-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpseudounipolar_neuron_class.py
670 lines (557 loc) · 29.9 KB
/
pseudounipolar_neuron_class.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
# class definitions to generate an MRG axon and trial objects
from neuron import h
import numpy as np
import pandas as pd
import os
from scipy.interpolate import LinearNDInterpolator
def line(x, a, b):
return a*x + b
def logfn(x, h, k):
return h*np.log(x) + k
def logreg(x, L, x0, k, b):
return (L/(1.0 + np.exp(-k*(x-x0)))) + b
h.celsius = 37.0
h.dt = 0.005 # ms
class pseudounipolar_neuron:
"""Class to create a pseudounipolar neuron
Geometric properties of the soma and neck are obtained from Devor et al. Structure and electrical properties are
obtained from the MRG model modified for sensory neurons in Gaines et al. Contains functions to generate each branch
of the neuron with user defined properties and set the XYZ position.
NOTE
----
Instantiating the pseudounipolar_neuron object DOES NOT automatically generate positions for the neuron segments.
Object function setXYZpos MUST be run on the pseudounipolar_neuron object AFTER it has been instantiated.
"""
def __init__(self, centralFiberD=5.7, peripheralFiberD=7.3, neckFiberD=7.3, numNodes_p = 100, numNodes_c = 100, somaSize = 80, femElec=''):
"""
Parameters
----------
centralFiberD : float
diameter of central branch
peripheralFiberD: float
diameter of peripheral branch
neckFiberD: float
diameter of neck branch
numNodes_p: int
number of nodes after abnormal segments in peripheral branch
numNodes_c : int
number of nodes after abnormal segments in central branch
femdf : dict
dictionary of FEM dataframes
femElec : str
electrode used to run simulation
"""
self.centralFiberD = centralFiberD
self.peripheralFiberD = peripheralFiberD
self.neckFiberD = neckFiberD
self.timestep = h.dt
# number of nodes after initial abnormal sections
self.numNodes_c = numNodes_c
self.numNodes_p = numNodes_p
self.somaSize = somaSize
# electrical parameters
self.rhoa = 0.7e6 # [Ohm-um] specific axoplasmic resistance
self.mycm = 0.1 # uF/cm2/lamella membrane//
self.mygm = 0.001 # S/cm2/lamella membrane//
self.paralength1 = 3.0
self.nodelength = 1.0
self.space_p1 = 0.002
self.space_p2 = 0.004
self.space_i = 0.004
self.sectionDF = pd.DataFrame(columns=['nodeIndex', 'sectionIndex', 'object', 'puniBranch', 'sectionType', 'x', 'y', 'z', 'L'])
neckRoot = self.createSomaSection()
tjnRoot = self.createNeckSections(neckRoot)
centralRoot = self.createCentralBranch(tjnRoot)
peripheralRoot = self.createPeripheralBranch(tjnRoot)
neckIdx = self.sectionDF['puniBranch'] == 'n'
self.stemLength = self.sectionDF[neckIdx]['L'].sum()
self.etype = femElec
def createSomaSection(self):
"""Create soma and iseg sections with modified sodium channel densities"""
soma = h.Section(name='soma')
soma.nseg = 1
soma.L = self.somaSize
soma.diam = self.somaSize
soma.Ra = self.rhoa / 10000
soma.cm = 2
soma.insert('node_sensory')
for seg in list(soma.allseg())[1:-1]:
seg.node_sensory.gnabar = seg.node_sensory.gnabar * (300/2000) # 0.45
seg.node_sensory.gnapbar = seg.node_sensory.gnapbar * (300/2000) # 0.0015
seg.node_sensory.el = -80
soma.insert('extracellular')
soma.e_extracellular = 0
soma.xraxial[0] = (self.rhoa*.01)/(np.pi*((((soma.diam/2)+self.space_p1)**2)-((soma.diam/2)**2)))
soma.xg[0] = 1e10
soma.xc[0] = 0.0
self.sectionDF = self.sectionDF.append({'sectionType': 'soma',
'nodeIndex': 0,
'sectionIndex': 0,
'puniBranch': 'n',
'x': 0, 'y': 0, 'z': 0, 'L': soma.L, 'object': soma}, ignore_index=True)
# stem unmyelinated segment
iseg = h.Section(name='iseg')
iseg.nseg = 33
iseg.L = 200.0
iseg.diam = (5.0/7.3)*self.neckFiberD # adjusting linearly for axon size
iseg.Ra = self.rhoa / 10000
iseg.cm = 2.0
iseg.insert('extracellular')
iseg.e_extracellular = 0
iseg.xraxial[0] = (self.rhoa*.01)/(np.pi*((((iseg.diam/2)+self.space_p1)**2)-((iseg.diam/2)**2)))
iseg.xg[0] = 1e10
iseg.xc[0] = 0.0
iseg.insert('node_sensory')
for idx, seg in enumerate(list(iseg.allseg())[1:-1]):
if idx == 0:
seg.node_sensory.gnabar = seg.node_sensory.gnabar * (1000/2000) # 0.75
seg.node_sensory.gnapbar = seg.node_sensory.gnapbar * (1000/2000) # 0.0025
else:
seg.node_sensory.gnabar = seg.node_sensory.gnabar * (600 / 2000) # 0.75
seg.node_sensory.gnapbar = seg.node_sensory.gnapbar * (600 / 2000) # 0.0025
seg.node_sensory.el = -80
self.sectionDF = self.sectionDF.append({'sectionType':'iseg',
'nodeIndex':0,
'sectionIndex':0,
'puniBranch':'n',
'x':0,'y':0,'z':0,'L':iseg.L,'object':iseg},ignore_index=True)
iseg.connect(soma)
return iseg
def createNeckSections(self, parentNode):
"""Create stem MRG axon sections with abnormal internode and myelination"""
pair1 = self.MRGaxon(axonnodes=1, fiberD=self.neckFiberD, prefix='n', deltaX=85, nlFactor=0.07747)
pair2 = self.MRGaxon(axonnodes=1, fiberD=self.neckFiberD, prefix='n', deltaX=130, nlFactor=0.48592)
pair3 = self.MRGaxon(axonnodes=1, fiberD=self.neckFiberD, prefix='n', deltaX=168, nlFactor=0.66197)
pair4 = self.MRGaxon(axonnodes=1, fiberD=self.neckFiberD, prefix='n', deltaX=201, nlFactor=1.0)
pair1[0].connect(parentNode)
pair2[0].connect(pair1[-1])
pair3[0].connect(pair2[-1])
pair4[0].connect(pair3[-1])
modelParams = self.getModelParamsContinuous(self.neckFiberD, None, 1, 'n')
tjNode = self.createNodeSection(modelParams, 'n', parentSection=pair4[-1], sectionType='tjnNode')
return tjNode
def createCentralBranch(self, parentSection):
"""Create central MRG axon sections with 3 abnormal sections"""
pair1 = self.MRGaxon(axonnodes=1, fiberD=self.centralFiberD, prefix='c', deltaX=358, spec='c1')
pair2 = self.MRGaxon(axonnodes=1, fiberD=self.centralFiberD, prefix='c', deltaX=780, spec='c2')
pair3 = self.MRGaxon(axonnodes=1, fiberD=self.centralFiberD, prefix='c', deltaX=1170, spec='c3')
fiber = self.MRGaxon(axonnodes=self.numNodes_c, fiberD=self.centralFiberD, prefix='c')
pair1[0].connect(parentSection)
pair2[0].connect(pair1[-1])
pair3[0].connect(pair2[-1])
fiber[0].connect(pair3[-1])
return fiber
def createPeripheralBranch(self, parentSection):
"""Create peripheral MRG axon sections with 3 abnormal sections"""
pair1 = self.MRGaxon(axonnodes=1, fiberD=self.peripheralFiberD, prefix='p', deltaX=461, spec='p1')
pair2 = self.MRGaxon(axonnodes=1, fiberD=self.peripheralFiberD, prefix='p', deltaX=670, spec='p2')
pair3 = self.MRGaxon(axonnodes=1, fiberD=self.peripheralFiberD, prefix='p', deltaX=1119, spec='p3')
fiber2 = self.MRGaxon(axonnodes=self.numNodes_p, fiberD=self.peripheralFiberD, prefix='p')
pair1[0].connect(parentSection)
pair2[0].connect(pair1[-1])
pair3[0].connect(pair2[-1])
fiber2[0].connect(pair3[-1])
return fiber2
def MRGaxon(self, axonnodes, fiberD, prefix, deltaX=None, nlFactor=1.0, spec='none'):
"""Create axon fiber using sections defined in MRG model
order of sections: Node-MYSA-FLUT-[STIN x 6]-FLUT-MYSA
note: each iteration ends in a MYSA section so that the output can be connected with a node parent
"""
modelParams = self.getModelParamsContinuous(fiberD, deltaX, nlFactor, prefix)
# construct axon
parentSection = []
for iNode in range(axonnodes):
node = self.createNodeSection(modelParams, prefix)
mysa1 = self.createMYSAsection(modelParams, prefix, node)
flut1 = self.createFLUTsection(modelParams, prefix, mysa1)
STINroot = flut1
for iSTIN in range(6):
STINroot = self.createSTINsection(modelParams, prefix, parentSection=STINroot)
flut2 = self.createFLUTsection(modelParams, prefix, STINroot)
mysa2 = self.createMYSAsection(modelParams, prefix, flut2)
if parentSection:
node.connect(parentSection)
parentSection = mysa2
return mysa2.wholetree()
def createNodeSection(self, modelparams, prefix, parentSection=None, sectionType='node'):
"""Create Node section with node_sensory mechanism from Gaines et. al"""
tmp = h.Section(name='node')
tmp.nseg = 1
tmp.L = self.nodelength
tmp.Ra = self.rhoa / 10000
tmp.cm = 2.0
tmp.diam = modelparams['nodeD']
# xc is a hoc nickname for xc[0], i.e. the capacitance in the extracellular layer that is immediately adjacent to the
# cell membrane. xc[1] is the capacitance in the outer extracellular layer. Similar comments apply to xg and xg[1].
tmp.insert('node_sensory')
tmp.insert('extracellular')
tmp.e_extracellular = 0
tmp.xraxial[0] = modelparams['Rpn0']
tmp.xg[0] = 1e10
tmp.xc[0] = 0.0
self.sectionDF = self.sectionDF.append({'sectionType': sectionType,
'nodeIndex':((self.sectionDF['sectionType']=='node')&((self.sectionDF['puniBranch']==prefix))).sum(),
'sectionIndex': (self.sectionDF['sectionType']=='node').sum(),
'puniBranch': prefix,
'x': 0, 'y': 0, 'z': 0, 'L':tmp.L,'object': tmp},ignore_index=True)
if parentSection:
tmp.connect(parentSection)
return tmp
def createMYSAsection(self, modelparams, prefix, parentSection=None):
"""Create MYSA section with mysa_sensory mechanism from Gaines et. al"""
tmp = h.Section(name='MYSA')
tmp.nseg = 1
tmp.L = self.paralength1
tmp.Ra = self.rhoa * (1 / (modelparams['paraD1'] / modelparams['fiberD']) ** 2.0) / 10000
tmp.cm = 2.0 * modelparams['paraD1'] / modelparams['fiberD'] #/ self.myelinLayers
tmp.diam = modelparams['fiberD']
tmp.insert('mysa_sensory')
tmp.insert('extracellular')
tmp.e_extracellular = 0
tmp.xraxial[0] = modelparams['Rpn1']
tmp.xg[0] = self.mygm / (modelparams['nl'] * 2)
tmp.xc[0] = self.mycm / (modelparams['nl'] * 2)
self.sectionDF = self.sectionDF.append({'sectionType': 'MYSA',
'nodeIndex': ((self.sectionDF['sectionType'] == 'node') & ((self.sectionDF['puniBranch'] == prefix))).sum(),
'sectionIndex': (self.sectionDF['sectionType']=='MYSA').sum(),
'puniBranch': prefix,
'x': 0, 'y': 0, 'z': 0,'L':tmp.L, 'object': tmp}, ignore_index=True)
if parentSection:
tmp.connect(parentSection)
return tmp
def createFLUTsection(self, modelparams, prefix, parentSection=None):
"""Create FLUT section with flut_sensory mechanism from Gaines et. al"""
tmp = h.Section(name='FLUT')
tmp.nseg = 1
tmp.L = modelparams['paralength2']
tmp.Ra = self.rhoa * (1 / (modelparams['paraD2'] / modelparams['fiberD']) ** 2.0) / 10000
tmp.cm = 2.0 * modelparams['paraD2'] / modelparams['fiberD'] #/ self.myelinLayers
tmp.diam = modelparams['fiberD']
tmp.insert('flut_sensory')
tmp.insert('extracellular')
tmp.e_extracellular = 0
tmp.xraxial[0] = modelparams['Rpn2']
tmp.xg[0] = self.mygm / (modelparams['nl'] * 2)
tmp.xc[0] = self.mycm / (modelparams['nl'] * 2)
self.sectionDF = self.sectionDF.append({'sectionType': 'FLUT',
'nodeIndex':((self.sectionDF['sectionType']=='node')&((self.sectionDF['puniBranch']==prefix))).sum(),
'sectionIndex': (self.sectionDF['sectionType']=='FLUT').sum(),
'puniBranch': prefix,
'x': 0, 'y': 0, 'z': 0,'L':tmp.L, 'object': tmp}, ignore_index=True)
if parentSection:
tmp.connect(parentSection)
return tmp
def createSTINsection(self, modelparams, prefix, parentSection=None):
"""Create STIN section with stin_sensory mechanism from Gaines et. al"""
tmp = h.Section(name='STIN')
tmp.nseg = 1
tmp.L = modelparams['interlength']
tmp.Ra = self.rhoa * (1 / (modelparams['axonD'] / modelparams['fiberD']) ** 2.0) / 10000
tmp.cm = 2 * modelparams['axonD'] / modelparams['fiberD'] #/ self.myelinLayers
tmp.diam = modelparams['fiberD']
#tmp.diam = modelparams['paraD2']
tmp.insert('stin_sensory')
# tmp.insert('pas')
# tmp.g_pas = 0.0001 * modelparams['axonD'] / modelparams['fiberD']
# tmp.e_pas = -80.0
tmp.insert('extracellular')
tmp.e_extracellular = 0
tmp.xraxial[0] = modelparams['Rpx']
tmp.xg[0] = self.mygm / (modelparams['nl'] * 2)
tmp.xc[0] = self.mycm / (modelparams['nl'] * 2)
self.sectionDF = self.sectionDF.append({'sectionType': 'STIN',
'nodeIndex':((self.sectionDF['sectionType']=='node')&((self.sectionDF['puniBranch']==prefix))).sum(),
'sectionIndex': (self.sectionDF['sectionType']=='STIN').sum(),
'puniBranch': prefix,
'x': 0, 'y': 0, 'z': 0,'L':tmp.L, 'object': tmp}, ignore_index=True)
if parentSection:
tmp.connect(parentSection)
return tmp
def getModelParamsContinuous(self, fiberD, deltax, nlfactor, branch):
"""Generate nterpolated model parameters for all fiberD values based on interpolation functions
applied to the discrete values defined in the MRG model."""
g = line(fiberD, 0.01716804, 0.5075587)
axonD = line(fiberD, 0.88904883, -1.9104369)
nodeD = line(fiberD, 0.34490792, -0.14841106)
paraD1 = line(fiberD, 0.34490792, -0.14841106)
paraD2 = line(fiberD, 0.88904883, -1.9104369)
if not deltax:
deltax = logreg(fiberD, 3.79906687e+03, 2.13820902e+00, 2.48122018e-01, -2.19548067e+03)
nl = int(round(logfn(fiberD, 65.89739004, -32.66582976))*nlfactor)
Rpn0 = (self.rhoa * .01) / (np.pi * ((((nodeD / 2) + self.space_p1) ** 2) - ((nodeD / 2) ** 2)))
Rpn1 = (self.rhoa * .01) / (np.pi * ((((paraD1 / 2) + self.space_p1) ** 2) - ((paraD1 / 2) ** 2)))
Rpn2 = (self.rhoa * .01) / (np.pi * ((((paraD2 / 2) + self.space_p2) ** 2) - ((paraD2 / 2) ** 2)))
Rpx = (self.rhoa * .01) / (np.pi * ((((axonD / 2) + self.space_i) ** 2) - ((axonD / 2) ** 2)))
if branch == 'n':
# smallest MRG fiberD has paralength2 of 35
paralength2 = 35
interlength = (deltax - self.nodelength - (2 * self.paralength1) - (2 * paralength2)) / 6
if deltax == 85:
nodeD = fiberD*(5.0/7.3)
else:
paralength2 = logreg(fiberD, 30.77203038, 10.53182692, 0.42725082, 31.47653035)
interlength = (deltax - self.nodelength - (2 * self.paralength1) - (2 * paralength2)) / 6
return {'fiberD':fiberD,'g':g, 'axonD':axonD, 'nodeD':nodeD, 'paraD1':paraD1, 'paraD2':paraD2,'deltax':deltax,
'paralength2':paralength2,'nl':nl, 'Rpn0':Rpn0, 'Rpn1':Rpn1, 'Rpn2':Rpn2, 'Rpx':Rpx,'interlength':interlength}
def getModelParameters(self, fiberD, deltax, nlfactor, branch):
""" ALTERNATE PARAMETER SETTING - BEST IF USED FOR DISCRETE fiberD VALUES DEFINED IN MRG MODEL.
Obtain model parameters for discrete fiberD values from the MRG model. Generate interpolated model parameters
for non-MRG fiberD values. Interpolation functions obtained from Gaines et. al."""
if fiberD == 5.7:
g = 0.605
axonD = 3.4
nodeD = 1.9
paraD1 = 1.9
paraD2 = 3.4
if not deltax:
deltax = 500.0
paralength2 = 35.0
nl = 80.0*nlfactor
elif fiberD == 7.3:
g = 0.630
axonD = 4.6
nodeD = 2.4
paraD1 = 2.4
paraD2 = 4.6
if not deltax:
deltax = 750.0
paralength2 = 38.0
nl = 100.0*nlfactor
elif fiberD == 8.7:
g = 0.661
axonD = 5.8
nodeD = 2.8
paraD1 = 2.8
paraD2 = 5.8
if not deltax:
deltax = 1000.0
paralength2 = 40.0
nl = 110.0*nlfactor
elif fiberD == 10.0:
g = 0.690
axonD = 6.9
nodeD = 3.3
paraD1 = 3.3
paraD2 = 6.9
if not deltax:
deltax = 1150.0
paralength2 = 46.0
nl = 120.0*nlfactor
elif fiberD == 11.5:
g = 0.700
axonD = 8.1
nodeD = 3.7
paraD1 = 3.7
paraD2 = 8.1
if not deltax:
deltax = 1250.0
paralength2 = 50.0
nl = 130.0*nlfactor
elif fiberD == 12.8:
g = 0.719
axonD = 9.2
nodeD = 4.2
paraD1 = 4.2
paraD2 = 9.2
if not deltax:
deltax = 1350.0
paralength2 = 54.0
nl = 135.0*nlfactor
elif fiberD == 14.0:
g = 0.739
axonD = 10.4
nodeD = 4.7
paraD1 = 4.7
paraD2 = 10.4
if not deltax:
deltax = 1400.0
paralength2 = 56.0
nl = 140.0*nlfactor
elif fiberD == 15.0:
g = 0.767
axonD = 11.5
nodeD = 5.0
paraD1 = 5.0
paraD2 = 11.5
if not deltax:
deltax = 1450.0
paralength2 = 58.0
nl = 145.0*nlfactor
elif fiberD == 16.0:
g = 0.791
axonD = 12.7
nodeD = 5.5
paraD1 = 5.5
paraD2 = 12.7
if not deltax:
deltax = 1500.0
paralength2 = 60.0
nl = 150.0*nlfactor
else: # interpolation from Gaines et al
g = 0.0172 * fiberD + 0.5076
axonD = 0.889 * fiberD - 1.9104
nodeD = 0.3449 * fiberD - 0.1484
paraD1 = 0.3527 * fiberD - 0.1804
paraD2 = 0.889 * fiberD - 1.9104
if not deltax:
deltax = 969.3 * np.log(fiberD) - 1144.6
paralength2 = 2.5811 * fiberD + 19.59
nl = (65.897 * np.log(fiberD) - 32.666)*nlfactor
Rpn0 = (self.rhoa * .01) / (np.pi * ((((nodeD / 2) + self.space_p1) ** 2) - ((nodeD / 2) ** 2)))
Rpn1 = (self.rhoa * .01) / (np.pi * ((((paraD1 / 2) + self.space_p1) ** 2) - ((paraD1 / 2) ** 2)))
Rpn2 = (self.rhoa * .01) / (np.pi * ((((paraD2 / 2) + self.space_p2) ** 2) - ((paraD2 / 2) ** 2)))
Rpx = (self.rhoa * .01) / (np.pi * ((((axonD / 2) + self.space_i) ** 2) - ((axonD / 2) ** 2)))
if branch == 'n':
# smallest MRG fiberD has paralength2 of 35
interlength = (deltax - self.nodelength - (2 * self.paralength1) - (2 * 35)) / 6
else:
interlength = (deltax - self.nodelength - (2 * self.paralength1) - (2 * paralength2)) / 6
return {'fiberD':fiberD,'g':g, 'axonD':axonD, 'nodeD':nodeD, 'paraD1':paraD1, 'paraD2':paraD2,'deltax':deltax,
'paralength2':paralength2,'nl':nl, 'Rpn0':Rpn0, 'Rpn1':Rpn1, 'Rpn2':Rpn2, 'Rpx':Rpx,'interlength':interlength}
def getSectionFromDF(self, sType, branch=None, nodeIdx=None, returnObj=True):
"""
Parameters:
sType: str
string associated with section type: 'soma', 'iseg', 'tjnNode', 'node', 'MYSA', 'FLUT', OR 'STIN'
branch: str
string associated with the branch of the pseudounipolar neuron: 'n' for neck or stem axon, 'c' for central axon, 'p' for peripheral axon
nodeIdx: int
index of the node with which the section is associated
returnObj: boolean
True if the section object is desired
False if the relevant row of the dataframe containing information about the section is desired
"""
if isinstance(nodeIdx,int) and branch:
dfRow = self.sectionDF[(self.sectionDF['sectionType'] == sType) &
(self.sectionDF['puniBranch'] == branch) &
(self.sectionDF['nodeIndex'] == nodeIdx)]
elif not(nodeIdx and branch) and sType:
dfRow = self.sectionDF[self.sectionDF['sectionType'] == sType]
else:
raise ValueError('missing input argument for branch and/or node index')
if returnObj:
return dfRow['object'].item()
else:
return dfRow
def setXYZpos(self, femdict, tjnPos=(0, 0, 0), neckAngle=90):
"""Verify XYZ position of T-junction is inside the DRG and assign xyz coordinates of all sections
x > 0 --> peripheral branch
x < 0 --> central branch
Parameters
----------
femdict : dict
dictionary of FEM dataframes; IF ALL NEURON POSITIONS ARE ALREADY VERIFIED WITHIN THE DRG, SET femdict = None
tjnPos: tuple
(x,y,z) position of t-junction
neckAngle: float
angle of the stem axon with the z-axis in the y-z plane
"""
neckIdx = self.sectionDF['puniBranch'] == 'n'
if femdict is not None:
femdata = femdict['DRG']
stemLength = self.sectionDF[neckIdx]['L'].sum()
FEM_xMax = tjnPos[0]
FEM_xPosMask = tjnPos[0] == femdata['x']
if np.any(FEM_xPosMask): # coordinate exists in the fem
FEM_yMax = femdata[FEM_xPosMask]['y'].max()
FEM_zMax = femdata[FEM_xPosMask]['z'].max()
else: # cooridnate needs interpolation
# for coord, label in zip(tjnPos[0], 'x'):
diffList = np.array(femdata['x'].unique()) - FEM_xMax
x_lowerLim = FEM_xMax + np.max(diffList[diffList < 0])
x_upperLim = FEM_xMax + np.min(diffList[diffList > 0])
y_upperLim = femdata[femdata['x'] == x_upperLim]['y'].max()
y_lowerLim = femdata[femdata['x'] == x_lowerLim]['y'].max()
FEM_yMax = np.interp(FEM_xMax, [x_lowerLim, x_upperLim], [y_lowerLim, y_upperLim])
z_upperLim = femdata[femdata['x'] == x_upperLim]['z'].max()
z_lowerLim = femdata[femdata['x'] == x_lowerLim]['z'].max()
FEM_zMax = np.interp(FEM_xMax, [x_lowerLim, x_upperLim], [z_lowerLim, z_upperLim])
if FEM_yMax == FEM_zMax:
DRG_CSradius_X = FEM_yMax # or FEM_zMax/2
else:
DRG_CSradius_X = np.min([FEM_yMax, FEM_zMax])
punRadius = DRG_CSradius_X - stemLength # the tjn node can be anywhere inside this circle
else:
stemLength = 0
DRG_CSradius_X = np.inf
punRadius = np.inf # if femdata hasnt been provided all positions are valid
# check that the input position is valid
# an elegant way would be to create a Polygon object and find if the point is inside the polygon, this will work too
neckAngle = np.deg2rad(neckAngle)
xshift = tjnPos[0]
yshift = tjnPos[1]
zshift = tjnPos[2]
if (np.sqrt(zshift ** 2 + yshift ** 2) <= punRadius): # tjn lies inside annulus
self.sectionDF.loc[neckIdx, 'x'] = xshift
self.sectionDF.loc[neckIdx, 'y'] = yshift + (self.sectionDF[neckIdx].loc[::-1, 'L'].cumsum()[::-1] - self.sectionDF[neckIdx]['L'] / 2) * np.round(np.sin(neckAngle), 4)
self.sectionDF.loc[neckIdx, 'z'] = zshift + (self.sectionDF[neckIdx].loc[::-1, 'L'].cumsum()[::-1] - self.sectionDF[neckIdx]['L'] / 2) * np.round(np.cos(neckAngle), 4)
outsideDRGidx = (self.sectionDF.loc[neckIdx, 'z']**2 + self.sectionDF.loc[neckIdx, 'y']**2)**(1/2) > DRG_CSradius_X
if outsideDRGidx.any():
insideDRGidx = (self.sectionDF.loc[neckIdx, 'z'] ** 2 + self.sectionDF.loc[neckIdx, 'y'] ** 2)**(1/2) < DRG_CSradius_X
self.sectionDF.loc[neckIdx & outsideDRGidx, 'z'] = self.sectionDF.loc[neckIdx & insideDRGidx, 'z'].iloc[0]
self.sectionDF.loc[neckIdx & outsideDRGidx, 'y'] = self.sectionDF.loc[neckIdx & insideDRGidx, 'y'].iloc[0]
centralIdx = self.sectionDF['puniBranch'] == 'c'
self.sectionDF.loc[centralIdx, 'x'] = xshift - (self.sectionDF[centralIdx].loc[::, 'L'].cumsum() - self.sectionDF[centralIdx]['L'] / 2)
self.sectionDF.loc[centralIdx, 'y'] = yshift
self.sectionDF.loc[centralIdx, 'z'] = zshift
periIdx = self.sectionDF['puniBranch'] == 'p'
self.sectionDF.loc[periIdx, 'x'] = xshift + (self.sectionDF[periIdx].loc[::, 'L'].cumsum() - self.sectionDF[periIdx]['L'] / 2)
self.sectionDF.loc[periIdx, 'y'] = yshift
self.sectionDF.loc[periIdx, 'z'] = zshift
return 1
elif (stemLength*np.sin(neckAngle) < DRG_CSradius_X) and (stemLength*np.cos(neckAngle) < DRG_CSradius_X): #soma coordinates are inside drg
self.sectionDF.loc[neckIdx, 'x'] = xshift
self.sectionDF.loc[neckIdx, 'y'] = yshift + (self.sectionDF[neckIdx].loc[::-1, 'L'].cumsum()[::-1] - self.sectionDF[neckIdx]['L'] / 2) * np.round(np.sin(neckAngle), 4)
self.sectionDF.loc[neckIdx, 'z'] = zshift + (self.sectionDF[neckIdx].loc[::-1, 'L'].cumsum()[::-1] - self.sectionDF[neckIdx]['L'] / 2) * np.round(np.cos(neckAngle), 4)
outsideDRGidx = (self.sectionDF.loc[neckIdx, 'z']**2 + self.sectionDF.loc[neckIdx, 'y']**2)**(1/2) > DRG_CSradius_X
if outsideDRGidx.any():
insideDRGidx = (self.sectionDF.loc[neckIdx, 'z'] ** 2 + self.sectionDF.loc[neckIdx, 'y'] ** 2)**(1/2) < DRG_CSradius_X
self.sectionDF.loc[neckIdx & outsideDRGidx, 'z'] = self.sectionDF.loc[neckIdx & insideDRGidx, 'z'].iloc[0]
self.sectionDF.loc[neckIdx & outsideDRGidx, 'y'] = self.sectionDF.loc[neckIdx & insideDRGidx, 'y'].iloc[0]
if tjnPos[1] >= 750:
yTube = 600
elif tjnPos[1] <= -750:
yTube = -600
else:
yTube = yshift
if tjnPos[2] >= 750:
zTube = 600
elif tjnPos[2] <= -750:
zTube = -600
else:
zTube = zshift
centralTube = (-3200, yTube, zTube)
peripheralTube = (3200, yTube, zTube)
rho_c = np.sqrt((tjnPos[0]-centralTube[0])**2 + (tjnPos[1]-centralTube[1])**2 + (tjnPos[2]-centralTube[2])**2)
a_c = (tjnPos[0]-centralTube[0])/rho_c
b_c = (tjnPos[1]-centralTube[1])/rho_c
c_c = (tjnPos[2]-centralTube[2])/rho_c
if centralTube[2] == zshift:
gamma_c = np.pi/2
else:
gamma_c = np.abs(np.arctan((centralTube[0] - xshift) / (centralTube[2]-zshift)))
centralIdx = self.sectionDF['puniBranch'] == 'c'
self.sectionDF.loc[centralIdx, 'x'] = xshift - (self.sectionDF[centralIdx].loc[::, 'L'].cumsum() - self.sectionDF[centralIdx]['L'] / 2) * np.round(np.sin(gamma_c), 4)
self.sectionDF.loc[centralIdx, 'y'] = centralTube[1] + b_c/a_c*(self.sectionDF.loc[centralIdx, 'x'] - centralTube[0])
self.sectionDF.loc[centralIdx, 'z'] = centralTube[2] + c_c/a_c*(self.sectionDF.loc[centralIdx, 'x'] - centralTube[0])
tubeIdx_c = self.sectionDF[centralIdx]['x'] <= centralTube[0]
self.sectionDF.loc[(centralIdx & tubeIdx_c),'z'] = centralTube[2]
self.sectionDF.loc[(centralIdx & tubeIdx_c), 'y'] = centralTube[1]
rho_p = np.sqrt((tjnPos[0] - peripheralTube[0]) ** 2 + (tjnPos[1] - peripheralTube[1]) ** 2 + (tjnPos[2] - peripheralTube[2]) ** 2)
a_p = (tjnPos[0] - peripheralTube[0]) / rho_p
b_p = (tjnPos[1] - peripheralTube[1]) / rho_p
c_p = (tjnPos[2] - peripheralTube[2]) / rho_p
if centralTube[2] == zshift:
gamma_p = np.pi / 2
else:
gamma_p = np.abs(np.arctan((peripheralTube[0] - xshift) / (peripheralTube[2] - zshift)))
periIdx = self.sectionDF['puniBranch'] == 'p'
self.sectionDF.loc[periIdx, 'x'] = xshift + (self.sectionDF[periIdx].loc[::, 'L'].cumsum() - self.sectionDF[periIdx]['L'] / 2) * np.round(np.sin(gamma_p), 4)
self.sectionDF.loc[periIdx, 'y'] = peripheralTube[1] + b_p / a_p * (self.sectionDF.loc[periIdx, 'x'] - peripheralTube[0])
self.sectionDF.loc[periIdx, 'z'] = peripheralTube[2] + c_p / a_p * (self.sectionDF.loc[periIdx, 'x'] - peripheralTube[0])
tubeIdx_p = self.sectionDF[periIdx]['x'] >= peripheralTube[0]
self.sectionDF.loc[(periIdx & tubeIdx_p),'z'] = centralTube[2]
self.sectionDF.loc[(periIdx & tubeIdx_p), 'y'] = centralTube[1]
return 1
else:
return 0