#!/usr/bin/python
from neuron import h
from neuron import gui
import numpy as np
import os
import pickle
import cPickle
from scipy.stats.stats import pearsonr
import sys
sys.path.append(os.getcwd())
import itertools
print ('Setting cell model...')
neuron_num = 9
synaptic_map_file_name = 'synapse_location_for_neuron_{}.pickle'.format(neuron_num)
rate_E = 1.75
rate_I = 10
h.load_file('model_files/{}.hoc'.format(neuron_num))
h.v_init = -80
h.dt = 0.025
def set_cell(neuron_num):
h.load_file('model_files/{}.hoc'.format(neuron_num))
h('''
objref somatic, apical, axonal, basal
somatic = new SectionList()
basal = new SectionList()
apical = new SectionList()
axonal = new SectionList()
celsius=37
proc distribute(){local x localobj sl
strdef stmp,distfunc,mech
sl = $o1
mech = $s2
distfunc = $s3
sprint(distfunc,"%%s %s(%%f) = %s",mech, distfunc)
forsec sl for(x) {
sprint (stmp,distfunc,secname(),x,distance(x))
execute(stmp)
}
}
create axon[2]''')
h('''
access axon[0]
{
L = 30
diam = 0.66
nseg = 30
all.append()
axonal.append()
}
access axon[1]
{
L = 30
diam = 0.33
nseg = 30
all.append()
axonal.append()
}
nSecAxonal = 2
soma connect axon[0](0), 1
axon[0] connect axon[1](0), 1
''')
h("celsius = 37.0")
for sec in h.all:
if sec.name().startswith('soma') or sec.name().startswith('axon'): continue
h("{}.nseg = {}".format(sec.name(), max(eval('h.{}.L'.format(sec.name())), 1)))
for sec in h.all:
if sec.name().startswith('dend_5') or sec.name().startswith('apic'):
h('access {}'.format(sec.name()))
h('apical.append()')
h.pop_section()
elif sec.name().startswith('dend_7'):
h('access {}'.format(sec.name()))
h('basal.append()')
h.pop_section()
elif 'soma' in sec.name() or sec.name().startswith('dend_0'):
h('access {}'.format(sec.name()))
h('somatic.append()')
h.pop_section()
elif 'axon' in sec.name():
h('access {}'.format(sec.name()))
h('axonal.append()')
h.pop_section()
else:
print 'no section list for {}'.format(sec.name())
h('''
forsec all {
insert pas
}
forsec all {
e_pas = -75
}
forsec all {
Ra = 150
}
forsec all {
cm = 1
}
forsec all {
g_pas = 1.0 / 12000.0
}
forsec apical {
cm = 1
}
forsec basal {
cm = 1
}
''')
def active_soma():
h('''
forsec somatic {
insert Ca_HVA
insert SKv3_1
insert SK_E2
insert Ca_LVAst
insert Ih
insert NaTs2_t
insert CaDynamics_E2
}
forsec axonal {
insert Ca_HVA
insert SKv3_1
insert SK_E2
insert CaDynamics_E2
insert Nap_Et2
insert K_Pst
insert K_Tst
insert Ca_LVAst
insert NaTa_t
}
forsec axonal {
ena = 50
ek = -85
}
forsec somatic {
ena = 50
}
forsec somatic {
ek = -85
}
access soma
distance()
distribute(axonal,"gNaTa_tbar_NaTa_t","(0.0 * %g + 1.0)*3.429725")
distribute(axonal,"gK_Tstbar_K_Tst","(0.0 * %g + 1.0)*0.001035")
distribute(axonal,"gamma_CaDynamics_E2","(0.0 * %g + 1.0)*0.016713")
distribute(axonal,"gNap_Et2bar_Nap_Et2","(0.0 * %g + 1.0)*0.009803")
distribute(axonal,"gSK_E2bar_SK_E2","(0.0 * %g + 1.0)*0.008085")
distribute(axonal,"gCa_HVAbar_Ca_HVA","(0.0 * %g + 1.0)*0.000306")
distribute(axonal,"gK_Pstbar_K_Pst","(0.0 * %g + 1.0)*0.959296")
distribute(axonal,"gSKv3_1bar_SKv3_1","(0.0 * %g + 1.0)*0.094971")
distribute(axonal,"decay_CaDynamics_E2","(0.0 * %g + 1.0)*384.114655")
distribute(axonal,"gCa_LVAstbar_Ca_LVAst","(0.0 * %g + 1.0)*0.000050")
distribute(somatic,"gamma_CaDynamics_E2","(0.0 * %g + 1.0)*0.000533")
distribute(somatic,"gSKv3_1bar_SKv3_1","(0.0 * %g + 1.0)*0.102517")
distribute(somatic,"gSK_E2bar_SK_E2","(0.0 * %g + 1.0)*0.099433")
distribute(somatic,"gCa_HVAbar_Ca_HVA","(0.0 * %g + 1.0)*0.000374")
distribute(somatic,"gNaTs2_tbar_NaTs2_t","(0.0 * %g + 1.0)*0.926705")
distribute(somatic,"gIhbar_Ih","(0.0 * %g + 1.0)*0.000080")
distribute(somatic,"decay_CaDynamics_E2","(0.0 * %g + 1.0)*342.544232")
distribute(somatic,"gCa_LVAstbar_Ca_LVAst","(0.0 * %g + 1.0)*0.000778")
''')
def active_dendrites_tuned():
h('''
forsec basal {
insert Ih
}
forsec apical {
insert Im
insert NaTs2_t
insert SKv3_1
insert Ih
}
forsec apical {
ena = 50
ek = -85
}
access soma
distance()
distribute(basal,"gIhbar_Ih","(0.0 * %g + 1.0)*0.000080")
distribute(apical,"gNaTs2_tbar_NaTs2_t","(0.0 * %g + 1.0)*0.008009") // based on ttx comparison
// distribute(apical,"gNaTs2_tbar_NaTs2_t","(0.0 * %g + 1.0)*0.012009") // original
distribute(apical,"gSKv3_1bar_SKv3_1","(0.0 * %g + 1.0)*0.000513")
distribute(apical,"gIhbar_Ih","(-0.869600 + 2.087000*exp((%g-0.000000)*0.020100))*0.000080") // based on attenuation comparison
// distribute(apical,"gIhbar_Ih","(-0.869600 + 2.087000*exp((%g-0.000000)*0.003100))*0.000080") // original
distribute(apical,"gImbar_Im","(0.0 * %g + 1.0)*0.000740")
''')
set_cell(neuron_num)
active_soma()
active_dendrites_tuned()
print ('Setting parameters...')
duration_of_simulation = 1000
number_of_inhibitory_synapses_on_soma = 100
(spine_locations, full_IS_locations) = pickle.load(open('model_files/{}'.format(synaptic_map_file_name), 'rb'))
full_IS_locations = list(full_IS_locations)
for loc in np.linspace(0, 1, number_of_inhibitory_synapses_on_soma):
full_IS_locations.append(('soma', loc, 'soma',0,0))
h.tstop = duration_of_simulation
number_of_E_synapses = len(spine_locations)
number_of_I_synapses = len(full_IS_locations)
print ('Set recording...')
voltage_vectors = []
print ('Placing synapses...')
background_E_events = np.random.poisson(rate_E / 1000.0, size=(number_of_E_synapses, duration_of_simulation))
background_I_events = np.random.poisson(rate_I / 1000.0, size=(number_of_I_synapses, duration_of_simulation))
eSynlist = []
eNetconlist = []
iSynlist = []
iNetconlist = []
E_vcs = []
I_vcs = []
E_vcs_events = []
I_vcs_events = []
def placeNMDA(location, conductance):
eSynlist.append(h.ProbAMPANMDA2_RATIO(float(location)))
eSynlist[-1].gmax = conductance
eSynlist[-1].mgVoltageCoeff = 0.08
eNetconlist.append(h.NetCon(E_vcs[-1], eSynlist[-1]))
eNetconlist[-1].weight[0] = 1
eNetconlist[-1].delay = 0
def placeGABA(location, conductance):
iSynlist.append(h.ProbUDFsyn2_lark(float(location)))
iSynlist[-1].tau_r = 0.18
iSynlist[-1].tau_d = 5
iSynlist[-1].e = - 80
iSynlist[-1].gmax = conductance
iNetconlist.append(h.NetCon(I_vcs[-1], iSynlist[-1]))
iNetconlist[-1].weight[0] = 1
iNetconlist[-1].delay = 0
for sec in h.all:
if sec.name().startswith('soma') or sec.name().startswith('axon'): continue
h("{}.nseg = {}".format(sec.name(), max(eval('h.{}.L'.format(sec.name())), 1)))
for sec in h.all:
voltage_vectors.append(h.Vector())
voltage_vectors[-1].record(eval('h.{}({})._ref_v'.format(sec.name(), 0.5)))
spine_ind = 0
spine_firing_times = []
for sec, loc, branch_type, spine_head_width, spine_volume, spine_length in spine_locations:
h("access {}".format(sec))
h("nseg = {}".format(max(h.L, 1)))
E_vcs.append(h.VecStim())
E_vcs_events.append(h.Vector())
spine_firing_times.append([])
events = np.where(background_E_events[spine_ind, :].flatten())[0] + 100
for event in events:
E_vcs_events[-1].append(event)
spine_firing_times[-1].append(event)
E_vcs[-1].play(E_vcs_events[-1])
placeNMDA(loc, 0.0004)
spine_ind += 1
h.pop_section()
IS_ind = 0
IS_firing_times = []
for sec, loc, branch_type, IS_volume, on_spine in full_IS_locations:
h("access {}".format(sec))
h("nseg = {}".format(max(h.L, 1)))
I_vcs.append(h.VecStim())
I_vcs_events.append(h.Vector())
IS_firing_times.append([])
events = np.where(background_I_events[IS_ind, :].flatten())[0] + 100
for event in events:
I_vcs_events[-1].append(event)
IS_firing_times[-1].append(event)
I_vcs[-1].play(I_vcs_events[-1])
h("access {}".format(sec))
placeGABA(loc, 0.001)
h.pop_section()
IS_ind += 1
firing_times = np.zeros(shape=(int(number_of_E_synapses), int(h.tstop) + 100))
for ind in range(len(spine_firing_times)):
firing_times[ind, spine_firing_times[ind]] = 1
print ('Running...')
soma_voltageVector = h.Vector()
soma_voltageVector.record(h.soma(0.5)._ref_v)
timeVector = h.Vector()
timeVector.record(h._ref_t)
E_densities = []
I_densities = []
h.stdinit()
h.run()
voltage_vectors = np.array(voltage_vectors)
timeVector = np.array(timeVector)