3dLMEr gltCode and glfCode

AFNI Message Board
1

Hello!

I am trying to write a script to analyze task data using 3dLMEr. We are modelling one between-subjects factor (Grp: setA / setB), and 2 within-subjects factors (PrevTrial: congruent / incongruent; CurrTrial: congruent/incongruent; this is an emotional Stroop task based on Etkin et al. 2006 https://pubmed.ncbi.nlm.nih.gov/16982430/). We are interested in looking at the difference between the response to incongruent trials depending on if the previous trial was congruent or incongruent (PrevTrial: congruent & CurrTrial: incongruent vs. PrevTrial: incongruent & CurrTrial: incongruent), as well as if this contrast is significantly different between our 2 groups. I was wondering if the gltCode and glfCode arguments here are doing what I was hoping they would?

3dLMEr -prefix emo_stroop -jobs 10															\
	-mask group_mask+tlrc																\
	-model 'Grp*PrevTrial*CurrentTrial+(1|Subj)+(1|Subj:PrevTrial)+(1|Subj:CurrTrial)'								\
	-gltCode con			'CurrTrial : 1*congruent'											\
	-gltCode incon			'CurrTrial : 1*incongruent'											\
	-gltCode incon-con		'CurrTrial : 1*incongruent -1*congruent'									\
	-gltCode low			'PrevTrial : 1*congruent CurrTrial : 1*incongruent'								\
	-gltCode high_1			'PrevTrial : 1*incongruent CurrTrial : 1*incongruent'								\
	-gltCode interf			'PrevTrial : 1*congruent & 1*incongruent CurrTrial : 1*incongruent'						\
	-glfCode A-B_interf	'Grp : 1*setA & 1*setB PrevTrial : 1*congruent & 1*incongruent CurrTrial : 1*incongruent'			\
	-dataTable @datatable.txt

I am most interested in making sure the gltCode 'interf' and the glfCode 'A-B_interf' are correct, as I think I have the others sorted out.

Thank you!

Ryann

2

The following line:

-gltCode interf	'PrevTrial : 1*congruent & 1*incongruent CurrTrial : 1*incongruent'						\

should be

-glfCode interf	'PrevTrial : 1*congruent & 1*incongruent CurrTrial : 1*incongruent'						\

However, it might be more informative to assess:

-gltCode interf	'PrevTrial : 1*congruent  -1*incongruent CurrTrial : 1*incongruent'						\

We are interested in looking at the difference between the response to incongruent trials depending on if the previous trial was congruent or incongruent (PrevTrial: congruent & CurrTrial: incongruent vs. PrevTrial: incongruent & CurrTrial: incongruent), as well as if this contrast is significantly different between our 2 groups.

You may consider

-gltCode A-B_inter1	'Grp : 1*setA -1*setB PrevTrial : 1*congruent -1*incongruent CurrTrial : 1*incongruent'			\

and

-glfCode A-B_inter2	'Grp : 1*setA -1*setB PrevTrial : 1*congruent & 1*incongruent CurrTrial : 1*incongruent'			\

Gang Chen