params.py

import os

"""
settings for simulation environment
"""
# PATH = os.getcwd() + "/data/hetero/unimodal/intra/"
PATH = os.getcwd() + "/data/"
# PATH = os.environ["TMPDIR"] + "/"
num_process = 3  # how many cores?
num_threads = 2  # how many process on each core?
timeprint = False  # print the simulation progress live?
endspiketime = 10000.0  # spike detector stops after this time
short_threshold = 99999  # any time longer than this will start recording voltages, anything shorter starts spikes

"""
neuron parameter
"""
C_m = 250.0  # capacitance (pF)
E_L = -70.0  # leak reversal potential (mV)
# tau_m = 15.0  # membrane time constant (ms) = C_m / g_L
V_reset = - 60.0  # reset potential
tau_ref = 2.0  # absolute refractory period
g_L = 16.7  # peak conductance (nS)
V_th = -50.0  # threshold (mV)
neuron_model = "iaf_cond_exp"

"""
synapse parameter
"""
tau_E = 5.0  # syn. decay time constant exci.
tau_I = 10.0  # syn. decay time constant inhi.
V_revr_E = 0.0  # exci. reversal potential
V_revr_I = -80.0  # inhi. reversal potential

"""
connection parameter
"""
d = 1.5  # synaptic transmission delay
gbar_E = 1.0  # exci. syn. cond.
gamma = 16  # scaling factor for inhi. syn. cond.
gbar_I = -1 * gamma * gbar_E  # inhi. syn. cond.
epsilon = 0.1  # internal recurrent rand. conn. prob., sparse connection
p_x_0 = epsilon  # background connection prob. for the input module
p_x_rest = 0.25 * epsilon  # same for other modules
p_ff = 0.75 * epsilon  # feed-forward intra connection prob.

"""
network parameter
"""
N = 10000  # total num. of neurons in each module
N_E = int(0.8 * N)  # exci.
N_I = int(0.2 * N)  # inhi.
N_E_speci = int(epsilon * N_E)  # num. of stimulus specific exci. neurons in each module, not needed for random network
N_I_speci = int(epsilon * N_I)  # inhi.
# K_E = int(epsilon * N_E) # exci.pre synapse number for each neuron
# K_I = int(epsilon * N_I) # inhi.
module_depth = 4

"""
background noise input
"""
v_x = 5.0  # intensity of a Poisson process
N_X = N_E  # num. of background neurons
K_x_0 = int(p_x_0 * N_X)  # num. of synapses for background noise for the input module
K_x_rest = int(p_x_rest * N_X)

"""
stimuli input
"""
num_stimulus = 10  # num. of stimuli
# delta = 3  # Poisson process rate factor for stimuli
v_stim = 4.0  # define v_stim independently of delta
t_onset = 1.0  # when the stimuli starts
t_asterisk = 200.0  # how long each Poisson process lasts
input_spike_len = 800  # record from this much of neurons to use for classifier