3dLMEr Error

AFNI Message Board
1

Hello people of AFNI!

I’m trying to run 3dLMEr on some data and I’m getting an error…

The command I’m running is:


3dLMEr -prefix antHippo.Cue \
       -mask group_mask.nii \
       -model 'Run*Learned+(1|Subj)' \
       -qVars 'Run' \
       -gltCode Succ 'Learned : 1*succ' \
       -gltCode Unsucc 'Learned : 1*unsucc' \
       -gltCode LearningGrp 'Learned : 1*succ -1*unsucc' \
       -dataTable @Cue_antHippo.txt

And the error I’m getting is:


Error in apply(comArr[, , kk, ], c(1, 2), runLME, dataframe = lop$dataStr,  : 
  object 'comArr' not found
Calls: aperm -> apply
Execution halted

According to the command output, all of the R packages are loading fine as well as the data table and the .nii input files…

I’m using AFNI version AFNI_21.0.04 ‘Titus’
I’m also using R version 3.6.2
My OS is CentOS Linux

The head of my table file (Cue_antHippo.txt) looks like this (note: there are a bunch of extra columns I’m not modeling at the moment):


Subj Acc RT Run Learned_Run Learned Acc_Total RT_Total InputFile
s04 0.594 631.594 1 no succ 0.875 490.003 sub-s04_ses-1_task-cdrl_run-1_space-MNI152NLin2009cAsym_desc-Cue_antHippo_corrZ.nii
s04 0.906 491.469 2 yes succ 0.875 490.003 sub-s04_ses-1_task-cdrl_run-2_space-MNI152NLin2009cAsym_desc-Cue_antHippo_corrZ.nii

Thanks for your help!

2

Update: I was able to run the model with the older 3dLME command. Not sure why the 3dLMEr wouldn’t work though.

Here is the command that worked:


3dLME -prefix antHippo.Cue \
       -mask group_mask.nii \
       -model "Run*Learned" \
       -ranEff '~1' \
       -qVars 'Run' \
       -num_glt 3 \
       -gltLabel 1 'Succ' -gltCode 1 'Learned : 1*succ' \
       -gltLabel 2 'Unsucc' -gltCode 2 'Learned : 1*unsucc' \
       -gltLabel 3 'LearningGrp' -gltCode 3 'Learned : 1*succ -1*unsucc' \
       -dataTable @Cue_antHippo.txt
3

One possibility is that there was a bug in 3dLMEr for that particular version. Update your AFNI:

@update.afni.binaries -d

and see if the most recent version would resolve the problem.