

rapidtide - a suite of programs for doing time lag correlation analysis on fMRI data
Apache License 2.0
75 stars 14 forks source link

`--passes` appears to be ignored #178

Closed tsalo closed 3 weeks ago

tsalo commented 4 weeks ago

Describe the bug I ran the 28andMe dataset through rapidtide with --passes 50, but the runoptions dict shows that passes is set to 3.

To Reproduce

rapidtide \
    /data/sub-01/ses-04/func/sub-01_ses-04_task-rest_space-MNI152NLin6Asym_res-2_desc-preproc_bold.nii.gz \
    /out/sub-01_ses-04_task-rest_space-MNI152NLin6Asym_res-2 \
    --delaymapping \
    --brainmask /data/sub-01/ses-04/func/sub-01_ses-04_task-rest_space-MNI152NLin6Asym_res-2_desc-brain_mask.nii.gz \
    --graymattermask /out/sub-01_ses-04_space-MNI152NLin6Asym_res-2_label-GM_mask.nii.gz \
    --whitemattermask /out/sub-01_ses-04_space-MNI152NLin6Asym_res-2_label-WM_mask.nii.gz \
    --motionfile /data/sub-01/ses-04/func/sub-01_ses-04_task-rest_desc-confounds_timeseries.tsv \
    --similaritymetric hybrid \
    --passes 50 \
    --despecklepasses 50 \
    --filterband lfo \
    --numtozero 3 \
    --datatstep 0.72 \
    --nprocs 10 \
    --noprogressbar

Expected behavior The --delaymapping analysis type shouldn't override individual parameters, so I figured that passes would be set to 50.

Desktop (please complete the following information):

bbfrederick commented 4 weeks ago

Yeah, that's definitely a bug! I'll look at it in a bit.

bbfrederick commented 3 weeks ago

This is fixed in v2.9.9.1. I also put in a check to stop despeckling when it's not changing anything.