-
Notifications
You must be signed in to change notification settings - Fork 27
/
limo_glm_handling.m
807 lines (734 loc) · 38.1 KB
/
limo_glm_handling.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
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
function limo_glm_handling(LIMO)
% data handling function for GLM
% this function calls limo_glm, limo_glm_boot to get the analysis done, and
% organize all the files around that - externalized from limo_eeg(4)
%
% FORMAT limo_glm_handling(LIMO)
%
% Cyril Pernet
% ------------------------------------------------------------------
% Copyright (C) LIMO Team 2020
cd(LIMO.dir);
warning on;
%% Compute GLM and save stats files
if strcmp(LIMO.design.status,'to do')
Yr = load('Yr'); Yr = Yr.(cell2mat(fieldnames(Yr)));
Yhat = load('Yhat'); Yhat = Yhat.(cell2mat(fieldnames(Yhat)));
Res = load('Res'); Res = Res.(cell2mat(fieldnames(Res)));
R2 = load('R2'); R2 = R2.(cell2mat(fieldnames(R2)));
Betas = load('Betas'); Betas = Betas.(cell2mat(fieldnames(Betas)));
% check method and change parameters accordingly
% -----------------------------------------------
if size(Yr,1) == 1 % in any cases, just one channel/component
array = 1;
else
if LIMO.Level == 2 % second level we can have missing data because of
array = [1:size(Yr,1)]'; %#ok<NBRAK> % bad channels for some subjects - adjust X
else % level 1 = skip empty channels
if strcmpi(LIMO.Analysis,'Time-Frequency')
array = find(~isnan(Yr(:,1,1,1)));
else
array = find(~isnan(Yr(:,1,1)));
end
end
end
if strcmpi(LIMO.design.method,'IRLS') % 1st or 2nd level
N = size(Yr,numel(size(Yr)));
if N < 50
LIMO.design.method = 'OLS';
warning('with %g observations detected, IRLS won''t converge, switching to OLS',N)
end
end
% check dimensions (3D vs 4D)
% --------------------------------------
if strcmpi(LIMO.Analysis,'Time-Frequency')
[~,n_freqs,n_times,~] = size(Yr);
Yr = limo_tf_4d_reshape(Yr);
Yhat = limo_tf_4d_reshape(Yhat); % reshape to 3D
Res = limo_tf_4d_reshape(Res);
R2 = limo_tf_4d_reshape(R2);
Betas = limo_tf_4d_reshape(Betas);
end
% ------------ prepare condition/covariates -------------------
if LIMO.design.nb_conditions ~=0
tmp_Condition_effect = NaN(size(Yr,1),size(Yr,2),length(LIMO.design.nb_conditions),2);
end
if LIMO.design.nb_interactions ~=0
tmp_Interaction_effect = NaN(size(Yr,1),size(Yr,2),length(LIMO.design.nb_interactions),2);
end
if LIMO.design.nb_continuous ~=0
tmp_Covariate_effect = NaN(size(Yr,1),size(Yr,2),LIMO.design.nb_continuous,2);
end
% ------------- prepare weight matrix -------------------------------------
if strcmp(LIMO.design.method,'WLS') || strcmp(LIMO.design.method,'OLS')
if strcmpi(LIMO.Analysis,'Time-Frequency')
W = ones(size(Yr,1),n_freqs,size(Yr,4));
else
W = ones(size(Yr,1),size(Yr,3));
end
elseif strcmp(LIMO.design.method,'IRLS')
W = ones(size(Yr));
end
% ------------ run limo_glm per channels ---------------------------
update = 1;
X = LIMO.design.X;
if isfield(LIMO,'model')
LIMO = rmfield(LIMO,'model');
end
warning off;
for e = 1:length(array)
channel = array(e);
if LIMO.Level == 2
fprintf('analyzing channel %g/%g \n',e,size(array,1));
Y = squeeze(Yr(channel,:,:));
index = find(~isnan(Y(1,:))); % which subjects to keep
if isempty(index)
index = 1:size(Y,2);
end
Y = Y(:,index);
LIMO.design.X = X(index,:);
model = limo_glm(Y',LIMO); warning on;
else % level 1 we should not have any NaNs because we use 'array'
if sum(squeeze(Yr(channel,1,:))) ~= 0
if strcmp(LIMO.Type,'Channels')
fprintf('analyzing channel %g/%g \n',e,size(array,1));
else
fprintf('analyzing component %g/%g \n',e,size(array,1));
end
index = 1:size(Yr,3);
model = limo_glm(squeeze(Yr(channel,:,:))',LIMO);
else
model = [];
end
end
if ~isempty(model)
% update the LIMO.mat
if update == 1 && strcmpi(LIMO.design.method,'OLS')
LIMO.model.model_df = model.df;
if LIMO.design.nb_conditions ~=0
LIMO.model.conditions_df = model.conditions.df;
end
if LIMO.design.nb_interactions ~=0
LIMO.model.interactions_df = model.interactions.df;
end
if LIMO.design.nb_continuous ~=0
LIMO.model.continuous_df = model.continuous.df;
end
update = 0;
elseif update == 1 && ~strcmpi(LIMO.design.method,'OLS') % each channel can have different weighting and thus different df
% store temporarily as cell everything
LIMO.model.model_df{channel} = model.df;
if LIMO.design.nb_conditions ~=0
LIMO.model.conditions_df{channel} = squeeze(model.conditions.df);
end
if LIMO.design.nb_interactions ~=0
LIMO.model.interactions_df{channel} = squeeze(model.interactions.df);
end
if LIMO.design.nb_continuous ~=0
LIMO.model.continuous_df{channel} = squeeze(model.continuous.df);
end
% 1 cell per channel
if e == size(array,1)
tmp = cell2mat(LIMO.model.model_df)'; % dim (elec*[df dfe]) x 1 or time
if size(tmp,2) == size(Yr,1)*2 % when dim 2 is shorter matlab can switch dim around in limo_glm :-(
tmp = tmp';
end
df = tmp(1:2:end,1); % a single value over time
dfe = tmp(2:2:end,:); % could be different over time
LIMO.model = rmfield(LIMO.model,'model_df');
LIMO.model.model_df = [df dfe]; clear tmp
if LIMO.design.nb_conditions ~=0
tmp = cell2mat(LIMO.model.conditions_df)'; % dim (elec*[df dfe]) * 1
if size(tmp,1) == size(Yr,1)*2
df = tmp(1:2:end,:); dfe = tmp(2:2:end,:);
elseif size(tmp,1) == size(Yr,1)
df = tmp(1,:); dfe = tmp(2,:);
end
LIMO.model = rmfield(LIMO.model,'conditions_df');
LIMO.model.conditions_df = [df dfe]; clear tmp
end
if LIMO.design.nb_interactions ~=0
tmp = cell2mat(LIMO.model.interactions_df)'; % dim (elec*[df dfe]) * 1
if size(tmp,1) == size(Yr,1)*2
df = tmp(1:2:end,:); dfe = tmp(2:2:end,:);
elseif size(tmp,1) == size(Yr,1)
df = tmp(1,:); dfe = tmp(2,:);
end
LIMO.model = rmfield(LIMO.model,'interactions_df');
LIMO.model.interactions_df = [df dfe]; clear tmp
end
if LIMO.design.nb_continuous ~=0
tmp = cell2mat(LIMO.model.continuous_df)'; % dim (elec*[df dfe]) * n
if size(tmp,1) == size(Yr,1)*2
df = tmp(1:2:end,1); dfe = tmp(2:2:end,:);
elseif size(tmp,1) == size(Yr,1)
df = tmp(:,1); dfe = tmp(:,2:end);
end
LIMO.model = rmfield(LIMO.model,'continuous_df');
LIMO.model.continuous_df = [df dfe]; clear tmp
end
end
end
% update the files to be stored on the disk
if strcmpi(LIMO.Analysis,'Time-Frequency')
if strcmp(LIMO.design.method,'IRLS')
W(channel,:,index) = model.W';
for ft=size(W,2):-1:1 % each freq*time has different weighting
WX = LIMO.design.X .* repmat(squeeze(W(channel,ft,:)),1,size(X,2));
fitted_data(:,ft) = (WX*squeeze(model.betas(:,ft,:)));
end
elseif strcmp(LIMO.design.method,'WLS')
W(channel,:,index) = model.W';
for f=n_freqs:-1:1 % each freq has different weighting
WX = LIMO.design.X .* repmat(squeeze(W(channel,f,:)),1,size(X,2));
fitted_data(1,f,:,:) = (WX*squeeze(model.betas(:,f,:)))';
end
fitted_data = squeeze(limo_tf_4d_reshape(fitted_data))';
% reshape beta freq to ft
for c=size(model.betas,1):-1:1
tmp(c,:) = reshape(model.betas(c,:,:), [n_freqs*n_times,1]);
end
model.betas = tmp; clear tmp
else % OLS, W is already ones
fitted_data = LIMO.design.X*model.betas;
end
else
if strcmp(LIMO.design.method,'IRLS')
W(channel,:,index) = model.W';
elseif strcmp(LIMO.design.method,'WLS')
if strcmpi(LIMO.Analysis,'Time-Frequency')
W(channel,:,index) = model.W;
else
W(channel,index) = model.W;
end
end
fitted_data = LIMO.design.X*model.betas;
end
% all these always 3D - reshape before saving
Yhat(channel,:,index) = fitted_data';
Res(channel,:,index) = squeeze(Yr(channel,:,index)) - fitted_data';
clear fitted_data
R2(channel,:,1) = model.R2_univariate;
R2(channel,:,2) = model.F;
R2(channel,:,3) = model.p;
Betas(channel,:,:) = model.betas';
if prod(LIMO.design.nb_conditions) ~=0
if length(LIMO.design.nb_conditions) == 1
tmp_Condition_effect(channel,:,1,1) = model.conditions.F;
tmp_Condition_effect(channel,:,1,2) = model.conditions.p;
else
for i=1:length(LIMO.design.nb_conditions)
tmp_Condition_effect(channel,:,i,1) = model.conditions.F(i,:);
tmp_Condition_effect(channel,:,i,2) = model.conditions.p(i,:);
end
end
end
if LIMO.design.fullfactorial == 1
for i=1:length(LIMO.design.nb_interactions)
tmp_Interaction_effect(channel,:,i,1) = model.interactions.F(i,:);
tmp_Interaction_effect(channel,:,i,2) = model.interactions.p(i,:);
end
end
if LIMO.design.nb_continuous ~=0
if LIMO.design.nb_continuous == 1
tmp_Covariate_effect(channel,:,1,1) = model.continuous.F;
tmp_Covariate_effect(channel,:,1,2) = model.continuous.p;
else
for i=1:LIMO.design.nb_continuous
tmp_Covariate_effect(channel,:,i,1) = model.continuous.F(:,i);
tmp_Covariate_effect(channel,:,i,2) = model.continuous.p(:,i);
end
end
end
clear model
end
end
warning on;
% save data on the disk and clean out
disp('saving data to disk')
LIMO.design.X = X;
LIMO.design.weights = W;
LIMO.design.status = 'done';
if ~isfield(LIMO.design,'name')
LIMO.design.name = 'GLM';
end
save(fullfile(LIMO.dir,'LIMO.mat'),'LIMO');
if strcmpi(LIMO.Analysis,'Time-Frequency')
Yhat = limo_tf_4d_reshape(Yhat);
Res = limo_tf_4d_reshape(Res);
R2 = limo_tf_4d_reshape(R2);
Betas = limo_tf_4d_reshape(Betas);
end
save(fullfile(LIMO.dir,'Yhat.mat'), 'Yhat', '-v7.3');
save(fullfile(LIMO.dir,'Res.mat'), 'Res', '-v7.3');
save(fullfile(LIMO.dir,'Betas.mat'), 'Betas', '-v7.3');
save(fullfile(LIMO.dir,'R2.mat'), 'R2', '-v7.3');
clear Yhat Res Betas R2
if prod(LIMO.design.nb_conditions) ~=0
for i=1:length(LIMO.design.nb_conditions)
name = sprintf('Condition_effect_%g.mat',i);
if size(tmp_Condition_effect,1) == 1
tmp = squeeze(tmp_Condition_effect(1,:,i,:));
Condition_effect = NaN(1,size(tmp_Condition_effect,2),2);
Condition_effect(1,:,:) = tmp;
else
Condition_effect = squeeze(tmp_Condition_effect(:,:,i,:));
end
if strcmpi(LIMO.Analysis,'Time-Frequency')
Condition_effect = limo_tf_4d_reshape(Condition_effect);
end
save(fullfile(LIMO.dir,name),'Condition_effect','-v7.3')
end
clear Condition_effect tmp_Condition_effect
end
if LIMO.design.fullfactorial == 1
for i=1:length(LIMO.design.nb_interactions)
name = sprintf('Interaction_effect_%g.mat',i);
if size(tmp_Interaction_effect,1) == 1
tmp = squeeze(tmp_Interaction_effect(1,:,i,:));
Interaction_effect = NaN(1,size(tmp_Interaction_effect,2),2);
Interaction_effect(1,:,:) = tmp;
else
Interaction_effect = squeeze(tmp_Interaction_effect(:,:,i,:));
end
if strcmpi(LIMO.Analysis,'Time-Frequency')
Interaction_effect = limo_tf_4d_reshape(Interaction_effect);
end
save(fullfile(LIMO.dir,name),'Interaction_effect','-v7.3')
end
clear Interaction_effect tmp_Interaction_effect
end
if LIMO.design.nb_continuous ~=0
for i=1:LIMO.design.nb_continuous
name = sprintf('Covariate_effect_%g.mat',i);
if size(tmp_Covariate_effect,1) == 1
tmp = squeeze(tmp_Covariate_effect(1,:,i,:));
Covariate_effect = NaN(1,size(tmp_Covariate_effect,2),2);
Covariate_effect(1,:,:) = tmp;
else
Covariate_effect = squeeze(tmp_Covariate_effect(:,:,i,:));
end
if strcmpi(LIMO.Analysis,'Time-Frequency')
Covariate_effect = limo_tf_4d_reshape(Covariate_effect);
end
save(fullfile(LIMO.dir,name),'Covariate_effect','-v7.3')
end
clear Covariate_effect tmp_Covariate_effect
end
clear file channel filename model reg dir i W
end
%% Bootstrap under H0
% ----------------------------------------------------------
% ----------------------------------------------------------
if LIMO.design.bootstrap ~=0
limo_check_ppool
% avoid overwriting / recomputing H0 if done
% (limo_eeg(4) called via the results interface)
if exist('H0','dir')
if ~exist(fullfile(LIMO.dir,'TFCE'),'dir') && LIMO.design.tfce == 1
overwrite_H0boot = limo_questdlg('H0 present for tfce, overwrite?','limo check','yes','no','no');
else
overwrite_H0boot = questdlg('overwrite H0?','limo check','yes','no','yes');
if strcmp(overwrite_H0boot,'no') || isempty(overwrite_H0boot)
limo_warndlg('Analysis stopped - not overwriting H0')
return
end
end
else
overwrite_H0boot = 'yes';
end
if strcmp(overwrite_H0boot,'yes')
try
mkdir H0;
fprintf('\n %%%%%%%%%%%%%%%%%%%%%%%% \n Bootstrapping GLM, ... \n %%%%%%%%%%%%%%%%%%%%%%%% \n')
Yr = load('Yr');
Yr = Yr.Yr; % reload in any cases - ensuring right dimensions
if size(Yr,1) == 1 % in any cases, just one channel/component
array = 1;
else
if LIMO.Level == 2 % second level we can have missing subjects because of
array = 1:size(Yr,1); % bad channels for some subjects - just adjust X
else % level 1 = skip empty channels - can't be missing trials
if strcmpi(LIMO.Analysis,'Time-Frequency')
array = find(~isnan(Yr(:,1,1,1)));
else
array = find(~isnan(Yr(:,1,1)));
end
end
end
if LIMO.design.bootstrap < 800
if LIMO.design.bootstrap == 101
fprintf('bootstrap set to 101, this is a testing hack, otherwise the minimum required would be 800\n')
else
fprintf('setting bootstrap to the minimum required, i.e. 800 instead of %g\n',LIMO.design.bootstrap)
LIMO.design.bootstrap = 800;
end
end
nboot = LIMO.design.bootstrap;
if LIMO.Level == 2
if strcmpi(LIMO.Analysis,'Time-Frequency')
boot_table = limo_create_boot_table(squeeze(Yr(:,1,:,:)),nboot);
else
boot_table = limo_create_boot_table(Yr,nboot);
end
else
if strcmpi(LIMO.Analysis,'Time-Frequency')
boot_table = randi(size(Yr,4),size(Yr,4),nboot);
else
boot_table = randi(size(Yr,3),size(Yr,3),nboot);
end
end
% make file of the right size to avoid reshaping 5D files
if strcmpi(LIMO.Analysis,'Time-Frequency')
H0_R2 = NaN(size(Yr,1), size(Yr,2), size(Yr,3), 3, nboot); % stores R, F and p values for each boot
H0_Betas = NaN(size(Yr,1), size(Yr,2), size(Yr,3), size(LIMO.design.X,2), nboot);
if LIMO.design.nb_conditions ~= 0
tmp_H0_Conditions = NaN(size(Yr,1), size(Yr,2), size(Yr,3), length(LIMO.design.nb_continuous), 2, nboot);
end
if LIMO.design.nb_interactions ~=0
tmp_H0_Interaction_effect = NaN(size(Yr,1), size(Yr,2), size(Yr,3),length(LIMO.design.nb_interactions), 2, nboot);
end
if LIMO.design.nb_continuous ~= 0
tmp_H0_Covariates = NaN(size(Yr,1), size(Yr,2), size(Yr,3), LIMO.design.nb_continuous, 2, nboot);
end
else
H0_R2 = NaN(size(Yr,1), size(Yr,2), 3, nboot); % stores R, F and p values for each boot
H0_Betas = NaN(size(Yr,1), size(Yr,2), size(LIMO.design.X,2), nboot);
if LIMO.design.nb_conditions ~= 0
tmp_H0_Conditions = NaN(size(Yr,1), size(Yr,2), length(LIMO.design.nb_continuous), 2, nboot);
end
if LIMO.design.nb_interactions ~=0
tmp_H0_Interaction_effect = NaN(size(Yr,1), size(Yr,2), length(LIMO.design.nb_interactions), 2, nboot);
end
if LIMO.design.nb_continuous ~= 0
tmp_H0_Covariates = NaN(size(Yr,1), size(Yr,2), LIMO.design.nb_continuous, 2, nboot);
end
end
% run the analysis, loop per channel
% limo_glm_boot then uses parfor to bootstrap the data under the null
warning off;
X = LIMO.design.X;
h = waitbar(0,'bootstraping data','name','% done');
for e = 1:length(array)
channel = array(e);
waitbar(e/size(array,2))
fprintf('bootstrapping channel %g \n',channel);
if LIMO.Level == 2
if strcmpi(LIMO.Analysis,'Time-Frequency')
if strcmp(LIMO.design.method,'WLS') || strcmp(LIMO.design.method,'OLS')
Y = squeeze(Yr(channel,:,:,:));
index = find(~isnan(Y(1,1,:))); % because across subjects, we can have missing data
for f=1:size(Yr,2)
Weights = squeeze(LIMO.design.weights(channel,f,index));
model{f} = limo_glm_boot(squeeze(Y(f,:,index))',X(index,:), Weights,...
LIMO.design.nb_conditions,LIMO.design.nb_interactions,LIMO.design.nb_continuous,...
LIMO.design.method,LIMO.Analysis,boot_table{channel});
end
elseif strcmp(LIMO.design.method,'IRLS')
Y = squeeze(Yr(channel,:,:,:));
index = find(~isnan(Y(1,1,:))); % because across subjects, we can have missing data
if numel(size(LIMO.design.weights)) == 3
LIMO.design.weights = limo_tf_4d_reshape(LIMO.design.weights,LIMO.data.size4D);
end
for f=1:size(Yr,2)
Weights = squeeze(LIMO.design.weights(channel,f,:,index));
model{f} = limo_glm_boot(squeeze(Y(f,:,index))',X(index,:), Weights,...
LIMO.design.nb_conditions,LIMO.design.nb_interactions,LIMO.design.nb_continuous,...
LIMO.design.method,LIMO.Analysis,boot_table{channel});
end
end
else
if strcmp(LIMO.design.method,'WLS') || strcmp(LIMO.design.method,'OLS')
Y = squeeze(Yr(channel,:,:));
index = find(~isnan(Y(1,:)));
Weights = squeeze(LIMO.design.weights(channel,index))';
model = limo_glm_boot(squeeze(Y(:,index))',X(index,:),Weights,...
LIMO.design.nb_conditions,LIMO.design.nb_interactions,LIMO.design.nb_continuous,...
LIMO.design.method,LIMO.Analysis,boot_table{channel});
elseif strcmp(LIMO.design.method,'IRLS')
Y = squeeze(Yr(channel,:,:));
index = find(~isnan(Y(1,:)));
Weights = squeeze(LIMO.design.weights(channel,:,index));
model = limo_glm_boot(squeeze(Y(:,index))',X(index,:),Weights,...
LIMO.design.nb_conditions,LIMO.design.nb_interactions,LIMO.design.nb_continuous,...
LIMO.design.method,LIMO.Analysis,boot_table{channel});
end
end
else % LIMO.Level == 1
if strcmpi(LIMO.Analysis,'Time-Frequency')
if strcmp(LIMO.design.method,'WLS') || strcmp(LIMO.design.method,'OLS')
for f=1:size(Yr,2)
LIMO.Weights = squeeze(LIMO.design.weights(channel,f,:));
model{f} = limo_glm_boot(squeeze(Yr(channel,f,:,:))',LIMO,boot_table);
end
elseif strcmp(LIMO.design.method,'IRLS')
if numel(size(LIMO.design.weights)) == 3
LIMO.design.weights = limo_tf_4d_reshape(LIMO.design.weights,LIMO.data.size4D);
end
for f=1:size(Yr,2)
LIMO.Weights = squeeze(LIMO.design.weights(channel,f,:,:));
model{f} = limo_glm_boot(squeeze(Yr(channel,f,:,:))',LIMO,boot_table);
end
end
else
if strcmp(LIMO.design.method,'WLS') || strcmp(LIMO.design.method,'OLS')
LIMO.Weights = squeeze(LIMO.design.weights(channel,:))';
model = limo_glm_boot(squeeze(Yr(channel,:,:))',LIMO,boot_table);
elseif strcmp(LIMO.design.method,'IRLS')
LIMO.Weights = squeeze(LIMO.design.weights(channel,:,:));
model = limo_glm_boot(squeeze(Yr(channel,:,:))',LIMO,boot_table);
end
end
end
% update the files to be stored on the disk
if strcmpi(LIMO.Analysis,'Time-Frequency')
for f=1:length(model)
for B = 1:nboot % now loop because we use cells
H0_Betas(channel,f,:,:,B) = model{f}.betas{B};
H0_R2(channel,f,:,1,B) = model{f}.R2_univariate{B};
H0_R2(channel,f,:,2,B) = model{f}.F{B};
H0_R2(channel,f,:,3,B) = model{f}.p{B};
if prod(LIMO.design.nb_conditions) ~=0
if length(LIMO.design.nb_conditions) == 1
tmp_H0_Conditions(channel,f,:,1,1,B) = model{f}.conditions.F{B};
tmp_H0_Conditions(channel,f,:,1,2,B) = model{f}.conditions.p{B};
else
for i=1:length(LIMO.design.nb_conditions)
tmp_H0_Conditions(channel,f,:,i,1,B) = model{f}.conditions.F{B}(i,:);
tmp_H0_Conditions(channel,f,:,i,2,B) = model{f}.conditions.p{B}(i,:);
end
end
end
if LIMO.design.fullfactorial == 1
if length(LIMO.design.nb_interactions) == 1
tmp_H0_Interaction_effect(channel,f,:,1,1,B) = model{f}.interactions.F{B};
tmp_H0_Interaction_effect(channel,f,:,1,2,B) = model{f}.interactions.p{B};
else
for i=1:length(LIMO.design.nb_interactions)
tmp_H0_Interaction_effect(channel,f,:,i,1,B) = model{f}.interactions.F{B}(i,:);
tmp_H0_Interaction_effect(channel,f,:,i,2,B) = model{f}.interactions.p{B}(i,:);
end
end
end
if LIMO.design.nb_continuous ~=0
if LIMO.design.nb_continuous == 1
tmp_H0_Covariates(channel,f,:,1,1,B) = model{f}.continuous.F{B};
tmp_H0_Covariates(channel,f,:,1,2,B) = model{f}.continuous.p{B};
else
for i=1:LIMO.design.nb_continuous
tmp_H0_Covariates(channel,f,:,i,1,B) = model{f}.continuous.F{B}(:,i);
tmp_H0_Covariates(channel,f,:,i,2,B) = model{f}.continuous.p{B}(:,i);
end
end
end
end
end
else % erp or spec
for B = 1:nboot
H0_Betas(channel,:,:,B) = model.betas{B};
H0_R2(channel,:,1,B) = model.R2_univariate{B};
H0_R2(channel,:,2,B) = model.F{B};
H0_R2(channel,:,3,B) = model.p{B};
if prod(LIMO.design.nb_conditions) ~=0
if length(LIMO.design.nb_conditions) == 1
tmp_H0_Conditions(channel,:,1,1,B) = model.conditions.F{B};
tmp_H0_Conditions(channel,:,1,2,B) = model.conditions.p{B};
else
for i=1:length(LIMO.design.nb_conditions)
tmp_H0_Conditions(channel,:,i,1,B) = model.conditions.F{B}(i,:);
tmp_H0_Conditions(channel,:,i,2,B) = model.conditions.p{B}(i,:);
end
end
end
if LIMO.design.fullfactorial == 1
if length(LIMO.design.nb_interactions) == 1
tmp_H0_Interaction_effect(channel,:,1,1,B) = model.interactions.F{B};
tmp_H0_Interaction_effect(channel,:,1,2,B) = model.interactions.p{B};
else
for i=1:length(LIMO.design.nb_interactions)
tmp_H0_Interaction_effect(channel,:,i,1,B) = model.interactions.F{B}(i,:);
tmp_H0_Interaction_effect(channel,:,i,2,B) = model.interactions.p{B}(i,:);
end
end
end
if LIMO.design.nb_continuous ~=0
if LIMO.design.nb_continuous == 1
tmp_H0_Covariates(channel,:,1,1,B) = model.continuous.F{B};
tmp_H0_Covariates(channel,:,1,2,B) = model.continuous.p{B};
else
for i=1:LIMO.design.nb_continuous
if all(size(squeeze(tmp_H0_Covariates(channel,:,i,1,B))) == size(squeeze(model.continuous.F{B}(:,i)))) || ...
all(size(squeeze(tmp_H0_Covariates(channel,:,i,1,B))) == size(squeeze(model.continuous.F{B}(:,i))'))
tmp_H0_Covariates(channel,:,i,1,B) = model.continuous.F{B}(:,i);
tmp_H0_Covariates(channel,:,i,2,B) = model.continuous.p{B}(:,i);
else
tmp_H0_Covariates(channel,:,i,1,B) = model.continuous.F{B}(i,:);
tmp_H0_Covariates(channel,:,i,2,B) = model.continuous.p{B}(i,:);
end
end
end
end
end
end
end
close(h); warning on;
clear Yr
% save data on the disk and clear out
save([LIMO.dir filesep 'H0' filesep 'H0_R2.mat'],'H0_R2','-v7.3');
save([LIMO.dir filesep 'H0' filesep 'boot_table.mat'],'boot_table');
save([LIMO.dir filesep 'H0' filesep 'H0_Betas.mat'],'H0_Betas','-v7.3');
clear H0_R2 boot_table H0_Betas
if prod(LIMO.design.nb_conditions) ~=0
for i=1:length(LIMO.design.nb_conditions)
name = sprintf('H0_Condition_effect_%g',i);
if strcmpi(LIMO.Analysis,'Time-Frequency')
tmp = squeeze(tmp_H0_Conditions(:,:,:,i,:,:));
else
tmp = squeeze(tmp_H0_Conditions(:,:,i,:,:));
end
if isfield(LIMO.design,'electrode')
if ~isempty(LIMO.design.electrode)
H0_Condition_effect = NaN([1 size(tmp)]);
if strcmpi(LIMO.Analysis,'Time-Frequency')
H0_Condition_effect(1,:,:,:,:) = tmp;
else
H0_Condition_effect(1,:,:,:) = tmp;
end
else
H0_Condition_effect = tmp;
end
else
H0_Condition_effect = tmp;
end
save(fullfile(LIMO.dir,['H0' filesep name]),'H0_Condition_effect','-v7.3');
clear tmp H0_Condition_effect
end
clear tmp_H0_Conditions
end
if LIMO.design.fullfactorial == 1
for i=1:length(LIMO.design.nb_interactions)
name = sprintf('H0_Interaction_effect_%g',i);
if strcmpi(LIMO.Analysis,'Time-Frequency')
tmp = squeeze(tmp_H0_Interaction_effect(:,:,:,i,:,:));
else
tmp = squeeze(tmp_H0_Interaction_effect(:,:,i,:,:));
end
if isfield(LIMO.design,'electrode')
if ~isempty(LIMO.design.electrode)
H0_Interaction_effect = NaN([1 size(tmp)]);
if strcmpi(LIMO.Analysis,'Time-Frequency')
H0_Interaction_effect(1,:,:,:,:) = tmp;
else
H0_Interaction_effect(1,:,:,:) = tmp;
end
else
H0_Interaction_effect = tmp;
end
else
H0_Interaction_effect = tmp;
end
save(fullfile(LIMO.dir,['H0' filesep name]),'H0_Interaction_effect','-v7.3');
clear H0_Interaction_effect
end
clear tmp_H0_Interaction_effect
end
if LIMO.design.nb_continuous ~=0
for i=1:LIMO.design.nb_continuous
name = sprintf('H0_Covariate_effect_%g',i);
if strcmpi(LIMO.Analysis,'Time-Frequency')
tmp = squeeze(tmp_H0_Covariates(:,:,:,i,:,:));
else
tmp = squeeze(tmp_H0_Covariates(:,:,i,:,:));
end
if isfield(LIMO.design,'electrode')
if ~isempty(LIMO.design.electrode)
H0_Covariate_effect = NaN([1 size(tmp)]);
if strcmpi(LIMO.Analysis,'Time-Frequency')
H0_Covariate_effect(1,:,:,:,:) = tmp;
else
H0_Covariate_effect(1,:,:,:) = tmp;
end
else
H0_Covariate_effect = tmp;
end
else
H0_Covariate_effect = tmp;
end
save(fullfile(LIMO.dir,['H0' filesep name]),'H0_Covariate_effect','-v7.3');
clear tmp H0_Covariate_effect
end
clear tmp_H0_Covariates
end
clear channel model H0_R2;
cd(LIMO.dir); disp(' ');
catch boot_error
disp('an error occured while attempting to bootstrap the data')
error('%s \n',boot_error.message);
end
end
end
%% TFCE
% --------------
if LIMO.design.tfce == 1
if exist(fullfile(LIMO.dir,'TFCE'),'dir')
if strcmp(limo_questdlg('TFCE directory detected, overwrite?','data check','Yes','No','No'),'No')
limo_warndlg('Analysis stopped - not overwriting TFCE')
return
end
end
% check if there is a neighbouring matrix
% (since TFCE integrates over clusters)
if ~isfield(LIMO.data,'neighbouring_matrix')
warning('no neighbouring matrix found, this is required for TFCE')
answer = questdlg('load or compute neighbouring matrix?','channel neighbouring definition','Load','Compute','Compute');
if strcmp(answer,'Load')
[file,newpath,whatsup] = uigetfile('*.mat','select neighbourghing matrix (or expected chanloc file)');
if whatsup == 0
disp('selection aborded');
return
else
tmp = load(fullfile(newpath,file));
fn = fieldnames(tmp);
index = find(ismember(fn,'channeighbstructmat'));
if isempty(index)
error('no neighbouring matrix ''channeighbstructmat'' found')
else
LIMO.data.neighbouring_matrix = getfield(tmp,fn{index});
save(fullfile(LIMO.dir,'LIMO.mat'),'LIMO');
end
end
else
[~, LIMO.data.neighbouring_matrix] = limo_expected_chanlocs(LIMO.data.data, LIMO.data.data_dir);
if isempty(LIMO.data.neighbouring_matrix)
limo_error('no neighbouring matrix returned, try creating with limo tools')
else
save(fullfile(LIMO.dir,'LIMO.mat'),'LIMO')
end
end
end
fprintf('\n %%%%%%%%%%%%%%%%%%%%%%%% \n Computing TFCE for GLM takes a while, be patient .. \n %%%%%%%%%%%%%%%%%%%%%%%% \n')
mkdir tfce;
% R2
limo_tfce_handling(fullfile(LIMO.dir,'R2.mat'),'checkfile','no');
% conditions
if prod(LIMO.design.nb_conditions) ~=0
for i=1:length(LIMO.design.nb_conditions)
name = sprintf('Condition_effect_%g.mat',i);
limo_tfce_handling(fullfile(LIMO.dir,name),'checkfile','no');
end
end
% interactions
if LIMO.design.fullfactorial == 1
for i=1:length(LIMO.design.fullfactorial)
name = sprintf('Interaction_effect_%g.mat',i);
limo_tfce_handling(fullfile(LIMO.dir,name),'checkfile','no');
end
end
% covariates / continuous regressors
if LIMO.design.nb_continuous ~=0
for i=1:LIMO.design.nb_continuous
name = sprintf('Covariate_effect_%g.mat',i);
limo_tfce_handling(fullfile(LIMO.dir,name),'checkfile','no')
end
end
end