forked from branelab/SimMEEG
-
Notifications
You must be signed in to change notification settings - Fork 0
/
run_source_modeling_org.m
767 lines (659 loc) · 46.2 KB
/
run_source_modeling_org.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
function run_source_modeling(varargin)
%% Run Source modeling
global h
% try
h.btn_3D_plot_peak_waves.Value=0;
h.next_inv_soln = length(h.inv_soln)+1;
h.current_inv_soln = h.next_inv_soln;
inv_soln = h.menu_inv_soln.String{h.menu_inv_soln.Value};
% fprintf('Inverse Modeling using %s\n',inv_soln);
% hm = msgbox(sprintf('Running Inverse Source Modeling\n\n %s',inv_soln));
if h.monte_carlo_flag == 1
h.waitfor_txt.String = sprintf('Inverse Modeling using %s\n',inv_soln); drawnow;
else
h.waitfor_panel.Visible='on';
h.waitfor_txt.String = sprintf('Inverse Modeling using %s\n',inv_soln); drawnow;
end
%% make-shift paramaters right now to test simulations
if h.menu_inv_headmodel.Value==1 % Whole Brain
if h.menu_sens_type.Value == 1 %MEG
h.inv_soln(h.next_inv_soln).leadfield = h.anatomy.leadfield_meg_vol;
h.inv_soln(h.next_inv_soln).headmodel = h.anatomy.headmodel_meg_vol;
h.inv_soln(h.next_inv_soln).sens = h.anatomy.sens_meg;
h.inv_soln(h.current_inv_soln).headmodel_type = 'Whole Brain';
h.anatomy.leadfield = h.anatomy.leadfield_meg_vol;
h.anatomy.headmodel = h.anatomy.headmodel_meg_vol;
h.anatomy.sens = h.anatomy.sens_meg;
elseif h.menu_sens_type.Value == 2 %EEG
h.inv_soln(h.next_inv_soln).leadfield = h.anatomy.leadfield_eeg_vol;
h.inv_soln(h.next_inv_soln).headmodel = h.anatomy.headmodel_eeg_vol;
h.inv_soln(h.next_inv_soln).sens = h.anatomy.sens_eeg;
h.inv_soln(h.current_inv_soln).headmodel_type = 'Whole Brain';
h.anatomy.leadfield = h.anatomy.leadfield_eeg_vol;
h.anatomy.headmodel = h.anatomy.headmodel_eeg_vol;
h.anatomy.sens = h.anatomy.sens_eeg;
end
h.inv_soln(h.next_inv_soln).headmodel_mesh = h.anatomy.mesh_volumes(3);
elseif h.menu_inv_headmodel.Value==2 % Cortical Surface
if h.menu_sens_type.Value == 1 %MEG
h.inv_soln(h.next_inv_soln).leadfield = h.anatomy.leadfield_meg_cortex;
h.inv_soln(h.next_inv_soln).headmodel = h.anatomy.headmodel_meg_cortex;
h.inv_soln(h.next_inv_soln).sens = h.anatomy.sens_meg;
h.inv_soln(h.current_inv_soln).headmodel_type = 'Cortical Surface';
h.anatomy.leadfield = h.anatomy.leadfield_meg_cortex;
h.anatomy.headmodel = h.anatomy.headmodel_meg_cortex;
h.anatomy.sens = h.anatomy.sens_meg;
elseif h.menu_sens_type.Value == 2 %EEG
h.inv_soln(h.next_inv_soln).leadfield = h.anatomy.leadfield_eeg_cortex;
h.inv_soln(h.next_inv_soln).headmodel = h.anatomy.headmodel_eeg_cortex;
h.inv_soln(h.next_inv_soln).sens = h.anatomy.sens_eeg;
h.inv_soln(h.current_inv_soln).headmodel_type = 'Cortical Surface';
h.anatomy.leadfield = h.anatomy.leadfield_eeg_cortex;
h.anatomy.headmodel = h.anatomy.headmodel_eeg_cortex;
h.anatomy.sens = h.anatomy.sens_eeg;
end
h.inv_soln(h.next_inv_soln).headmodel_mesh = h.anatomy.mesh_volumes(4);
end
inside_idx = find(h.inv_soln(h.current_inv_soln).leadfield.inside==1);
h.cfg.study.bl_bmf.inside_idx = inside_idx;
mHref = []; mref = [];
%% data
data = h.sim_data.sens_final; % [samples x channels x trials]
%% convert BrainSim data to Field Trip data format
ft_data = convert_bs2ft_data(data,h.anatomy,h.cfg);
h.ft_data = ft_data;
%% removing bad_chans from data and leadfields
if isfield(h.anatomy.sens,'good_sensors')
if ~isempty(h.anatomy.sens.bad_sensors)
data = data(:,h.anatomy.sens.good_sensors,:);
h.inv_soln(h.next_inv_soln).leadfield.H=h.inv_soln(h.next_inv_soln).leadfield.H(h.anatomy.sens.good_sensors,:,:);
fprintf('Rank = %.f\n',rank(squeeze(data(:,:,1))));
end
end
%% getting Active & Control intervals
h.cfg.study.bl_bmf.act_int = str2num(h.edit_act_int.String); % active interval set by user
h.cfg.study.bl_bmf.ctrl_int = str2num(h.edit_ctrl_int.String);
act_s = round( (h.cfg.study.bl_bmf.act_int-h.cfg.study.lat_sim(1)) *h.cfg.study.srate); act_s(act_s==0)=1;
ctrl_s = round( (h.cfg.study.bl_bmf.ctrl_int-h.cfg.study.lat_sim(1)) *h.cfg.study.srate); ctrl_s(ctrl_s==0)=1;
h.cfg.study.bl_bmf.act_samps = act_s(1):act_s(end);
h.cfg.study.bl_bmf.ctrl_samps = ctrl_s(1):ctrl_s(end);
bs = round( (h.cfg.study.base_int-h.cfg.study.lat_sim(1))*h.cfg.study.srate);
h.cfg.study.h.cfg.study.base_samps = bs(1):bs(2); % params.study.h.cfg.study.base_samps; % -300 to 0 ms
h.cfg.study.bl_bmf.slice_orient = [1 1 1]; h.cfg.study.bl_bmf.sFaceAlpha = .25;
h.cfg.study.bl_bmf.vw_angle = [0 90];
h.inv_soln(h.current_inv_soln).params = h.cfg.study.bl_bmf;
%% checking rank of data. If it doesn't match the num_chans then white noise will be added until rank sufficient.
xs=round(size(h.cfg.study.bl_bmf.ctrl_samps,2)/2);
R=[]; t=0;
[R,N,Rbar,~,~,~]=BRANELab_calc_cov(data,h.cfg.study.bl_bmf.act_samps,h.cfg.study.bl_bmf.ctrl_samps(xs+1:end));
r2=rank(R);
xstd=min(min(std(data)))/size(data,2);
while r2<size(h.inv_soln(h.next_inv_soln).leadfield.H,1)
t=t+1;
data=data+(2*xstd*randn(size(data)));
[R,N,Rbar,~,~,~]=BRANELab_calc_cov(data,h.cfg.study.bl_bmf.act_samps,h.cfg.study.bl_bmf.ctrl_samps(xs+1:end));
r2=rank(R);
fprintf('Rank = %.f\tChannel Count=%.f\n',r2,size(data,2));
if 100*(2*t*xstd)/max(max(std(data)))>10; fprintf('WARNING! Insufficient Rank in Data or Data is empty\n'); end % stopping infinite loop
% fprintf('Rank = %.f\n',r2);
fprintf('Percent white noise added for sufficient rank = %.3f %%\n\n',100*(2*t*xstd)/max(max(std(data))));
end
h.cfg.study.bl_bmf.noise_alpha = str2num(h.edit_SPA_noise_alpha.String);
%% Preprocesing Data for Field Trip
if h.menu_inv_soln.Value>3 && h.menu_inv_soln.Value<=7 % field Trip's Inv Solutions
%% using same noise covariance for all inverse solns;
soln=[];
soln.lambda = .05; % 1e-9 is good for phase-coupling for MNE according to [Ana-Sofía Hincapiéa et al., 2017 Neuroimage 156: 29-42; Ana-Sofía Hincapiéa et al., Computational Intelligence and Neuroscience Volume 2016, Article ID 3979547, 11 pages] = scalar value, regularisation parameter for the noise covariance matrix (default = 0)
% soln.lambdanoise = 1e-9;
soln.powmethod = 'lambda1'; % = can be 'trace' or 'lambda1'
soln.feedback = 'none'; %' = give ft_progress indication, can be 'text', 'gui' or 'none' (default)
soln.fixedori= 'yes'; % = use fixed or free orientation, can be 'yes' or 'no'
soln.projectnoise = 'yes'; % = project noise estimate through filter, can be 'yes' or 'no'
soln.projectmom = 'yes'; % = project the dipole moment timecourse on the direction of maximal power, can be 'yes' or 'no'
soln.keepfilter = 'yes'; % = remember the beamformer filter, can be 'yes' or 'no'
soln.keepleadfield = 'yes'; % = remember the forward computation, can be 'yes' or 'no'
soln.keepmom = 'no'; % = remember the estimated dipole moment timeseries, can be 'yes' or 'no'
soln.keepcov = 'no'; % = remember the estimated dipole covariance, can be 'yes' or 'no'
soln.kurtosis = 'no'; % = compute the kurtosis of the dipole timeseries, can be 'yes' or 'no'
% These options influence the forward computation of the leadfield
soln.reducerank= 'no'; % = reduce the leadfield rank, can be 'no' or a number (e.g. 2)
soln.normalize= 'yes'; % = normalize the leadfield
soln.normalizeparam= 0.5; % = parameter for depth normalization (default = 0.5)
soln.prewhiten = 'yes'; % = 'no' or 'yes', prewhiten the leadfield matrix with the noise covariance matrix C
% soln.realfilter = 'yes';
elseif h.menu_inv_soln.Value>=8 && h.menu_inv_headmodel.Value==1 % Moiseev's MCMV & TRAP-MUSIC Inv Solutions
% making transformation grid for Moiseev's programs
if ~isfield(h.anatomy,'moiseev') % only Volume leadfields allowed
vx_pos = h.inv_soln(h.next_inv_soln).leadfield.voxel_pos;
%% creating grid "vx_grid" as per Alex's description "nVoxels = nX * nY * nZ, and the ordering is such that the last index (i.e. Z direction) changes most fast, and the first (i.e. X) - most slow"
vx_res = diff(vx_pos(:,1)); vx_res = min(abs(vx_res(vx_res~=0))); % voxel resolution
% making 3D grid
xg = min(vx_pos(:,1))-vx_res:vx_res:max(vx_pos(:,1))+vx_res;
yg = min(vx_pos(:,2))-vx_res:vx_res:max(vx_pos(:,2))+vx_res;
zg = min(vx_pos(:,3))-vx_res:vx_res:max(vx_pos(:,3))+vx_res;
%% Moiseev's programs' (C++) box order
v=0; h.anatomy.moiseev.vx_grid=[];
for d1=1:length(xg)
for d2=1:length(yg)
for d3=1:length(zg)
v=v+1;
h.anatomy.moiseev.vx_grid(v,:) = [xg(d1) yg(d2) zg(d3)];
end
end
end
h.anatomy.moiseev.dims = [length(xg) length(yg) length(zg)];
% iV = idx3Dto1D(vx_grid, dims, 1)
h.anatomy.moiseev.v_idx = find_nearest_voxel(vx_pos,h.anatomy.moiseev.vx_grid);
grid_idx = zeros(size(h.anatomy.moiseev.vx_grid,1),1); % inside brain indices of leadfields
grid_idx(h.anatomy.moiseev.v_idx)=1; h.anatomy.moiseev.lstFlag = grid_idx;
h.anatomy.moiseev.inside_idx = find(h.anatomy.moiseev.lstFlag==1);
% figure(998); clf; hold on; scatter3(h.anatomy.moiseev.vx_grid(grid_idx==0,1),h.anatomy.moiseev.vx_grid(grid_idx==0,2),h.anatomy.moiseev.vx_grid(grid_idx==0,3),'k.');
% scatter3(h.anatomy.moiseev.vx_grid(grid_idx==1,1),h.anatomy.moiseev.vx_grid(grid_idx==1,2),h.anatomy.moiseev.vx_grid(grid_idx==1,3),'ro');
%% matlab box order seed in SimMEEG
v=0; h.anatomy.moiseev.simmeeg_grid=[];
for d1=1:length(zg)
for d2=1:length(yg)
for d3=1:length(xg)
v=v+1;
h.anatomy.moiseev.simmeeg_grid(v,:) = [xg(d3) yg(d2) zg(d1)];
end
end
end
h.anatomy.moiseev.simmeeg_idx = find_nearest_voxel(vx_pos,h.anatomy.moiseev.simmeeg_grid);
grid_idx = zeros(size(h.anatomy.moiseev.simmeeg_grid,1),1); % inside brain indices of leadfields
grid_idx(h.anatomy.moiseev.simmeeg_idx)=1; h.anatomy.moiseev.simmeeg_inside = grid_idx;
h.anatomy.moiseev.simmeeg_inside_idx = find(h.anatomy.moiseev.simmeeg_inside==1);
% figure(999); clf; hold on;
% scatter3(h.anatomy.moiseev.simmeeg_grid(h.anatomy.moiseev.simmeeg_inside==0,1),h.anatomy.moiseev.simmeeg_grid(h.anatomy.moiseev.simmeeg_inside==0,2),h.anatomy.moiseev.simmeeg_grid(h.anatomy.moiseev.simmeeg_inside==0,3),'k.');
% scatter3(h.anatomy.moiseev.simmeeg_grid(h.anatomy.moiseev.simmeeg_inside==1,1),h.anatomy.moiseev.simmeeg_grid(h.anatomy.moiseev.simmeeg_inside==1,2),h.anatomy.moiseev.simmeeg_grid(h.anatomy.moiseev.simmeeg_inside==1,3),'ro');
else
end
if h.menu_inv_headmodel.Value==1
%% Covariance matrix
act_samps = h.cfg.study.bl_bmf.act_samps;
ctrl_samps = h.cfg.study.bl_bmf.ctrl_samps;
xs=round(size(ctrl_samps,2)/2);
%% Covariance for Active vs. Control interval
[R,N,Rbar,Nbar,Rinv,Ninv]=BRANELab_calc_cov(data,act_samps,ctrl_samps(xs+1:end));
%% Covariance for Noise estimate from control samples
% splitting ctrl interval in half to calculate null distribution of noise
[nR,nN,nRbar,nNbar,nRinv,nNinv]=BRANELab_calc_cov(data,ctrl_samps(1:xs),ctrl_samps(xs+1:end));
% [nR,nN,nRbar,nNbar,nRinv,nNinv]=BRANELab_calc_cov(h.sim_data.sens_noise_scaled,ctrl_samps(1:xs),ctrl_samps(xs+1:end));
%% reconfigure leadfields indices to match the moissev grid data
% H = permute(h.inv_soln(h.current_inv_soln).leadfield.H,[3 2 1]); arrH = nan(size(h.anatomy.moiseev.lstFlag,1),size(H,2),size(H,3));
H = permute(h.inv_soln(h.next_inv_soln).leadfield.H,[3 2 1]); arrH = nan(size(h.anatomy.moiseev.lstFlag,1),size(H,2),size(H,3));
arrH(h.anatomy.moiseev.v_idx,:,:) = H;
% other parameters
maxSrc = str2num(h.edit_SPA_max_sources.String); % Maximum number of r sources to be found
gap = 3; % distance between voxels for finding peaks (should be eual to 15 mm) <-- imbedded function in Moiseev code
else
msgbox('Only Volume Head Models are allowed for Moiseev Beamformers');
return
end
elseif h.menu_inv_soln.Value>=8 && h.menu_inv_headmodel.Value==2 % Moiseev's MCMV & TRAP-MUSIC Inv Solutions
msgbox('Only Volume Head Models are allowed for Moiseev Beamformers');
return
end
%% %%%%% Inverse Solutions %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
switch inv_soln
case 'SPA'
%% SPA
h.cfg.study.bl_bmf.loc_flag = h.menu_SPA_loc_flag.Value-1;
h.cfg.study.bl_bmf.noise_alpha = str2num(h.edit_SPA_noise_alpha.String);
[SPA]=BRANElab_LCMV_beamformer_SPA(h.inv_soln(h.next_inv_soln).leadfield.H,data,...
h.cfg.study.bl_bmf.act_samps,h.cfg.study.bl_bmf.ctrl_samps,h.cfg.study.bl_bmf.loc_flag,h.cfg.study.bl_bmf.noise_alpha);
h.inv_soln(h.next_inv_soln).Type = 'SPA';
h.inv_soln(h.next_inv_soln).soln = SPA;
case 'SIA'
h.cfg.study.bl_bmf.loc_flag = h.menu_SPA_loc_flag.Value-1;
h.cfg.study.bl_bmf.plot_flag=0; h.cfg.study.bl_bmf.text_flag=1;
h.cfg.study.bl_bmf.perc_crit = 2;
%% SIA
max_sources = str2num(h.edit_SPA_max_sources.String); %floor(size(h.inv_soln(h.next_inv_soln).leadfield.H,1)/5); % 1 dipole source has 5 degrees of freedom (3 spatial and 2 orientations), thus dividing number of channels/sensors by 5
[SIA]=BRANElab_MCMV_beamformer_SIA(h.inv_soln(h.next_inv_soln).leadfield.H,[],[],data,...
h.cfg.study.bl_bmf.act_samps,h.cfg.study.bl_bmf.ctrl_samps,...
h.inv_soln(h.next_inv_soln).leadfield.voxel_pos,h.cfg.study.bl_bmf.loc_flag,h.cfg.study.bl_bmf.plot_flag,h.cfg.study.bl_bmf.perc_crit,h.cfg.study.bl_bmf.noise_alpha,max_sources);
h.inv_soln(h.next_inv_soln).Type = 'SIA';
h.inv_soln(h.next_inv_soln).soln = SIA;
h.inv_soln(h.next_inv_soln).soln.P.img_org = h.inv_soln(h.next_inv_soln).soln.P.img;
h.inv_soln(h.next_inv_soln).soln.P.img = h.inv_soln(h.next_inv_soln).soln.P.nulled_img; % shifting to nulled img
h.inv_soln(h.next_inv_soln).soln.wts_org = h.inv_soln(h.next_inv_soln).soln.wts;
h.inv_soln(h.next_inv_soln).soln.wts = h.inv_soln(h.next_inv_soln).soln.nulled_wts;
case 'MIA'
h.cfg.study.bl_bmf.loc_flag = h.menu_SPA_loc_flag.Value-1;
h.cfg.study.bl_bmf.plot_flag=0; h.cfg.study.bl_bmf.text_flag=1;
h.cfg.study.bl_bmf.perc_crit = 2;
%% MIA
max_sources = str2num(h.edit_SPA_max_sources.String); %floor(size(h.inv_soln(h.next_inv_soln).leadfield.H,1)/5); % 1 dipole source has 5 degrees of freedom (3 spatial and 2 orientations), thus dividing number of channels/sensors by 5
[MIA]=BRANElab_MCMV_beamformer_MIA(h.inv_soln(h.next_inv_soln).leadfield.H,[],[],data,...
h.cfg.study.bl_bmf.act_samps,h.cfg.study.bl_bmf.ctrl_samps,...
h.inv_soln(h.next_inv_soln).leadfield.voxel_pos,h.cfg.study.bl_bmf.loc_flag,h.cfg.study.bl_bmf.plot_flag,h.cfg.study.bl_bmf.perc_crit,h.cfg.study.bl_bmf.noise_alpha,h.cfg.study.bl_bmf.text_flag,h.anatomy,max_sources);
h.inv_soln(h.next_inv_soln).Type = 'MIA';
h.inv_soln(h.next_inv_soln).soln = MIA;
h.inv_soln(h.next_inv_soln).soln.P.img_org = h.inv_soln(h.next_inv_soln).soln.P.img;
h.inv_soln(h.next_inv_soln).soln.P.img = h.inv_soln(h.next_inv_soln).soln.P.nulled_img; % shifting to nulled img
h.inv_soln(h.next_inv_soln).soln.wts_org = h.inv_soln(h.next_inv_soln).soln.wts;
h.inv_soln(h.next_inv_soln).soln.wts = h.inv_soln(h.next_inv_soln).soln.nulled_wts;
case 'LCMV'
act_int = str2num(h.edit_act_int.String);
ctrl_int = str2num(h.edit_ctrl_int.String);
inside_idx = find(h.inv_soln(h.current_inv_soln).leadfield.inside==1);
%% averaging data and getting covariance matrix
cfg = [];
cfg.channel = h.anatomy.sens.label(h.anatomy.sens.good_sensors);
cfg.baseline = h.sim_data.cfg.study.base_int;
ft_data = ft_timelockbaseline(cfg, ft_data);
cfg = [];
cfg.channel = h.anatomy.sens.label(h.anatomy.sens.good_sensors);
cfg.covariance='yes';
cfg.covariancewindow = [ctrl_int(1) act_int(2)]; % calculating covariance across entire ctrl and act interval just as stated in Field Trip's Tutorial
avg_data = ft_timelockanalysis(cfg,ft_data);
%% Calculating LCMV
cfg = [];
cfg.lcmv = soln;
cfg.keepleadfield = cfg.lcmv.keepleadfield;
cfg.reducerank = cfg.lcmv.reducerank;
cfg.normalize = cfg.lcmv.normalize;
cfg.normalizeparam = cfg.lcmv.normalizeparam;
cfg.lcmv=rmfield(cfg.lcmv,'keepleadfield'); cfg.lcmv=rmfield(cfg.lcmv,'reducerank');
cfg.lcmv=rmfield(cfg.lcmv,'normalize'); cfg.lcmv=rmfield(cfg.lcmv,'normalizeparam');
cfg.method = 'lcmv';
cfg.lcmv.fixedori = 'yes'; % 'yes' = get single orientation using svd
cfg.weightnorm = 'arraygain'; % ''; %'nai'; %'unitnoisegain'; %'unitgain'; %'arraygain'; % 'nai'; % based on equation 4.47 from Sekihara & Nagarajan (2008)
cfg.grad = h.inv_soln(h.next_inv_soln).sens;
cfg.sourcemodel = h.inv_soln(h.next_inv_soln).leadfield;
cfg.headmodel = h.inv_soln(h.next_inv_soln).headmodel;
cfg.lcmv.lambda = '5%';
LCMV = ft_sourceanalysis(cfg, avg_data);
h.LCMV = LCMV;
%% creating LCMV power image and estimating orientations
LCMV.P.img= LCMV.avg.pow(inside_idx)./LCMV.avg.noise(inside_idx); % simple ratio
% LCMV.P.img= (LCMV.avg.pow(inside_idx)-LCMV.avg.noise(inside_idx)) ./LCMV.avg.noise(inside_idx); % deviation ratio ???
% LCMV.P.img= (LCMV.avg.pow(inside_idx)-LCMV.avg.noise(inside_idx)) ./ (LCMV.avg.pow(inside_idx)+LCMV.avg.noise(inside_idx)); % deviation ratio ???
nd=sort(LCMV.P.img); LCMV.null_thresh=nd(ceil(length(nd)*h.cfg.study.bl_bmf.noise_alpha)); %img(img<pthresh)=0;
ori=cell2mat(LCMV.avg.ori(inside_idx)); LCMV.ori=ori';
wts = cell2mat(LCMV.avg.filter(inside_idx)); LCMV.wts = wts';
% img=LCMV.P.img; ori=LCMV.ori; null_thresh=max(img)*.55;
% min_max=[min(img) max(img)*.95]; vol_types=1;
% seed_idx = 1:3; ln_wdth = 1; ln_wdth2 = 2;
% figure(1004); clf; [peak_voxels,p_idx]=bl_plot_lcmv_peak_img_FT_new(img,ori,null_thresh,8,h.inv_soln(h.current_inv_soln).leadfield.voxel_pos,jet(255),min_max,h.anatomy.vol.bnd(3),[],...
% h.cfg.study.bl_bmf.vw_angle,1,vol_types,h.inv_soln(h.current_inv_soln).leadfield.pos,inside_idx,h.cfg.study.bl_bmf.slice_orient,h.cfg.study.bl_bmf.sFaceAlpha);
% hold on; mrk_size=150; s1=scatter3(h.cfg.source.vx_locs(seed_idx,1),h.cfg.source.vx_locs(seed_idx,2),h.cfg.source.vx_locs(seed_idx,3),'k+','sizedata',mrk_size,'linewidth',ln_wdth); s2=scatter3(h.cfg.source.vx_locs(seed_idx,1),h.cfg.source.vx_locs(seed_idx,2),h.cfg.source.vx_locs(seed_idx,3),'ks','sizedata',mrk_size,'linewidth',ln_wdth2);
% title('FT LCMV'); colorbar; caxis(min_max);
% removing 'avg' to reduce storage space
LCMV = rmfield(LCMV,{'avg'}); %MNE.avg = rmfield(MNE.avg,{'mom','filter','noisecov'});
h.inv_soln(h.next_inv_soln).Type = 'LCMV';
h.inv_soln(h.next_inv_soln).soln = LCMV;
case 'eLORETA'
act_int = str2num(h.edit_act_int.String);
ctrl_int = str2num(h.edit_ctrl_int.String);
inside_idx = find(h.inv_soln(h.current_inv_soln).leadfield.inside==1);
%% averaging data and getting covariance matrix
cfg = []; cfg.baseline = h.sim_data.cfg.study.base_int;
cfg.channel = h.anatomy.sens.label(h.anatomy.sens.good_sensors);
ft_data = ft_timelockbaseline(cfg, ft_data);
cfg = [];
cfg.channel = h.anatomy.sens.label(h.anatomy.sens.good_sensors);
cfg.covariance='yes';
cfg.covariancewindow = [ctrl_int(1) act_int(2)]; % calculating covariance across entire ctrl and act interval just as stated in Field Trip's Tutorial
avg_data = ft_timelockanalysis(cfg,ft_data);
%% FT eLORETA
cfg=[];
cfg.method = 'eloreta';
cfg.eloreta = soln; % using same additional options for all inverse solns
cfg.eloreta.lambda = 1;
cfg.keepleadfield = cfg.eloreta.keepleadfield;
cfg.reducerank = cfg.eloreta.reducerank;
cfg.normalize = cfg.eloreta.normalize;
cfg.normalizeparam = cfg.eloreta.normalizeparam;
cfg.eloreta=rmfield(cfg.eloreta,'keepleadfield'); cfg.eloreta=rmfield(cfg.eloreta,'reducerank');
cfg.eloreta=rmfield(cfg.eloreta,'normalize'); cfg.eloreta=rmfield(cfg.eloreta,'normalizeparam');
cfg.grad = h.inv_soln(h.next_inv_soln).sens;
cfg.sourcemodel = h.inv_soln(h.next_inv_soln).leadfield;
cfg.headmodel = h.inv_soln(h.next_inv_soln).headmodel;
eloreta = ft_sourceanalysis(cfg,avg_data);
%% calculating image power from filter weights;
base_samps=h.sim_data.cfg.study.base_samps; % -300 to 0 ms
act_samps=h.cfg.study.bl_bmf.act_samps;
ctrl_samps=h.cfg.study.bl_bmf.ctrl_samps;
ctrl_samps1 = ctrl_samps(1:round(length(ctrl_samps)/2));
ctrl_samps2 = ctrl_samps(round(length(ctrl_samps)/2)+1:end);
y=cell2mat(eloreta.avg.filter); eloreta.wts=permute(reshape(y,[size(y,1) size(y,2)/length(inside_idx) length(inside_idx)]),[3 1 2]);
% for ox = 1:size(eloreta.wts,2)
% eloreta.swf(:,ox,:)=avg_data.avg'*squeeze(eloreta.wts(:,ox,:))';
% end
% eloreta.wts = y';
eloreta.wts = permute(eloreta.wts,[3 1 2]);
eloreta.P.img = squeeze(eloreta.avg.pow(inside_idx)); % source image based on recalcuated swf data
ori = cell2mat(eloreta.avg.ori(inside_idx));
eloreta.ori = ori';
pow = eloreta.avg.pow;
eloreta = rmfield(eloreta,{'avg'}); %MNE.avg = rmfield(MNE.avg,{'mom','filter','noisecov'});
eloreta.avg.pow = single(pow);
h.inv_soln(h.next_inv_soln).Type = 'eLORETA';
h.inv_soln(h.next_inv_soln).soln = eloreta;
% %% PLotting eLORETA on new figure
% img=eloreta.P.img; ori=eloreta.ori; null_thresh=max(img)*.15;
% min_max=[min(img) max(img)*.85]; %min_max = [0 5];
% sFaceAlpha=0; vol_types=1;
% figure(1006); clf; [peak_voxels,p_idx]=bl_plot_lcmv_peak_img_FT_new(img,ori,null_thresh,15,h.inv_soln(h.current_inv_soln).leadfield.voxel_pos,jet(255),min_max,h.anatomy.vol.bnd(3),[],...
% h.cfg.study.bl_bmf.vw_angle,1,vol_types,h.inv_soln(h.current_inv_soln).leadfield.pos,h.cfg.study.bl_bmf.inside_idx,h.cfg.study.bl_bmf.slice_orient,h.cfg.study.bl_bmf.sFaceAlpha);
% seed_idx = 1:3; ln_wdth = 1; ln_wdth2 = 2;
% hold on; mrk_size=150; s1=scatter3(h.cfg.source.vx_locs(seed_idx,1),h.cfg.source.vx_locs(seed_idx,2),h.cfg.source.vx_locs(seed_idx,3),'k+','sizedata',mrk_size,'linewidth',ln_wdth); s2=scatter3(h.cfg.source.vx_locs(seed_idx,1),h.cfg.source.vx_locs(seed_idx,2),h.cfg.source.vx_locs(seed_idx,3),'ks','sizedata',mrk_size,'linewidth',ln_wdth2);
%
% title('FT eLORETA'); colorbar; caxis(map_scale);
case 'sLORETA'
act_int = str2num(h.edit_act_int.String);
ctrl_int = str2num(h.edit_ctrl_int.String);
inside_idx = find(h.inv_soln(h.current_inv_soln).leadfield.inside==1);
%% averaging data and getting covariance matrix
cfg = []; cfg.baseline = h.sim_data.cfg.study.base_int;
cfg.channel = h.anatomy.sens.label(h.anatomy.sens.good_sensors);
ft_data = ft_timelockbaseline(cfg, ft_data);
cfg = [];
cfg.channel = h.anatomy.sens.label(h.anatomy.sens.good_sensors);
cfg.covariance='yes';
cfg.covariancewindow = [ctrl_int(1) act_int(2)]; % calculating covariance across entire ctrl and act interval just as stated in Field Trip's Tutorial
avg_data = ft_timelockanalysis(cfg,ft_data);
%% FT sLORETA
cfg=[];
cfg.method = 'sloreta';
cfg.sloreta = soln; % using same additional options for all inverse solns
cfg.sloreta.lambda = '5%';
cfg.keepleadfield = cfg.sloreta.keepleadfield;
cfg.reducerank = cfg.sloreta.reducerank;
cfg.normalize = cfg.sloreta.normalize;
cfg.normalizeparam = cfg.sloreta.normalizeparam;
cfg.sloreta=rmfield(cfg.sloreta,'keepleadfield'); cfg.sloreta=rmfield(cfg.sloreta,'reducerank');
cfg.sloreta=rmfield(cfg.sloreta,'normalize'); cfg.sloreta=rmfield(cfg.sloreta,'normalizeparam');
cfg.grad = h.inv_soln(h.next_inv_soln).sens;
cfg.sourcemodel = h.inv_soln(h.next_inv_soln).leadfield;
cfg.headmodel = h.inv_soln(h.next_inv_soln).headmodel;
sloreta = ft_sourceanalysis(cfg,avg_data);
%% calculating image power from filter weights;
base_samps=h.sim_data.cfg.study.base_samps; % -300 to 0 ms
act_samps=h.cfg.study.bl_bmf.act_samps;
ctrl_samps=h.cfg.study.bl_bmf.ctrl_samps;
ctrl_samps1 = ctrl_samps(1:round(length(ctrl_samps)/2));
ctrl_samps2 = ctrl_samps(round(length(ctrl_samps)/2)+1:end);
y=cell2mat(sloreta.avg.filter); %sloreta.wts=permute(reshape(y,[3 size(y,2)/length(inside_idx) length(inside_idx)]),[3 1 2]);
sloreta.wts = y';
% sloreta.wts = permute(sloreta.wts,[3 1 2]);
% sloreta.P.img=squeeze(sloreta.avg.pow(inside_idx)); % source image based on recalcuated swf data
sloreta.P.img=squeeze(sloreta.avg.pow(inside_idx)./sloreta.avg.noise(inside_idx)); % source image based on recalculated swf data
ori = cell2mat(sloreta.avg.ori(inside_idx));
sloreta.ori = ori';
sloreta = rmfield(sloreta,{'avg'}); %MNE.avg = rmfield(MNE.avg,{'mom','filter','noisecov'});
h.inv_soln(h.next_inv_soln).Type = 'sLORETA';
h.inv_soln(h.next_inv_soln).soln = sloreta;
% %% PLotting sloreta on new figure
% img=sloreta.P.img; ori=sloreta.ori; null_thresh=sloreta.null_thresh*2;
% min_max=[min(img) max(img)*.15]; %min_max = [0 5];
% sFaceAlpha=0; vol_types=1;
% figure(1006); clf; [peak_voxels,p_idx]=bl_plot_lcmv_peak_img_FT_new(img,ori,null_thresh,15,h.inv_soln(h.current_inv_soln).leadfield.voxel_pos,jet(255),min_max,h.anatomy.vol.bnd(3),[],...
% h.cfg.study.bl_bmf.vw_angle,1,vol_types,h.inv_soln(h.current_inv_soln).leadfield.pos,h.cfg.study.bl_bmf.inside_idx,h.cfg.study.bl_bmf.slice_orient,h.cfg.study.bl_bmf.sFaceAlpha);
% seed_idx = 1:3; ln_wdth = 1; ln_wdth2 = 2;
% hold on; mrk_size=150; s1=scatter3(h.cfg.source.vx_locs(seed_idx,1),h.cfg.source.vx_locs(seed_idx,2),h.cfg.source.vx_locs(seed_idx,3),'k+','sizedata',mrk_size,'linewidth',ln_wdth); s2=scatter3(h.cfg.source.vx_locs(seed_idx,1),h.cfg.source.vx_locs(seed_idx,2),h.cfg.source.vx_locs(seed_idx,3),'ks','sizedata',mrk_size,'linewidth',ln_wdth2);
%
% title('FT sLORETA'); colorbar; caxis(map_scale);
case 'MNE'
act_int = str2num(h.edit_act_int.String);
ctrl_int = str2num(h.edit_ctrl_int.String);
inside_idx = find(h.inv_soln(h.current_inv_soln).leadfield.inside==1);
%% averaging data and getting covariance matrix
cfg = []; cfg.baseline = h.sim_data.cfg.study.base_int;
cfg.channel = h.anatomy.sens.label(h.anatomy.sens.good_sensors);
ft_data = ft_timelockbaseline(cfg, ft_data);
cfg = [];
cfg.channel = h.anatomy.sens.label(h.anatomy.sens.good_sensors);
cfg.covariance='yes';
% cfg.covariancewindow = [ctrl_int(1) act_int(2)]; % calculating covariance across entire ctrl and act interval just as stated in Field Trip's Tutorial
cfg.covariancewindow = [-inf 0];
avg_data = ft_timelockanalysis(cfg,ft_data);
%% FT MNE
cfg = [];
cfg.mne = soln;
cfg.keepleadfield = cfg.mne.keepleadfield;
cfg.reducerank = cfg.mne.reducerank;
cfg.normalize = cfg.mne.normalize;
cfg.normalizeparam = cfg.mne.normalizeparam;
cfg.mne=rmfield(cfg.mne,'keepleadfield'); cfg.mne=rmfield(cfg.mne,'reducerank');
cfg.mne=rmfield(cfg.mne,'normalize'); cfg.mne=rmfield(cfg.mne,'normalizeparam');
cfg.method = 'mne';
cfg.mne.fixedori = 'yes'; % 'yes' = get single orientation using svd
cfg.weightnorm = 'nai'; % based on equation 4.47 from Sekihara & Nagarajan (2008)
cfg.grad = h.inv_soln(h.next_inv_soln).sens;
cfg.sourcemodel = h.inv_soln(h.next_inv_soln).leadfield;
cfg.headmodel = h.inv_soln(h.next_inv_soln).headmodel;
cfg.mne.prewhiten = 'yes';
cfg.mne.lambda = 3;
cfg.mne.scalesourcecov = 'yes';
% cfg.mne.cov=soln.cov; %params.ctrl_cov; % must use noise covariance for MNE to return reasonable results based on simulated data
% avg_data.cov=soln.cov; %params.ctrl_cov;
% cfg.mne.powmethod = 'trace'; % = can be 'trace' or 'lambda1'
% cfg.mne.lambda = 1e-9; % 1e-9 is good for phase-coupling for MNE according to [Ana-Sofía Hincapiéa et al., 2017 Neuroimage 156: 29-42; Ana-Sofía Hincapiéa et al., Computational Intelligence and Neuroscience Volume 2016, Article ID 3979547, 11 pages] = scalar value, regularisation parameter for the noise covariance matrix (default = 0)
% cfg.mne.lambda = 1e-7; % 1e-9 is good for power localization for MNE according to [Ana-Sofía Hincapiéa et al., 2017 Neuroimage 156: 29-42; Ana-Sofía Hincapiéa et al., Computational Intelligence and Neuroscience Volume 2016, Article ID 3979547, 11 pages] = scalar value, regularisation parameter for the noise covariance matrix (default = 0)
% cfg.mne.scalesourcecov = 'yes';
MNE = ft_sourceanalysis(cfg,avg_data);
% calculating image power from filter weights;
y=cell2mat(permute(MNE.avg.filter,[2 1])); MNE.wts=permute(reshape(y,[3 size(y,1)/3 size(y,2)]),[2 1 3]);
MNE.swf=[];MNE.swf_snr=[];MNE.swf_pwr=[];MNE.noise_pwr=[];
for ox=1:3
MNE.swf(:,ox,:)=avg_data.avg'*squeeze(MNE.wts(:,ox,:))';
% swf_snr(:,ox,:)=20*log10(bsxfun(@rdivide,swf(:,ox,:),nanmean(swf(h.cfg.study.base_samps,ox,:),1)));
% MNE.swf_snr(:,ox,:)=(bsxfun(@rdivide,MNE.swf(:,ox,:),nanmean(MNE.swf(h.cfg.study.base_samps,ox,:),1))); % SNR in percent from baseline
% MNE.swf_snr(:,ox,:)=(bsxfun(@rdivide,MNE.swf(:,ox,:),nanstd(MNE.swf(h.cfg.study.base_samps,ox,:),1))); % normalized to baseline
MNE.swf_pwr(ox,:)=rms(MNE.swf(h.cfg.study.bl_bmf.act_samps,ox,:),1)./rms(MNE.swf(h.cfg.study.bl_bmf.ctrl_samps,ox,:),1);
% MNE.noise_pwr(ox,:)=rms(MNE.swf(h.cfg.study.bl_bmf.ctrl_samps,ox,:),1)./rms(MNE.swf(h.cfg.study.bl_bmf.ctrl_samps,ox,:),1);
end
MNE.wts = permute(MNE.wts,[3 1 2]);
MNE.P.img = abs(squeeze(nanmean(rms(MNE.swf(h.cfg.study.bl_bmf.act_samps,:,:),2))));
% MNE.P.img = MNE.avg.pow(inside_idx,s);
% MNE.P.img = abs(squeeze(rms(MNE.swf(s,:,:),2)));
% MNE.P.img=squeeze(max(abs(MNE.avg.pow(h.cfg.study.bl_bmf.inside_idx,h.cfg.study.bl_bmf.act_samps))'))';
% MNE.P.img=squeeze(max(max(MNE.swf_snr(h.cfg.study.bl_bmf.act_samps,:,:))));
% MNE.P.img=squeeze(rms(MNE.swf_pwr,1))'; % averaging across orientations
% MNE.P.img=squeeze(rms(MNE.swf_pwr./MNE.noise_pwr,1))'; % averaging across orientations
nd=sort(MNE.P.img); MNE.null_thresh=nd(ceil(length(nd)*h.cfg.study.bl_bmf.noise_alpha)); %img(img<pthresh)=0;
% Approximating orientations for plotting purposes only.
x=MNE.swf_pwr(1,:)./max(MNE.swf_pwr,[],1);
y=MNE.swf_pwr(2,:)./max(MNE.swf_pwr,[],1);
z=MNE.swf_pwr(3,:)./max(MNE.swf_pwr,[],1);
MNE.ori=[x;y;z]';
%% removing swf and swf_pwr to save memory storage
MNE = rmfield(MNE,{'swf','swf_pwr'}); MNE.avg = rmfield(MNE.avg,{'mom','filter','noisecov'});
MNE.avg.pow = single(MNE.avg.pow);
h.inv_soln(h.next_inv_soln).Type = 'MNE';
h.inv_soln(h.next_inv_soln).soln = MNE;
%% Plotting MNE
% img=MNE.P.img; ori=MNE.ori; null_thresh=max(img)*.25; %MNE.null_thresh*2;
%
% min_max=[min(img) max(img)*.15]; %min_max = [0 5];
% sFaceAlpha=0; vol_types=1;
% figure(1006); clf; [peak_voxels,p_idx]=bl_plot_lcmv_peak_img_FT_new(img,ori,null_thresh,15,h.inv_soln(h.current_inv_soln).leadfield.voxel_pos,jet(255),min_max,h.anatomy.vol.bnd(3),[],...
% h.cfg.study.bl_bmf.vw_angle,1,vol_types,h.inv_soln(h.current_inv_soln).leadfield.pos,h.cfg.study.bl_bmf.inside_idx,h.cfg.study.bl_bmf.slice_orient,h.cfg.study.bl_bmf.sFaceAlpha);
% seed_idx = 1:3; ln_wdth = 1; ln_wdth2 = 2;
% hold on; mrk_size=150; s1=scatter3(h.cfg.source.vx_locs(seed_idx,1),h.cfg.source.vx_locs(seed_idx,2),h.cfg.source.vx_locs(seed_idx,3),'k+','sizedata',mrk_size,'linewidth',ln_wdth); s2=scatter3(h.cfg.source.vx_locs(seed_idx,1),h.cfg.source.vx_locs(seed_idx,2),h.cfg.source.vx_locs(seed_idx,3),'ks','sizedata',mrk_size,'linewidth',ln_wdth2);
% title('FT MNE'); colorbar; caxis(min_max);
%
case 'sMCMV' % Alex Moiseev's sub-space MCMV
% subspace MCMV
sInSMCMV.beamType = h.menu_SPA_loc_flag.String{h.menu_SPA_loc_flag.Value};
sInSMCMV.R = R;
sInSMCMV.arrN = nN; %N;
sInSMCMV.arrH = arrH;
sInSMCMV.lstFlag = h.anatomy.moiseev.lstFlag;
sInSMCMV.dims = h.anatomy.moiseev.dims;
sInSMCMV.nSrc = maxSrc;
sInSMCMV.Cavg = Rbar; % This is C-bar - only needed for MER
sInSMCMV.bMCMVS = true; % Should be ALWAYS set to true
sInSMCMV.pVal = 1; % p-value (kind of) for the peak to be considered real. Set to 1 to get all peaks
sInSMCMV.gap = gap;
sInSMCMV.bVerbose = true;
sInSMCMV.bPlotLambda = false;
sInSMCMV.bRAPBeam = false; % Choose RAP (true) or SMCMV (false). You can try both
sInSMCMV.bDoTRAP = false; % Don't ask - just set to false :)
sOutSMCMV = doSMCMV(sInSMCMV);
%% just finding max peaks in each images run -- assuming that each image in fImg corresponds to a single peak
mcmv_voxel=[]; img_max=[]; pidx=[];
for v=1:maxSrc
[img_max(v),pidx(v)] = max(sOutSMCMV.fImg(v,h.anatomy.moiseev.inside_idx)');
mcmv_voxel(v,:) = h.anatomy.moiseev.vx_grid(h.anatomy.moiseev.inside_idx(pidx(v)),:);
end
% finding voxel locations within original anatomical leadfield grid
m_idx = find_nearest_voxel(mcmv_voxel,h.inv_soln(h.current_inv_soln).leadfield.voxel_pos(h.inv_soln(h.current_inv_soln).leadfield.inside==1,:));
% true_idx = h.cfg.source.vx_idx;
% figure(998); clf; set(gcf,'color','w'); hold on;
% opt.vol_nums=1; vol = h.anatomy.mesh_volumes(3);
% [p1]=bl_plot_mesh(vol,opt); view(-90,90);
% scatter3(vx_pos(true_idx,1),vx_pos(true_idx,2),vx_pos(true_idx,3),'ks','SizeData',220,'linewidth',2);
% scatter3(vx_pos(m_idx,1),vx_pos(m_idx,2),vx_pos(m_idx,3),'r.','SizeData',220,'linewidth',2);
% weights wts=signal wts_noise = noise
h.inv_soln(h.next_inv_soln).soln.wts = zeros(size(H,3),size(H,1));
h.inv_soln(h.next_inv_soln).soln.wts(:,m_idx) = sOutSMCMV.Wr;
% h.inv_soln(h.next_inv_soln).soln.wts_noise = zeros(size(H,3),size(H,1));
h.inv_soln(h.next_inv_soln).soln.wts_noise = sOutSMCMV.Wr;
% P.img
h.inv_soln(h.next_inv_soln).soln.P.img = zeros(size(H,1),1);
h.inv_soln(h.next_inv_soln).soln.P.img(m_idx) = real(img_max);
h.inv_soln(h.next_inv_soln).soln.type = sInSMCMV.beamType;
% MCMV indices & Orientations
h.inv_soln(h.next_inv_soln).soln.MCMV_idx = m_idx;
h.inv_soln(h.next_inv_soln).soln.ori = zeros(size(H,1),3);
h.inv_soln(h.next_inv_soln).soln.ori(m_idx,:) = sOutSMCMV.U3D;
sOutSMCMV = rmfield(sOutSMCMV,'arrH'); % to reduce storage
h.inv_soln(h.next_inv_soln).Type = inv_soln;
h.inv_soln(h.next_inv_soln).soln.sOutSMCMV = sOutSMCMV;
case 'bRAPBeam' % Alex Moiseev's bRAP MCMV beamformer
% subspace MCMV
sInSMCMV.beamType = h.menu_SPA_loc_flag.String{h.menu_SPA_loc_flag.Value};
sInSMCMV.R = R;
sInSMCMV.arrN = N;
sInSMCMV.arrH = arrH;
sInSMCMV.lstFlag = h.anatomy.moiseev.lstFlag;
sInSMCMV.dims = h.anatomy.moiseev.dims;
sInSMCMV.nSrc = maxSrc;
sInSMCMV.Cavg = Rbar; % This is C-bar - only needed for MER
sInSMCMV.bMCMVS = true; % Should be ALWAYS set to true
sInSMCMV.pVal = 1; % p-value (kind of) for the peak to be considered real. Set to 1 to get all peaks
sInSMCMV.gap = gap;
sInSMCMV.bVerbose = true;
sInSMCMV.bPlotLambda = false;
sInSMCMV.bRAPBeam = true; % Choose RAP (true) or SMCMV (false). You can try both
sInSMCMV.bDoTRAP = false; % Don't ask - just set to false :)
sOutSMCMV = doSMCMV(sInSMCMV);
%% just finding max peaks in each images run -- assuming that each image in fImg corresponds to a single peak
mcmv_voxel=[]; img_max=[]; pidx=[];
for v=1:maxSrc
[img_max(v),pidx(v)] = max(sOutSMCMV.fImg(v,h.anatomy.moiseev.inside_idx)');
mcmv_voxel(v,:) = h.anatomy.moiseev.vx_grid(h.anatomy.moiseev.inside_idx(pidx(v)),:);
end
% finding voxel locations within original anatomical leadfield grid
m_idx = find_nearest_voxel(mcmv_voxel,h.inv_soln(h.current_inv_soln).leadfield.voxel_pos(h.inv_soln(h.current_inv_soln).leadfield.inside==1,:));
% weights wts=signal wts_noise = noise
h.inv_soln(h.next_inv_soln).soln.wts = zeros(size(H,3),size(H,1));
h.inv_soln(h.next_inv_soln).soln.wts(:,m_idx) = sOutSMCMV.Wr;
% h.inv_soln(h.next_inv_soln).soln.wts_noise = zeros(size(H,3),size(H,1));
h.inv_soln(h.next_inv_soln).soln.wts_noise = sOutSMCMV.Wr;
% P.img
h.inv_soln(h.next_inv_soln).soln.P.img = zeros(size(H,1),1);
h.inv_soln(h.next_inv_soln).soln.P.img(m_idx) = real(img_max);
h.inv_soln(h.next_inv_soln).soln.type = sInSMCMV.beamType;
% MCMV indices & Orientations
h.inv_soln(h.next_inv_soln).soln.MCMV_idx = m_idx;
h.inv_soln(h.next_inv_soln).soln.ori = zeros(size(H,1),3);
h.inv_soln(h.next_inv_soln).soln.ori(m_idx,:) = sOutSMCMV.U3D;
sOutSMCMV = rmfield(sOutSMCMV,'arrH'); % to reduce storage
h.inv_soln(h.next_inv_soln).Type = inv_soln;
h.inv_soln(h.next_inv_soln).soln.sOutSMCMV = sOutSMCMV;
case 'TrapMUSIC' % Alex Moiseev's code for TRAP MUSIC
% TRAP MUSIC
sInTRAP.R = R; % Full covariance
sInTRAP.arrN = N; % Noise covariance
sInTRAP.arrH = arrH; % Lead fields
sInTRAP.lstFlag = h.anatomy.moiseev.lstFlag; % Flags
sInTRAP.dims = h.anatomy.moiseev.dims; % ROI dimensions in voxels
sInTRAP.nSrc = maxSrc; % Max number of sources to extract
sInTRAP.gap = gap; % Min allowed number nodes between the peaks (default 2)
sInTRAP.bPlotLambda = false; % Set to false
sInTRAP.bVerbose = true; % Enables more printouts
sOutTRAP = trapMUSIC(sInTRAP);
%% just finding max peaks in each images run -- assuming that each image in fImg corresponds to a single peak
mcmv_voxel=[]; img_max=[]; pidx=[];
for v=1:maxSrc
[img_max(v),pidx(v)] = max(sOutTRAP.fImg(v,h.anatomy.moiseev.inside_idx)');
mcmv_voxel(v,:) = h.anatomy.moiseev.vx_grid(h.anatomy.moiseev.inside_idx(pidx(v)),:);
end
% finding voxel locations within original anatomical leadfield grid
m_idx = find_nearest_voxel(mcmv_voxel,h.inv_soln(h.current_inv_soln).leadfield.voxel_pos(h.inv_soln(h.current_inv_soln).leadfield.inside==1,:));
% weights wts=signal wts_noise = noise
h.inv_soln(h.next_inv_soln).soln.wts = zeros(size(H,3),size(H,1));
h.inv_soln(h.next_inv_soln).soln.wts(:,m_idx) = sOutTRAP.Wr;
% h.inv_soln(h.next_inv_soln).soln.wts_noise = zeros(size(H,3),size(H,1));
h.inv_soln(h.next_inv_soln).soln.wts_noise = sOutTRAP.Wr;
% P.img
h.inv_soln(h.next_inv_soln).soln.P.img = zeros(size(H,1),1);
h.inv_soln(h.next_inv_soln).soln.P.img(m_idx) = real(img_max);
h.inv_soln(h.next_inv_soln).soln.type = inv_soln;
% MCMV indices & Orientations
h.inv_soln(h.next_inv_soln).soln.MCMV_idx = m_idx;
h.inv_soln(h.next_inv_soln).soln.ori = zeros(size(H,1),3);
h.inv_soln(h.next_inv_soln).soln.ori(m_idx,:) = sOutTRAP.U3D;
sOutTRAP = rmfield(sOutTRAP,'arrH'); % to reduce storage
h.inv_soln(h.next_inv_soln).Type = inv_soln;
h.inv_soln(h.next_inv_soln).soln.sOutTRAP = sOutTRAP;
% figure(998); clf; set(gcf,'color','w'); hold on;
% opt.vol_nums=1; vol = h.anatomy.mesh_volumes(3);
% true_idx = h.cfg.source.vx_idx
% [p1]=bl_plot_mesh(vol,opt); view(-90,90);
% scatter3(vx_pos(true_idx,1),vx_pos(true_idx,2),vx_pos(true_idx,3),'ks','SizeData',220,'linewidth',2);
% scatter3(vx_pos(m_idx,1),vx_pos(m_idx,2),vx_pos(m_idx,3),'r.','SizeData',220,'linewidth',2);
end
%% Listbox Name
h.inv_soln(h.next_inv_soln).ListBox_name = sprintf('%s %s (%.f-%.f ms) %s',h.inv_soln(h.next_inv_soln).Type, h.menu_SPA_loc_flag.String{h.menu_SPA_loc_flag.Value} ,h.inv_soln(h.current_inv_soln).params.act_int*1000,h.inv_soln(h.current_inv_soln).headmodel_type);
%% original map that gets changed with spatiotemporal mapping
h.inv_soln(h.current_inv_soln).org_img = h.inv_soln(h.current_inv_soln).soln.P.img;
min_max = round([min(h.inv_soln(h.current_inv_soln).soln.P.img) max(h.inv_soln(h.current_inv_soln).soln.P.img)],3,'significant');
if min_max(1)==min_max(2); min_max(2)=min_max(1)+1;
elseif min_max(1)>min_max(2); min_max(2)=min_max(1)+1;
end
h.inv_soln(h.current_inv_soln).soln.plot_min_max = real(min_max);
h.inv_soln(h.current_inv_soln).soln.plot_thresh = round(h.inv_soln(h.current_inv_soln).soln.plot_min_max(2)*.5,3,'significant'); %h.inv_soln(h.current_inv_soln).soln.null_thresh;
h.edit_3D_min_max.String = sprintf('%.3f %.3f',h.inv_soln(h.current_inv_soln).soln.plot_min_max);
% updating threshold slider
h.slider_3D_image_thresh.Max = h.inv_soln(h.current_inv_soln).soln.plot_min_max(2);
h.slider_3D_image_thresh.Min = h.inv_soln(h.current_inv_soln).soln.plot_min_max(1);
h.slider_3D_image_thresh_text_max.String = num2str(h.slider_3D_image_thresh.Max);
if h.slider_3D_image_thresh.Value>h.slider_3D_image_thresh.Max
h.slider_3D_image_thresh.Value = h.slider_3D_image_thresh.Max;
elseif h.slider_3D_image_thresh.Value < h.slider_3D_image_thresh.Min
h.slider_3D_image_thresh.Value = h.slider_3D_image_thresh.Min;
end
bs_update_3D_listbox();
bs_plot_inv_soln;
fprintf('Inverse Modeling Completed for %s\n',inv_soln);
%% converting to single precision for reduce memory storage
h.inv_soln(h.next_inv_soln).soln.wts = single(h.inv_soln(h.next_inv_soln).soln.wts);
h.inv_soln(h.next_inv_soln).soln.ori = single(h.inv_soln(h.next_inv_soln).soln.ori);
h.inv_soln(h.next_inv_soln).soln.ori = single(h.inv_soln(h.next_inv_soln).soln.ori);
h.inv_soln(h.next_inv_soln).soln.P.img = single(h.inv_soln(h.next_inv_soln).soln.P.img);
% reducing redundancy to save storage space
if isfield(h.inv_soln(h.next_inv_soln).leadfield,'cfg')
h.inv_soln(h.next_inv_soln).leadfield = rmfield(h.inv_soln(h.next_inv_soln).leadfield,{'cfg'});
end
% if isfield(h.inv_soln(h.next_inv_soln).leadfield,'leadfield')
% h.inv_soln(h.next_inv_soln).leadfield = rmfield(h.inv_soln(h.next_inv_soln).leadfield,{'leadfield'});
% end
% if isfield(h.inv_soln(h.next_inv_soln).leadfield,'H')
% h.inv_soln(h.next_inv_soln).leadfield = rmfield(h.inv_soln(h.next_inv_soln).leadfield,{'H'});
% end
h.inv_soln(h.next_inv_soln).leadfield = double2single(h.inv_soln(h.next_inv_soln).leadfield);
h.inv_soln(h.current_inv_soln).TFR_results =[];
% catch me
% display(me)
% errordlg('ERROR! Inverse Source Modeling Failed');
% end
if h.monte_carlo_flag ~= 1
h.waitfor_panel.Visible='off'; h.waitfor_txt.String = sprintf('Default Message');
end
% close(hm);