ME preprocessing...error?

cltrtech · September 24, 2026, 7:02pm

AFNI version info: Precompiled binary linux_ubuntu_16_64: Feb 24 2026 (Version AFNI_26.0.09 'Pupienus Maximus')

Hey afni wizards,

I'm testing some ME preprocessing options and keep running into the oddest message before the processing halts. I'm not sure if there's an issue or not. Some months ago, I ran a pipeline that (I think) didn't seem drastically different on a separate and the processing completed with the usual completion messages (APQ webpages ready for review). This time around, not so much for some reason. To preface my comments, when I first ran the afni_proc.py script below, I received echo run difference errors and resolved them (started with [95, 94, 94] and ended with [94, 94, 94]):

afni_proc.py                                                                 \
    -subj_id                $subj.me                                 \
    -dsets_me_echo          $data_dir/short_me1+orig.HEAD                    \
    -dsets_me_echo          $data_dir/me2+orig.HEAD                          \
    -dsets_me_echo          $data_dir/me3+orig.HEAD                          \
    -echo_times             0.015 0.04408 0.07316                            \
    -reg_echo               2                                                \
    -copy_anat              $data_dir/mpraget1+orig                          \

    -blocks                 tshift align tlrc volreg blur mask combine       \
                            regress                                          \
    -blip_forward_dset      $data_dir/blip_forward_sbref+orig.HEAD           \
    -blip_reverse_dset      $data_dir/blip_reverse_sbref+orig.HEAD           \
    -volreg_align_to        MIN_OUTLIER                                      \
    -volreg_align_e2a                                                        \
    -blur_size              3.25                                             \
    -mask_epi_anat          yes                                              \
    -combine_method         OC                                               \
    -html_review_style      pythonic
    
# -execute

it seems to fail in the combine step. Removing regress doesn't really change the surprising halt. Below is the end of the output.proc file (no regress):

@compute_OC_weights -echo_times 0.015 0.04408 0.07316 -echo_dsets pb04.Ext.nlsbref_me.r01.e01.blur+orig.HEAD pb04.Ext.nlsbref_me.r01.e02.blur+orig.HEAD pb04.Ext.nlsbref_me.r01.e03.blur+orig.HEAD -prefix oc.weights.r01 -work_dir oc.work.r01
++ times seem to be in seconds, will scale to ms

++ num echo dsets = 3
++ echo times = 15 44.08 73.16
++ weight prefix = oc.weights.r01
++ combine prefix =
++ work_dir = oc.work.r01
++ T2* limit = 300
++ NT = 94

-- have 3 echoes
++ 3dTstat: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: KR Hammett & RW Cox
++ 3dTstat: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: KR Hammett & RW Cox
++ 3dTstat: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: KR Hammett & RW Cox
++ 3dcalc: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: A cast of thousands
++ 3dcalc: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: A cast of thousands
++ 3dcalc: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: A cast of thousands
++ 3dTcat: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ elapsed time = 0.7 s
[7m*+ WARNING:[0m no -stim_label given for stim #1 ==> label = 'Stim#1'
++ 3dDeconvolve: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: B. Douglas Ward, et al.
++ loading dataset s3.tcat.logs+orig
++ STAT automask has 738906 voxels (out of 1296000 = 57.0%)
++ Skipping check for initial transients
[7m*+ WARNING:[0m Input polort=0; Longest run=3.0 s; Recommended minimum polort=1
++ Number of time points: 3 (no censoring)

Number of parameters: 2 [1 baseline ; 1 signal]
++ Memory required for output bricks = 31,104,000 bytes (about 31 million)
++ Wrote matrix values to file x.xmat.1D

++ ========= Things you can do with the matrix file =========
++ (a) Linear regression with ARMA(1,1) modeling of serial correlation:

3dREMLfit -matrix x.xmat.1D -input s3.tcat.logs+orig
-Rbuck s4.bucket_REML -Rvar s4.bucket_REMLvar -Rfitts s4.fitts_REML -verb

++ N.B.: 3dREMLfit command above written to file s4.REML_cmd
++ (b) Visualization/analysis of the matrix via ExamineXmat.R
++ (c) Synthesis of sub-model datasets using 3dSynthesize
++ ==========================================================
++ ----- Signal+Baseline matrix condition (3x2): 1.99128 ++ VERY GOOD ++
++ ----- Signal-only matrix condition (3x1): 1 ++ VERY GOOD ++
++ ----- Baseline-only matrix condition (3x1): 1 ++ VERY GOOD ++
++ ----- polort-only matrix condition (3x1): 1 ++ VERY GOOD ++
++ +++++ Matrix inverse average error = 1.28977e-14 ++ VERY GOOD ++
++ Matrix setup time = 0.00 s
++ Calculations starting; elapsed time=0.437
++ voxel loop:0123456789.0123456789.0123456789.0123456789.0123456789.
++ Calculations finished; elapsed time=21.474
[7m*+ WARNING:[0m Smallest FDR q [0 Full_Fstat] = 0.985575 ==> few true single voxel detections
++ Wrote bucket dataset into ./s4.bucket+orig.BRIK

created 1 FDR curves in bucket header
++ Wrote 3D+time dataset into ./s4.fitts+orig.BRIK
++ #Flops=8.94242e+07 Average Dot Product=1.84615
++ 3dbucket: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ 3dcalc: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: A cast of thousands
++ 3dcalc: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: A cast of thousands
++ 3dcalc: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: A cast of thousands
++ 3dcalc: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: A cast of thousands
++ Calculating 3D+time[3] dataset from 3D datasets and time series, with dt=1 s
++ 3dTstat: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: KR Hammett & RW Cox
++ 3dcalc: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: A cast of thousands
++ 3dTstat: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: KR Hammett & RW Cox
++ 3dcalc: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: A cast of thousands
++ 3dcalc: AFNI version=AFNI_26.0.09 (Feb 24 2026) [64-bit]
++ Authored by: A cast of thousands
cd: Too many arguments.

Checking @compute_OC_weights — AFNI, SUMA and FATCAT: v26.2.09, it seems to halt in 3dcalc. I tried ru@compute_OC_weight@compute_OC_weightsning @compute_OC_weights:

@compute_OC_weights -echo_times "0.015 0.04408 0.07316" -echo_dsets pb04.Ext.sbref_me.r01.e01.blur+orig.HEAD pb04.Ext.sbref_me.r01.e02.blur+orig.HEAD pb04.Ext.sbref_me.r01.e03.blur+orig.HEAD -prefix oc.weights.r01 -work_dir oc.work.r01

from the results folder and obtained the error so it seems to occur then, but there are many 3dcalc@compute_OC_weightss in @compute_OC_weights ;D so I somewhat gave up with debugging for now. Would anyone be able to help at all?

Best,

Chris

rickr · September 24, 2026, 7:53pm

Hi Chris,

For that sort of 'cd' error, I suspect that you just have some spaces in the current directory name (as shown by pwd). I will add a fix to account for that (at least in that program).

Thanks,

-rick

cltrtech · September 24, 2026, 8:58pm

Hello Rick and thanks for the quick reply!

The folder containing the data and results, and scripts, does in fact contain a space in the directory. After changing the folder name, pre-processing runs as expected! Thanks alot!

I did notice this strange error (bold text) before calling afni_proc.py again:

** ERROR: Nothing to do
-- template = 'TT_N27+tlrc', exists = 1
-- will use min outlier volume as motion base
-- including default: -radial_correlate_blocks regress
-- including default: -find_var_line_blocks tcat
-- tcat: reps is now 94
-- multi-echo data: have 3 echoes across 1 run(s)
++ updating polort to 2, from run len 222.8 s
-- volreg: using base dset vr_base_min_outlier+orig
++ volreg: applying blip/volreg/epi2anat xforms to isotropic 2 mm voxels
-- applying anat warps to 1 dataset(s): mpraget1
++ mask: using epi_anat mask in place of EPI one
-- have 1 ROI dict entries ...
++ will compute regress TSNR stats for dsets: brain
-- no 3dClustSim (since no blur estimation)
-- using default: will not apply EPI Automask
(see 'MASKING NOTE' from the -help for details)

not entirely sure what that means, but the script ran without issue.

Thanks again!!

rickr · September 24, 2026, 9:27pm

I'm glad it's working. That "Nothing to do" error might be coming from an afni_open command, but I'm not sure. Was there any text leading up to that?

Thanks,

-rick

cltrtech · September 28, 2026, 3:25pm

There was no text leading to that. It's prompted before running the output processing script in command line. I haven't checked all of the processing results but nothing glaring jumps out at me so far as it concerns errors

rickr · September 28, 2026, 5:13pm

While that "Nothing to do" error might be coming from afni_open or something with atlas masks, it is not clear, so I will let it go for now with that comment (in case we see it again).

Thanks,

-rick