-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathmaskTaskEphysAnalysis.py
1524 lines (1394 loc) · 62.7 KB
/
maskTaskEphysAnalysis.py
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
# -*- coding: utf-8 -*-
"""
Created on Mon Mar 11 12:34:44 2019
@author: svc_ccg
"""
import copy
import math
import os
import pickle
import numpy as np
import scipy.signal
import scipy.ndimage
import scipy.stats
import matplotlib
matplotlib.rcParams['pdf.fonttype']=42
import matplotlib.pyplot as plt
from sklearn.svm import LinearSVC
from sklearn.ensemble import RandomForestClassifier
import fileIO
from maskTaskAnalysisUtils import MaskTaskData,getPsth,loadDatData
obj = MaskTaskData()
obj.loadEphysData(led=False)
obj.loadKilosortData()
obj.loadBehavData()
obj.loadRFData()
obj.saveToHdf5()
obj.loadFromHdf5()
exps = []
for f in fileIO.getFiles('choose experiments',fileType='*.hdf5'):
obj = MaskTaskData()
obj.loadFromHdf5(f)
# obj.loadBehavData(obj.behavDataPath)
# obj.loadKilosortData(os.path.join(os.path.dirname(obj.datFilePath),'kilosort_filtered'))
exps.append(obj)
unitPos = []
peakToTrough = []
for obj in exps:
units = obj.goodUnits
unitPos.extend([obj.units[u]['position'][1]/1000 for u in units])
peakToTrough.extend([obj.units[u]['peakToTrough'] for u in units])
unitPos = np.array(unitPos)
peakToTrough = np.array(peakToTrough)
fsThresh = 0.4
fs = peakToTrough < fsThresh
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
bw = 0.1
bins = np.arange(0,peakToTrough.max()+bw,bw)
h = np.histogram(peakToTrough,bins=bins)[0]
ax.bar(x=bins[:-1]+bw/2,height=h,width=bw,color='k')
ymax = plt.get(ax,'ylim')[1]
ax.plot([fsThresh]*2,[0,ymax],'--',color='0.5')
for x,lbl in zip((0.2,1),('FS','RS')):
ax.text(x,ymax,lbl,color='k',fontsize=16,ha='center',va='top')
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_xlim([0,1.6])
ax.set_ylim([0,ymax])
ax.set_xlabel('Spike Peak-to-Trough Duration (ms)',fontsize=16)
ax.set_ylabel('# Units',fontsize=16)
plt.tight_layout()
fpRate = np.array([obj.units[u]['fpRate'] for obj in exps for u in obj.sortedUnits if obj.units[u]['label']!='noise' and len(obj.units[u]['samples'])/(obj.totalSamples/obj.sampleRate)>0.1])
fpRate[fpRate>1] = 1
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
bw = 0.1
bins = np.arange(0,1.05,bw)
h = np.histogram(fpRate,bins=bins)[0]
ax.bar(x=bins[:-1]+bw/2,height=h,width=bw,color='k')
ymax = plt.get(ax,'ylim')[1]
ax.plot([0.5]*2,[0,ymax],'--',color='0.5')
for x,lbl in zip((0.25,0.75),('Accepted','Rejected')):
ax.text(x,ymax,lbl,color='k',fontsize=16,ha='center',va='top')
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_xlim([0,1])
ax.set_ylim([0,ymax])
ax.set_xlabel('False-Positive Rate',fontsize=16)
ax.set_ylabel('# Units',fontsize=16)
plt.tight_layout()
# plot response to visual stimuli without opto
trialTime = [(obj.openLoopFrames+obj.responseWindowFrames)/obj.frameRate for obj in exps]
assert(all([t==trialTime[0] for t in trialTime]))
trialTime = trialTime[0]
frameRate = 120
binSize = 1/frameRate
activeThresh = 0.1 # spikes/s
respThresh = 5 # stdev
stimLabels = ('maskOnly','mask','targetOnly','catch')
hemiLabels = ('contra','ipsi')
rewardDir = (-1,1)
behavRespLabels = ('all','go','nogo')
preTime = 0.5
postTime = 2
windowDur = preTime+trialTime+postTime
ntrials = {stim: {hemi: {resp: {} for resp in behavRespLabels} for hemi in hemiLabels} for stim in stimLabels}
trialPsth = copy.deepcopy(ntrials)
psth = copy.deepcopy(ntrials)
hasSpikes = copy.deepcopy(psth)
hasResp = copy.deepcopy(psth)
peakResp = copy.deepcopy(psth)
timeToFirstSpike = copy.deepcopy(psth)
for obj in exps:
ephysHemi = hemiLabels if hasattr(obj,'hemi') and obj.hemi=='right' else hemiLabels[::-1]
validTrials = ~obj.longFrameTrials
for stim in stimLabels:
for rd,hemi in zip((1,-1),ephysHemi):
stimTrials = obj.trialType==stim
if stim in ('targetOnly','mask'):
stimTrials = stimTrials & (obj.rewardDir==rd)
for resp in behavRespLabels:
if resp=='all':
respTrials = np.ones(obj.ntrials,dtype=bool)
else:
respTrials = ~np.isnan(obj.responseDir) if resp=='go' else np.isnan(obj.responseDir)
for mo in np.unique(obj.maskOnset[stimTrials]):
moTrials = obj.maskOnset==mo
trials = validTrials & stimTrials & respTrials & (obj.maskOnset==mo)
if mo not in ntrials[stim][hemi][resp]:
ntrials[stim][hemi][resp][mo] = 0
trialPsth[stim][hemi][resp][mo] = []
psth[stim][hemi][resp][mo] = []
hasSpikes[stim][hemi][resp][mo] = []
hasResp[stim][hemi][resp][mo] = []
peakResp[stim][hemi][resp][mo] = []
timeToFirstSpike[stim][hemi][resp][mo] = []
ntrials[stim][hemi][resp][mo] += trials.sum()
startTimes = obj.frameSamples[obj.stimStart[trials]+obj.frameDisplayLag]/obj.sampleRate-preTime
for u in obj.goodUnits:
spikeTimes = obj.units[u]['samples']/obj.sampleRate
p,t = getPsth(spikeTimes,startTimes,windowDur,binSize=binSize,avg=False)
trialPsth[stim][hemi][resp][mo].append(p)
m = p.mean(axis=0)
t -= preTime
psth[stim][hemi][resp][mo].append(m)
analysisWindow = (t>0.04) & (t<0.1)
hasSpikes[stim][hemi][resp][mo].append(m[analysisWindow].mean() > activeThresh)
b = p[:,t<0].mean(axis=1)
r = p[:,analysisWindow].mean(axis=1)
peak = (m-b.mean())[analysisWindow].max()
pval = 1 if np.sum(r-b)==0 else scipy.stats.wilcoxon(b,r)[1]
hasResp[stim][hemi][resp][mo].append(peak>respThresh*m[t<0].std() and pval<0.05)
peakResp[stim][hemi][resp][mo].append(peak)
lat = []
for st in startTimes:
firstSpike = np.where((spikeTimes > st+0.03) & (spikeTimes < st+0.15))[0]
if len(firstSpike)>0:
lat.append(spikeTimes[firstSpike[0]]-st)
else:
lat.append(np.nan)
timeToFirstSpike[stim][hemi][resp][mo].append(np.nanmedian(lat)*1000)
activeUnits = np.array(hasSpikes['targetOnly']['contra']['all'][0]) | np.array(hasSpikes['maskOnly']['contra']['all'][0])
targetRespUnits = np.array(hasResp['targetOnly']['contra']['all'][0])
maskRespUnits = np.array(hasResp['maskOnly']['contra']['all'][0])
respUnits = targetRespUnits | maskRespUnits
xlim = [-0.1,0.65]
for units in (respUnits,):
axs = []
ymin = ymax = 0
for resp in ('all',): #behavRespLabels:
fig = plt.figure(figsize=(10,5))
fig.text(0.5,0.99,'n='+str(units.sum())+' units',ha='center',va='top',fontsize=12)
for i,hemi in enumerate(hemiLabels):
ax = fig.add_subplot(1,2,i+1)
axs.append(ax)
for stim,clr in zip(stimLabels,('0.5','r','k','m')):
mskOn = psth[stim][hemi][resp].keys()
if stim=='mask' and len(mskOn)>1:
cmap = np.ones((len(mskOn),3))
cint = 1/len(mskOn)
cmap[:,1:] = np.arange(0,1.01-cint,cint)[:,None]
else:
cmap = [clr]
for mo,c in zip(mskOn,cmap):
p = np.array(psth[stim][hemi][resp][mo])[units]
m = np.mean(p,axis=0)
s = np.std(p,axis=0)/(len(p)**0.5)
lbl = 'target+mask, SOA '+str(round(1000*mo/frameRate,1))+' ms' if stim=='mask' else stim
rlbl = '' if resp=='all' else ' '+resp
lbl += ' ('+str(ntrials[stim][hemi][resp][mo])+rlbl+' trials)'
ax.plot(t,m,color=c,label=lbl)
# ax.fill_between(t,m+s,m-s,color=c,alpha=0.25)
ymin = min(ymin,np.min(m[(t>=xlim[0]) & (t<=xlim[1])]))
ymax = max(ymax,np.max(m[(t>=xlim[0]) & (t<=xlim[1])]))
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False)
ax.set_xlim(xlim)
ax.set_xlabel('Time from stimulus onset (s)')
if i==0:
ax.set_ylabel('Response (spikes/s)')
ax.legend(loc='upper left',frameon=False,fontsize=8)
ax.set_title('target '+hemi)
for ax in axs:
ax.set_ylim([1.05*ymin,1.05*ymax])
# save psth
units = respUnits & ~fs
popPsth = {stim: {hemi: {} for hemi in hemiLabels} for stim in stimLabels}
for stim in stimLabels:
for hemi in hemiLabels:
p = psth[stim][hemi]['all']
for mo in p.keys():
popPsth[stim][hemi][mo] = np.array(p[mo])[units]
popPsth['t'] = t
pkl = fileIO.saveFile(fileType='*.pkl')
pickle.dump(popPsth,open(pkl,'wb'))
# peak response and latency
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ymax = 0
for stim,clr,lbl in zip(('targetOnly','maskOnly'),'kg',('Target Only','Mask Only')):
p = np.array(psth[stim]['contra']['all'][0])[respUnits]
m = p.mean(axis=0)
s = p.std(axis=0)/(len(p)**0.5)
ax.plot(t*1000,m,color=clr,label=lbl)
ax.fill_between(t*1000,m+s,m-s,color=clr,alpha=0.25)
ymax = max(ymax,np.max(m+s))
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_xlim([0,150])
ax.set_ylim([0,42])
ax.set_xlabel('Time From Stimulus Onset (ms)',fontsize=16)
ax.set_ylabel('Spikes/s',fontsize=16)
ax.legend(loc='upper right',fontsize=14)
plt.tight_layout()
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
for stim,clr in zip(('targetOnly','maskOnly'),'kg'):
d = np.array(timeToFirstSpike[stim]['contra']['all'][0])[respUnits]
s = np.sort(d)
c = [np.sum(s<=i)/len(s) for i in s]
ax.plot(s,c,color=clr,lw=2,label=stim)
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_xlim([0,150])
ax.set_ylim([0,1.02])
ax.set_xlabel('Median Time To First Spike (ms)',fontsize=16)
ax.set_ylabel('Cumulative Probability',fontsize=16)
#ax.legend(loc='upper right',fontsize=14)
plt.tight_layout()
for data,amax,albl in zip((peakResp,timeToFirstSpike),(200,150),('Peak Response (spikes/s)','Median Time To First Spike (ms)')):
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ymax = 0
ax.plot([0,1000],[0,1000],'--',color='0.8')
for cellType,clr,lbl in zip((~fs,fs),'km',('RS','FS')):
d = [np.array(data[stim]['contra']['all'][0])[respUnits & cellType] for stim in ('targetOnly','maskOnly')]
ax.plot(d[0],d[1],'o',color=clr,label=lbl)
ymax = max(ymax,np.nanmax(d))
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
aticks = np.arange(0,amax+1,50)
ax.set_xticks(aticks)
ax.set_yticks(aticks)
ax.set_xlim([0,amax])
ax.set_ylim([0,amax])
ax.set_aspect('equal')
ax.set_xlabel('Target',fontsize=16)
ax.set_ylabel('Mask',fontsize=16)
ax.set_title(albl,fontsize=16)
if 'Peak Response' in albl:
ax.legend(loc='lower right',fontsize=14)
plt.tight_layout()
for data,lbl in zip((peakResp,timeToFirstSpike),('peak response','time to first spike')):
d = [np.array(data[stim]['contra']['all'][0])[respUnits] for stim in ('targetOnly','maskOnly')]
m = [np.nanmedian(i) for i in d]
pval = scipy.stats.ranksums(d[0],d[1])[1]
print(lbl,('target','mask'),m,pval)
for data,lbl in zip((peakResp,timeToFirstSpike),('peak response','time to first spike')):
for stim in ('targetOnly','maskOnly'):
d = [np.array(data[stim]['contra']['all'][0])[respUnits & ct] for ct in (~fs,fs)]
m = [np.nanmedian(i) for i in d]
pval = scipy.stats.ranksums(d[0],d[1])[1]
print(lbl,stim,('RS','FS'),m,pval)
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
for cellType,clr,lbl in zip((~fs,fs),'km',('RS','FS')):
d = [np.array(data['maskOnly']['contra']['all'][0])[respUnits & cellType] for data in (timeToFirstSpike,peakResp)]
ax.plot(d[0],d[1],'o',color=clr,label=lbl)
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_xlabel('Time To First Spike (ms)',fontsize=16)
ax.set_ylabel('Peak Response To Mask (spikes/s)',fontsize=16)
ax.legend(fontsize=12)
plt.tight_layout()
# compare contra, ipsi, and predicted response to target + mask
for units in (respUnits,): #targetRespUnits,maskRespUnits & ~targetRespUnits):
units = units & ~fs
fig = plt.figure(figsize=(4,6))
axs = []
ymin = 0
ymax = 0
i = 0
target = np.array(psth['targetOnly']['contra']['all'][0])[units]
for stim in ('targetOnly','mask','maskOnly'):
for mo in psth[stim]['contra']['all']:
ax = fig.add_subplot(6,1,i+1)
for hemi,clr in zip(hemiLabels,'rb'):
p = np.array(psth[stim][hemi]['all'][mo])[units]
b = p-p[:,t<0].mean(axis=1)[:,None]
m = np.mean(b,axis=0)
s = np.std(b,axis=0)/(len(p)**0.5)
if stim=='targetOnly':
lbl = hemi+'lateral target'
elif stim=='maskOnly':
if hemi=='contra':
clr = 'k'
lbl = 'no target'
else:
continue
else:
lbl = None
ax.plot(t*1000,m,color=clr,label=lbl)
ax.fill_between(t*1000,m+s,m-s,color=clr,alpha=0.25)
ymin = min(ymin,np.min(m-s))
ymax = max(ymax,np.max(m+s))
if stim=='mask':
mask = np.array(psth['mask']['ipsi']['all'][mo])[units]
p = target+mask
b = p-p[:,t<0].mean(axis=1)[:,None]
m = np.mean(b,axis=0)
s = np.std(b,axis=0)/(len(p)**0.5)
# ax.plot(t*1000,m,ls='--',color='0.5',label='linear sum\ntarget + mask')
# ax.fill_between(t*1000,m+s,m-s,color='0.5',alpha=0.25)
ymin = min(ymin,np.min(m-s))
ymax = max(ymax,np.max(m+s))
if i==5:
ax.set_xlabel('Time From Stimulus Onset (ms)',fontsize=12)
else:
ax.set_xticklabels([])
if i==0:
ax.set_ylabel('Spikes/s',fontsize=12)
title = stim
if stim=='targetOnly':
title = 'target only'
elif stim=='maskOnly':
title = 'mask only'
elif stim=='mask':
title = title+' onset\n'+str(int(round(mo/120*1000)))+' ms'
ax.text(0.15,0.99,title,transform=ax.transAxes,color='k',fontsize=10,ha='center',va='top')
if i in (0,1,5):
ax.legend(fontsize=8)
axs.append(ax)
i += 1
for ax in axs:
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=10)
ax.set_xlim([0,200])
ax.set_ylim([1.05*ymin,1.05*ymax])
plt.tight_layout()
# cumulative spike count and excess contralateral spikes
maskOnset = np.unique(exps[0].maskOnset[~np.isnan(exps[0].maskOnset)])
maskOnsetTicks = np.concatenate((maskOnset,(8,10)))/frameRate*1000
maskOnsetLabels = ['mask only']+[str(int(round(onset)))+' ms' for onset in maskOnsetTicks[1:-2]] + ['target only','no stim']
maskOnsetClrs = np.zeros((5,3))
maskOnsetClrs[:-1] = plt.cm.plasma(np.linspace(0,0.85,5-1))[::-1,:3]
units = respUnits & ~fs
analysisWindow = (t>0.033) & (t<0.2)
cumContra = []
cumIpsi =[]
for stim in stimLabels:
for mo in maskOnset:
if stim!='mask' and mo>0 or stim=='mask' and mo==0:
continue
cumSpikes = []
for hemi in ('contra','ipsi'):
p = np.array(psth[stim][hemi]['all'][mo])[units]
b = p-p[:,t<0].mean(axis=1)[:,None]
cumSpikes.append(np.cumsum(b[:,analysisWindow],axis=1)*binSize)
cumContra.append(cumSpikes[0])
cumIpsi.append(cumSpikes[1])
for i,lbl in enumerate(maskOnsetLabels):
d = cumContra[i].max(axis=1)
print(lbl,np.median(d),np.min(d),np.max(d))
d = np.array(cumContra)[1:5].max(axis=2)
print(np.median(d),np.min(d),np.max(d))
for resp,hemi in zip((cumContra,cumIpsi),('Contralateral','Ipsilateral')):
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
for d,clr,lbl in zip(resp[1:-1],maskOnsetClrs,maskOnsetLabels[1:-1]):
m = d.mean(axis=0)
s = d.std(axis=0)/(len(d)**0.5)
ax.plot(t[analysisWindow]*1000,m,color=clr,label=lbl)
ax.fill_between(t[analysisWindow]*1000,m+s,m-s,color=clr,alpha=0.25)
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_xlim([33,200])
ax.set_yticks([0,0.5,1,1.5])
ax.set_ylim([-0.05,1.5])
ax.set_xlabel('Time Relative to Target Onset (ms)',fontsize=16)
ax.set_ylabel('Cumulative Spikes Per Neuron\n('+hemi+' Target)',fontsize=16)
# ax.set_title(hemi+' Target',fontsize=14)
if hemi=='Contralateral':
leg = ax.legend(title='mask onset',loc='upper left',fontsize=12)
plt.setp(leg.get_title(),fontsize=12)
plt.tight_layout()
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
for contra,ipsi,clr,lbl in zip(cumContra[1:-1],cumIpsi[1:-1],maskOnsetClrs,maskOnsetLabels[1:-1]):
d = contra-ipsi
m = d.mean(axis=0)
s = d.std(axis=0)/(len(d)**0.5)
ax.plot(t[analysisWindow]*1000,m,color=clr,label=lbl)
ax.fill_between(t[analysisWindow]*1000,m+s,m-s,color=clr,alpha=0.25)
ax.plot([33,200],[0,0],'k--')
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_yticks([0,0.5])
ax.set_xlim([33,200])
ax.set_xlabel('Time Relative to Target Onset (ms)',fontsize=16)
ax.set_ylabel('Difference in Cumulative Spikes\n(Contra - Ipsi Target)',fontsize=16)
plt.tight_layout()
# behavior analysis
rtBins = ((0,200),(200,650))
behavOutcomeLabels = ('all','right','left','correct','incorrect','all resp','no resp') + rtBins
behavTrialPsth = {stim: {hemi: {resp: {} for resp in behavOutcomeLabels} for hemi in hemiLabels} for stim in stimLabels}
behavPsth = copy.deepcopy(behavTrialPsth)
reacTime = copy.deepcopy(behavTrialPsth)
for i,obj in enumerate(exps):
ephysHemi,R,L = (hemiLabels,-1,1) if hasattr(obj,'hemi') and obj.hemi=='right' else (hemiLabels[::-1],1,-1)
validTrials = (~obj.longFrameTrials) & obj.engaged & (~obj.earlyMove)
for stim in stimLabels:
for rd,hemi in zip((1,-1),ephysHemi):
stimTrials = obj.trialType==stim
if stim in ('targetOnly','mask'):
stimTrials = stimTrials & (obj.rewardDir==rd)
for resp in behavOutcomeLabels:
if resp=='right':
respTrials = obj.responseDir==R
elif resp=='left':
respTrials = obj.responseDir==L
elif resp=='correct':
respTrials = obj.response==1
elif resp=='incorrect':
respTrials = obj.response==-1
elif resp=='all resp':
respTrials = ~np.isnan(obj.responseDir)
elif resp=='no resp':
respTrials = obj.response==0
else:
respTrials = np.ones(len(stimTrials),dtype=bool)
for mo in np.unique(obj.maskOnset[stimTrials]):
trials = validTrials & stimTrials & respTrials & (obj.maskOnset==mo)
if resp in rtBins:
trials = trials & (obj.reactionTime>=resp[0]) & (obj.reactionTime < resp[1])
if mo not in behavTrialPsth[stim][hemi][resp]:
behavTrialPsth[stim][hemi][resp][mo] = []
behavPsth[stim][hemi][resp][mo] = []
reacTime[stim][hemi][resp][mo] = []
reacTime[stim][hemi][resp][mo].append(obj.reactionTime[trials])
startTimes = obj.frameSamples[obj.stimStart[trials]+obj.frameDisplayLag]/obj.sampleRate - preTime
for u in obj.goodUnits:
spikeTimes = obj.units[u]['samples']/obj.sampleRate
p,t = getPsth(spikeTimes,startTimes,windowDur,binSize=binSize,avg=False)
behavTrialPsth[stim][hemi][resp][mo].append(p)
behavPsth[stim][hemi][resp][mo].append(p.mean(axis=0))
t -= preTime
expInd = np.array([i for i,obj in enumerate(exps) for _ in enumerate(obj.goodUnits)])
isOptoExpUnit = np.array([np.any(~np.isnan(obj.optoOnset)) for obj in exps for unit in obj.goodUnits])
behavUnits = respUnits & ~fs
baselineWindow = (t>-0.2) & (t<0.033)
analysisWindow = (t>0.033) & (t<0.2)
maskOnset = [0,2]
# right vs left
figs = []
axs = []
ymin = ymax = 0
for stim in stimLabels:
mo = 2 if stim=='mask' else 0
for i,hemi in enumerate(hemiLabels):
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
xmax = 0
for resp,clr in zip(('left','right'),('g','m')):
p = np.array(behavPsth[stim][hemi][resp][mo])[behavUnits]
b = p-p[:,baselineWindow].mean(axis=1)[:,None]
m = np.nanmean(b,axis=0)
s = np.nanstd(b,axis=0)/(b.shape[0]**0.5)
lbl = 'turn '+resp
if stim in ('targetOnly','mask'):
c = 'correct' if (hemi=='contra' and resp=='left') or (hemi=='ipsi' and resp=='right') else 'incorrect'
lbl += ' ('+c+')'
ax.plot(t*1000,m,color=clr,lw=2,label=lbl)
ax.fill_between(t*1000,m+s,m-s,color=clr,alpha=0.25)
if stim != 'catch':
ymin = min(ymin,np.min(m-s))
ymax = max(ymax,np.max(m+s))
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_xticks(np.arange(0,250,50))
ax.set_xlim([0,200])
ax.set_xlabel('Time Relative to Target Onset (ms)',fontsize=16)
ax.set_ylabel('Spikes/s',fontsize=16)
title = stim if stim in ('maskOnly','catch') else stim+', '+hemi+' target'
ax.set_title(title,fontsize=14)
ax.legend(loc='upper right',fontsize=14)
axs.append(ax)
figs.append(fig)
if stim in ('maskOnly','catch'):
break
for a in axs:
a.set_ylim([1.02*ymin,1.02*ymax])
for f in figs:
f.tight_layout()
for stim in stimLabels:
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.plot([0,200],[0,0],':',color='k')
mo = 2 if stim=='mask' else 0
for hemi,clr in zip(hemiLabels,('r','b')):
clr = clr if stim in ('targetOnly','mask') else 'k'
r = []
for resp in ('left','right'):
p = np.array(behavPsth[stim][hemi][resp][mo])[behavUnits]
b = p-p[:,baselineWindow].mean(axis=1)[:,None]
r.append(np.cumsum(b[:,analysisWindow],axis=1)*binSize)
d = np.array(r[0])-np.array(r[1])
m = np.nanmean(d,axis=0)
s = np.nanstd(d)/(len(d)**0.5)
ax.plot(t[analysisWindow]*1000,m,color=clr,lw=2,label=hemi+'lateral target')
ax.fill_between(t[analysisWindow]*1000,m+s,m-s,color=clr,alpha=0.25)
if stim in ('maskOnly','catch'):
break
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_yticks([-0.5,0,0.5,1])
ax.set_xlim([33,200])
ax.set_ylim([-0.2,1])
ax.set_xlabel('Time Relative to Target Onset (ms)',fontsize=16)
ax.set_ylabel('Difference in Cumulative Spikes\n(Turn Left - Turn Right)',fontsize=16)
ax.set_title(stim,fontsize=14)
if stim in ('targetOnly','mask'):
ax.legend(loc='upper left',fontsize=14)
plt.tight_layout()
# correct vs incorrect
figs = []
axs = []
ymin = ymax = 0
for mo,moLbl in zip(maskOnset,('target only','target, mask onset 17 ms')):
fig = plt.figure()
stim = 'mask' if mo>0 else 'targetOnly'
for i,hemi in enumerate(hemiLabels):
ax = fig.add_subplot(2,1,i+1)
xmax = 0
for resp,clr in zip(('correct','incorrect'),('g','m')):
p = np.array(behavPsth[stim][hemi][resp][mo])[behavUnits]
b = p-p[:,baselineWindow].mean(axis=1)[:,None]
m = np.nanmean(b,axis=0)
s = np.nanstd(b,axis=0)/(b.shape[0]**0.5)
ax.plot(t*1000,m,color=clr,lw=2,label=resp)
ax.fill_between(t*1000,m+s,m-s,color=clr,alpha=0.25)
ymin = min(ymin,np.min(m-s))
ymax = max(ymax,np.max(m+s))
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=12)
ax.set_xticks(np.arange(0,250,50))
ax.set_xlim([0,200])
if i==1:
ax.set_xlabel('Time Relative to Target Onset (ms)',fontsize=14)
ax.set_ylabel('Spikes/s',fontsize=14)
ax.set_title(hemi+' '+moLbl,fontsize=14)
if i==0:
ax.legend()
axs.append(ax)
figs.append(fig)
for a in axs:
a.set_ylim([1.02*ymin,1.02*ymax])
for f in figs:
f.tight_layout()
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.plot([0,200],[0,0],':',color='k')
for mo,moLbl in zip(maskOnset,('target only','target, mask onset 17 ms')):
stim = 'mask' if mo>0 else 'targetOnly'
for hemi,ls in zip(hemiLabels,('-','--')):
r = []
for resp in ('correct','incorrect'):
p = np.array(behavPsth[stim][hemi][resp][mo])[behavUnits]
b = p-p[:,baselineWindow].mean(axis=1)[:,None]
r.append(np.cumsum(b[:,analysisWindow],axis=1)*binSize)
d = np.array(r[0])-np.array(r[1])
m = np.nanmean(d,axis=0)
s = np.nanstd(d)/(len(d)**0.5)
if mo==0:
clr = 'k' if hemi=='contra' else '0.5'
else:
clr = 'r' if hemi=='contra' else [1,0.5,0.5]
lbl = hemi+' '+moLbl
ax.plot(t[analysisWindow]*1000,m,color=clr,ls=ls,label=lbl)
ax.fill_between(t[analysisWindow]*1000,m+s,m-s,color=clr,alpha=0.25)
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_yticks([-0.5,0,0.5,1])
ax.set_xlim([33,200])
ax.set_ylim([-0.7,1])
ax.set_xlabel('Time Relative to Target Onset (ms)',fontsize=16)
ax.set_ylabel('Difference in Cumulative Spikes\n(Correct - Incorrect Trials)',fontsize=16)
leg = ax.legend(loc='upper left',fontsize=10)
plt.tight_layout()
for mo,moLbl in zip(maskOnset,('target only','mask onset 17 ms')):
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.plot([0,200],[0,0],':',color='k')
stim = 'mask' if mo>0 else 'targetOnly'
# for resp,clr in zip(('all',),'k'):
for resp,clr in zip(('all resp','correct','incorrect'),'kgm'):
r = []
for hemi in hemiLabels:
p = np.array(behavPsth[stim][hemi][resp][mo])[behavUnits]
b = p-p[:,baselineWindow].mean(axis=1)[:,None]
r.append(np.cumsum(b[:,analysisWindow],axis=1)*binSize)
d = np.array(r[0])-np.array(r[1])
m = np.nanmean(d,axis=0)
s = np.nanstd(d)/(len(d)**0.5)
lbl = 'all responses' if resp=='all resp' else resp
ax.plot(t[analysisWindow]*1000,m,color=clr,label=lbl)
ax.fill_between(t[analysisWindow]*1000,m+s,m-s,color=clr,alpha=0.25)
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_yticks([-0.5,0,0.5,1])
ax.set_xlim([33,200])
ax.set_ylim([-0.7,0.8])
ax.set_xlabel('Time Relative to Target Onset (ms)',fontsize=16)
ax.set_ylabel('Difference in Cumulative Spikes\n(Contra - Ipsi Target)',fontsize=16)
ax.set_title(moLbl,fontsize=16)
leg = ax.legend(loc='lower left',fontsize=14)
plt.tight_layout()
# fast vs slow reaction
for mo,moLbl in zip(maskOnset,('target only','target, mask onset 17 ms')):
fig = plt.figure()
stim = 'mask' if mo>0 else 'targetOnly'
for i,hemi in enumerate(hemiLabels):
ax = fig.add_subplot(2,1,i+1)
xmax = 0
for resp,clr,lbl in zip(rtBins,'gm',('fast','slow')):
stim = 'mask' if mo>0 else 'targetOnly'
p = np.array(behavPsth[stim][hemi][resp][mo])[behavUnits]
b = p-p[:,baselineWindow].mean(axis=1)[:,None]
m = np.nanmean(b,axis=0)
s = np.nanstd(b,axis=0)/(b.shape[0]**0.5)
ax.plot(t*1000,m,color=clr,lw=2,label=lbl+' reaction')
ax.fill_between(t*1000,m+s,m-s,color=clr,alpha=0.25)
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=12)
ax.set_xticks(np.arange(0,250,50))
ax.set_xlim([0,200])
if i==1:
ax.set_xlabel('Time Relative to Target Onset (ms)',fontsize=14)
ax.set_ylabel('Spikes/s',fontsize=14)
ax.set_title(hemi+' '+moLbl,fontsize=14)
if i==0:
ax.legend()
plt.tight_layout()
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.plot([0,200],[0,0],':',color='0.5')
for mo,clr,moLbl in zip(maskOnset,'kr',('target only','mask onset 17 ms')):
stim = 'mask' if mo>0 else 'targetOnly'
for hemi,ls in zip(hemiLabels,('-','--')):
r = []
for resp in rtBins:
p = np.array(behavPsth[stim][hemi][resp][mo])[behavUnits]
b = p-p[:,baselineWindow].mean(axis=1)[:,None]
r.append(np.cumsum(b[:,analysisWindow],axis=1)*binSize)
d = np.array(r[0])-np.array(r[1])
m = np.nanmean(d,axis=0)
s = np.nanstd(d)/(len(d)**0.5)
ax.plot(t[analysisWindow]*1000,m,color=clr,ls=ls,label=moLbl)
ax.fill_between(t[analysisWindow]*1000,m+s,m-s,color=clr,alpha=0.25)
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_yticks([-0.5,0,0.5,1])
ax.set_xlim([33,200])
ax.set_ylim([-0.7,1])
ax.set_xlabel('Time Relative to Target Onset (ms)',fontsize=16)
ax.set_ylabel('Difference in Cumulative Spikes\n(Fast - Slow Reaction Trials)',fontsize=16)
leg = ax.legend(loc='upper left',fontsize=10)
plt.tight_layout()
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.plot([0,200],[0,0],':',color='k')
for mo,moLbl in zip(maskOnset,('target only','mask onset 17 ms')):
stim = 'mask' if mo>0 else 'targetOnly'
for resp,lbl,ls in zip(rtBins,('fast','slow'),('-','--')):
r = []
for hemi in hemiLabels:
p = np.array(behavPsth[stim][hemi][resp][mo])[behavUnits]
b = p-p[:,baselineWindow].mean(axis=1)[:,None]
r.append(np.cumsum(b[:,analysisWindow],axis=1)*binSize)
d = np.array(r[0])-np.array(r[1])
m = np.nanmean(d,axis=0)
s = np.nanstd(d)/(len(d)**0.5)
if mo==0:
clr = 'k' if lbl=='fast' else '0.5'
else:
clr = 'r' if lbl=='fast' else [1,0.5,0.5]
lbl = moLbl + ', ' + lbl
ax.plot(t[analysisWindow]*1000,m,color=clr,ls=ls,label=lbl)
ax.fill_between(t[analysisWindow]*1000,m+s,m-s,color=clr,alpha=0.25)
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_yticks([-0.5,0,0.5,1])
ax.set_xlim([33,200])
ax.set_ylim([-0.7,1])
ax.set_xlabel('Time Relative to Target Onset (ms)',fontsize=16)
ax.set_ylabel('Difference in Cumulative Spikes\n(Contra - Ipsi Target)',fontsize=16)
leg = ax.legend(loc='lower left',fontsize=10)
plt.tight_layout()
# save behav psth
resp = 'all resp'
popPsth = {stim: {hemi: {} for hemi in hemiLabels} for stim in stimLabels}
for stim in stimLabels:
for hemi in hemiLabels:
p = behavPsth[stim][hemi][resp] if stim in behavPsth else psth[stim][hemi]['all']
for mo in (m for m in p.keys() if m in maskOnset):
popPsth[stim][hemi][mo] = np.array(p[mo])[behavUnits]
popPsth['t'] = t
pkl = fileIO.saveFile(fileType='*.pkl')
pickle.dump(popPsth,open(pkl,'wb'))
# decoding
def getDecoderResult(decoderMethod,var,varLabels,units,dPsth,trainInd,testInd,numTrials,randomizeTrials,analysisWindow,dataType,offsets,mskOn):
# X = features, y = decoding label, m = mask onset label
X = {trialSet: [] for trialSet in ('train','test')}
y = []
m = []
for trialSet,trialInd in zip(('train','test'),(trainInd,testInd)):
for i,u in enumerate(units):
X[trialSet].append([])
for vind,vlbl in enumerate(varLabels):
for mo in mskOn:
stim = 'targetOnly' if mo==0 else 'mask'
if var=='side':
p = dPsth[stim][vlbl]['all']
ind = trialInd[vlbl][mo][u]
trials = np.random.choice(ind,numTrials,replace=True) if randomizeTrials else ind[:numTrials]
else:
pContra = dPsth[stim]['contra'][vlbl]
pIpsi = dPsth[stim]['ipsi'][vlbl]
trials = np.stack([np.random.choice(trialInd[vlbl][mo][hemi][u],numTrials,replace=True) for hemi in ('contra','ipsi')],axis=1)
for trial in trials:
if var=='side':
X[trialSet][-1].append(p[mo][unitInd[u]][trial][analysisWindow])
else:
X[trialSet][-1].append(pContra[mo][unitInd[u]][trial[0]][analysisWindow]-pIpsi[mo][unitInd[u]][trial[1]][analysisWindow])
if i==0 and trialSet=='train':
y.append(vind)
m.append(mo)
X[trialSet] = np.concatenate(X[trialSet],axis=1)
y = np.array(y)
m = np.array(m)
if decoderMethod=='RF':
decoder = RandomForestClassifier(n_estimators=100)
featureMethod = 'feature_importances_'
else:
decoder = LinearSVC(C=1.0,max_iter=1e4)
featureMethod = 'coef_'
trainScore = np.full((3,len(offsets),len(maskOnset)),np.nan)
testScore = trainScore.copy()
for i,lbl in enumerate(dataType):
for j,offset in enumerate(offsets):
if offset is None:
if lbl=='psth':
Xtrain,Xtest = X['train'],X['test']
else:
break
else:
ind = offset if lbl=='bin' else slice(0,offset+1)
Xdata = []
for trialSet in ('train','test'):
d = np.reshape(X[trialSet],(len(y),len(units),-1))[:,:,ind]
if lbl=='count':
d = d.sum(axis=2)*binSize
Xdata.append(np.reshape(d,(len(y),-1)))
Xtrain,Xtest = Xdata
decoder.fit(Xtrain,y)
for k,mo in enumerate(mskOn):
ind = m==mo
trainScore[i][j][k] = decoder.score(Xtrain[ind],y[ind])
testScore[i][j][k] = decoder.score(Xtest[ind],y[ind])
if lbl=='psth' and offset==offsets[-1]:
features = np.mean(np.reshape(np.absolute(getattr(decoder,featureMethod)),(len(units),-1)),axis=0)
return trainScore,testScore,features
maskOnset = [0,2,3,4,6]
analysisWindow = (t>0) & (t<0.2)
unitInd = np.where(respUnits & ~fs)[0]
nUnits = len(unitInd)
unitSessionInd = np.array([i for i,obj in enumerate(exps) for _ in enumerate(obj.goodUnits)])[unitInd]
unitSampleSize = {}
unitSampleSize['sessionCorr'] = [np.sum(unitSessionInd==i) for i in range(len(exps))]
unitSampleSize['sessionRand'] = unitSampleSize['sessionCorr']
unitSampleSize['pooled'] = [1,5,10,20,40,nUnits]
decoderOffset = np.arange(analysisWindow.sum())
trainTestIters = 100
trialsPerIter = 100
decoderMethod = 'SVM' # 'SVM' or 'RF'
decodingVariable = ('side','choice')
unitSource = list(unitSampleSize.keys())
dataType = ('psth','bin','count')
trainScore = {var: {src: np.full((trainTestIters,len(unitSampleSize[src]),len(dataType),len(decoderOffset),len(maskOnset)),np.nan) for src in unitSource} for var in decodingVariable}
testScore = copy.deepcopy(trainScore)
decoderFeatures = {var: {src: np.full((trainTestIters,len(unitSampleSize[src]),analysisWindow.sum()),np.nan) for src in unitSource} for var in decodingVariable}
for var,usource,dPsth,mskOn in zip(decodingVariable,(unitSource,('pooled',)),(trialPsth,behavTrialPsth),([0,2,3,4,6],[2])):
if var=='choice':
continue
for iterInd in range(trainTestIters):
# assign trials to training and testing sets
if var=='side':
varLabels = hemiLabels
trainInd = {vlbl: {mo: [] for mo in mskOn} for vlbl in varLabels}
trialKeys = ('all',)
else:
varLabels = ('correct','incorrect')
trainInd = {vlbl: {mo: {hemi: [] for hemi in hemiLabels} for mo in mskOn} for vlbl in varLabels}
trialKeys = hemiLabels
testInd = copy.deepcopy(trainInd)
for vlbl in varLabels:
for mo in mskOn:
stim = 'targetOnly' if mo==0 else 'mask'
for session in range(len(exps)):
units = unitInd[unitSessionInd==session]
for key in trialKeys:
p = dPsth[stim][vlbl][key] if var=='side' else dPsth[stim][key][vlbl]
n = len(p[mo][units[0]])
trials = np.arange(n)
train = np.random.choice(trials,n//2,replace=False)
test = np.setdiff1d(trials,train)
for u in units:
if var=='side':
trainInd[vlbl][mo].append(train)
testInd[vlbl][mo].append(test)
else:
trainInd[vlbl][mo][key].append(train)
testInd[vlbl][mo][key].append(test)
for src in usource:
for s,sampleSize in enumerate(unitSampleSize[src]):
if src=='pooled':
if var=='choice':
unitSamples = [[u for u in range(nUnits) if all([len(ind[vlbl][mo][hemi][u])>0 for ind in (trainInd,testInd) for vlbl in varLabels for mo in mskOn for hemi in hemiLabels])]]
elif sampleSize==nUnits:
unitSamples = [np.arange(nUnits)]
elif sampleSize==1:
unitSamples = [[u] for u in range(nUnits)]
else:
# >99% chance each neuron is chosen at least once
nsamples = int(math.ceil(math.log(0.01)/math.log(1-sampleSize/nUnits)))
unitSamples = [np.random.choice(nUnits,sampleSize,replace=False) for _ in range(nsamples)]
numTrials = trialsPerIter
randomizeTrials = True
else:
unitSamples = [np.where(unitSessionInd==s)[0]]
numTrials = 10000
for trialSet,trialInd in zip(('train','test'),(trainInd,testInd)):
for hemi,rd in zip(hemiLabels,rewardDir):
for mo in mskOn:
numTrials = min(numTrials,len(trialInd[vlbl][mo][unitSamples[0][0]]))
randomizeTrials = False if src=='sessionCorr' else True
trainSamples = []
testSamples = []
coefSamples = []
for units in unitSamples:
offsets = [None] if src=='pooled' and sampleSize<max(unitSampleSize[src]) else decoderOffset
train,test,coef = getDecoderResult(decoderMethod,var,varLabels,units,dPsth,trainInd,testInd,numTrials,randomizeTrials,analysisWindow,dataType,offsets,mskOn)
trainSamples.append(train)
testSamples.append(test)
coefSamples.append(coef)
trainScore[var][src][iterInd,s] = np.mean(trainSamples,axis=0)
testScore[var][src][iterInd,s] = np.mean(testSamples,axis=0)
decoderFeatures[var][src][iterInd,s] = np.mean(coefSamples,axis=0)
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
for score,clr,trialSet in zip((trainScore,testScore),('0.5','k'),('train','test')):
# for src,mrk,mfc in zip(unitSource,'soo',('none','none',clr)):
for src,mrk,mfc in zip(unitSource[1:],'oo',('none',clr)):
overallScore = score['side'][src][:,:,0,-1].mean(axis=-1)
mean = overallScore.mean(axis=0)
sem = overallScore.std(axis=0)/(trainTestIters**0.5)
# if src=='sessionCorr':
# srcLabel = 'session'
# elif src=='sessionRand':
# srcLabel = 'session, shuffled'
# else:
# srcLabel = src
srcLabel = 'individual' if src=='sessionRand' else 'pooled'
lbl = trialSet+', '+srcLabel
ax.plot(unitSampleSize[src],mean,mrk,mec=clr,mfc=mfc,mew=2,ms=12,label=lbl)
# for x,m,s in zip(unitSampleSize[src],mean,sem):
# ax.plot([x,x],[m-s,m+s],clr)
for side in ('right','top'):
ax.spines[side].set_visible(False)
ax.tick_params(direction='out',top=False,right=False,labelsize=14)
ax.set_ylim([0.5,1.02])
ax.set_xlabel('Number of Units',fontsize=16)
ax.set_ylabel('Decoder Accuracy',fontsize=16)
ax.legend()
plt.tight_layout()