[Question] How to design within-between interaction ANOVA ?

[sato16]

Dear LIMO Developers,

I would like to perform a 3-way ANOVA analysis with 2 conditions (Recognized/Unrecognized) * 2 sessions (1,2) within-subjects, across 2 groups (nap/break). For this test, I used only two participants in each group (nap/break) for a minimum test.

I followed the steps below to conduct this analysis but encountered an error at the 2nd-level analysis. I suspect there might be an issue with the study design settings or the 2nd-level configuration. I would greatly appreciate your advice.

1. Loading the .set data for each condition
   I loaded data for 2 conditions * 2 sessions from 4 participants (2 nap, 2 break), resulting in a total of 16 datasets.

2. Making study design
   I made study design using the following command.
   STUDY = std_makedesign(STUDY, ALLEEG, 1,... 'name',studyName,'delfiles','off','defaultdesign','off',... 'variable1','condition','values1',{'Recognized','Unrecognized'},... 'vartype1','categorical','variable2','group','values2',{'break','nap'},... 'vartype2','categorical','variable3','session','values3',{1,2},... 'vartype3','categorical','subjselect',subSelectID);

3. Pre-comp
   For the minimum test, I analyzed only the ERP data using the following command.
   [STUDY, ALLEEG] = std_precomp(STUDY, ALLEEG, {},'savetrials','on',... 'interp','on',... 'recompute','on',... 'erp','on','erpparams',{'rmbase',[-1000 0] },... 'spec','off',... 'erpim','off',... 'ersp','off',... 'itc','off');

4. Running the 1st level
   I performed the first-level analysis using the command below.
   STUDY = pop_limo(STUDY, ALLEEG, 'method','WLS','measure','daterp','timelim',[-50 650],'erase','on','splitreg','off','interaction','off');

Ideally, I thought this should be done with 4 conditions within each subject, but as noted in the warning(below image), each condition was ignored. Based on the documentation from here, it seems that conditions should NEVER be used for LIMO 1st level analysis. Therefore, I think there should be no issues.

The study design that can be viewed from EEGLAB looks like the image below.

(Additionally, I referred to this page and tried to create and use a cat.txt file with values like 1, 2, 3, 4, 1, 2, 3, 4... on each line, but I could not find the correct timing to load it from GUI.)

5. Running the 2nd level
   In response to the question, "How many independent groups of subjects or sessions per subject?" I answered 2 (for the nap/break groups).
   For the "Enter repeated factor levels" screen, I input the following values.

In the subsequent "Which parameter to test?" screen, I was unsure how to set the beta parameters. Since there are two groups, I tried [1:2], but it resulted in an error.

The error is as follows (apologies for the Japanese):

インデックスが配列要素数を超えています。インデックスは 0 を超えてはなりません。

エラー: limo_random_select>reorganize_params (行 2310)
{{ 'style' 'popupmenu' 'string' betas(parameters) 'value' iRow }} ];

エラー: limo_random_select (行 1335)
paramTmp = reorganize_params(parameters(i,:), betas,factor_names, factor_nb);

エラー: limo_random_effect>ANOVA_Callback (行 262)
limo_random_select(answer,handles.chan_file,'nboot',handles.b,'tfce',handles.tfce,'type','Channels');

エラー: gui_mainfcn (行 95)
feval(varargin{:});

エラー: limo_random_effect (行 36)
gui_mainfcn(gui_State, varargin{:});

エラー: matlab.graphics.internal.figfile.FigFile/read>@(hObject,eventdata)limo_random_effect('ANOVA_Callback',hObject,eventdata,guidata(hObject))

UIControl Callback の実行中にエラーが発生しました。

I think I have made a mistake in the configuration, but I am unsure how to fix it.
I would appreciate any guidance you can provide.

Best regards,
Fumiaki