-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsc1sc2_connectivity_old.m
1491 lines (1359 loc) · 64.2 KB
/
sc1sc2_connectivity_old.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
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
function varargout=sc1sc2_connectivity(what,varargin)
% Matlab script to do connectivity analyses for sc1 and sc2 experiments
% Does time series extraction
%==========================================================================
type=[];
% % (1) Directories
% rootDir = '/Users/jdiedrichsen/Data/super_cerebellum';
rootDir = '/Volumes/MotorControl/data/super_cerebellum_new';
wbDir = fullfile(rootDir,'sc1','surfaceWB');
sc1Dir = [rootDir '/sc1'];
sc2Dir = [rootDir '/sc2'];
behavDir = ['data'];
imagingDir = ['imaging_data'];
imagingDirRaw = ['imaging_data_raw'];
dicomDir = ['imaging_data_dicom'];
anatomicalDir = ['anatomicals'];
suitDir = ['suit'];
caretDir = ['surfaceCaret'];
regDir = ['RegionOfInterest'];
connDir = ['connectivity_cerebellum'];
%==========================================================================
% % (2) Hemisphere and Region Names
numReg = length(regName);
Hem = {'L','R'};
hemname = {'CortexLeft','CortexRight'};
subj_name = {'s01','s02','s03','s04','s05','s06','s07','s08','s09','s10','s11',...
's12','s13','s14','s15','s16','s17','s18','s19','s20','s21','s22','s23','s24',...
's25','s26','s27','s28','s29','s30','s31'};
returnSubjs=[2,3,4,6,8,9,10,12,14,15,17,18,19,20,21,22,24,25,26,27,28,29,30,31];
expStr = {'sc1','sc2'};
%==========================================================================
switch(what)
case 'make_bucknermargin' % Generates the Buckner Margin ROI
kernel = 8;
threshold = 0.05;
for s=2:22
Vcort=spm_vol(fullfile(sc1Dir,regDir,'data',subj_name{s},'cortical_mask_grey_corr.nii'));
Vcereb=spm_vol(fullfile(sc1Dir,suitDir,'anatomicals',subj_name{s},'cereb_prob_corr_grey.nii'));
% smooth cerebellar image using 6 mm kernel
Vcort.dat = spm_read_vols(Vcort);
X = spm_read_vols(Vcereb);
Vcereb.dat = zeros(size(X));
spm_smooth(X,Vcereb.dat,kernel);
linind = find(Vcort.dat>0);
[i,j,k]=ind2sub(Vcort.dim,linind);
[i1,j1,k1]=spmj_affine_transform(i,j,k,inv(Vcereb.mat)*Vcort.mat);
Vcereb.dt =[64 0]; Vcereb.pinfo=[1 0 0]';
Data = spm_sample_vol(Vcereb,i1,j1,k1,1);
Vcort.dat(linind(Data<threshold))=0; % Set these parts to zero
Vcort.fname = (fullfile(sc1Dir,regDir,'data',subj_name{s},'bucknermargin.nii'));
spm_write_vol(Vcort,Vcort.dat);
end;
case 'TS_correlation_map' % Seed based correlation analysis for 1 location - for visualization
sn=varargin{1}; % subjNum
glm=varargin{2}; % glmNum
location=[-8 -50 -11]; % [x,y,z]
glmDir =[sc1Dir sprintf('/GLM_firstlevel_%d',glm)];
subjs=length(sn);
for s=1:subjs,
glmDirSubj=fullfile(glmDir, subj_name{sn(s)});
load(fullfile(glmDirSubj,'SPM_light.mat'));
T=load(fullfile(glmDirSubj,'SPM_info.mat'));
% load data
tic;
SPM=spmj_move_rawdata(SPM,fullfile(imagingDir,subj_name{sn(s)}));
% Get the raw data files
V=SPM.xY.VY;
fprintf('Moved %d\n',toc);
% Determine locations in image
VresMS = spm_vol(fullfile(glmDirSubj,'ResMS.nii'));
XresMS =spm_read_vols(VresMS);
linIndx = find(XresMS>0);
[i,j,k]=ind2sub(VresMS.dim,linIndx);
SS1 = zeros(2,length(i));
SS2 = zeros(2,1);
SSC = zeros(2,length(i));
[i1,j1,k1]=spmj_affine_transform(location(1),location(2),location(3),inv(VresMS.mat));
tar=[round(i1) round(j1) round(k1)];
target = findrow([i j k],tar);
if (isempty(target))
error('target not in volume');
end;
% Now loop over runs and get the
tic;
for r=1:16
fprintf('%d.',r);
row = SPM.Sess(r).row;
col = SPM.Sess(r).col; % Regressors of interest
col1 = [SPM.Sess(r).col SPM.xX.iB(r)];
Y = zeros(length(row),length(i));
for t=1:length(row)
Y(t,:) = spm_sample_vol(V(row(t)),i,j,k,0)'; % Data is N x P
end;
Yfilt = SPM.xX.W(row,row)*Y;
B = SPM.xX.pKX(col1,row)*Yfilt; %-Parameter estimates
Yhat = SPM.xX.xKXs.X(row,col1)*B; %- predicted values
Yres = Yfilt - Yhat; %- residuals
Ya{1}=bsxfun(@minus,Yhat,mean(Yhat));
Ya{2}=bsxfun(@minus,Yres,mean(Yres));
for ty=1:2
SS1(ty,:)=SS1(ty,:)+sum(Ya{ty}.^2);
SS2(ty,:)=SS2(ty,:)+sum(Ya{ty}(:,target).^2);
SSC(ty,:)=SSC(ty,:)+Ya{ty}(:,target)'*Ya{ty};
end;
end;
DE = sqrt(bsxfun(@times,SS1,SS2));
CR = SSC./DE;
for ty=1:2
Vo= VresMS;
Vo.fname = fullfile(glmDirSubj,sprintf('correlation_map_%d.nii',ty));
Vo.dt = [16 0];
Xo = zeros(Vo.dim);
Xo (linIndx)=CR(ty,:);
spm_write_vol(Vo,Xo);
end;
end;
case 'TS_get_meants' % Get univariately pre-whitened mean times series for each region
% sc1_connectivity('TS_get_meants',[2 3 6 8 9 10 12 17:22],'sc2',4,'162_tessellation_hem');
% sc1_connectivity('TS_get_meants',[2:22],'sc1',4,'162_tessellation_hem');
sn=varargin{1}; % subjNum
exper=varargin{2}; % Experiment string 'sc1' or 'sc2'
glm=varargin{3}; % glmNum
type=varargin{4}; % 'cortical_lobes','whole_brain','yeo','desikan','cerebellum','yeo_cerebellum'
if glm == 6; % Select the right directory for GLMs different from GLM4
imagingDir=[imagingDir '_aggr'];
end
B = [];
glmDir =fullfile(rootDir,exper,sprintf('/GLM_firstlevel_%d',glm));
subjs=length(sn);
for s=1:subjs,
glmDirSubj=fullfile(glmDir, subj_name{sn(s)});
load(fullfile(glmDirSubj,'SPM_light.mat'));
% load(fullfile(glmDirSubj,'SPM.mat'));
T=load(fullfile(glmDirSubj,'SPM_info.mat'));
% load data
tic;
load(fullfile(sc1Dir,regDir,'data',subj_name{sn(s)},sprintf('regions_%s.mat',type))); % 'regions' are defined in 'ROI_define'
SPM=spmj_move_rawdata(SPM,fullfile(rootDir,'sc1',imagingDir,subj_name{sn(s)}));
% Get the raw data files
V=SPM.xY.VY;
VresMS = spm_vol(fullfile(glmDirSubj,'ResMS.nii'));
fprintf('Moved %d\n',toc);
% Get time series data
tic;
Y = region_getdata(V,R); % Data is N x P
resMS = region_getdata(VresMS,R);
fprintf('Data %d\n',toc);
% Spatially prewhiten and average
Data=zeros(numel(V),numel(R));
for r=1:numel(R), % R is the output 'regions' structure from 'ROI_define'
Y{r}=bsxfun(@rdivide,Y{r},sqrt(resMS{r}));
Data(:,r)=nanmean(Y{r},2);
end;
clear Y;
% Redo regression
reg_interest=[SPM.xX.iH SPM.xX.iC];
Yfilt = SPM.xX.W*Data;
B = SPM.xX.pKX*Yfilt; %-Parameter estimates
Yres = spm_sp('r',SPM.xX.xKXs,Yfilt); %-Residuals
% Decompose betas into mean and residuals
Z=indicatorMatrix('identity',T.cond);
Bm = Z*pinv(Z)*B(reg_interest,:); % Mean betas
Br = B(reg_interest,:)-Bm;
Yhatm = SPM.xX.xKXs.X(:,reg_interest)*Bm; %- predicted values
Yhatr = SPM.xX.xKXs.X(:,reg_interest)*Br; %- predicted values
filename=(fullfile(rootDir,exper,regDir,sprintf('glm%d',glm),subj_name{sn(s)},sprintf('ts_%s.mat',type)));
save(filename,'Data','Yres','Yhatm','Yhatr','B');
fprintf('ts saved for %s (%s) for %s \n',subj_name{sn(s)},sprintf('glm%d',glm),type);
end
case 'TS_get_ts' % Get univariately pre-whitened time series for each voxel of a region
% sc1_connectivity('TS_get_ts',[2:22],'sc2',4,'Cerebellum_grey');
% sc1_connectivity('TS_get_ts',[2 3 6 8 9 10 12 17:22],'sc2',4,'Cerebellum_grey');
sn=varargin{1}; % subjNum
exper=varargin{2}; % Experiment string 'sc1' or 'sc2
glm=varargin{3}; % glmNum
type=varargin{4}; % 'Cerebellum_grey'
B = [];
glmDir =fullfile(rootDir,exper,sprintf('GLM_firstlevel_%d',glm));
subjs=length(sn);
for s=1:subjs,
glmDirSubj=fullfile(glmDir, subj_name{sn(s)});
load(fullfile(glmDirSubj,'SPM_light.mat'));
% load(fullfile(glmDirSubj,'SPM.mat'));
T=load(fullfile(glmDirSubj,'SPM_info.mat'));
% load data
tic;
load(fullfile(sc1Dir,regDir,'data',subj_name{sn(s)},sprintf('regions_%s.mat',type))); % 'regions' are defined in 'ROI_define'
SPM=spmj_move_rawdata(SPM,fullfile(rootDir,'sc1',imagingDir,subj_name{sn(s)}));
% Get the raw data files
V=SPM.xY.VY;
VresMS = spm_vol(fullfile(glmDirSubj,'ResMS.nii'));
fprintf('Moved %d\n',toc);
% Get time series data
tic;
Y = region_getdata(V,R{1}); % Data is N x P
resMS = region_getdata(VresMS,R{1});
fprintf('Data %d\n',toc);
% Spatially prewhiten
Y=bsxfun(@rdivide,Y,sqrt(resMS));
% % Redo regression
reg_interest=[SPM.xX.iH SPM.xX.iC];
Yfilt = SPM.xX.W*Y;
B = SPM.xX.pKX*Yfilt; %-Parameter estimates
Yres = spm_sp('r',SPM.xX.xKXs,Yfilt); %-Residuals
% Decompose betas into mean and residuals
Z=indicatorMatrix('identity',T.cond);
Bm = Z*pinv(Z)*B(reg_interest,:); % Mean betas
Br = B(reg_interest,:)-Bm;
Yhatm = SPM.xX.xKXs.X(:,reg_interest)*Bm; %- predicted values
Yhatr = SPM.xX.xKXs.X(:,reg_interest)*Br; %- predicted values
filename=(fullfile(rootDir,exper,regDir,sprintf('glm%d',glm),subj_name{sn(s)},sprintf('ts_%s.mat',type)));
save(filename,'Yres','Yhatm','Yhatr','B');
fprintf('ts saved for %s (%s) for %s \n',subj_name{sn(s)},sprintf('glm%d',glm),type);
end
case 'TS_allsubj' % Make a structure of all cortical time series of all subject - also seperate out residual from
sn=goodsubj; % Take only good subjects
glm=4; % glmNum
type='162_tessellation_hem'; % 'cortical_lobes','whole_brain','yeo','desikan','cerebellum','yeo_cerebellum'
session_specific = 1;
expDir = sc1Dir;
vararginoptions(varargin,{'sn','glm4','type','session_specific'});
glmDir =fullfile(expDir,sprintf('/GLM_firstlevel_%d',glm));
numSubj=length(sn);
for s=1:numSubj
fprintf('%d\n',s);
% load condition data
glmDirSubj=fullfile(glmDir, subj_name{sn(s)});
load(fullfile(glmDirSubj,'SPM_light.mat'));
T=load(fullfile(glmDirSubj,'SPM_info.mat'));
% load data
filename=(fullfile(expDir,regDir,sprintf('glm%d',glm),subj_name{sn(s)},sprintf('ts_%s.mat',type)));
S=load(filename);
reg_interest=[SPM.xX.iH SPM.xX.iC];
% Decompose betas into mean and residuals
if (session_specific)
Z=indicatorMatrix('hierarchicalI',[T.sess T.cond]);
else
Z=indicatorMatrix('identity',T.cond);
end;
B(:,:,s)=S.B;
Yres(:,:,s)=S.Yres;
Bm = Z*pinv(Z)*S.B(reg_interest,:); % Mean betas
Br = S.B(reg_interest,:)-Bm;
Yhatm(:,:,s) = SPM.xX.xKXs.X(:,reg_interest)*Bm; %- predicted values
Yhatr(:,:,s) = SPM.xX.xKXs.X(:,reg_interest)*Br; %- predicted values
end;
if (session_specific)
save(fullfile(expDir,regDir,sprintf('glm%d',glm),sprintf('ts_%s_all_se.mat',type)),'Yres','Yhatm','Yhatr','B');
else
save(fullfile(expDir,regDir,sprintf('glm%d',glm),sprintf('ts_%s_all.mat',type)),'Yres','Yhatm','Yhatr','B');
end;
case 'get_meanBeta_cerebellum' % Take the time series structure and extract mean beta for cross-session modeling
% This stores the betas like the occurr in the GLM
type = '162_tessellation_hem'; % 'Cerebellum_grey';
exper = {'sc1','sc2'};
sn = goodsubj;
glm = 4;
vararginoptions(varargin,{'sn','glm','type'});
% Loop over experiments
for e=1:2
myRegDir = fullfile(rootDir,exper{e},regDir);
for s=1:length(sn);
fprintf('%d\n',sn(s));
filename = (fullfile(myRegDir,sprintf('glm%d',glm),subj_name{sn(s)},sprintf('ts_%s.mat',type)));
D = load(filename);
glmDirSubj = fullfile(rootDir,exper{e},sprintf('GLM_firstlevel_%d',glm),subj_name{sn(s)});
T = load(fullfile(glmDirSubj,'SPM_info.mat'));
if (strcmp(type,'162_tessellation_hem'))
Tcort=load(fullfile(rootDir,'sc1',regDir,'data','162_reorder.mat'));
Tcort=getrow(Tcort,Tcort.good==1);
D.B = D.B(:,Tcort.regIndx);
D.Yres = D.Yres(:,Tcort.regIndx);
end;
% Now generate mean estimates per session
% Note that in the new SPM_info Instruction is coded as cond =0
T.cond=T.cond+1; % This is only for generating the avergaging matrix
for se=1:2
X=indicatorMatrix('identity_p',T.cond.*(T.sess==se));
B{s,se}=pinv(X)*D.B(1:size(X,1),:); % This calculates average beta across runs, skipping intercepts
ResMS{s,1} = sum(D.Yres.^2)./size(D.Yres,1); % Residual mean square for this voxel/region
end;
end;
save(fullfile(myRegDir,sprintf('glm%d',glm),sprintf('mbeta_%s_all.mat',type)),'B','ResMS');
end;
case 'sample_DesignMatrix'
exper=varargin{1};
glm = varargin{2};
wdir = fullfile(rootDir,expStr{exper},sprintf('GLM_firstlevel_%d',glm));
load(fullfile(wdir,'s02','SPM_light.mat'));
Sess = SPM.Sess; % Session Structure
X = SPM.xX.X; % Design matrix
wX = SPM.xX.xKXs.X; % Filtered design matrix
save(fullfile(wdir,'SampleDesignMatrix.mat'),'Sess','X','wX');
case 'cortical_covariances' % Covariances between cortical areas in predicted time series
sn = goodsubj;
glm = 4;
type = '162_tessellation_hem';
vararginoptions(varargin,{'sn','glm','type'});
% Figure out the right ones to use
X=load(fullfile(sc1Dir,regDir,'data','162_reorder.mat'));
Xx=getrow(X,X.newIndx);
Xx=getrow(Xx,Xx.good==1);
% Calculate correlations
for s=1:length(sn)
% load Individual data
filename=(fullfile(sc1Dir,regDir,sprintf('glm%d',glm),subj_name{sn(s)},sprintf('ts_%s.mat',type)));
load(filename);
Y = Yhatm + Yhatr; % Predicted timeseries(Both mean and variable)
predT=Y(:,Xx.regIndx2); % Remove the bad regions
% Covariance and correlation
COV(:,:,s)=cov(predT);
COR(:,:,s)=corrcov(COV(:,:,s));
% Variance inflation factor for each
T.SN(s,1)=s;
T.VAR(s,:)=diag(COV(:,:,s)); % Variance of the task-related betas: How much activity could we induce?
T.VIF(s,:)=diag(inv(COR(:,:,s)))'; % Variance inflation factor: How much does the region suffer from co-linearity?
T.VARB(s,:)=diag(inv(COV(:,:,s))); % Combination of the last two: How well would we estimate connectivity weights from this region?
end;
% Plot group mean in flatmap
data{1}=sqrt(mean(T.VAR,1));
data{2}=sqrt(mean(T.VIF,1));
data{3}=sqrt(mean(T.VARB,1));
plottitle{1}='Variance';
plottitle{2}='Variance Inflation Factor';
plottitle{3}='Combined';
figure;
set(gcf,'color','w');
ppos=1;
for i=1:3
subplot(3,2,ppos);
ii=find(Xx.hem==1)';
sc1sc2_connectivity('plot_cortex',1,data{i}(ii));
axis off;
tmp=get(gca,'position');
set(gca,'position',[tmp(1) tmp(2) 0.9*tmp(3) tmp(4)]);
title(plottitle{i})
ppos=ppos+1;
subplot(3,2,ppos);
ii=find(Xx.hem==2)';
sc1sc2_connectivity('plot_cortex',2,data{i}(ii));
axis off;
tmp=get(gca,'position');
set(gca,'position',[tmp(1)-0.1 tmp(2) 1.2*tmp(3) tmp(4)]);
ppos=ppos+1;
colorbar;
end
set(gcf,'paperposition',[10 10 7 7])
wysiwyg
case 'cortical_pattern_consistency_all' % Consistency based on betas
T=load(fullfile(regDir,'glm4','ts_162_tessellation_hem_all.mat'));
X=load(fullfile(regDir,'data','162_reorder.mat'));
ii = X.regIndx(X.good==1);
X=getrow(X,X.good==1);
% Project down to condition space
for s=1:numSubj
Yc(:,:,s)=X*pinv(X)*B(:,:,s);
end;
C1=intersubj_corr(Y);
C2=intersubj_corr(Yc);
C3=intersubj_corr(Y-Yc);
subplot(3,1,1);
imagesc(C1);
title('alldata');
subplot(3,1,2);
imagesc(C2);
title('means');
subplot(3,1,3);
imagesc(C3);
title('residuals');
varargout={C1,C2};
case 'cortical_pattern_consistency_session' % Between sessions / subjects
load(fullfile(encodeDir,'162_tesselation_hem_all.mat'));
numSubj=length(goodsubj);
sessNum = (partNum>8)+1;
for se=1:2
indx = find(sessNum==se);
X=indicatorMatrix('identity',condNum(indx,:));
for subj = 1:numSubj
Yp(:,:,subj+(se-1)*numSubj)=X*pinv(X)*Y(indx,:,subj);
Yf(:,:,subj+(se-1)*numSubj)=Y(indx,:,subj);
end;
end;
Cf=intersubj_corr(Yf);
Cp=intersubj_corr(Yp);
Cr=intersubj_corr(Yf-Yp);
subplot(3,2,1);
imagesc(Cf);
title('betas');
subplot(3,2,2);
[a1,a2,a3]=bwse_corr(Cf);
barplot([],[a2 a1 a3],'XTickLabel',{'DiffSess','DiffSubj','DiSeDiSu'});
subplot(3,2,3);
imagesc(Cp);
[a1,a2,a3]=bwse_corr(Cp);
title('means');
subplot(3,2,4);
barplot([],[a2 a1 a3],'XTickLabel',{'DiffSess','DiffSubj','DiSeDiSu'});
subplot(3,2,5);
imagesc(Cr);
[a1,a2,a3]=bwse_corr(Cr);
title('residuals');
subplot(3,2,6);
barplot([],[a2 a1 a3],'XTickLabel',{'DiffSess','DiffSubj','DiSeDiSu'});
case 'cortical_pattern_consistency_ts' % Consistency between time series
T=load(fullfile(regDir,'glm4','ts_162_tessellation_hem_all.mat'));
type = {'Yhatm','Yhatr','Yres'};
X=load(fullfile(regDir,'data','162_reorder.mat'));
ii = X.regIndx(X.good==1);
for t=1:3
Data = T.(type{t})(:,ii,:);
Data = bsxfun(@minus,Data,mean(Data,1));
Cor=intersubj_corr(Data);
subplot(1,3,t);
imagesc(Cor);
fprintf('%s: %2.2f\n',type{t},1-mean(squareform(1-Cor)));
end;
set(gcf,'PaperPosition',[2 2 6 2]);
case 'TS_power'
% Plot power spectrum of signal in cortical ROIs
sn=[2:22]; % subjNum
glm=4; % glmNum
type='162_tessellation_hem'; % 'cortical_lobes','whole_brain','yeo','desikan','cerebellum','yeo_cerebellum'
vararginoptions(varargin,{'sn','glm4','type','session_specific'});
glmDir =[baseDir sprintf('/GLM_firstlevel_%d',glm)];
numSubj=length(sn);
for s=1:numSubj
fprintf('%d\n',s);
% load data
glmDirSubj=fullfile(glmDir, subj_name{sn(s)});
load(fullfile(glmDirSubj,'SPM_light.mat'));
T=load(fullfile(glmDirSubj,'SPM_info.mat'));
% load data
filename=(fullfile(regDir,sprintf('glm%d',glm),subj_name{sn(s)},sprintf('ts_%s.mat',type)));
load(filename);
TS.Y=[];
for b=1:length(SPM.Sess);
row=SPM.Sess(b).row;
TS.Y=[TS.Y Data(row,:)];
end;
TS.t=[0:length(row)-1]';
TS.Y=bsxfun(@minus,TS.Y,mean(TS.Y));
F=tsFT(TS);
F.P=F.F.*conj(F.F);
mid=find(F.om>0);
mid=mid(1);
semilogy(F.om(mid:end,1),mean(F.P(mid:end,:),2));
set(gca,'YLim',[1000 300000],'Box','off');
keyboard;
end;
case 'TS_interregion_corr' % How consistent is the cortical connectivity pattern through the three signal sources
T=load(fullfile(sc1Dir,regDir,'glm4','ts_162_tessellation_hem_all_se.mat'));
numsubj = size(T.Yres,3);
N = size(T.Yres,1);
T.Yhat = T.Yhatm + T.Yhatr;
T.Y = T.Yhat + T.Yres;
X=load(fullfile(sc1Dir,regDir,'data','162_reorder.mat'));
Xx=getrow(X,X.newIndx);
ii = Xx.regIndx(Xx.good==1);
Xx=getrow(Xx,Xx.good==1);
type ={'Yres','Yhat','Y'};
blockSize = N/16;
Z=kron(eye(16),ones(blockSize,1));
for t=1:3
Data = T.(type{t})(:,ii,:);
for s=1:21
Data(:,:,s) = Data(:,:,s) - Z*pinv(Z)*Data(:,:,s); % subtract run means
meanSD(s,t)=mean((sum(Data(:,:,s).^2)/N)); % How much of the mean signal is this?
end;
C{t} = zeros(length(ii),length(ii),numsubj,2);
R{t} = zeros(length(ii),length(ii),numsubj,2);
for s=1:numsubj
for se=1:2
idx = [1:N/2]+(se-1)*N/2;
C{t}(:,:,s,se)=cov(Data(idx,:,s));
R{t}(:,:,s,se)=corrcov(C{t}(:,:,s,se));
end;
end;
end;
DD=[];
pairs=[1 1;2 2;3 3;1 2;1 3;2 3];
pairtype=[1 2 3 4 5 6];
for p=1:size(pairs,1);
for s=1:numsubj
% fprintf('%d %d\n',s,p);
D.SN(s,1) = s;
D.pair(s,1:2) = pairs(p,:);
r1 = mycorr([rsa_vectorizeRDM(R{pairs(p,1)}(:,:,s,1))',rsa_vectorizeRDM(R{pairs(p,2)}(:,:,s,2))']);
r2 = mycorr([rsa_vectorizeRDM(R{pairs(p,2)}(:,:,s,1))',rsa_vectorizeRDM(R{pairs(p,1)}(:,:,s,2))']);
D.corr(s,1)=(r1+r2)/2;
D.pairtype(s,1) = pairtype(p);
end;
DD=addstruct(DD,D);
end;
subplot(2,2,4);
barplot(DD.pairtype,DD.corr);
set(gca,'XTickLabel',type)
title('Correlation of Correlation');
% Plots for lobes
borders=find([diff(Xx.lobe)~=0;0] & Xx.lobe>0);
for t=1:3
subplot(2,2,t);
mR = mean(mean(R{t},3),4);
imagesc(mR);
drawline(borders);
drawline(borders,'dir','horz');
end;
set(gcf,'PaperPosition',[2 2 6 6]);
wysiwyg;
% save(fullfile(regDir,'glm4','Covariance_by_session.mat'),...
% 'C','R','Xx');
varargout={DD,meanSD};
case 'get_mbeta_all' % Gets the mean betas for each session and experiment
exper = {'sc1','sc2'};
glm = 4;
yname = 'Cerebellum_grey';
xname = '162_tessellation_hem';
incInstr = 0; % Include instruction?
vararginoptions(varargin,{'exper','glm','yname','xname','incInstr'});
% Load the betas for all the tasks
for e=1:2
myRegDir = fullfile(rootDir,exper{e},regDir);
YD{e}=load(fullfile(myRegDir,sprintf('glm%d',glm),sprintf('mbeta_%s_all.mat',yname)));
XD{e}=load(fullfile(myRegDir,sprintf('glm%d',glm),sprintf('mbeta_%s_all.mat',xname)));
end;
% Make an integrated structure, either including instruction or
% rest
if (incInstr)
T=dload(fullfile(rootDir,'sc1_sc2_taskConds_GLM.txt'));
else
T=dload(fullfile(rootDir,'sc1_sc2_taskConds.txt'));
end;
% Make an integrated Structure
T1=T;
T2=T;
T1.sess=ones(length(T.condNum),1)*1;
T2.sess=ones(length(T.condNum),1)*2;
T=addstruct(T1,T2);
for s=1:size(YD{1}.B,1)
X{s}=[];
Y{s}=[];
for se=1:2
if (incInstr==0) % If no instruction, add rest and center within session
for e=1:2
YD{e}.B{s,se}=[YD{e}.B{s,se}(2:end,:);zeros(1,size(YD{e}.B{s,se},2))];
XD{e}.B{s,se}=[XD{e}.B{s,se}(2:end,:);zeros(1,size(XD{e}.B{s,se},2))];
YD{e}.B{s,se}=bsxfun(@minus,YD{e}.B{s,se},mean(YD{e}.B{s,se}));
XD{e}.B{s,se}=bsxfun(@minus,XD{e}.B{s,se},mean(XD{e}.B{s,se}));
end;
end;
Y{s}=[Y{s};YD{1}.B{s,se};YD{2}.B{s,se}];
X{s}=[X{s};XD{1}.B{s,se};XD{2}.B{s,se}];
end;
end;
varargout = {X,Y,T};
case 'mbeta_reliability' % Get beta reliability only on common task conditions
% In the cerebellum (Y)
glm =4;
vararginoptions(varargin,{'glm'});
[X,Y,T]=sc1sc2_connectivity('get_mbeta_all','glm',glm);
indx=find(T.overlap==1);
T=getrow(T,indx);
[condn,ind] = unique(T.condNumUni);
condNames = T.condNames(ind);
RR=[];
for s=1:length(Y) % Loop over subjects
D=Y{s}(indx,:); % Get the data
B=indicatorMatrix('identity',T.StudyNum*2+T.sess-2);
D=D-B*pinv(B)*D; % Subtract the mean pattern within each study and session
in=~isnan(sum(D));
for c=1:length(condn)
R.cond=condn(c);
R.sn=s;
i11 = find(T.condNumUni==condn(c) & T.StudyNum==1 & T.sess==1);
i12 = find(T.condNumUni==condn(c) & T.StudyNum==1 & T.sess==2);
i21 = find(T.condNumUni==condn(c) & T.StudyNum==2 & T.sess==1);
i22 = find(T.condNumUni==condn(c) & T.StudyNum==2 & T.sess==2);
R.Rw1 = [corrN(D(i11,in)',D(i12,in)')];
R.Rw2 = [corrN(D(i21,in)',D(i22,in)')];
R.Rbetween = [corrN(D(i11,in)',D(i21,in)') corrN(D(i11,in)',D(i22,in)') ...
corrN(D(i12,in)',D(i21,in)') corrN(D(i12,in)',D(i22,in)')];
RR=addstruct(RR,R);
end;
end;
RR.Rw = ssqrt(RR.Rw1.*RR.Rw2);
RR.Rb = mean(RR.Rbetween,2);
subplot(1,2,1);
scatterplot(RR.Rw,RR.Rb,'identity');
xlabel('Within-experiment reliability');
xlabel('Between-experiment reliability');
subplot(1,2,2);
barplot(RR.cond,[RR.Rw1 RR.Rw2 RR.Rb],'leg',{'Exp1','Exp2','crossExp'});
ylabel('Within-subject Reliability');
set(gca,'XTickLabel',condNames,'XTickLabelRotation',60);
varargout={RR,T};
case 'conn_mbeta' % Estimate connectivty model on the meanbetas
% L2=[100 500 1000 2000 3000 4000];
% sc1sc2_connectivity('conn_mbeta','method','ridgeFixed','lambdaL1',L2*0,'lambdaL2',L2,'name','mb4_162_ridge','exper',1);
% sc1sc2_connectivity('conn_mbeta','method','ridgeFixed','lambdaL1',L2*0,'lambdaL2',L2,'name','mb4_162_ridge','exper',2);
% L2=[100 1000 2000];
% sc1sc2_connectivity('conn_mbeta','method','cplexqpL1L2','lambdaL1',L2*0,'lambdaL2',L2,'name','mb4_162_nn','exper',1);
T=dload(fullfile(rootDir,'sc1_sc2_taskConds.txt'));
sn=goodsubj;
RR= [];
method= 'ridgeFixed'; % linRegress, ridgeFixed, nonNegExp, cplexqp, lasso, winnerTakeAll
yname = 'Cerebellum_grey';
xname = '162_tessellation_hem';
name = 'mb4_162_ridge'; % mb4_162_ridge or ts4_yeo_nneg or mb5_162_lasso
trainMode = 'crossed'; % training can be done in a crossed or uncrossed fashion
lambdaL1 = 2500;
lambdaL2 = 2500;
overwrite = 0; % Overwrite old results or add to existing structure?
exper = 1; % Task set to train
incInstr = 0; % For future use: include Instruction
vararginoptions(varargin,{'sn','xname','type','exper','method','glm',...
'name','trainMode','lambdaL1','lambdaL2','incInstr','overwrite'});
% Get data
[X,Y,S]=sc1sc2_connectivity('get_mbeta_all','incInstr',incInstr);
% Save the fit
outDir = fullfile(rootDir,expStr{exper},connDir,name);
dircheck(outDir);
% Find the data that we want to use for fitting the connectivity
% model
SI1 = find(ismember(S.StudyNum,exper) & S.sess==1);
SI2 = find(ismember(S.StudyNum,exper) & S.sess==2);
trainXindx=[SI1;SI2];
switch (trainMode)
case 'crossed'
trainYindx=[SI2;SI1];
case 'uncrossed'
trainYindx=[SI1;SI2];
end;
% Estimate the model and store the information attached to it
for s=1:length(sn)
% Find relevant file
fprintf('%d\n',sn(s));
outName=fullfile(outDir,sprintf('%s_%s.mat',name,subj_name{sn(s)}));
if (exist(outName) && overwrite==0);
RR=load(outName);
else
RR=[];
end;
% Get data
subjn=find(goodsubj==sn(s));
xx = X{subjn}(trainXindx,:);
yy = Y{subjn}(trainYindx,:);
% Run for all regularisation parameters
for l=1:length(lambdaL1)
R.SN = sn(s);
[W,R.fR2m,R.fRm] = sc_connect_fit(yy,xx,method,'lambda',[lambdaL1(l) lambdaL2(l)]);
R.W={W};
R.lambda = [lambdaL1(l) lambdaL2(l)];
R.method = {method};
R.incInstr = incInstr;
R.trainMode = {trainMode};
R.xname = {xname};
RR=addstruct(RR,R);
end;
save(outName,'-struct','RR');
end;
case 'run_connB'
L1=[10 25 50 100];
L2=[250 500 1000 1750 3500];
sc1sc2_connectivity('conn_mbeta','method','cplexqpL1L2','lambdaL1',L1,'lambdaL2',L1*0,'name','mb4_162_nn','exper',1);
sc1sc2_connectivity('conn_mbeta','method','cplexqpL1L2','lambdaL1',L1,'lambdaL2',L1*0,'name','mb4_162_nn','exper',2);
sc1sc2_connectivity('conn_mbeta','method','cplexqpL1L2','lambdaL1',L2*0,'lambdaL2',L2,'name','mb4_162_nn','exper',1);
sc1sc2_connectivity('conn_mbeta','method','cplexqpL1L2','lambdaL1',L2*0,'lambdaL2',L2,'name','mb4_162_nn','exper',2);
case 'conn_ts' % Run encoding model on time series
% sc1_connectivity('conn_ts',[2:22],'method','winnerTakeAll_nonNeg','name','glm4_162_nnWTA','lambdaL1',0,'lambdaL2',0);
% sc1_connectivity('conn_ts',goodsubj,'method','winnerTakeAll','name','glm4_162_WTA','lambdaL1',0,'lambdaL2',0);
% sc1_connectivity('conn_ts',[2:22],'method','ridgeFixed','name','glm4_162_ridge','lambdaL1',[0 0 0 0 0 0],'lambdaL2',[0 500 1000 2500 5000 7500]);
% sc1_connectivity('conn_ts',[2:22],'method','cplexqpL1L2','name','glm4_162_L1L2c','lambdaL1',[2500 7500 7500],'lambdaL2',[2500 7500 2500]);
sn=goodsubj;%varargin{1};
glm=4; % usually 4
method= 'ridgeFixed'; % cplexqpL1L2, linRegress, nonNegExp, cplexqp, lasso, winnerTakeAll
yname = 'cerebellum_grey';
xname = '162_tessellation_hem';
name = 'glm4_162_ridgeFixed'; % Name of the output
type = {'Y','Yhat'}; % Run with three Y
exper = 'sc1';
lambdaL1 = 0; %[2500 7500 7500]; % Initially set to 0
lambdaL2 = 2500; %[2500 7500 2500]; % Inititally select one value
partial = [];
vararginoptions({varargin{2:end}},{'xname','type','method','glm','partial','name','lambdaL1','lambdaL2'});
outDir = fullfile(rootDir,exper,connDir,name);
dircheck(outDir);
myRegDir = fullfile(rootDir,exper,regDir);
% Build matrix to remove block mean
SPM=load(fullfile(rootDir,exper,'GLM_firstlevel_4','SampleDesignMatrix.mat'));
for b=1:16
B(SPM.Sess(b).row,b)=1;
end;
for s=sn
TT=[];
T=[];
% Load X
file = fullfile(myRegDir,sprintf('glm%d',glm),subj_name{s},sprintf('ts_%s.mat',xname));
XX=load(file);
if (strcmp(xname,'162_tessellation_hem'))
Tcort=load(fullfile(myRegDir,'data','162_reorder.mat'));
Tcort=getrow(Tcort,Tcort.good==1);
XX.Yhatm = XX.Yhatm(:,Tcort.regIndx,:);
XX.Yhatr = XX.Yhatr(:,Tcort.regIndx,:);
XX.Yres = XX.Yres(:,Tcort.regIndx,:);
end;
if (strcmp(xname,'yeo'))
XX.Yhatm = XX.Yhatm(:,2:18,:);
XX.Yhatr = XX.Yhatr(:,2:18,:);
XX.Yres = XX.Yres(:,2:18,:);
end;
XX.Yhat = XX.Yhatm + XX.Yhatr;
XX.Y = XX.Yres + XX.Yhat;
if (~isempty(partial))
PP=load(fullfile(rootDir,exper,regDir,regDir,'glm4',subj_name{sn(s)},sprintf('ts_%s.mat',partial)));
PP.Yhat = PP.Yhatm + PP.Yhatr;
PP.Y = PP.Yhat + PP.Yres;
end;
% Load Y
Y=load(fullfile(myRegDir,'glm4',subj_name{s},sprintf('ts_%s.mat',yname)));
Y.Yhat = Y.Yhatm + Y.Yhatr;
Y.Y = Y.Yhat + Y.Yres;
% Loop over the different types
for t=1:length(type);
yy = Y.(type{t});
xx = XX.(type{t});
% remove mean
yy = yy - B*pinv(B)*yy;
xx = xx - B*pinv(B)*xx;
% Normalize to unity
%ySS = sum((Y.Y).^2);
% yy = bsxfun(@rdivide,yy,sqrt(ySS));
%xSS = sum((X.Y).^2);
% xx = bsxfun(@rdivide,X.Yhat,sqrt(xSS));
% X.Yres = bsxfun(@rdivide,X.Yres,sqrt(xSS));
% Regress out the partial ROI
if (~isempty(partial))
pp = PP.(type{t});
pp = pp - B*pinv(B)*pp;
yy = yy - pp*pinv(pp)*yy; % Subtract from cerebellar data only
end;
for l=1:length(lambdaL1);
T.SN = s;
T.typeNum = t;
T.type = {type{t}};
T.lambda = [lambdaL1(l) lambdaL2(l)];
fprintf('Fitting s%2.2d %s lambdaL1 = %2.2f lambdaL2 = %2.2f\n',s,type{t},lambdaL1(l),lambdaL2(l));
[W,T.fR2m,T.fRm,T.fR2v,T.fRv]=sc_connect_fit(yy,xx,method,'lambda',[lambdaL1(l) lambdaL2(l)]);
T.W = {W};
TT=addstruct(TT,T);
end;
end;
outName=fullfile(outDir,sprintf('%s_s%2.2d.mat',name,s));
save(outName,'-struct','TT');
fprintf('encode model: cerebellar voxels predicted for %s \n',subj_name{s});
end
case 'numberOfConditions' % Build predictive models on a random assortment of conditions - evaluate them on the other experiment
T = dload(fullfile(rootDir,'sc1_sc2_taskConds.txt'));
RR=[];
for nc=2:29
for n=1:50
CondSet = sample_wor([1:29],nc,1);
M=sc1_connectivity('conn_mbeta','trainMode','uncrossed','subset',T.StudyNum==1 & ismember(T.condNum,CondSet));
R=sc1_connectivity('evaluate',M,'subset',T.StudyNum==2,'splitby',ismember(T.condNumUni,CondSet));
vec = ones(length(R.SN),1);
R.exp = 2*vec;
R.numSet = n*vec;
R.numCond = nc*vec;
R.CondSet = repmat(ismember([1:29],CondSet),length(R.SN),1);
RR=addstruct(RR,R);
R=sc1_connectivity('conn_evaluate',M,'subset',T.StudyNum==1,'splitby',ismember(T.condNumUni,CondSet));
vec = ones(length(R.SN),1);
R.exp = vec;
R.numSet = n*vec;
R.numCond = nc*vec;
R.CondSet = repmat(ismember([1:29],CondSet),length(R.SN),1);
RR=addstruct(RR,R);
end;
end;
save(fullfile(sc1Dir,'connectivity_cerebellum','evaluation','numberOfConditions_glm4_1_2.mat'),'-struct','RR');
varargout={RR};
case 'singleTaskPlot'
R=varargin{1};
T=dload(fullfile(rootDir,'sc1_sc2_taskConds.txt'));
T=getrow(T,T.StudyNum==2);
[D.Rcv,D.cond]=pivottable(R.cond,[],R.Rcv,'mean');
[D.Ry,D.cond]=pivottable(R.cond,[],R.Ry,'mean');
[D.Rp,D.cond]=pivottable(R.cond,[],R.Rp,'mean');
barplot(D.cond,D.Rcv./sqrt(D.Ry.*D.Rp));
set(gca,'XTickLabel',T.condNames,'XTickLabelRotation',70);
keyboard;
varargout={R,D};
case 'map_cortex' % Map of where projections come from
sn=varargin{1}; % [2:22]
name = varargin{2};
glm=4; % usually 4
parcel = 'yeo17';
type = {'Yres','Yhat'};
vararginoptions({varargin{3:end}},{'type','glm','parcel'});
lambda=0;
encodeDirSp = fullfile(encodeDir,sprintf('glm%d',glm),'encode_4');
for s=1:length(sn)
T=load(fullfile(encodeDirSp,sprintf('encode_%s_s%2.2d.mat',name,sn(s))));
for t=1:2
data{t}(s,:) = mean(T.W{t}');
end;
end;
for t=1:2
md =mean(data{t},1);
max(md)
for h=1:2
subplot(2,2,h+(t-1)*2);
switch(parcel)
case '162'
dat = md([1:142]+(h-1)*142);
cscale = [0 0.005];
case 'yeo17'
dat = md;
cscale = [0 0.05];
end;
sc1_imana_jd('plot_cortex',h,dat,'parcel',parcel,'cscale',cscale);
end;
end;
case 'plot_cortex'
h=varargin{1}; %
data=varargin{2};
cscale = [];
parcel = '162';
vararginoptions({varargin{3:end}},{'cscale','parcel'});
switch (parcel)
case '162'
D=caret_load(fullfile(sc1Dir,caretDir,'fsaverage_sym',hemName{h},sprintf([hem{h} '.tessel162_new.metric'])));
Tcort=load(fullfile(sc1Dir,regDir,'data','162_reorder.mat'));
Tcort=getrow(Tcort,Tcort.good==1 & Tcort.hem==h);
indx = Tcort.regIndx-158*(h-1);
case 'yeo17'
D=caret_load(fullfile(sc1Dir,caretDir,'fsaverage_sym',hemName{h},sprintf([hem{h} '.yeo17.paint'])));
indx=[2:18]';
end;
FData=nan(size(D.data));
for i=1:length(data)
FData(D.data==indx(i),1)=data(i);
end;
coord=fullfile(sc1Dir,caretDir,'fsaverage_sym',hemName{h},[hem{h} '.FLAT.coord']);
topo=fullfile(sc1Dir,caretDir,'fsaverage_sym',hemName{h},[hem{h} '.CUT.topo']);
if (h==1)
caret_plotflatmap('coord',coord,'topo',topo,'data',FData,'xlims',[-200 150],'ylims',[-140 140],'cscale',cscale);
else
caret_plotflatmap('coord',coord,'topo',topo,'data',FData,'xlims',[-150 200],'ylims',[-140 140],'cscale',cscale);
end;
case 'map_cerebellum' % Maps certain stats from an individual to SUIT and then to flat map
data = varargin{1};
sn = varargin{2};
% Determine the voxels we want to resample in SUIT space
V=spm_vol(fullfile(sc1Dir,suitDir,'anatomicals','cerebellarGreySUIT.nii'));
X= spm_read_vols(V);
grey_threshold = 0.1; % gray matter threshold
linIn1=find(X>grey_threshold);
[i1,j1,k1]= ind2sub(V.dim,linIn1');
[x1,y1,z1]= spmj_affine_transform(i1,j1,k1,V.mat);
% Determine voxel locations from the original ROI
load(fullfile(sc1Dir,regDir,'data',subj_name{sn},'regions_cerebellum_grey.mat')); % 'regions' are defined in 'ROI_define'
Vmask = spm_vol(fullfile(sc1Dir,suitDir,'glm4',subj_name{sn},'maskbrainSUITGrey.nii'));
[i3,j3,k3]=spmj_affine_transform(R{1}.data(:,1),R{1}.data(:,2),R{1}.data(:,3),inv(Vmask.mat));
linIn3 = sub2ind(Vmask.dim,round(i3),round(j3),round(k3));
% transform SUIT coords into anatomical space of the individual
flowfield = fullfile(sc1Dir,suitDir,'anatomicals',subj_name{sn},'u_a_c_anatomical_seg1.nii');
affine = fullfile(sc1Dir,suitDir,'anatomicals',subj_name{sn},'Affine_c_anatomical_seg1.mat');
[Def,Aff]=spmdefs_get_dartel(flowfield,affine);
[x2,y2,z2]=spmdefs_transform(Def,Aff,x1,y1,z1);
[i2,j2,k2]=spmj_affine_transform(x2,y2,z2,inv(Vmask.mat));
% resample the weights into SUIT space
for r=1:size(data,1)
Vout=Vmask;
Vout.dat = zeros(Vout.dim);
Vout.dat(linIn3)=data(r,:);
Vout.dt = [64 0];
Vout.pinfo = [1 0 0]';
DataSUIT(r,:)=spm_sample_vol(Vout,i2,j2,k2,1);
V.dat = zeros(V.dim);
Vres(r)=V;
Vres(r).dat(linIn1)=DataSUIT(r,:); % Offset by one to account for 1 being medial wall
Vres(r).fname=sprintf('data_%2.2d.nii',r);
Vres(r).pinfo=[1 0 0]';
end;
% Now map to surface-based representation
D = suit_map2surf(Vres,'stats','mean');
varargout={D,Vres};
case 'evaluate' % Evaluates predictive performance of connectivity model
% M: is the model structure with the connectivity weights in it (M.W)
% 'subset': what data should the evaluation be based upon?
% 'splitby': by which variable should the evaluation be split?
% 'meanSub': Mean pattern subtraction before evaluation?
M = varargin{1}; % This is a structure with the fitted data
T = dload(fullfile(rootDir,'sc1_sc2_taskConds.txt'));
subset = T.StudyNum==2;
splitby= T.condNum;
yname = 'Cerebellum_grey';
xname = '162_tessellation_hem';
meanSub = 0; % Mean pattern subtraction before prediction?
vararginoptions(varargin(2:end),{'subset','splitby','meanSub','xname','yname'});