-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathpreprocess_data.m
372 lines (279 loc) · 12.9 KB
/
preprocess_data.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
function preprocess_data(sbjs, data_dir, results_dir)
addpath(genpath('/Users/nguyentiendung/Desktop/Studium/Charite/Research/Tutorials/BuenosAires/'));
addpath('/Users/nguyentiendung/Desktop/Studium/Charite/matlab/eeglab')
data_dir = '/Users/nguyentiendung/Desktop/Studium/Charite/Research/Tutorials/BuenosAires/data/EEG/';
results_dir = '/Users/nguyentiendung/Desktop/Studium/Charite/Research/Tutorials/BuenosAires_project/results/';
%%% Creating result folder (Plots will be saved here)
resultDir = ['../analysis_output/preprocessing/'];
if ~isdir(resultDir)
mkdir(resultDir)
end
%%% Creating output data (matfile) folder
outputDataDir = [resultDir, 'data/'];
if ~isdir(outputDataDir)
mkdir(outputDataDir)
end
load metadata
fs = 128;
fres = fs;
frqs = sfreqs(fres, fs);
% maxfreq_ind = max(find(frqs <= 45));
% frqs = frqs(1:maxfreq_ind);
%% Preprocessing Settings
saveOutputData = 1; % 1: To save preprocessd data and preprocessing details 0: Not save
%% %%%%%%%%%%%%%%%%%
lpFilter = 63; % low-pass filter cut-off
hpFilter = 1; % high-pass filter cut-off
bsFilter = [49 51]; % band-stop filter range
filterOrder = 2; % Butterworth filter order
dsRatio = 4; % downsampling rate
epochLeng = 2; % epoch length in seconds
mincomp = 30;
% Updating 'info' variable
c = fix(clock);
info = [];
info.prepStartTime = [' ',num2str(c(3)),'.',num2str(c(2)),'.',num2str(c(1)),' ',num2str(c(4)),':',num2str(c(5))];
info.prep.lpFilter = lpFilter;
info.prep.hpFilter = hpFilter;
info.prep.filterOrder = filterOrder;
[ALLEEG, EEG, CURRENTSET, ALLCOM] = eeglab;
for isbj = intersect(1:length(fnames), sbjs)
isbj
% set isbj = 14;
load([data_dir fnames{isbj}]);
EEG.epoch = 1;
% Updating "info" variable
info.prep.Nchan0 = EEG.nbchan; info.data.fs0 = EEG.srate;
info.data.lengthPoints = EEG.pnts;
info.prep.ref0 = 'common'; info.prep.chanLabels = {EEG.chanlocs.labels};
%% Plotting Channel locations
EEG = eeg_checkset( EEG );
figure; topoplot([],EEG.chanlocs, 'style', 'blank', 'electrodes', 'labelpoint', 'chaninfo', EEG.chaninfo);
saveas(gcf,[resultDir,'topo_labels'],'jpg');
EEG = eeg_checkset( EEG );
figure; topoplot([],EEG.chanlocs, 'style', 'blank', 'electrodes', 'numpoint', 'chaninfo', EEG.chaninfo);
saveas(gcf,[resultDir,'topo_indices'],'jpg');
%% Plotting Spectra of raw data
EEG = eeg_checkset( EEG );
figure; pop_spectopo(EEG, 1, [0 EEG.xmax*1000], 'EEG' , 'percent', 15, 'freq', [10 20], 'freqrange',[0 63],'electrodes','on');
saveas(gcf,[resultDir,'spectra_raw'],'jpg');
%% zero-phase filtering
[b_low, a_low] = butter(filterOrder, lpFilter/(EEG.srate/2), 'low');
[b_high, a_high] = butter(filterOrder, hpFilter/(EEG.srate/2), 'high');
[b_notch, a_notch] = butter(filterOrder, bsFilter/(EEG.srate/2),'stop');
a_all = poly([roots(a_low);roots(a_high); roots(a_notch)]);
b_all = conv(conv(b_low,b_high), b_notch);
% fvtool(b_all, a_all)
EEG.data = filtfilt(b_all, a_all, double(EEG.data)')';
%% plain downsampling to 100 Hz.
% that's not 100 Hz but 128 Hz
EEG.data = EEG.data(:, 1:dsRatio:end);
EEG.srate = EEG.srate/dsRatio;
EEG.pnts = size(EEG.data,2);
EEG.xmax = EEG.xmin + (EEG.pnts-1)/EEG.srate;
EEG.times = linspace(EEG.xmin*1000, EEG.xmax*1000, EEG.pnts);
EEG = eeg_checkset( EEG );
%% Epoching: Trasforming data from Nchan*NtpAll to Nchan*NtpEpoch*Nepoch
%epochLeng = 2; % epoch length in seconds
[Nchan, Ntp]=size(EEG.data);
EEG = pop_select( EEG,'point',[1 (Ntp - mod(Ntp,EEG.srate*epochLeng))] );
Nepoch = EEG.pnts / (EEG.srate*epochLeng);
EEG = eeg_checkset(EEG);
%EEG.data = EEG.data(:, 1:(Ntp - mod(Ntp,fsNew*8)) );
for ievent = 1: Nepoch
EEG.event(ievent).type = num2str(epochLeng);
EEG.event(ievent).latency = (ievent-1)*(EEG.srate*epochLeng)+1;
EEG.event(ievent).duration = epochLeng;
end
EEG = eeg_checkset( EEG );
EEG = pop_epoch( EEG, { num2str(epochLeng) }, [0 epochLeng], 'epochinfo', 'yes');
%% Plotting Spectra of filtered and downsampled data
EEG = eeg_checkset( EEG );
figure; pop_spectopo(EEG, 1, [0 EEG.xmax*1000*Nepoch], 'EEG' , 'percent', 15, 'freq', [10 20 ], 'freqrange',[0 63],'electrodes','on');
saveas(gcf,[resultDir,'spectra_prep1'],'jpg');
% Updating 'info' variable
% info.subject.ID = isbj;
% info.subject.eyeCondition = EYE;
% info.subject.medicalCondition = MED;
% info.subject.fileDirectory = dataDir;
info.subject.outputDirectory = resultDir;
info.prep.fs = EEG.srate; % sampling frequency after down-sampling
info.prep.Nepochs0 = size(EEG.data,3);
info.prep.epochPoints = size(EEG.data, 2);
info.prep.Nchannels0 = size(EEG.data, 1);
info.prep.dsRation = dsRatio;
info.prep.ref = '';
%% Detecting Bad channels/epochs
par.HFAnalysis.run = 1 ;
par.HFAnalysis.channelRejection = 1;
info.outlier = k1_detect_bad_epoch_channel(EEG,par);
%% Detecting channels/epochs rejected because of Strong alpha activity
[b, a] = butter(2, [5 12]/(EEG.srate/2),'stop');
XNoAlpha = reshape(filtfilt(b, a, reshape(double(EEG.data), Nchan, [])')',Nchan,[],Nepoch);
par = [];
par.srate = EEG.srate;
par.deviationAnalysis.run = 1;
par.HFAnalysis.run = 1;
par.deviationAnalysis.threshold = 4.5;
par.HFAnalysis.channelRejection =1;
infoAlpha = k1_detect_bad_epoch_channel(XNoAlpha, par);
info.outlier.epochRejectFinal = info.outlier.epochRejectFinal & infoAlpha.epochRejectFinal; % |
%% Plotting Rejceted channels/epochs
epoch = find(info.outlier.epochRejectFinal);
epochchanind = cell(1, length(epoch)); %{[],[3,9,12],[6:12,16,18]};
rejepochcol = [.6 .8 .9];
rejepoch = zeros(1, EEG.trials);
rejepoch(epoch) = ones(1, length(epoch));
rejepochE = zeros(EEG.nbchan, EEG.trials);
for i = 1:length(find(rejepoch))
rejepochE(epochchanind{i},epoch(i))=ones(size(epochchanind{i}));
end
winrej = trial2eegplot(rejepoch, rejepochE, EEG.pnts, rejepochcol);
colors = cell(1,Nchan); colors(1,:) = {'k'};
colors(1,info.outlier.chanRejectFinal) = {'r'};
eegplot(EEG.data, 'srate', EEG.srate, 'title', 'Rejected Channels Epochs', ...
'limits', [EEG.xmin EEG.xmax]*1000, 'color', colors, 'winrej',winrej, 'eloc_file', EEG.chanlocs);
%% Additional plots for testing channel/epoch rejection
if 1
figure, imagesc(info.outlier.epoch.epochDeviation);
title('Epoch-Epoch-Deviation')
xlabel('Epoch','fontweight','b'); ylabel('Channel','fontweight','b');
% xticks(5:5:Nepoch);
% yticks(1:2:Nchan);
saveas(gcf,[resultDir,'epoch_epoch_deviation'],'jpg');
figure, imagesc(info.outlier.epoch.chanDeviation);
title('Epoch-Chan-Deviation')
xlabel('Epoch','fontweight','b'); ylabel('Channel','fontweight','b');
% xticks(5:5:Nepoch);
% yticks(1:2:Nchan);
saveas(gcf,[resultDir,'epoch_chan_deviation'],'jpg');
figure, imagesc(info.outlier.epoch.epochHFNoise);
title('Epoch-Epoch-HFNOISE')
xlabel('Epoch','fontweight','b'); ylabel('Channel','fontweight','b');
% xticks(5:5:Nepoch);
% yticks(1:2:Nchan);
saveas(gcf,[resultDir,'epoch_epoch_hfnoise'],'jpg');
figure, imagesc(info.outlier.epoch.chanHFNoise);
title('Epoch-Chan-HFNOISE')
xlabel('Epoch','fontweight','b'); ylabel('Channel','fontweight','b');
% xticks(5:5:Nepoch);
% yticks(1:2:Nchan);
saveas(gcf,[resultDir,'epoch_chan_hfnoise'],'jpg');
chanEpochMatrix = zeros(Nchan, Nepoch);
chanEpochMatrix(info.outlier.chanRejectFinal, :) = 1;
chanEpochMatrix(:, info.outlier.epochRejectFinal) = 1;
figure, imagesc(chanEpochMatrix);
title('Detected outlier channels epochs')
xlabel('Epoch','fontweight','b'); ylabel('Channel','fontweight','b');
% xticks(5:5:Nepoch);
% yticks(1:2:Nchan);
saveas(gcf,[resultDir,'outlier_chans_epochs'],'jpg');
end
%% Rejecting bad channels/epochs before ICA
info.originalChanlocs = EEG.chanlocs;
EEG = pop_select( EEG,'notrial',find(info.outlier.epochRejectFinal) ,'nochannel',find(info.outlier.chanRejectFinal));
%% Computing ICA weights accoding to INFOMAX implemented in EEGLAB
npca = sum(max(svd(EEG.data(:, :)))./sort(svd(EEG.data(:, :)), 'descend') < 1e4);
if npca < mincomp
warning(['Effective rank lower than ' num2str(mincomp) '. Lowering minum number of components to ' num2str(npca) '.'])
end
EEG = pop_runica(EEG, 'extended',1,'interupt','off', 'pca', npca);
%TO DO %[weights,sphere,meanvar,bias,signs,lrates,data,y] = runica(EEG.data);
%% find subset of electrodes that is close to those defined by 10/20 system on which MARA is trained
load electrode_montages_MARA
d = eucl([EasyCap128.Channel.Loc]', [BioSemi128.Channel.Loc]');
[di mi] = min(d, [], 2);
[so inso] = sort(di);
nmatch = sum(so*1000 < 8);
clab_BioSemi128 = {BioSemi128.Channel(mi(inso(1:nmatch))).Name};
clab_EasyCap128 = {EasyCap128.Channel(inso(1:nmatch)).Name};
[clabs_int, ia, ib] = intersect(clabs, clab_BioSemi128);
chanlocs_bak = EEG.chanlocs;
for ii = 1:length(EEG.chanlocs)
EEG.chanlocs(ii).labels = ['X' num2str(ii)];
end
for ii = 1:length(clabs_int)
EEG.chanlocs(ia(ii)).labels = clab_EasyCap128{ib(ii)};
end
%% Detecting bad ICs based on MARA
[ALLEEG,EEG,CURRENTSET] = processMARA (ALLEEG,EEG,CURRENTSET );
EEG.chanlocs = chanlocs_bak;
%% Ensuring that at least mincomp ICs are retained (According to MARA artefact probabiliy)
% unless npca is even lower
if sum(~EEG.reject.gcompreject) < min(mincomp, npca)
[~, indxICascending] = sort(EEG.reject.MARAinfo.posterior_artefactprob);
EEG.reject.gcompreject(indxICascending(1:min(mincomp, npca))) = 0;
end
%% Visualizing ICs and saving plots
%pop_visualizeMARAfeatures_modified(EEG.reject.gcompreject, EEG.reject.MARAinfo);
EEG.saveDir = resultDir;
% pop_selectcomps_MARA_save(EEG);
% Updating 'info' variable
info.rejectedIC = EEG.reject.gcompreject;
c = fix(clock);
info.prepEndTime = [' ',num2str(c(3)),'.',num2str(c(2)),'.',num2str(c(1)),' ',num2str(c(4)),':',num2str(c(5))];
%% Rejecting bad ICs and back-projecting the retained ICs to Sensor space
[EEG, com] = pop_subcomp(EEG,find(EEG.reject.gcompreject));
%% Plotting Spectra after ICA analysis
EEG = eeg_checkset( EEG );
figure; pop_spectopo(EEG, 1, [0 EEG.xmax*1000*Nepoch], 'EEG' , 'percent', 15, 'freq', [10 20 ], 'freqrange',[0 63],'electrodes','on');
saveas(gcf,[resultDir,'spectra_prep2'],'jpg');
%% Detecting Bad epochs, again, after ICA
par = [];
par.HFAnalysis.run = 1 ;
par.HFAnalysis.channelRejection = 0;
info.outlier2 = k1_detect_bad_epoch_channel(EEG,par);
%% Plotting Detected Bad epochs after ICA
chanEpochMatrix = zeros(Nchan, Nepoch);
chanEpochMatrix(info.outlier2.chanRejectFinal, :) = 1;
chanEpochMatrix(:, info.outlier2.epochRejectFinal) = 1;
figure, imagesc(chanEpochMatrix);
title('Detected outlier channels epochs after ICA')
xlabel('Epoch','fontweight','b'); ylabel('Channel','fontweight','b');
% xticks(5:5:Nepoch);
% yticks(1:2:Nchan);
saveas(gcf,[resultDir,'outlier2_chans_epochs'],'jpg');
%% Rejecting bad epochs found after ICA
% info.originalChanlocs = EEG.chanlocs;
EEG = pop_select( EEG,'notrial',find(info.outlier2.epochRejectFinal) );
%% Saving data before channel interpolation
if saveOutputData
% in EEGLAB format
EEG = eeg_checkset( EEG );
pop_saveset( EEG, 'filename',['prep_sub_',num2str(isbj),'.set'],'filepath',outputDataDir);
% in .mat format
%save(['prep_sub_',num2str(isbj)], 'EEG')
end
%% Interpolating rejected channels
EEG = pop_interp(EEG, info.originalChanlocs, 'spherical');
%% Plotting Spectra after Interpolation
EEG = eeg_checkset( EEG );
figure; pop_spectopo(EEG, 1, [0 EEG.xmax*1000*Nepoch], 'EEG' , 'percent', 15, 'freq', [10 20], 'freqrange',[0 63],'electrodes','on');
saveas(gcf,[resultDir,'spectra_prep3'],'jpg');
%% Saving Interpolated data
if saveOutputData
% in EEGLAB format
EEG = eeg_checkset( EEG );
pop_saveset( EEG, 'filename',['interp_prep_sub_',num2str(isbj),'.set'],'filepath',outputDataDir);
% in .mat format
%save(['prep_sub_',num2str(isbj)], 'EEG')
end
%% Saving preprocessing details
if saveOutputData
save([outputDataDir,'info_preprocessing'], 'info')
end
%% Making report
k1_mk_reprot_latex(info,resultDir);
%% Converting tex report file to PDF
% command = ['/usr/local/texlive/2017/bin/x86_64-darwin/pdflatex ',...
% '-output-directory ' resultDir, ' ',resultDir 'report.tex;'];
command = ['cd ', resultDir,'; ', '/usr/local/texlive/2017/bin/x86_64-darwin/pdflatex ',...
'report.tex;'];
system(command);
%% Opening Report file on MAC
system(['open -a Preview ',resultDir,'report.pdf'])
% mkdir([results_dir 'sensorspace_spectral6'])
% save([results_dir 'sensorspace_spectral6/sbj' num2str(isbj)], ...
% 'fs', 'fres', 'frqs', 'TRGC', 'GC', 'PS', 'CS', 'nlags', ...
% 'maxfreq_ind', '-v7.3');
end