/
netParams.py
244 lines (201 loc) · 9.34 KB
/
netParams.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
'''
NetPyNE version of Potjans and Diesmann thalamocortical network
netParams.py -- contains the network parameters (netParams object)
'''
from netpyne import specs
import numpy as np
from cfg import cfg
############################################################
#
# NETWORK PARAMETERS
#
############################################################
# Reescaling function (move to separate module?)
def Reescale(ScaleFactor, C, N_Full, w_p, f_ext, tau_syn, Inp, InpDC):
if ScaleFactor<1.0:
# This is a good approximation of the F_out param for the Balanced option "True".
# Note for the Balanced=False option, it should be possible to calculate a better approximation.
F_out=np.array([0.860, 2.600, 4.306, 5.396, 8.142, 8.188, 0.941, 7.3])
Ncon=np.vstack(np.column_stack(0 for i in range(0,8)) for i in range(0,8))
for r in range(0,8):
for c in range(0,8):
Ncon[r][c]=(np.log(1.-C[r][c])/np.log(1. -1./(N_Full[r]*N_Full[c])) ) /N_Full[r]
w=w_p*np.column_stack(np.vstack( [1.0, -4.0] for i in range(0,8)) for i in range(0,4))
w[0][2]=2.0*w[0][2]
x1_all = w * Ncon * F_out
x1_sum = np.sum(x1_all, axis=1)
x1_ext = w_p * Inp * f_ext
I_ext=np.column_stack(0 for i in range(0,8))
I_ext = 0.001 * tau_syn * (
(1. - np.sqrt(ScaleFactor)) * x1_sum +
(1. - np.sqrt(ScaleFactor)) * x1_ext)
InpDC=np.sqrt(ScaleFactor)*InpDC*w_p*f_ext*tau_syn*0.001 #pA
w_p=w_p/np.sqrt(ScaleFactor) #pA
InpDC=InpDC+I_ext
N_=[int(ScaleFactor*N) for N in N_Full]
else:
InpDC=InpDC*w_p*f_ext*tau_syn*0.001
N_=N_Full
return InpDC, N_, w_p
###########################################################
# Network Constants
###########################################################
# Frequency of external input
f_ext=8.0 # (Hz)
# Postsynaptic current time constant
tau_syn=0.5 # (ms)
# Membrane time constant
tau_m=10 # (ms)
# Other constants defined inside .mod
'''
#Absolute refractory period
tauref =2 (ms)
#Reset potential
Vreset = -65 (mV) : -49 (mV) :
#Fixed firing threshold
Vteta = -50 (mV)'''
# Membrane capacity
C_m=250 #pF
# Mean amplitude of the postsynaptic potential (in mV).
w_v=0.15
# Mean amplitude of the postsynaptic potential (in pA).
w_p= (((C_m) ** (-1) * tau_m * tau_syn / (
tau_syn - tau_m) * ((tau_m / tau_syn) ** (
- tau_m / (tau_m - tau_syn)) - (tau_m / tau_syn) ** (
- tau_syn / (tau_m - tau_syn)))) ** (-1))
w_p=w_v*w_p #(pA)
#C probability of connection
C=np.array([[0.1009, 0.1689, 0.0437, 0.0818, 0.0323, 0., 0.0076, 0.],
[0.1346, 0.1371, 0.0316, 0.0515, 0.0755, 0., 0.0042, 0.],
[0.0077, 0.0059, 0.0497, 0.135, 0.0067, 0.0003, 0.0453, 0.],
[0.0691, 0.0029, 0.0794, 0.1597, 0.0033, 0., 0.1057, 0.],
[0.1004, 0.0622, 0.0505, 0.0057, 0.0831, 0.3726, 0.0204, 0.],
[0.0548, 0.0269, 0.0257, 0.0022, 0.06, 0.3158, 0.0086, 0.],
[0.0156, 0.0066, 0.0211, 0.0166, 0.0572, 0.0197, 0.0396, 0.2252],
[0.0364, 0.001, 0.0034, 0.0005, 0.0277, 0.008, 0.0658, 0.1443]])
#Population size N
L=['L2e', 'L2i', 'L4e', 'L4i', 'L5e', 'L5i', 'L6e', 'L6i']
N_Full=np.array([20683, 5834, 21915, 5479, 4850, 1065, 14395, 2948, 902])
# Number of Input per Layer
Inp=np.array([1600, 1500, 2100, 1900, 2000, 1900, 2900, 2100])
if cfg.Balanced == False:
InpUnb=np.array([2000, 1850, 2000, 1850, 2000, 1850, 2000, 1850])
###########################################################
# Reescaling calculation
###########################################################
if cfg.DC == True:
InpDC=Inp
if cfg.Balanced== False:
InpDC=InpUnb
else:
InpDC=np.zeros(8)
InpPoiss=Inp*cfg.ScaleFactor
if cfg.Balanced== False:
InpPoiss=InpUnb*cfg.ScaleFactor
InpDC, N_, w_p = Reescale(cfg.ScaleFactor, C, N_Full, w_p, f_ext, tau_syn, Inp, InpDC)
############################################################
# NetPyNE Network Parameters (netParams)
############################################################
netParams = specs.NetParams() # object of class NetParams to store the network parameters
netParams.delayMin_e = 1.5
netParams.ddelay = 0.5
netParams.delayMin_i = 0.75
netParams.weightMin = w_p
netParams.dweight = 0.1
############################################################
# Populations parameters
############################################################
# population locations
# from Schmidt et al 2018, PLoS Comp Bio, Macaque V1
netParams.sizeX = 300 # x-dimension (horizontal length) size in um
netParams.sizeY = 1470 # y-dimension (vertical height or cortical depth) size in um
netParams.sizeZ = 300 # z-dimension (horizontal depth) size in um
netParams.shape = 'cylinder' # cylindrical (column-like) volume
popDepths = [[0.08, 0.27], [0.08, 0.27], [0.27, 0.58], [0.27, 0.58], [0.58, 0.73], [0.58, 0.73], [0.73, 1.0], [0.73, 1.0]]
# create populations
for i in range(0,8):
netParams.popParams[L[i]] = {'cellType': str(L[i]), 'numCells': int(N_[i]), 'cellModel': 'IntFire_PD', 'm':0, 'Iext':float(InpDC[i]), 'ynormRange': popDepths[i] }
# To atualization of Point Neurons
netParams.popParams['bkg_IF'] = {'numCells': 1, 'cellModel': 'NetStim','rate': 40000, 'start':0.0, 'noise': 0.0, 'delay':0}
############################################################
# External input parameters
############################################################
if cfg.DC == False: # External Input as Poisson
for r in range(0,8):
netParams.popParams['poiss'+str(L[r])] = {
'numCells': N_[r],
'cellModel': 'NetStim',
'rate': InpPoiss[r]*f_ext,
'start': 0.0,
'noise': 1.0,
'delay': 0}
auxConn=np.array([range(0,N_[r],1),range(0,N_[r],1)])
netParams.connParams['poiss->'+str(L[r])] = {
'preConds': {'pop': 'poiss'+str(L[r])},
'postConds': {'pop': L[r]},
'connList': auxConn.T,
'weight': 'max(0, weightMin+normal(0,dweight*weightMin))',
'delay': 0.5} # 1 delay
# Thalamus Input: increased of 15Hz that lasts 10 ms
# 0.15 fires in 10 ms each 902 cells -> number of spikes = T*f*N_ = 0.15*902 -> 1 spike each N_*0.15
if cfg.TH == True:
fth=15 #Hz
Tth=10 #ms
InTH=[0, 0, 93, 84, 0, 0, 47, 34]
for r in [2,3,6,7]:
nTH=int(np.sqrt(cfg.ScaleFactor)*InTH[r]*fth*Tth/1000)
netParams.popParams['bkg_TH'+str(L[r])] = {'numCells': N_[r], 'cellModel': 'NetStim','rate': 2*(1000*nTH)/Tth , 'start': 200.0, 'noise': 1.0, 'number': nTH, 'delay':0}
auxConn=np.array([range(0,N_[r],1),range(0,N_[r],1)])
netParams.connParams['bkg_TH->'+str(L[r])] = {
'preConds': {'pop': 'bkg_TH'+str(L[r])},
'postConds': {'pop': L[r]},
'connList': auxConn.T,
'weight':'max(0, weightMin +normal(0,dweight*weightMin))',
'delay': 0.5} # 1 delay
############################################################
# Connectivity parameters
############################################################
for r in range(0,8):
for c in range(0,8):
if (c % 2) == 0:
if c == 2 and r == 0:
netParams.connParams[str(L[c])+'->'+str(L[r])] = {
'preConds': {'pop': L[c]}, # conditions of presyn cells
'postConds': {'pop': L[r]}, # conditions of postsyn cells
'divergence': cfg.ScaleFactor*(np.log(1.-C[r][c])/np.log(1. -1./(N_Full[r]*N_Full[c])) ) /N_Full[c],
'weight':'2*max(0, weightMin +normal(0,dweight*weightMin))', # synaptic weight
'delay':'max(0.1, delayMin_e +normal(0,ddelay*delayMin_e))', # transmission delay (ms)
}
else:
netParams.connParams[str(L[c])+'->'+str(L[r])] = {
'preConds': {'pop': L[c]}, # conditions of presyn cells
'postConds': {'pop': L[r]}, # conditions of postsyn cells
'divergence': cfg.ScaleFactor*(np.log(1.-C[r][c])/np.log(1. -1./(N_Full[r]*N_Full[c])) ) /N_Full[c],
'weight':'max(0, weightMin +normal(0,dweight*weightMin))', # synaptic weight
'delay':'max(0.1, delayMin_e +normal(0,ddelay*delayMin_e))', # transmission delay (ms)
} # synaptic mechanism
else:
netParams.connParams[str(L[c])+'->'+str(L[r])] = {
'preConds': {'pop': L[c]}, # conditions of presyn cells
'postConds': {'pop': L[r]}, # conditions of postsyn cells
'divergence': cfg.ScaleFactor*(np.log(1.-C[r][c])/np.log(1. -1./(N_Full[r]*N_Full[c])) ) /N_Full[c],
'weight':'-4*max(0, weightMin +normal(0,dweight*weightMin))', # synaptic weight
'delay':'max(0.1, delayMin_i +normal(0,ddelay*delayMin_i))', # transmission delay (ms)
} # synaptic mechanism
netParams.connParams['S2->M'] = {
'preConds': {'pop': 'bkg_IF'},
'postConds': {'cellModel': 'IntFire_PD'},
'probability': 1,
'weight': 0,
'delay': 0.5}
############################################################
# Update cfg plotting options based on network rescaling
############################################################
# raster 10% of cells
scale = 10 #max(1,int(41.444*cfg.ScaleFactor))
include = [(pop, list(range(0, netParams.popParams[pop]['numCells'], scale))) for pop in L]
cfg.analysis['plotRaster']['include'] = include
# plot statistics for 10% of cells
scale = 10 #max(1,int(sum(N_[:8])/1000))
include = [(pop, range(0, netParams.popParams[pop]['numCells'], scale)) for pop in L]
cfg.analysis['plotSpikeStats']['include'] = include