[FIX] onsets and confounds are saved in the proper dir when there are several tasks

scripts/group_rm_anova.m

@@ -0,0 +1,132 @@
+% attempt at a full factorial design
+%
+%
+% (C) Copyright 2019 Remi Gau
+
+clear;
+clc;
+
+run ../../initEnv.m;
+
+% in participants.tsv
+group_field = 'Group';
+
+opt = opt_stats_group_level();
+
+opt = checkOptions(opt);
+
+opt.verbosity = 2;
+
+opt.pipeline.type = 'stats';
+
+opt.model.bm = BidsModel('file', opt.model.file);
+
+opt.space = opt.model.bm.Input.space;
+opt.taskName = opt.model.bm.Input.task;
+
+[BIDS, opt] = getData(opt, opt.dir.preproc);
+
+% for folder naming
+node_name = 'rm_anova';
+contrasts = 'ALL';
+groups = 'ALL';
+
+% Levels names for each factor
+group_labels = unique(BIDS.raw.participants.content.(group_field));
+task_labels = opt.taskName;
+odor_labels = {'almond', 'eucalyptus'};
+side_labels = {'left', 'right'};
+subject_labels = opt.subjects;
+
+% list of constrasts and the levels they correspond to
+contrasts_list = {'olfid_almond_left'; ...
+                  'olfid_almond_right'; ...
+                  'olfid_eucalyptus_left'; ...
+                  'olfid_eucalyptus_right'; ...
+                  'olfloc_almond_left'; ...
+                  'olfloc_almond_right'; ...
+                  'olfloc_eucalyptus_left'; ...
+                  'olfloc_eucalyptus_right'};
+
+contrast_levels = cartesian(1:numel(task_labels), ...
+                            1:numel(odor_labels), ...
+                            1:numel(side_labels));
+
+%%
+opt.dir.output = fullfile(opt.dir.stats, 'derivatives', 'cpp_spm-groupStats');
+opt.dir.jobs = fullfile(opt.dir.output, 'jobs',  strjoin(opt.taskName, ''));
+
+% collect con images
+for iSub = 1:numel(subject_labels)
+  sub_label = subject_labels{iSub};
+  contrasts_images_filenames{iSub} = findSubjectConImage(opt, ...
+                                                         sub_label, ...
+                                                         contrasts_list); %#ok<SAGROW>
+end
+
+%% set batch
+factor = @(w, x, y, z) struct('name', w, ...
+                              'levels', x, ...
+                              'dept', y, ...
+                              'variance', 1, ...
+                              'gmsca', 0, ...
+                              'ancova', 0);
+
+factorial_design.des.fd.fact(1) = factor('group', numel(group_labels), false);
+factorial_design.des.fd.fact(end + 1) = factor('task', numel(task_labels), true);
+factorial_design.des.fd.fact(end + 1) = factor('odor', numel(odor_labels), true);
+factorial_design.des.fd.fact(end + 1) = factor('side', numel(side_labels), true);
+factorial_design.des.fd.fact(end + 1) = factor('subject', numel(subject_labels), false);
+
+factorial_design.dir = {getRFXdir(opt, node_name, contrasts, groups)};
+
+row = 1;
+
+for iSub = 1:numel(subject_labels)
+
+  sub_label = opt.subjects{iSub};
+
+  sub_idx = strcmp(BIDS.raw.participants.content.participant_id, ['sub-' sub_label]);
+  participant_group = BIDS.raw.participants.content.(group_field){sub_idx};
+
+  group_index = ismember(participant_group, group_labels);
+
+  for iCon = 1:numel(contrasts_list)
+    icell = struct('levels', [group_index contrast_levels(iCon, :) iSub], ...
+                   'scans', {contrasts_images_filenames{iSub}(iCon)});
+    factorial_design.des.fd.icell(row) = icell;
+    row = row + 1;
+  end
+
+end
+
+factorial_design.des.fd.contrasts = 1;
+
+factorial_design.cov = struct('c', {}, ...
+                              'cname', {}, ...
+                              'iCFI', {}, ...
+                              'iCC', {});
+factorial_design.multi_cov = struct('files', {}, ...
+                                    'iCFI', {}, ...
+                                    'iCC', {});
+
+factorial_design.masking.tm.tm_none = 1;
+factorial_design.masking.im = 1;
+factorial_design.masking.em = {''};
+
+factorial_design.globalc.g_omit = 1;
+
+factorial_design.globalm.gmsca.gmsca_no = 1;
+factorial_design.globalm.glonorm = 1;
+
+matlabbatch{1}.spm.stats.factorial_design = factorial_design;
+
+matlabbatch = setBatchEstimateModel(matlabbatch, ...
+                                    opt, ...
+                                    node_name, ...
+                                    contrasts, ...
+                                    groups);
+
+%% run
+
+saveAndRunWorkflow(matlabbatch, 'RM ANOVA', opt);

src/batches/stats/setBatchFactorialDesign.m

@@ -36,7 +36,7 @@
 
   [BIDS, opt] = getData(opt, opt.dir.preproc);
 
-  printBatchName('specify group level fmri model', opt);
+  printBatchName(sprintf('specify group level fmri model for node "%s"', nodeName), opt);
 
   contrasts = getContrastsListForFactorialDesign(opt, nodeName);
 

src/subject_level/createAndReturnCounfoundMatFile.m

@@ -44,6 +44,10 @@
   bf = bids.File(tsvFile);
   bf.extension = '.mat';
 
+  % reset task query to original value
+  % in case we are merging several tasks in one GLM
+  opt.query.task = opt.taskName;
+
   ffxDir = getFFXdir(bf.entities.sub, opt);
   counfoundMatFile = fullfile(ffxDir, bf.filename);
 

src/subject_level/createAndReturnOnsetFile.m

@@ -37,6 +37,11 @@
 
   % move file into the FFX directory
   [~, filename, ext] = spm_fileparts(onsetFilename);
+
+  % reset task query to original value
+  % in case we are merging several tasks in one GLM
+  opt.query.task = opt.taskName;
+
   ffxDir = getFFXdir(subLabel, opt);
   movefile(onsetFilename, ffxDir);
 

src/workflows/stats/bidsFFX.m

@@ -219,7 +219,9 @@ function checkRootNode(opt)
 end
 
 function matlabbatch = setAction(action, matlabbatch, BIDS, opt, subLabel)
+
   outputDir = getFFXdir(subLabel, opt);
+
   switch action
     case 'specify'
       matlabbatch = setBatchSubjectLevelGLMSpec(matlabbatch, BIDS, opt, subLabel);

src/workflows/stats/bidsRFX.m

@@ -69,6 +69,7 @@
   end
 
   if ismember(lower(action), {'rfx', 'contrasts'})
+    % TODO add possibility to pass several nodeNames at once
     if ~isempty(nodeName)
       datasetNodes = opt.model.bm.get_nodes('Name', nodeName);
     else