Create run_pre_processing.m

fahrenfort · May 30, 2022 · d8a7b16 · d8a7b16
1 parent ab6bf43
commit d8a7b16
Showing 1 changed file with 143 additions and 0 deletions.
diff --git a/eeg_preprocessing/example_scripts/import_and_preprocess/run_pre_processing.m b/eeg_preprocessing/example_scripts/import_and_preprocess/run_pre_processing.m
@@ -0,0 +1,143 @@
+%% trigger labels and epoch settings
+clear;
+
+% file names
+filenames= {'PP001.set'     'PP002.set' ...
+            'PP003.set'     'PP004.set' ...
+            'PP005.set'     'PP006.set' };
+
+% all cues
+cue_face = [1:3 11:13 21:23 31:33];
+cue_house = [4:6 14:16 24:26 34:36];
+cue_letter = [7:9 17:19 27:29 37:39];
+
+% all stimuli
+stim_face = [1 4 7 11 14 17 21 24 27 31 34 37];
+stim_house = [2 5 8 12 15 18 22 25 28 32 35 38];
+stim_letter = [3 6 9 13 16 19 23 26 29 33 36 39];
+
+% targets and non-targets in prediction session
+upright = 1:9;                        % non-target (do not press)
+upsidedown = 11:19;                   % target (press space bar)
+predicted = [1 5 9 11 16 19];         % targets and non-targets (no relation to button press)
+nonpredicted = [2:4 6:8 12:14 16:18]; % targets and non-targets (no relation to button press)
+triggers = [predicted nonpredicted];
+
+% epoch time period
+eeglab_epochtime = [-2.5 1];
+
+%% pipeline to follow:
+% - import into EEGLAB
+% - highpass filter of .01 Hz
+% - epoch (-2.5 1) ms
+% - remove blinks using ICA
+%
+% During analysis:
+% - reject trials with muscle artifacts in a window from 0-500 ms (large deviating power in the 110-140 Hz range)
+% - baseline correction on (-2300,-2000) ms -> the fixation period 
+% - lots of decoding on lots of stuff
+
+%% read data into eeg lab
+files2do = {};
+basedir = '/Volumes/backup/Predictive_EEG/RAW_logfiles_bdfs/BDFs';
+save2dir = '/Volumes/backup/Predictive_EEG/EEGLAB_DATA/unfiltered_unepoched';
+filz=dir([basedir filesep '*.bdf']); % find all set files for each subject
+for cFiles=29:numel(filz)
+    filename = [basedir filesep filz(cFiles).name];
+    if sum(strncmp(filz(cFiles).name,files2do,numel(filz(cFiles).name)-4))>0 || isempty(files2do)
+        EEG = readbdf_into_eeglab_vu(filename);
+        % downsample a bit, 512 is really not necessary
+        EEG = pop_resample(EEG, 256);
+        % get the name
+        [~, fname, ~ ] = fileparts(filename);
+        EEG = pop_saveset(EEG, 'filename',strrep(fname,'.','_'),'filepath',save2dir);
+        events = [EEG.event(:).type];
+        trig_events = sum(ismember(events, triggers));
+    end
+end
+
+%% filter and epoch data on server
+sourcedir =  '$HOME/predictive/EEGLAB_DATA/unfiltered_unepoched';
+targetdir =  '$HOME/predictive/EEGLAB_DATA/highpass_.01';
+
+triglabels = prediction;
+filter_eeg_and_epoch(sourcedir,filenames,targetdir,0.01,0,eeglab_epochtime(1),eeglab_epochtime(2),cond_string(triggers));
+
+%% run ICA
+filepath = '$HOME/predictive/EEGLAB_DATA/highpass_.01';
+outpath = '$HOME/predictive/EEGLAB_DATA/IC';
+for cFiles=1:numel(filenames)
+    filename = filenames{cFiles};
+    compute_ICs_new(filepath,filename,outpath,'no','no','no','yes');
+end
+
+%% Remove ICs 
+sourcedir = '$HOME/predictive/EEGLAB_DATA/ICA';
+targetdir = '$HOME/predictive/EEGLAB_DATA/noblinks';
+for cFiles=1:numel(filenames)
+    remove_ICs(sourcedir,filenames,targetdir,'0,0,1,0');
+end
+
+% %% Remove eyeblinks RUN IN MATLAB ON LISA
+% clear;
+% sourcedir = '/home/johannes/N2pc_Eimer2/EEGLAB_DATA/noblinks';
+% targetdir = '/home/johannes/N2pc_Eimer2/EEGLAB_DATA/noblinks_noeye';
+% filz=dir([sourcedir filesep '*.set']); % find all set files
+% for cFiles=1:numel(filz)
+%     filename = filz(cFiles).name;
+%     fullfile = [sourcedir filesep filename];
+%     EEG = pop_loadset('filename',filename,'filepath',sourcedir);
+%     events = pop_exportepoch(EEG);
+%     % make sure you get the labels for the ocular channels right
+%     channelnames = {EEG.chanlocs(:).labels};
+%     timewindow = [0 500]; % time window to look at
+%     % compute pure horizontal eye movement channel
+%     EEG2 = EEG;
+%     EEG2.data(ismember(channelnames,{'HEOG_R'}),:,:) = EEG2.data(ismember(channelnames,{'HEOG_R'}),:,:) - EEG2.data(ismember(channelnames,{'HEOG_L'}),:,:);
+%     % find eye-movements
+%     rejected_trials = pop_artstepX(EEG2,'channel',find(ismember(channelnames,{'HEOG_R'})),'flag',1,'threshold',30,'Twindow', timewindow, 'Windowsize',  100, 'Windowstep',  50 );
+%     clear EEG2;
+%     rejected_trials = logical(rejected_trials);
+%     EEG.reject.rejmanual = rejected_trials;
+%     % add 10000
+%     events(rejected_trials) = events(rejected_trials) + 10000;
+%     EEG = pop_importepoch(EEG,events,{'eventtype'},'typefield','eventtype');
+%     % save
+%     pop_saveset(EEG, 'filename',filename,'filepath',targetdir);
+%     filenames{cFiles,1} = filename;
+%     percentage_eye_movements(cFiles,1) = sum(rejected_trials)/numel(rejected_trials);
+% end
+
+% %% EXAMPLE Recode incorrect responses by adding 1000 for missed and incorrect trials
+% sourcedir = '/Users/VU-MBP/Dropbox/Work/Experimenten/N2pc_Eimer2/EEGLAB_DATA/highpass_.1';
+% targetdir = '/Users/VU-MBP/Dropbox/Work/Experimenten/N2pc_Eimer2/EEGLAB_DATA/recoded';
+% filz = dir([sourcedir filesep '*.set']);
+% clear percentage;
+% for cFiles = 1:numel(filz)
+%     fname = filz(cFiles).name;
+%     disp(fname);
+%     EEG = pop_loadset('filename',fname,'filepath',sourcedir);
+%     n = 0;
+%     for cEpoch= 1 : numel(EEG.epoch)
+%         % then recode incorrect response trials by adding 1000
+%         if sum(strcmp(EEG.epoch(cEpoch).eventtype,'100')) == 0 % 100 are the correct responses, so this identifies incorrect or absent responses
+%             %disp(['recoding event ' num2str(cEpoch)]);
+%             indxoftrig = cell2mat(EEG.epoch(cEpoch).eventlatency) == 0;
+%             if ~isempty(indxoftrig)
+%                 n = n + 1;
+%                 eventcode = string2double(EEG.event(EEG.epoch(cEpoch).event(indxoftrig)).type);
+%                 EEG.event(EEG.epoch(cEpoch).event(indxoftrig)).type = num2str(eventcode+1000);
+%             end
+%         end
+%     end
+%     disp(['recoded ' num2str(n) ' events.']);
+%     percentage(cFiles) = n/numel(EEG.epoch);
+%     EEG = eeg_checkset(EEG,'eventconsistency');
+%     EEG = pop_saveset(EEG, 'filename',fname,'filepath',targetdir);
+% end
+
+% %% EXAMPLE count events for check
+% for cEpoch = 1 : numel(EEG.epoch)
+%     indxoftrig = cell2mat(EEG.epoch(cEpoch).eventlatency) == 0;
+%     event_val(cEpoch) = string2double(EEG.event(EEG.epoch(cEpoch).event(indxoftrig)).type);
+% end