3dLME script error data.frame: undefined columns selected

AFNI Message Board
1

Hi,

I am trying to run a 3dLME script (simple ANOVA) with my variable for comparison being ‘latent class.’ I keep getting the following error:

Error in [.data.frame(lop$dataStr, , jj) : undefined columns selected
Calls: process.LME.opts → [ → [.data.frame
Execution halted

My script is as follows:

3dLME -prefix test -jobs 1
-model “latent_class+bl_completed_timestamp+Sex_best+Phil_IMP_PDS+IMP_handedness+Site_1+Site_2+site_3+site_4+site_5+site_6+site_7+IMP_IQ_PR+IMP_IQ_VC+IMP_Phil_totalGMV.14”
-qVars “bl_completed_timestamp,Phil_IMP_PDS,IMP_IQ_PR,IMP_IQ_VC,IMP_Phil_totalGMV.14”
-ranEff ’~1’
-SS_type 3
-dataTable
Subj bl_completed_timestamp Sex_best Phil_IMP_PDS IMP_handedness Site_1 Site_2 site_3 site_4 site_5 site_6 site_7 IMP_IQ_PR IMP_IQ_VC IMP_Phil_totalGMV.14 latent_class InputFile
000000022453 5351 0 2.2 1 0 0 0 0 0 0 0 114 116 769991 1 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000022453.nii
000000075717 5350 0 3 1 0 1 0 0 0 0 0 100 109 791158 3 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000075717.nii
000000106601 5060 0 2.4 0 0 0 0 0 0 0 1 125 105 772073 3 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000106601.nii
000000106871 5608 1 2.8 1 0 0 0 0 0 0 0 110 99 729178 3 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000106871.nii
000000112288 5258 0 2.4 1 0 0 0 0 0 0 1 112 99 720459 4 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000112288.nii
000000240546 5285 0 2.8 1 0 0 0 1 0 0 0 131 140 808224 1 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000240546.nii
000000297685 5275 1 3.8 1 0 1 0 0 0 0 0 104 87 697298 1 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000297685.nii
000000308867 5172 0 2.8 1 0 1 0 0 0 0 0 92 105 743821 2 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000308867.nii
000000397377 5299 1 2.6 1 0 0 0 1 0 0 0 104 115 580802 3 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000397377.nii
000000458976 5194 1 3 0 0 0 0 0 0 0 0 116 116 706129 4 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000458976.nii
000000459848 5255 1 2.2 1 0 0 0 0 0 0 1 119 93 717422 3 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000459848.nii
000000469693 5249 1 3.2 1 0 0 0 0 1 0 0 112 117 571626 4 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000469693.nii
000000540905 5251 1 2 1 0 0 0 1 0 0 0 119 142 684633 3 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000540905.nii
000000602062 5166 0 3 1 0 0 0 0 0 0 1 92 109 791640 3 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000602062.nii
000000613223 5345 0 2.6 1 0 0 0 0 0 0 0 125 109 852181 2 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000613223.nii
000000642263 5503 1 2.6 1 0 0 0 0 0 1 0 104 117 718236 2 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000642263.nii
000000663084 5450 1 2.6 0 0 0 1 0 0 0 0 125 124 661362 4 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000663084.nii
000000724314 5128 0 2.6 1 0 0 0 0 0 0 1 117 124 744396 3 /Volumes/Backup_2/Test_Folder/s888mwc1mprage000000724314.nii

Any ideas on how to fix this issue?
~BFair

2

Is “latent_class” a subject-grouping (between-subjects) variable? If so, it would be easier to just use 3dttest++. I hope that you have more subjects than what’s copied in the script here.

3

Hi Gang,

Yes, latent_class is a subject-grouping variable. I do have more subjects…but I wanted to make sure the script would run with fewer subjects before adding the rest. I would like to refrain from using multiple independent T-tests in order to avoid Type I errors. Any idea on how to fix this error to use 3dLME?

Best,
~BFair

4

What are those variables with values of 0 and 1? Are they actually categorical variables (factors)? If so, it’s better to explicitly treat them as factors instead of quantitative variables. In addition, center those real quantitative variables properly:

https://afni.nimh.nih.gov/sscc/gangc/centering.html

Since there are no within-subject variables in your case, 3dLME is not an ideal program for the situation. Instead use 3dMVM.

5

Hi Gang,

Thank you for your response. The variables listed as 0’s and 1’s are categorical (they dictate the site of data collection), thus I did not list them under qVars. Do I need to list them as a factor somehow?

Also, I am only interested in the covariate effect of my quantitative variables, so I believe I do not need to center them.

I thought my quantitative variables were also my within-subject variables. The only variable I am trying to compare between groups is latent_class. The rest of my variables are covariates.

Best,
~Brittany