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
|
unit regression;
//only for Delphi - not Freepascal
//Unit for running multiple regression
interface
uses
{$H+}
{$IFNDEF UNIX} Windows, {$ENDIF}
{$IFDEF FPC} utypes,regmult,{$ELSE}
utypes,regmult,
{$ENDIF}define_types,Classes,nifti_hdr,sysutils,nifti_img,
StatThdsUtil,Forms,Distr,Dialogs,npmform;
function GetValReg (var lnSubj,lnFactors: integer; var X : PMatrix; var lImageNames: TStrings; var lPredictorList: TStringList): boolean;
function ARegressNPMAnalyze (var lImages: TStrings; var lMaskHdr: TMRIcroHdr; var X: PMatrix; lnFactors: integer; var lPredictorList: TStringList; lOutname: string): boolean;
function Regress2NPMAnalyze (var lImages: TStrings; var lMaskHdr: TMRIcroHdr; lOutname: string; var lXadditional: PMatrix; lnAdditionalFactors: integer ): boolean;
function TtoR(t,df: double): double;
implementation
uses valformat,hdr,math;
(*function readCSV (lFilename: string; var lnObservations,lnFactors : integer; var X : PMatrix; var Y: PVector): boolean;
var
lNumStr: string;
F: TextFile;
lTempFloat: double;
lCh: char;
lPos,MaxC,R,C:integer;
lError: boolean;
begin
result := false;
if not fileexists(lFilename) then exit;
AssignFile(F, lFilename);
FileMode := 0; //Set file access to read only
//First pass: determine column height/width
Reset(F);
C := 0;
MaxC := 0;
R := 0;
while not Eof(F) do begin
//read next line
//read next line
Read(F, lCh);
if lCh = '#' then
while not (lCh in [#10,#13]) do
Read(F, lCh)
else if not (lCh in [#10,#13,#9,',']) then begin
lNumStr := lNumStr + lCh;
end else if lNumStr <> '' then begin
lNumStr := '';
inc(C);
if C > MaxC then
MaxC := C;
if (lCh in [#10,#13]) then begin
C := 0;
inc(R);
end; //eoln
end; //if lNumStr <> '' and not tab
end;
if lNumStr <> '' then //july06- read data immediately prior to EOF
inc(R);
if (R < 2) or (MaxC < 5) then begin
showmessage('problems reading CSV');
exit;
end;
lnObservations := MaxC;
lnFactors := R -1;
DimVector(Y, lnObservations);
DimMatrix(X, lnFactors, lnObservations);
//second pass
Reset(F);
C := 0;
MaxC := 0;
R := 1;
lNumStr := '';
while not Eof(F) do begin
//read next line
Read(F, lCh);
if lCh = '#' then
while not (lCh in [#10,#13]) do
Read(F, lCh)
else if not (lCh in [#10,#13,#9,',']) then begin
lNumStr := lNumStr + lCh;
end else if lNumStr <> '' then begin
if lNumStr = '-' then begin
lTempFloat := 0;
end else begin //number
try
lTempFloat := strtofloat(lNumStr);
except
on EConvertError do begin
if not lError then
showmessage('Empty cells? Error reading VAL file row:'+inttostr(R)+' col:'+inttostr(C)+' - Unable to convert the string '+lNumStr+' to a number');
lError := true;
lTempFloat := nan;
end;
end;
if R = 1 then
Y^[C+1] := lTempFloat
else
X^[R-1]^[C+1] := lTempFloat;
//xxx := lTempFloat;//DataGrid.Cells[ C, kMaxFactors+R-1 ] := (lNumStr) ;
end;
lNumStr := '';
inc(C);
if C > MaxC then
MaxC := C;
if (lCh in [#10,#13]) then begin
C := 0;
inc(R);
end; //eoln
end; //if lNumStr <> '' and not tab
end;
CloseFile(F);
FileMode := 2; //Set file access to read/write
result := true;
end;
function TestMultReg: boolean;
var
i,lnFactors, lnObservations: integer;
X : PMatrix;
Y: PVector;
lOutT,lOutSlope: DoubleP0;
lStart: dword;
begin
if not readCSV('C:\xio.csv',lnObservations,lnFactors,X,Y ) then exit;
//showmessage('alpha');
getmem(lOutT, (lnFactors+1)* sizeof(double));
getmem(lOutSlope, (lnFactors+1)* sizeof(double));
lStart := gettickcount;
for i := 1 to 10000 do
MultipleRegressionVec (lnObservations,lnFactors, X, Y, lOutT,lOutSlope);
fx(gettickcount-lstart);
if MultipleRegressionVec (lnObservations,lnFactors, X, Y, lOutT,lOutSlope) then begin
for i := 0 to lnFactors do
fx(lOutT^[i],lOutSlope^[i]);
end;
freemem(lOutT);
freemem(lOutSlope);
DelMatrix(X, lnFactors, lnObservations);
DelVector(Y, lnObservations);
end; *)
(*procedure rx(lnObs,lnFactors: integer; X: PMatrix; lObs: Doublep0);
var
n,f: integer;
str: string;
begin
for n := 1 to lnObs do begin
str := floattostr(lObs^[n-1]);
for f := 1 to lnFactors do
str := str+','+floattostr(X^[f]^[n]);
MainForm.NPMmsg(str);
end;//each obs
str := '----------';
end;//proc RX*)
function Sign(value: double): double;
begin
if value > 0 then
result := 1
else if value < 0 then
result := -1
else
result := 0;
end;
function TtoR(t,df: double): double;
CONST
eps=3.0e-7;
begin
result := 0;
if (t = 0) or (df = 0) then
exit;
result := sign(t)/ sqrt( (df/(t*t+eps)) +1 );
end;
{$DEFINE SaveT} //if SaveT then t-score map will be saved
{$DEFINE SaveRnotZ} //if SaveRnotZ then r-value map will be saved, but not Z-score map
function Regress2NPMAnalyze (var lImages: TStrings; var lMaskHdr: TMRIcroHdr; lOutname: string; var lXadditional: PMatrix; lnAdditionalFactors: integer ): boolean;
//lImages is list 1..N of 1st images followed by 1..N of corresponding control images
//example c1.img, c2.img,c3.img,e1.img,e2.img,e3.img
//lImages.Count must be even
label
667;
const
kMaxFact = 80;
var
lOutNameMod,lFactName,lRunName: string;
lMaskImg,lPlankImg,lOutImgMn: SingleP;
lOutImgR: array [1..kMaxFact] of SingleP;
lTotalMemory: int64;
lnFactors,lnObservations,lnObservationsDiv2,lPlank,lVolVox,lPos,lMinMask,lMaxMask,lnPlanks,lVoxPerPlank,
lDF,lPos2,lPos2Offset,lStartVox,lEndVox,lPlankImgPos,lnTests,lnVoxTested,lPosPct,lFact,lnStatFact: integer;
l1st, lSum, lMn: double;
lVar: boolean;
lObsp: pointer;
lObs: Doublep0;
lStatHdr: TNIfTIhdr;
lFdata: file;
lnPermute: integer;
lRanOrderp: pointer;
lRanOrder: Doublep0;
lZP: Pointer;
lZra : DoubleP0;
X : PMatrix;
begin
lnFactors := 1+lnAdditionalFactors;
lnPermute := MainForm.ReadPermute;
if odd(lImages.Count) then begin
showmessage('Regress2NPMAnalyze must be passed an even number of images: the first half of the list is the experimental images, followed by corresponding control images.');
exit;
end;
lnObservations := lImages.Count;
lnObservationsDiv2 := lImages.Count div 2;
lDF := lnObservationsDiv2-lnFactors-1;
if lDF < 1 then begin
showmessage('Regress2NPMAnalyze: DF must be >0 (DF=[Num-Factors-1]) Num='+inttostr(lnObservationsDiv2)+' Factors='+inttostr(lnFactors) );
exit;
end;
DimMatrix(X, lnFactors, lnObservationsDiv2);
//fx(lnAdditionalFactors);
if lnAdditionalFactors > 0 then begin
for lPos2 := 1 to lnAdditionalFactors do begin
for lPos := 1 to lnObservationsDiv2 do begin
X^[lPos2+1]^[lPos] := lXadditional^[lPos2]^[lPos];
//fx(lPos2+1,lPos, X^[lPos2+1]^[lPos]);
end;
end; //pos 2
end; //additional factros
//Memo1.Lines.Add('Permutations = ' +IntToStr(lnPermute));
MainForm.Memo1.Lines.Add('Analysis began = ' +TimeToStr(Now));
lTotalMemory := 0;
lVolVox := lMaskHdr.NIFTIhdr.dim[1]*lMaskHdr.NIFTIhdr.dim[2]* lMaskHdr.NIFTIhdr.dim[3];
if (lVolVox < 1) then goto 667;
lnStatFact := lnFactors + 1; //factors + overall model
if lnStatFact > (kMaxFact-1) then begin //-1 because factors + model
MainForm.Memo1.Lines.Add('ERROR: Can not analyze more than = ' +inttostr(kMaxFact-1)+' factors');
goto 667;
end;
//load mask
getmem(lMaskImg,lVolVox*sizeof(single));
if not LoadImg(lMaskHdr.ImgFileName, lMaskImg, 1, lVolVox,round(gOffsetRA[0]),1,lMaskHdr.NIFTIhdr.datatype,lVolVox) then begin
MainForm.Memo1.Lines.Add('Unable to load mask ' +lMaskHdr.ImgFileName);
goto 667;
end;
//next find start and end of mask
lPos := 0;
repeat
inc(lPos);
until (lMaskImg^[lPos] > 0) or (lPos = lVolVox);
lMinMask := lPos;
lPos := lVolVox+1;
repeat
dec(lPos);
until (lMaskImg^[lPos] > 0) or (lPos = 1);
lMaxMask := lPos;
if lMaxMask = 1 then begin
MainForm.Memo1.Lines.Add('Mask appears empty' +lMaskHdr.ImgFileName);
goto 667;
end;
MainForm.Memo1.Lines.Add('Mask has voxels from '+inttostr(lMinMask)+'..'+inttostr(lMaxMask));
lVoxPerPlank := kPlankSz div lnObservations div sizeof(single) ;
if (lVoxPerPlank = 0) then goto 667; //no data
lTotalMemory := ((lMaxMask+1)-lMinMask) * lnObservations;
if (lTotalMemory = 0) then goto 667; //no data
lnPlanks := trunc(lTotalMemory/(lVoxPerPlank*lnObservations) ) + 1;
MainForm.Memo1.Lines.Add('Memory planks = ' +Floattostr(lTotalMemory/(lVoxPerPlank*lnObservations)));
MainForm.Memo1.Lines.Add('Max voxels per Plank = ' +Floattostr(lVoxPerPlank));
if (lnPlanks = 1) then
getmem(lPlankImg,lTotalMemory*sizeof(single)) //assumes 4bpp
else
getmem(lPlankImg,kPlankSz);
lStartVox := lMinMask;
lEndVox := lMinMask-1;
lnVoxTested := 0;
for lPos := 1 to lnObservations do
if gScaleRA[lPos] = 0 then
gScaleRA[lPos] := 1;
createArray64(lObsp,lObs,lnObservations);
getmem(lOutImgMn,lVolVox* sizeof(single));
for lPos := 1 to lVolVox do
lOutImgMn^[lPos] := 0;
for lFact := 1 to (lnStatFact) do begin //+1 as we include full model
getmem(lOutImgR[lFact],lVolVox* sizeof(single));
for lPos := 1 to lVolVox do
lOutImgR[lFact]^[lPos] := 0;
end;
createArray64(lZp,lZra,lnFactors+1); //+1 as we include full model
//InitPermute (lImages.Count, lnPermute, lPermuteMaxT, lPermuteMinT,lPermuteMaxTW, lPermuteMinTW,lPermuteMaxWMW, lPermuteMinWMW, lRanOrderp, lRanOrder);
for lPlank := 1 to lnPlanks do begin
MainForm.Memo1.Lines.Add('Computing plank = ' +Inttostr(lPlank));
MainForm.Refresh;
lEndVox := lEndVox + lVoxPerPlank;
if lEndVox > lMaxMask then begin
lVoxPerPlank := lVoxPerPlank - (lEndVox-lMaxMask);
lEndVox := lMaxMask;
end;
lPlankImgPos := 1;
for lPos := 1 to lnObservations do begin
if not LoadImg(lImages[lPos-1], lPlankImg, lStartVox, lEndVox,round(gOffsetRA[lPos]),lPlankImgPos,gDataTypeRA[lPos],lVolVox) then
goto 667;
lPlankImgPos := lPlankImgPos + lVoxPerPlank;
end;//for each image
lPosPct := lVoxPerPlank div 100;
for lPos2 := 1 to lVoxPerPlank do begin
if (lPos2 mod lPosPct) = 0 then begin
MainForm.ProgressBar1.Position := round((lPos2/lVoxPerPlank)*100);
Application.Processmessages;
end;
lPos2Offset := lPos2+lStartVox-1;
if lMaskImg^[lPos2Offset] <> 0 then begin
inc(lnVoxTested);
lSum := 0;
//check for variance
lVar := false;
lPos := 1;
l1st := (gScaleRA[lPos]*lPlankImg^[((lPos-1)* lVoxPerPlank)+lPos2])+gInterceptRA[lPos];
for lPos := 1 to lnObservations do
lObs^[lPos-1] := (gScaleRA[lPos]*lPlankImg^[((lPos-1)* lVoxPerPlank)+lPos2])+gInterceptRA[lPos];
for lPos := 1 to lnObservationsDiv2 do begin
lSum := lSum + lObs^[lPos-1];
if (not lVar) and (lObs^[lPos-1]<>l1st) then
lVar := true;
//lSumOfSqrs := lSumOfSqrs + sqr(lObs[lPos-1]);
X^[1]^[lPos] := lObs^[lnObservationsDiv2+lPos-1];
end;
lOutImgMn^[lPos2Offset] := lSum/lnObservationsDiv2;
if lVar then begin
MultipleRegression (lnObservationsDiv2,lnFactors, X, lObs, lZra);
//if lPos2Offset = 359948 then rx(lnObservationsDiv2,lnFactors,X,lObs);
for lFact := 1 to lnStatFact do
lOutImgR[lFact]^[lPos2Offset] := lZra^[lFact-1];
end;
//StatPermute (lttest,lwelch,lWMW,lImages.Count, lnGroup1,lnPermute,lPermuteMaxT, lPermuteMinT,lPermuteMaxTW, lPermuteMinTW,lPermuteMaxWMW, lPermuteMinWMW, lObs,lRanOrder);
end; //in brain mask - compute
end;
lStartVox := lEndVox + 1;
end;
//next report findings
MainForm.Memo1.Lines.Add('Voxels tested = ' +Inttostr(lnVoxTested));
MainForm.reportBonferroni('Std',lnVoxTested);
//next: save data
if lnFactors = 1 then
lRunName := 'reg'
else
lRunName := '';
//savedata
MakeHdr (lMaskHdr.NIFTIhdr,lStatHdr);
//save mean
lOutNameMod := ChangeFilePostfixExt(lOutName,'Mn'+lRunName,'.hdr');
if not FileExistsEX(lOutNameMod) then
NIFTIhdr_SaveHdrImg(lOutNameMod,lStatHdr,true,not IsNifTiMagic(lMaskHdr.NIFTIhdr),true,lOutImgMn,1);
//save regression
for lFact := 1 to (lnStatFact) do begin
if (lFact > lnFactors) and (lnFactors = 1) then
lFactName := 'intercept'+'reg' //for analysis of multiple single regressions
else if (lFact > lnFactors) then
lFactName := 'intercept'
else
lFactName := 'reg'+inttostr(lFact);
MakeHdr (lMaskHdr.NIFTIhdr,lStatHdr);
{$IFDEF SaveT} //if SaveTRnotZ then t-score and r-score maps will be created, but no Z-score maps
//the next bit is optional - save data as T-values instead of Z-scores
// this allows direct comparison with SPM...
MakeStatHdr (lMaskHdr.NIFTIhdr,lStatHdr,-6, 6,lDF,0,lnVoxTested,kNIFTI_INTENT_TTEST,inttostr(lnVoxTested) );
lOutNameMod := ChangeFilePostfixExt(lOutName, 'wlsT'+lFactName,'.hdr');
NIFTIhdr_SaveHdrImg(lOutNameMod,lStatHdr,true,not IsNifTiMagic(lMaskHdr.NIFTIhdr),true,lOutImgR[lFact],1);
{$ENDIF}
{$IFDEF SaveRnotZ}
MakeStatHdr (lMaskHdr.NIFTIhdr,lStatHdr,-6, 6,lDF,0,lnVoxTested,kNIFTI_INTENT_CORREL,inttostr(lnVoxTested) );
for lPos := 1 to lVolVox do
lOutImgR[lFact]^[lPos] := TtoR (lOutImgR[lFact]^[lPos],lDF);
lOutNameMod := ChangeFilePostfixExt(lOutName, 'wlsR'+lFactName,'.hdr');
NIFTIhdr_SaveHdrImg(lOutNameMod,lStatHdr,true,not IsNifTiMagic(lMaskHdr.NIFTIhdr),true,lOutImgR[lFact],1);
{$ELSE}
//next - save Zscores
MakeStatHdr (lMaskHdr.NIFTIhdr,lStatHdr,-6, 6,lDF,0,lnVoxTested,kNIFTI_INTENT_ZSCORE,inttostr(lnVoxTested) );
//{ DoF = Nb points - Nb parameters }
for lPos := 1 to lVolVox do
lOutImgR[lFact]^[lPos] := TtoZ (lOutImgR[lFact]^[lPos],lDF);
MainForm.reportFDR ('wls'+lFactName, lVolVox, lnVoxTested, lOutImgR[lFact]);
lOutNameMod := ChangeFilePostfixExt(lOutName, 'wls'+lFactName,'.hdr');
NIFTIhdr_SaveHdrImg(lOutNameMod,lStatHdr,true,not IsNifTiMagic(lMaskHdr.NIFTIhdr),true,lOutImgR[lFact],1);
{$ENDIF}
freemem(lOutImgR[lFact]);
end;
//next: close images
Freemem(lZp);
freemem(lOutImgMn);
freemem(lObsp);
freemem(lMaskImg);
freemem(lPlankImg);
MainForm.Memo1.Lines.Add('Analysis finished = ' +TimeToStr(Now));
lOutNameMod := ChangeFilePostfixExt(lOutName,'Notes'+lRunName,'.txt');
MainForm.Memo1.Lines.SaveToFile(lOutNameMod);
MainForm.ProgressBar1.Position := 0;
DelMatrix(X, lnFactors, lnObservationsDiv2);
exit;
667: //you only get here if you aborted ... free memory and report error
DelMatrix(X, 1, lnObservationsDiv2);
if lVolVox > 1 then freemem(lMaskImg);
if lTotalMemory > 1 then freemem(lPlankImg);
MainForm.Memo1.Lines.Add('Unable to complete analysis.');
MainForm.ProgressBar1.Position := 0;
end;
{$DEFINE NoThread}
function ARegressNPMAnalyze (var lImages: TStrings; var lMaskHdr: TMRIcroHdr; var X: PMatrix; lnFactors: integer; var lPredictorList: TStringList; lOutname: string): boolean;
{$IFNDEF Thread}
const
kMaxFact = 80;
{$ENDIF}
label
667;
var
lOutNameMod,lFactName,lRunName: string;
lMaskImg,lPlankImg,lOutImgMn: SingleP;
{$IFDEF Thread}
lOutImgR: TRegRA;
{$ELSE}
lOutImgR: array [1..kMaxFact] of SingleP;
{$ENDIF}
lTotalMemory: int64;
lnObservations,lPlank,lVolVox,lPos,lMinMask,lMaxMask,lnPlanks,lVoxPerPlank,
//lPos2,lPos2Offset,
lDF,lStartVox,lEndVox,lPlankImgPos,lnTests,lnVoxTested,lFact,lnStatFact: integer;
//l1st, lSum, lMn: double;
//lVar: boolean;
//lObsp: pointer;lObs: Doublep0;
lStatHdr: TNIfTIhdr;
lFdata: file;
{$IFDEF Thread}
lThread,lThreadStart,lThreadEnd,lThreadInc: integer;
{$ELSE}
lObsP,lZP: Pointer;
lObs,lZra : DoubleP0;
lSum,l1st: double;
lVar: boolean;
lPos2,lPosPct,lPos2Offset: integer;
{$ENDIF}
begin
lnObservations := lImages.Count;
lDF := lnObservations-lnFactors-1;
if lDF < 1 then begin
showmessage('Regress2NPMAnalyze: DF must be >0 (DF=[Num-Factors-1]) Num='+inttostr(lnObservations)+' Factors='+inttostr(lnFactors) );
exit;
end;
MainForm.NPMmsg('Analysis began = ' +TimeToStr(Now));
lTotalMemory := 0;
lVolVox := lMaskHdr.NIFTIhdr.dim[1]*lMaskHdr.NIFTIhdr.dim[2]* lMaskHdr.NIFTIhdr.dim[3];
if (lVolVox < 1) then goto 667;
lnStatFact := lnFactors + 1; //factors + overall model
if lnStatFact > (kMaxFact-1) then begin //-1 because factors + model
MainForm.NPMmsg('ERROR: Can not analyze more than = ' +inttostr(kMaxFact-1)+' factors');
goto 667;
end;
//load mask
getmem(lMaskImg,lVolVox*sizeof(single));
if not LoadImg(lMaskHdr.ImgFileName, lMaskImg, 1, lVolVox,round(gOffsetRA[0]),1,lMaskHdr.NIFTIhdr.datatype,lVolVox) then begin
MainForm.NPMmsg('Unable to load mask ' +lMaskHdr.ImgFileName);
goto 667;
end;
//next find start and end of mask
lPos := 0;
repeat
inc(lPos);
until (lMaskImg^[lPos] > 0) or (lPos = lVolVox);
lMinMask := lPos;
lPos := lVolVox+1;
repeat
dec(lPos);
until (lMaskImg^[lPos] > 0) or (lPos = 1);
lMaxMask := lPos;
if lMaxMask = 1 then begin
MainForm.NPMmsg('Mask appears empty' +lMaskHdr.ImgFileName);
goto 667;
end;
MainForm.NPMmsg('Mask has voxels from '+inttostr(lMinMask)+'..'+inttostr(lMaxMask));
lVoxPerPlank := kPlankSz div lnObservations div sizeof(single) ;
if (lVoxPerPlank = 0) then goto 667; //no data
lTotalMemory := ((lMaxMask+1)-lMinMask) * lnObservations;
if (lTotalMemory = 0) then goto 667; //no data
lnPlanks := trunc(lTotalMemory/(lVoxPerPlank*lnObservations) ) + 1;
MainForm.NPMmsg('Memory planks = ' +Floattostr(lTotalMemory/(lVoxPerPlank*lnObservations)));
MainForm.NPMmsg('Max voxels per Plank = ' +Floattostr(lVoxPerPlank));
if (lnPlanks = 1) then
getmem(lPlankImg,lTotalMemory* sizeof(single)) //assumes 4bpp
else
getmem(lPlankImg,kPlankSz);
lStartVox := lMinMask;
lEndVox := lMinMask-1;
//lnVoxTested := 0;
for lPos := 1 to lnObservations do
if gScaleRA[lPos] = 0 then
gScaleRA[lPos] := 1;
//createArray64(lObsp,lObs,lnObservations);
getmem(lOutImgMn,lVolVox* sizeof(single));
for lPos := 1 to lVolVox do
lOutImgMn^[lPos] := 0;
for lFact := 1 to (lnStatFact) do begin //+1 as we include full model
getmem(lOutImgR[lFact],lVolVox* sizeof(single));
for lPos := 1 to lVolVox do
lOutImgR[lFact]^[lPos] := 0;
end;
//createArray64(lZp,lZra,lnFactors+1); //+1 as we include full model
{$IFDEF Thread}
ClearThreadDataPvals(gnCPUThreads,0) ;
{$ELSE}
lnVoxTested := 0;
{$ENDIF}
for lPlank := 1 to lnPlanks do begin
MainForm.NPMmsg('Computing plank = ' +Inttostr(lPlank));
MainForm.Refresh;
lEndVox := lEndVox + lVoxPerPlank;
if lEndVox > lMaxMask then begin
lVoxPerPlank := lVoxPerPlank - (lEndVox-lMaxMask);
lEndVox := lMaxMask;
end;
lPlankImgPos := 1;
for lPos := 1 to lnObservations do begin
if not LoadImg(lImages[lPos-1], lPlankImg, lStartVox, lEndVox,round(gOffsetRA[lPos]),lPlankImgPos,gDataTypeRA[lPos],lVolVox) then
goto 667;
lPlankImgPos := lPlankImgPos + lVoxPerPlank;
end;//for each image
{$IFDEF Thread}
lThreadStart := 1;
lThreadInc := lVoxPerPlank div gnCPUThreads;
lThreadEnd := lThreadInc;
Application.processmessages;
for lThread := 1 to gnCPUThreads do begin
if lThread = gnCPUThreads then
lThreadEnd := lVoxPerPlank; //avoid integer rounding error
with TLinThreadStat.Create (X,ProgressBar1, lnFactors,lThread,lThreadStart,lThreadEnd,lStartVox,lVoxPerPlank,lnObservations, lMaskImg,lPlankImg,lOutImgMn,lOutImgR) do
{$IFDEF FPC} OnTerminate := @ThreadDone; {$ELSE}OnTerminate := ThreadDone;{$ENDIF}
inc(gThreadsRunning);
Msg('Thread ' +Inttostr(gThreadsRunning)+' = '+inttostr(lThreadStart)+'..'+inttostr(lThreadEnd));
lThreadStart := lThreadEnd + 1;
lThreadEnd :=lThreadEnd + lThreadInc;
end; //for each thread
repeat
Application.processmessages;
until gThreadsRunning = 0;
Application.processmessages;
{$ELSE} //not threaded
createArray64(lZp,lZra,lnFactors+1); //+1 as we include full model
createArray64(lObsp,lObs,lnObservations);
lPosPct := lVoxPerPlank div 100;
for lPos2 := 1 to lVoxPerPlank do begin
if (lPos2 mod lPosPct) = 0 then begin
MainForm.ProgressBar1.Position := round((lPos2/lVoxPerPlank)*100);
Application.Processmessages;
end;
lPos2Offset := lPos2+lStartVox-1;
if lMaskImg^[lPos2Offset] <> 0 then begin
inc(lnVoxTested);
lSum := 0;
//check for variance
lVar := false;
lPos := 1;
l1st := (gScaleRA[lPos]*lPlankImg^[((lPos-1)* lVoxPerPlank)+lPos2])+gInterceptRA[lPos];
for lPos := 1 to lnObservations do begin
lObs^[lPos-1] := (gScaleRA[lPos]*lPlankImg^[((lPos-1)* lVoxPerPlank)+lPos2])+gInterceptRA[lPos];
lSum := lSum + lObs^[lPos-1];
if (not lVar) and (lObs^[lPos-1]<>l1st) then
lVar := true;
end;
lOutImgMn^[lPos2Offset] := lSum/lnObservations;
if {lPos2Offset = 762287 }lVar then begin
MultipleRegression (lnObservations,lnFactors, X, lObs, lZra);
//if {lZra^[0] < -5.548} lPos2Offset = 762287 then
// ReportRegression (lPos2Offset,lnObservations,lnFactors, X, lObs, lZra );
for lFact := 1 to lnStatFact do
lOutImgR[lFact]^[lPos2Offset] := lZra^[lFact-1];
end;
end; //in brain mask - compute
end; //for each voxel
Freemem(lZp);
Freemem(lObsp);
{$ENDIF} //if threaded else not threaded
lStartVox := lEndVox + 1;
end; //for each plank
{$IFDEF Thread}
lnVoxTested := SumThreadDataLite(gnCPUThreads);
{$ENDIF}
//next report findings
MainForm.NPMmsg('Voxels tested = ' +Inttostr(lnVoxTested));
MainForm.reportBonferroni('Std',lnVoxTested);
//next: save data
if lnFactors = 1 then
lRunName := lPredictorList[0]
else
lRunName := '';
//savedata
MakeHdr (lMaskHdr.NIFTIhdr,lStatHdr);
//save mean
lOutNameMod := ChangeFilePostfixExt(lOutName,'Mean'+lRunName,'.hdr');
if not FileExistsEX(lOutNameMod) then
NIFTIhdr_SaveHdrImg(lOutNameMod,lStatHdr,true,not IsNifTiMagic(lMaskHdr.NIFTIhdr),true,lOutImgMn,1);
//save regression
for lFact := 1 to (lnStatFact) do begin
if (lFact > lnFactors) and (lnFactors = 1) then begin
//nothing
end else begin
if (lFact > lnFactors) and (lnFactors = 1) then
lFactName := 'intercept'+lPredictorList[0] //for analysis of multiple single regressions
else if (lFact > lnFactors) then
lFactName := 'model'
else
lFactName := lPredictorList[lFact-1];
MakeHdr (lMaskHdr.NIFTIhdr,lStatHdr);
//NEXT : optional save t-maps
//MakeStatHdr (lMaskHdr.NIFTIhdr,lStatHdr,-6, 6,lDF,0,lnVoxTested,kNIFTI_INTENT_TTEST,inttostr(lnVoxTested) );
//lOutNameMod := ChangeFilePostfixExt(lOutName, 'wlsT'+lFactName,'.hdr');
//NIFTIhdr_SaveHdrImg(lOutNameMod,lStatHdr,true,not IsNifTiMagic(lMaskHdr.NIFTIhdr),true,lOutImgR[lFact],1);
//END: t-maps
//next - Z scores
MakeStatHdr (lMaskHdr.NIFTIhdr,lStatHdr,-6, 6,lDF,0,lnVoxTested,kNIFTI_INTENT_ZSCORE,inttostr(lnVoxTested) );
//{ DoF = Nb points - Nb parameters }
for lPos := 1 to lVolVox do
lOutImgR[lFact]^[lPos] := TtoZ (lOutImgR[lFact]^[lPos],lDF);
MainForm.reportFDR ('wls'+lFactName, lVolVox, lnVoxTested, lOutImgR[lFact]);
lOutNameMod := ChangeFilePostfixExt(lOutName, 'wls'+lFactName,'.hdr');
NIFTIhdr_SaveHdrImg(lOutNameMod,lStatHdr,true,not IsNifTiMagic(lMaskHdr.NIFTIhdr),true,lOutImgR[lFact],1);
freemem(lOutImgR[lFact]);
end;//if..else intercept and lnFactors = 1
end;//for each statfactor
//next: close images
//Freemem(lZp);
freemem(lOutImgMn);
//freemem(lObsp);
freemem(lMaskImg);
freemem(lPlankImg);
MainForm.NPMmsg('Analysis finished = ' +TimeToStr(Now));
lOutNameMod := ChangeFilePostfixExt(lOutName,'Notes'+lRunName,'.txt');
MainForm.MsgSave(lOutNameMod);
MainForm.ProgressBar1.Position := 0;
exit;
667: //you only get here if you aborted ... free memory and report error
if lVolVox > 1 then freemem(lMaskImg);
if lTotalMemory > 1 then freemem(lPlankImg);
MainForm.NPMmsg('Unable to complete analysis.');
MainForm.ProgressBar1.Position := 0;
end;
function GetValReg (var lnSubj,lnFactors: integer; var X : PMatrix; var lImageNames: TStrings; var lPredictorList: TStringList): boolean;
var
lVALFilename,lTemplateName: string;
lnRow,lnColWObs,lnCritPct,lInc,lRow,lCol: integer;
lDesignUnspecified : boolean;
lFileList:TStringList;
lInRA: DoubleP0;
lInP: Pointer;
begin
result := false;
lnSubj := 0;
if not MainForm.OpenDialogExecute('Select MRIcron VAL file',false,false,'MRIcron VAL (*.val)|*.val') then begin
showmessage('NPM aborted: VAL file selection failed.');
exit;
end; //if not selected
lVALFilename := MainForm.OpenHdrDlg.Filename;
MainForm.Memo1.Lines.Add( 'VAL filename: '+lVALFilename);
lFileList := TStringList.Create;
if not OpenValFile (lVALFilename,lTemplateName, lnRow,lnFactors,lnColWObs,lnCritPct,
lDesignUnspecified,lPredictorList,lFileList, lInP) then
exit;
if lnRow > 1 then begin
lnSubj := lnRow -1; //top row is predictor
{$IFDEF FPC}
lInRA := align(lInP,16);
{$ELSE}
lInRA := DoubleP0($fffffff0 and (integer(lInP)+15));
//lInRA := DoubleP0((integer(lInP) and $FFFFFFF0)+16);
{$ENDIF}
DimMatrix(X, lnFactors, lnSubj);
for lCol := 1 to lnFactors do begin
for lRow := 1 to lnSubj do begin
//MainForm.Memo1.Lines.Add(inttostr( (lRow*lnColWObsAndCovary)-4+lCol ));
X^[lCol]^[lRow] := lInRA^[(lRow*lnColWObs)-lnColWObs-1+lCol];
end;
end;
MainForm.Memo1.Lines.Add(inttostr(lnFactors)+' '+inttostr(lnSubj));
for lInc := 1 to lnSubj do
lImageNames.add(ExtractFileDirWithPathDelim(lVALFilename)+lFileList.Strings[lInc-1]);
result := true;
end else
result := false;
lFileList.free;
Freemem(lInP);
end;
end.
|