# Data about barrels
def define_barrels(premodel):
import numpy as np
Nbarrels = premodel['Nbarrel']
if Nbarrels != 1:
BarrelStruct = premodel['barrelstruct']
else:
BarrelStruct = np.array([premodel['barrelstruct']])
Barrels = {}
for nb in range(Nbarrels): ## one dimensional barrels (Nby = 1) - otherwise, it may change the access to BarrelStruct[nx][ny], for example, with proper loops on nx, ny
Barrels['Barrel'+str(nb+1)] = {
'Nthalamic': BarrelStruct[nb].Nthalamic,
'Xpos': BarrelStruct[nb].xpos,
'Ypos': BarrelStruct[nb].ypos
}
if BarrelStruct[nb].mainbarrel == 1:
Barrels['Barrel'+str(nb+1)]['Type'] = 'Principal'
elif BarrelStruct[nb].mainbarrel == 2:
Barrels['Barrel'+str(nb+1)]['Type'] = 'Surround'
else:
print('Barrel category not recognized')
Nthalamic = sum(Barrels['Barrel'+str(nb+1)]['Nthalamic'] for nb in range(Nbarrels))
return Nbarrels, Nthalamic, Barrels
# Data about cells and conns
def define_pops(model, settings):
vr = settings['vr']
v0 = settings['v0']
u0 = settings['u0']
dyn_thres = settings['dyn_thres']
Neuron_Para = model['ConData'].Neuron_Para
Neuron_Vt = model['ConData'].Neuron_Vt
Cellinfo_In = model['ConData'].Cellinfo_In
Cellinfo_All = model['ConData'].Cellinfo_All
NtIn = model['ConData'].NtIn
NtAll = model['ConData'].NtAll
Ncells = Cellinfo_All.shape[0]
Pops = {'1':{'Label':'Pyr_SP', 'Layer':'l4', 'Type':'exc'},
'2':{'Label':'Pyr_SS', 'Layer':'l4', 'Type':'exc'},
'3':{'Label':'Inh_FS', 'Layer':'l4', 'Type':'inh'},
'4':{'Label':'Inh_RSNP', 'Layer':'l4', 'Type':'inh'},
'5':{'Label':'Pyr', 'Layer':'l23', 'Type':'exc'},
'6':{'Label':'Inh_FSBS', 'Layer':'l23', 'Type':'inh'}, # PV
'7':{'Label':'Inh_FSCH', 'Layer':'l23', 'Type':'inh'}, # PV
'8':{'Label':'Inh_BSPV', 'Layer':'l23', 'Type':'inh'}, # PV
'9':{'Label':'Inh_Mar', 'Layer':'l23', 'Type':'inh'}, # SOM
'10':{'Label':'Inh_Bit', 'Layer':'l23', 'Type':'inh'}, # SOM
'11':{'Label':'Inh_DBC', 'Layer':'l23', 'Type':'inh'}, # VIP
'12':{'Label':'Inh_Bip', 'Layer':'l23', 'Type':'inh'}, # VIP
'13':{'Label':'Inh_Bip', 'Layer':'l23', 'Type':'inh'}, # CR
'14':{'Label':'Inh_SBC', 'Layer':'l23', 'Type':'inh'}, # CR
'15':{'Label':'Inh_NG', 'Layer':'l23', 'Type':'inh'}} # AC
Npops = len(Pops)
N = [[n for n in range(Ncells) if Cellinfo_All[n,3]==(ntype+1)] for ntype in range(Npops)]
for ntype in range(Npops):
Pops[str(ntype+1)]['ids'] = N[ntype] #ids starting from 0 - differs in 1 from rows in matlab
if len(N[ntype])==NtAll[ntype]:
Pops[str(ntype+1)]['Ncells'] = len(N[ntype])
else:
print('Check number of cells per population - Pop',ntype+1)
# reading position information, in NetPyNE paradigm (y-axis as depth)
Pops[str(ntype+1)]['x'] = Cellinfo_All[N[ntype],0]
Pops[str(ntype+1)]['y'] = Cellinfo_All[N[ntype],2] # here, y is the depth
Pops[str(ntype+1)]['z'] = Cellinfo_All[N[ntype],1]
# reading barrel
Pops[str(ntype+1)]['Barrel'] = Cellinfo_All[N[ntype],4]
# reading parameters related to individual intrinsic dynamics (Izhikevich model)
Pops[str(ntype+1)]['vr'] = vr
Pops[str(ntype+1)]['v0'] = v0
Pops[str(ntype+1)]['u0'] = u0
Pops[str(ntype+1)]['vt0'] = Neuron_Vt.avg[ntype] # NOT USED IN THE ADAPTATIVE THRESHOLD MODEL
Pops[str(ntype+1)]['a'] = Neuron_Para[N[ntype],0]
Pops[str(ntype+1)]['b'] = Neuron_Para[N[ntype],1]
Pops[str(ntype+1)]['c'] = Neuron_Para[N[ntype],2]
Pops[str(ntype+1)]['d'] = Neuron_Para[N[ntype],3]
if dyn_thres == 1:
Pops[str(ntype+1)]['alpha'] = Neuron_Vt.alpha[ntype]
Pops[str(ntype+1)]['Vi'] = Neuron_Vt.Vi[ntype]
Pops[str(ntype+1)]['Vmin'] = Neuron_Vt.Vmin[ntype]
Pops[str(ntype+1)]['Ka'] = Neuron_Vt.Ka[ntype]
Pops[str(ntype+1)]['Ki'] = Neuron_Vt.Ki[ntype]
Pops[str(ntype+1)]['tau'] = Neuron_Vt.tau[ntype]
# Setting-up list, based on above properties, for instantiate individual neurons in NetPyNE
Pops[str(ntype+1)]['list'] = [{'x': Pops[str(ntype+1)]['x'][n],
'y': Pops[str(ntype+1)]['y'][n],
'z': Pops[str(ntype+1)]['z'][n],
'params':{'vr': Pops[str(ntype+1)]['vr'],
'v0': Pops[str(ntype+1)]['v0'],
'u0': Pops[str(ntype+1)]['u0'],
'vt0': Pops[str(ntype+1)]['vt0'],
'dyn_th': dyn_thres,
'a': Pops[str(ntype+1)]['a'][n],
'b': Pops[str(ntype+1)]['b'][n],
'c': Pops[str(ntype+1)]['c'][n],
'd': Pops[str(ntype+1)]['d'][n]}
} for n in range(len(N[ntype]))]
# adding properties for dynamical threshold evolution
if dyn_thres == 1:
for n in range(len(N[ntype])):
Pops[str(ntype+1)]['list'][n]['params']['alpha'] = Pops[str(ntype+1)]['alpha']
Pops[str(ntype+1)]['list'][n]['params']['Vi'] = Pops[str(ntype+1)]['Vi']
Pops[str(ntype+1)]['list'][n]['params']['Vmin'] = Pops[str(ntype+1)]['Vmin']
Pops[str(ntype+1)]['list'][n]['params']['Ka'] = Pops[str(ntype+1)]['Ka']
Pops[str(ntype+1)]['list'][n]['params']['Ki'] = Pops[str(ntype+1)]['Ki']
Pops[str(ntype+1)]['list'][n]['params']['tau'] = Pops[str(ntype+1)]['tau']
return Ncells, Npops, N, Pops
def define_conns(model, settings):
from random import random
fr = settings['fr']
PMat_IntoAll = model['ConData'].PMat_IntoAll
PMat_AlltoAll = model['ConData'].PMat_AlltoAll
# Assembling conns - Connections List: [Pre, Post], gIDs are relative to the regions (thalamus or cortex). This is consistent to the connParam rule set by "connList"
# Thalamus -> Cortex (IntoAll)
Exc_ThtoAll = {}
Exc_ThtoAll['connList'] = [ [PMat_IntoAll.preCell[nc][1],PMat_IntoAll.preCell[nc][0]] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
Exc_ThtoAll['connList_gID'] = [ [PMat_IntoAll.preCell[nc][1]-1,PMat_IntoAll.preCell[nc][0]-1] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
# weights (nA or equiv. mA/cm2 if thought of as distributed)
Exc_ThtoAll['weightList'] = [ 0.001*PMat_IntoAll.Am[nc] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
Exc_ThtoAll['CVList'] = [ 0.001*PMat_IntoAll.CV[nc] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
#delays
Exc_ThtoAll['delayList'] = [ PMat_IntoAll.Delay[nc] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
# short-term learning & failure
Exc_ThtoAll['STDList'] = [ PMat_IntoAll.STD[nc] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
Exc_ThtoAll['FailList'] = [ PMat_IntoAll.Fail[nc] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
Exc_ThtoAll['PlasList'] = [ PMat_IntoAll.Plas[nc] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
# temporal characteristics
Exc_ThtoAll['Trise'] = [ PMat_IntoAll.Trise[PMat_IntoAll.preCell[nc][0]-1] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
Exc_ThtoAll['Tfall'] = [ PMat_IntoAll.Tfall[PMat_IntoAll.preCell[nc][0]-1] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
Exc_ThtoAll['CN'] = [ PMat_IntoAll.CN[PMat_IntoAll.preCell[nc][0]-1] for nc in range(len(PMat_IntoAll.preCell)) if PMat_IntoAll.preCell[nc][2]==1]
if len(Exc_ThtoAll['connList']) != len(PMat_IntoAll.preCell):
print("Please, consider also inhibitory projections from thalamus to cortex\n")
# Cortex -> Cortex (AlltoAll)
idx_e = []
idx_i = []
for nc in range(len(PMat_AlltoAll.preCell)):
if random() < fr:
if PMat_AlltoAll.preCell[nc][2]==1:
idx_e.append(nc)
elif PMat_AlltoAll.preCell[nc][2]==-1:
idx_i.append(nc)
else:
print("Connection without specification of exc/inh")
# List of excitatory connections
Exc_AlltoAll = {}
Exc_AlltoAll['connList'] = [ [PMat_AlltoAll.preCell[idx_e[nc]][1],PMat_AlltoAll.preCell[idx_e[nc]][0]] for nc in range(len(idx_e)) ]
Exc_AlltoAll['connList_gID'] = [ [PMat_AlltoAll.preCell[idx_e[nc]][1]-1,PMat_AlltoAll.preCell[idx_e[nc]][0]-1] for nc in range(len(idx_e)) ]
Exc_AlltoAll['weightList'] = [ 0.001*PMat_AlltoAll.Am[idx_e[nc]] for nc in range(len(idx_e)) ]
Exc_AlltoAll['CVList'] = [ 0.001*PMat_AlltoAll.CV[idx_e[nc]] for nc in range(len(idx_e)) ]
Exc_AlltoAll['delayList'] = [ PMat_AlltoAll.Delay[idx_e[nc]] for nc in range(len(idx_e)) ]
Exc_AlltoAll['STDList'] = [ PMat_AlltoAll.STD[idx_e[nc]] for nc in range(len(idx_e)) ]
Exc_AlltoAll['FailList'] = [ PMat_AlltoAll.Fail[idx_e[nc]] for nc in range(len(idx_e)) ]
Exc_AlltoAll['PlasList'] = [ PMat_AlltoAll.Plas[idx_e[nc]] for nc in range(len(idx_e)) ]
Exc_AlltoAll['Trise'] = [ PMat_AlltoAll.Trise[PMat_AlltoAll.preCell[idx_e[nc]][0]-1][PMat_AlltoAll.preID[idx_e[nc]][1]-1] for nc in range(len(idx_e)) ]
Exc_AlltoAll['Tfall'] = [ PMat_AlltoAll.Tfall[PMat_AlltoAll.preCell[idx_e[nc]][0]-1][PMat_AlltoAll.preID[idx_e[nc]][1]-1] for nc in range(len(idx_e)) ]
Exc_AlltoAll['CN'] = [ PMat_AlltoAll.CN[PMat_AlltoAll.preCell[idx_e[nc]][0]-1][PMat_AlltoAll.preID[idx_e[nc]][1]-1] for nc in range(len(idx_e)) ]
# List of inhibitory connections
Inh_AlltoAll = {}
Inh_AlltoAll['connList'] = [ [PMat_AlltoAll.preCell[idx_i[nc]][1],PMat_AlltoAll.preCell[idx_i[nc]][0]] for nc in range(len(idx_i)) ]
Inh_AlltoAll['connList_gID'] = [ [PMat_AlltoAll.preCell[idx_i[nc]][1]-1,PMat_AlltoAll.preCell[idx_i[nc]][0]-1] for nc in range(len(idx_i)) ]
Inh_AlltoAll['weightList'] = [ 0.001*PMat_AlltoAll.Am[idx_i[nc]] for nc in range(len(idx_i)) ]
Inh_AlltoAll['CVList'] = [ 0.001*PMat_AlltoAll.CV[idx_i[nc]] for nc in range(len(idx_i)) ]
Inh_AlltoAll['delayList'] = [ PMat_AlltoAll.Delay[idx_i[nc]] for nc in range(len(idx_i)) ]
Inh_AlltoAll['STDList'] = [ PMat_AlltoAll.STD[idx_i[nc]] for nc in range(len(idx_i)) ]
Inh_AlltoAll['FailList'] = [ PMat_AlltoAll.Fail[idx_i[nc]] for nc in range(len(idx_i)) ]
Inh_AlltoAll['PlasList'] = [ PMat_AlltoAll.Plas[idx_i[nc]] for nc in range(len(idx_i)) ]
Inh_AlltoAll['Trise'] = [ PMat_AlltoAll.Trise[PMat_AlltoAll.preCell[idx_i[nc]][0]-1][PMat_AlltoAll.preID[idx_i[nc]][1]-1] for nc in range(len(idx_i)) ]
Inh_AlltoAll['Tfall'] = [ PMat_AlltoAll.Tfall[PMat_AlltoAll.preCell[idx_i[nc]][0]-1][PMat_AlltoAll.preID[idx_i[nc]][1]-1] for nc in range(len(idx_i)) ]
Inh_AlltoAll['CN'] = [ PMat_AlltoAll.CN[PMat_AlltoAll.preCell[idx_i[nc]][0]-1][PMat_AlltoAll.preID[idx_i[nc]][1]-1] for nc in range(len(idx_i)) ]
return Exc_ThtoAll, Exc_AlltoAll, Inh_AlltoAll
