Multi-level voxel-wise covariate

AFNI Message Board
1

AFNI version info (afni -ver): 24.1.15

Hi,

I am trying to run a mixed-effects model with a multi-level covariate. Specifically, one of my factors (emotion) has 3 levels, which apply to both my outcome measure and my voxel-wise covariate (since I am using time point 2 as outcome measure and time point 1 as a voxel-wise covariate). I tried to code it in 3dLME as below, but I get an error:

3dLME \\
			-prefix followup_3dLME_ROI \\
			-resid followup_3dLME_ROI_resid \\
			-jobs 4 \\
			-mask ROI+tlrc \\
			-model "Group* Drug* Emotion+Age+Sex+Basecov" \\
			-ranEff "~1" \\
			-qVars "Age" \\
			-vVars "Basecov" \\
			-SS_type 3 \\
			-dataTable @LME_dataTable.txt

Error in Stat[, , kk, ] <- aperm(parApply(cl, inData[, , kk, ], c(1, 2), : number of items to replace is not a multiple of replacement length Execution halted

If I remove "Basecov" from the -model and -vVars arguments, the function runs fine.

How can I run this analysis in 3dLME or another AFNI program, please?

Many thanks,
Paulina

2

Paulina,

What are the primary effects of interest in this analysis? Are Group, Drug, Age, and Sex considered as between-individual variables, while Emotion and Basecov are within-individual variables? Additionally, does each participant have three separate 3D input files for the Basecov variable?

Gang Chen

3

Hi Gang,

Many thanks for your reply!

Primary effects of interest here are main effects of Group & Drug and the Group:Drug interaction.

Between-subject variables are Group, Drug, Age & Sex; within-subject variable is Emotion which applies to the outcome measure and the voxel-wise covariate Basecov.

Each participant has three separate 3D input files for the Basecov variable.

Best wishes,
Paulina

4

I emailed you earlier requesting additional details about the analysis.