Correct second level contrast in gui + how output effect size?

AFNI Message Board
, ,
1

AFNI version info (afni -ver):
22

I have preprocessed my data with fMRIprep added into AFNI for smoothing and scaling (mean BOLD signal is 100). ROI are a social brain mask from neurosynth.

I am interested in two main contrasts (full second-level code below)...

  1. Brain differences during exclusion (exclusion block > equal play) between groups
  2. interaction between groups and age during exclusion (exclusion block > equal play)

My model includes also scanner site (2 locations)

  1. Are these the correct contrasts to identify brain differences between groups during exclusion vs. equal play (ignore the numbers as they are slightly different to the code below). My understanding is that the first "Olay" box should be the coefficient and the second "Thr" box should be the t-test? This is very different to SPM, so want to make sure I am doing this correctly?

Screenshot 2023-10-17 100257

  1. this is the correct contrast for the age x group comparison?

Screenshot 2023-10-17 100510

  1. If I also ran a group-specific contrast of to better understand the interaction findings such as ...
group: 1*HC condition : 1*exclude -1*equal age :'

Would an increased brain response be associated with increased age for this group? or the other way around?

  1. I need to output the effect size for these contrasts. I specified -GES but don't understand how to output effect size for these contrasts? Could someone explain for a beginner where these values live in AFNI and how to output?

Full code below.

3dMVM -prefix mask_Oct_14_2023 \
    -bsVars "grp*age*site" \
    -mask '/work/06953/jes6785/atlas_masks/resample_bin_combine_soc_reward_2.nii.gz' \
    -wsVars "condition" \
    -qVars "age" \
    -qVarsCenters '18.4' \
    -GES \
    -num_glt 2 \
    -gltLabel 1 EXC_RISKvHC exc_v_equ  -gltCode 1 'grp : 1*RISK -1*HC condition : 1*exclude -1*equal'  \
    -gltLabel 2 RISKvHC_exc_v_equ_AGE  -gltCode 2 'grp : 1*RISK -1*HC condition : 1*exclude -1*equal age :'  \
   -dataTable      \
    Subj	grp	age	site	condition	InputFile  \
    a002	HC	18.4	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a002b+tlrc[4] \
    a004	HC	14.6	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a004b+tlrc[4] \
    a042	HC	19	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a042b+tlrc[4] \
    a116	HC	20.5	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a116b+tlrc[4] \
    a130	HC	19.8	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a130b+tlrc[4] \
    a134	HC	18.4	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a134b+tlrc[4] \
    a146	HC	15.8	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a146b+tlrc[4] \
    b010	HC	18.3	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b010b+tlrc[4] \
    b100	HC	20.2	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b100b+tlrc[4] \
    b132	HC	15.2	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b132b+tlrc[4] \
    b140	HC	20.3	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b140b+tlrc[4] \
    a030	RISK	16.4	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a030b+tlrc[4] \
    a032	RISK	14.9	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a032b+tlrc[4] \
    a034	RISK	14.2	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a034b+tlrc[4] \
    a052	RISK	15.8	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a052b+tlrc[4] \
    a058	RISK	17.7	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a058b+tlrc[4] \
    a060	RISK	16	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a060b+tlrc[4] \
    b014	RISK	14.4	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b014b+tlrc[4] \
    b022	RISK	15.4	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b022b+tlrc[4] \
    b048	RISK	18.2	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b048b+tlrc[4] \
    b054	RISK	17.8	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b054b+tlrc[4] \
    b074	RISK	19.5	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b074b+tlrc[4] \
    b144	RISK	19.6	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b144b+tlrc[4] \
    b148	RISK	14.7	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b148b+tlrc[4] \
    b150	RISK	17.5	UC	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b150b+tlrc[4] \
    b154	RISK	17.2	UC	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b154b+tlrc[4] \
    a001	HC	21.1	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a001+tlrc[4] \
    a017	HC	21.5	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a017+tlrc[4] \
    a025	HC	18.9	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a025+tlrc[4] \
    a045	HC	18.5	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a045+tlrc[4] \
    a101	HC	20.5	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a101+tlrc[4] \
    a105	HC	20.9	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a105+tlrc[4] \
    a119	HC	19.8	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a119+tlrc[4] \
    a129	HC	20.9	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a129+tlrc[4] \
    a131	HC	20.7	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a131+tlrc[4] \
    a145	HC	21.3	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a145+tlrc[4] \
    b007	HC	21.9	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b007+tlrc[4] \
    b009	HC	18.8	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b009+tlrc[4] \
    b117	HC	18.6	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b117+tlrc[4] \
    b149	HC	14	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b149+tlrc[4] \
    b161	HC	19.7	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b161+tlrc[4] \
    b163	HC	20.7	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b163+tlrc[4] \
    b165	HC	20.5	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b165+tlrc[4] \
    b167	HC	21.1	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b167+tlrc[4] \
    b173	HC	21.7	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b173+tlrc[4] \
    a041	RISK	20	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a041+tlrc[4] \
    a073	RISK	19.7	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a073+tlrc[4] \
    a087	RISK	19.2	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a087+tlrc[4] \
    a135	RISK	17.3	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a135+tlrc[4] \
    a141	RISK	17.7	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a141+tlrc[4] \
    a147	RISK	18.9	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a147+tlrc[4] \
    a153	RISK	19	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a153+tlrc[4] \
    b029	RISK	18.8	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b029+tlrc[4] \
    b065	RISK	17.2	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b065+tlrc[4] \
    b071	RISK	20.3	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b071+tlrc[4] \
    b075	RISK	19.9	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b075+tlrc[4] \
    b083	RISK	14.3	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b083+tlrc[4] \
    b095	RISK	18.6	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b095+tlrc[4] \
    b103	RISK	19.7	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b103+tlrc[4] \
    b123	RISK	20.5	UT	equal		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b123+tlrc[4] \
    b143	RISK	14.1	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b143+tlrc[4] \
    b155	RISK	17.3	UT	equal 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b155+tlrc[4] \
    a002	HC	18.4	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a002b+tlrc[10] \
    a004	HC	14.6	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a004b+tlrc[10] \
    a042	HC	19	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a042b+tlrc[10] \
    a116	HC	20.5	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a116b+tlrc[10] \
    a130	HC	19.8	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a130b+tlrc[10] \
    a134	HC	18.4	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a134b+tlrc[10] \
    a146	HC	15.8	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a146b+tlrc[10] \
    b010	HC	18.3	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b010b+tlrc[10] \
    b100	HC	20.2	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b100b+tlrc[10] \
    b132	HC	15.2	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b132b+tlrc[10] \
    b140	HC	20.3	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b140b+tlrc[10] \
    a030	RISK	16.4	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a030b+tlrc[10] \
    a032	RISK	14.9	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a032b+tlrc[10] \
    a034	RISK	14.2	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a034b+tlrc[10] \
    a052	RISK	15.8	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a052b+tlrc[10] \
    a058	RISK	17.7	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a058b+tlrc[10] \
    a060	RISK	16	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a060b+tlrc[10] \
    b014	RISK	14.4	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b014b+tlrc[10] \
    b022	RISK	15.4	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b022b+tlrc[10] \
    b048	RISK	18.2	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b048b+tlrc[10] \
    b054	RISK	17.8	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b054b+tlrc[10] \
    b074	RISK	19.5	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b074b+tlrc[10] \
    b144	RISK	19.6	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b144b+tlrc[10] \
    b148	RISK	14.7	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b148b+tlrc[10] \
    b150	RISK	17.5	UC	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b150b+tlrc[10] \
    b154	RISK	17.2	UC	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b154b+tlrc[10] \
    a001	HC	21.1	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a001+tlrc[10] \
    a017	HC	21.5	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a017+tlrc[10] \
    a025	HC	18.9	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a025+tlrc[10] \
    a045	HC	18.5	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a045+tlrc[10] \
    a101	HC	20.5	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a101+tlrc[10] \
    a105	HC	20.9	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a105+tlrc[10] \
    a119	HC	19.8	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a119+tlrc[10] \
    a129	HC	20.9	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a129+tlrc[10] \
    a131	HC	20.7	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a131+tlrc[10] \
    a145	HC	21.3	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a145+tlrc[10] \
    b007	HC	21.9	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b007+tlrc[10] \
    b009	HC	18.8	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b009+tlrc[10] \
    b117	HC	18.6	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b117+tlrc[10] \
    b149	HC	14	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b149+tlrc[10] \
    b161	HC	19.7	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b161+tlrc[10] \
    b163	HC	20.7	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b163+tlrc[10] \
    b165	HC	20.5	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b165+tlrc[10] \
    b167	HC	21.1	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b167+tlrc[10] \
    b173	HC	21.7	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b173+tlrc[10] \
    a041	RISK	20	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a041+tlrc[10] \
    a073	RISK	19.7	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a073+tlrc[10] \
    a087	RISK	19.2	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a087+tlrc[10] \
    a135	RISK	17.3	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a135+tlrc[10] \
    a141	RISK	17.7	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a141+tlrc[10] \
    a147	RISK	18.9	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a147+tlrc[10] \
    a153	RISK	19	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a153+tlrc[10] \
    b029	RISK	18.8	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b029+tlrc[10] \
    b065	RISK	17.2	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b065+tlrc[10] \
    b071	RISK	20.3	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b071+tlrc[10] \
    b075	RISK	19.9	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b075+tlrc[10] \
    b083	RISK	14.3	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b083+tlrc[10] \
    b095	RISK	18.6	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b095+tlrc[10] \
    b103	RISK	19.7	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b103+tlrc[10] \
    b123	RISK	20.5	UT	exclude		/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b123+tlrc[10] \
    b143	RISK	14.1	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b143+tlrc[10] \
    b155	RISK	17.3	UT	exclude 	/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.b155+tlrc[10]
2

Howdy-

I will let more statistical minded folk (like @Gang ) chime in on those aspects, but I note it would be helpful if you could provide your full afni_proc.py (AP) command, to show how you made the contrasts and regression modeling factors, among other details.

thanks,
pt

3

Thanks for the feedback. I didn't use afniproc, but here is my level 1 3dDeconvolve for each participant. Happy to add more info if needed. My main contrasts of interest for the level two analysis were coefficient outputs for exclusion and equal play. In the level-one model, I did output exclusion > equal play to assess at the individual level but decided to model at the second level instead.

3dDeconvolve -input /mnt/analysis/sub-${subj}/func/sub-${subj}_task-cyberball_space-T1w_desc-preproc_bold_6.0blurmean+scale.nii.gz \
-mask /mnt/analysis/sub-${subj}/func/sub-${subj}_task-cyberball_space-MNI152NLin2009cAsym_desc-brain_mask.nii.gz \
-censor /mnt/analysis/beh/${subj}_fwd_censor.txt \
-polort 4 -num_glt 3 -num_stimts 3 \
-stim_times_AM1 3 ${behavioral_file_location}tf_${subj}_fixation.txt ${fix_mean} -stim_label 3 'fixation' \
-stim_times_AM1 1 ${behavioral_file_location}tf_${subj}_full_exclude.txt ${full_ex} -stim_label 1 'full_exclude' \
-stim_times_AM1 2 ${behavioral_file_location}tf_${subj}_equal_play.txt ${equal_mean} -stim_label 2 'equal_play' \
-ortvec ${behavioral_file_location}sub_${subj}_yeo_confounds.txt ORT_LABEL \
-gltsym 'SYM: +full_exclude -equal_play' -glt_label 1 'full_exclude > equal_equal' \
-gltsym 'SYM: +full_exclude -fixation' -glt_label 2 'full_exclude > fixation' \
-gltsym 'SYM: +equal_play -full_exclude' -glt_label 3 'equal_play > full_exclude' \
-fout -tout -x1D X.xmat.1D -xjpeg X.jpg                                 \
-x1D_uncensored X.nocensor.xmat.1D                                       \
-fitts fitts.$subj                                                       \
-errts errts.${subj}                                                     \
-bucket stats.$subj
4

OK, thanks, that still shows all the contrasts directly, yep. I am just spacing that command out so that GC and others can see it more clearly:

3dDeconvolve                                                                 \
    -input           /mnt/analysis/sub-${subj}/func/sub-${subj}_task-cyberball_space-T1w_desc-preproc_bold_6.0blurmean+scale.nii.gz \
    -mask            /mnt/analysis/sub-${subj}/func/sub-${subj}_task-cyberball_space-MNI152NLin2009cAsym_desc-brain_mask.nii.gz \
    -censor          /mnt/analysis/beh/${subj}_fwd_censor.txt                \
    -polort          4                                                       \
    -num_glt         3                                                       \
    -num_stimts      3                                                       \
    -stim_times_AM1  3 ${behavioral_file_location}tf_${subj}_fixation.txt    \
                     ${fix_mean}                                             \
    -stim_label      3 'fixation'                                            \
    -stim_times_AM1  1                                                       \
                     ${behavioral_file_location}tf_${subj}_full_exclude.txt  \
                     ${full_ex}                                              \
    -stim_label      1 'full_exclude'                                        \
    -stim_times_AM1  2 ${behavioral_file_location}tf_${subj}_equal_play.txt  \
                     ${equal_mean}                                           \
    -stim_label      2 'equal_play'                                          \
    -ortvec          ${behavioral_file_location}sub_${subj}_yeo_confounds.txt \
                     ORT_LABEL                                               \
    -gltsym          'SYM: +full_exclude -equal_play'                        \
    -glt_label       1 'full_exclude > equal_equal'                          \
    -gltsym          'SYM: +full_exclude -fixation'                          \
    -glt_label       2 'full_exclude > fixation'                             \
    -gltsym          'SYM: +equal_play -full_exclude'                        \
    -glt_label       3 'equal_play > full_exclude'                           \
    -fout                                                                    \
    -tout                                                                    \
    -x1D             X.xmat.1D                                               \
    -xjpeg           X.jpg                                                   \
    -x1D_uncensored  X.nocensor.xmat.1D                                      \
    -fitts           fitts.$subj                                             \
    -errts           errts.${subj}                                           \
    -bucket          stats.$subj

--pt

5

Thanks, wondering if anyone has an answer to the above?

6

Your concern regarding the 3dMVM output could be attributed to a specification issue. It's possible that the input files were not read correctly into 3dMVM. When the data table is directly incorporated into the script, ensure to enclose the sub-brick selector in quotes. For instance, modify the input file in the first row from:

/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a002b+tlrc[4]

to:

/scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a002b+tlrc'[4]'

Quotation marks are unnecessary when the data table is provided to 3dMVM as a standalone file. Additionally, I recommend using sub-brick labels (e.g., '[Pos#0_Coef]') instead of sub-brick numbers (e.g., 4 in your case) for improved clarity and reduced risk of errors.

  1. If I also ran a group-specific contrast of to better understand the interaction findings such as ...
'group: 1*HC condition : 1*exclude -1*equal age :'

Would an increased brain response be associated with increased age for this group? or the other way around?

The interpretation depends on the sign of the contrast. A positive result would indicate that older age is associated with a stronger response to "exclude" compared to "equal".

I need to extract the effect size for these contrasts.

The first sub-brick in the output for each contrast specified through -gltCode represents the effect magnitude. Combined with the standard error (embedded in the t-statistic value,) it serves as an appropriate indicator of effect size.

Gang

7

Thanks for the Gang, I really appreciate your help.

A few follow-up questions - you mention...

"The first sub-brick in the output for each contrast specified through -gltCode represents the effect magnitude. Combined with the standard error (embedded in the
t
-statistic value,) it serves as an appropriate indicator of effect size."
Could you clarify this with an example, I don't entirely understand?
Here is an example output...

Screenshot 2023-10-18 101344

Based on what you said above, would be effect size be .005182 for that voxel location? And if so, what type of statistic would this be (Cohen, R2)? I need to use this for a sample size calculation. Sorry, this isn't my area of strength.

Re: your comment about the quotation marks. Can I ask, why doesn't the program throw an error when it isn't reading the input files correctly? When I rerun the second level analysis, with the updated /scratch/06953/jes6785/SOBP/CYB_SOBP_SEPT_2023/level_1_output/stats.a002b+tlrc'[4]' - I got more of a delay when reading the input files, but how do I know it actually worked as it's the exact same as before otherwise.

8

What type of statistic would this be (Cohen's d, R²)? I need to use this for a sample size calculation.

Regarding the effect size, the point estimate at the voxel of focus is 0.005182. The corresponding t-statistic is 3.478042. Therefore, the standard error for the effect magnitude is calculated as follows:

Standard Error = 0.005182 / 3.478042 = 0.001489919

These two values, 0.005182 and 0.001489919, can be utilized for power analysis. In case it's relevant, here is a paper that addresses the complexity of sample sizes in neuroimaging.

Can I ask why the program doesn't throw an error when it isn't reading the input files correctly?

The program might simply read in the first sub-brick of each input file.

I experienced a delay when reading the input files, but how do I confirm if it actually worked, given that the output seems exactly the same as before?

The delay is likely due to the time spent on finding the location of the relevant data array in the input data. To ensure it works correctly, you may extract the desired sub-brick using 3dbucket and then use that as input.

Gang