Functional balanced networks with synaptic plasticity (Sadeh et al, 2015)

 Download zip file 
Help downloading and running models
Accession:182784
The model investigates the impact of learning on functional sensory networks. It uses large-scale recurrent networks of excitatory and inhibitory spiking neurons equipped with synaptic plasticity. It explains enhancement of orientation selectivity and emergence of feature-specific connectivity in visual cortex of rodents during development, as reported in experiments.
Reference:
1 . Sadeh S, Clopath C, Rotter S (2015) Emergence of Functional Specificity in Balanced Networks with Synaptic Plasticity. PLoS Comput Biol 11:e1004307 [PubMed]
Citations  Citation Browser
Model Information (Click on a link to find other models with that property)
Model Type: Realistic Network;
Brain Region(s)/Organism:
Cell Type(s): Abstract integrate-and-fire leaky neuron;
Channel(s):
Gap Junctions:
Receptor(s):
Gene(s):
Transmitter(s): Gaba; Glutamate;
Simulation Environment: Python;
Model Concept(s): Synaptic Plasticity; Long-term Synaptic Plasticity; Learning; Sensory processing; Homeostasis; Orientation selectivity; Vision;
Implementer(s): Sadeh, Sadra [s.sadeh at ucl.ac.uk];
Search NeuronDB for information about:  Gaba; Glutamate;
##################################################################################
# params.py -- Default set of parameters in (Table 1 of):
#
# Ref: Sadeh, Clopath and Rotter (PLOS Computational Biology, 2015).
# Emergence of Functional Specificity in Balanced Networks with Synaptic Plasticity.
#
# Author: Sadra Sadeh <s.sadeh@ucl.ac.uk> // Created: 2014-2015
##################################################################################

import numpy as np
import pylab as pl

##############################
###### ---- Default parameters

# -- stimulation params
block_no = 40 # number of batches for learning
stim_no = 20. # number of stimuli with different orientations in each batch

trial_time = .1 * 1000. # presentation time of each orientation (ms)
sim_time = stim_no * trial_time # total simulation time for each batch (ms)
dt = 1.  # time resolution (ms)
x_len = sim_time/dt
t = np.arange(0, sim_time, dt)

# -- network params
n = 500 # total number of neurons
f = .8 # fraction of excitatory neurons
ne = int(f*n) # number of exc neurons
ni = n - ne   # number of inh neurons

# connection probability
eps_ee = .3 # exc to exc
eps_ei = .3 # exc to inh
eps_ie = 1. # inh to exc
eps_ii = 1. # inh to inh

J = .5 # EPSP (mV)
g = 8. # inhibition dominance ratio (IPSP = -g EPSP)

# -- single neuron (Leaky Integrate-and-Fire) params
tm = 20. # membrane time constant (ms)
vth = 20. # threshold voltage (mV)

# coefficient matrix of the subthreshold dynamics (needed for Exact Integration)
A = -1./np.concatenate((tm*np.ones(ne) , tm*np.ones(ni))) 
#B = np.exp(A*dt)

# -- feedforward input
b_rate = 2000. # aggregare baseline rate of the ffw input to a neuron
m = .2 # modulation ratio of the input
m_exc = m # exc ffw modulation ratio
m_inh = m/10. # inh ffw modulation ratio

# initial preferred orientations
po_init = np.concatenate(( np.arange(0, np.pi, np.pi/ne) , np.arange(0, np.pi, np.pi/ni) ))
# stimulus orientation (before and after plasticity)
th = np.pi/2

# -- plst param
tm_plst = 10. # tau - (ms)
tp_plst = 7.  # tau + (ms)
tx_plst = 15. # tau x (ms)

Bm_plst =  np.exp((-1./tm_plst) *dt)  
Bp_plst =  np.exp((-1./tp_plst) *dt) 
Bx_plst =  np.exp((-1./tx_plst) *dt) 

w_min = 0. # lower bound on all weights
w_max = 2. # upper bound on exc weights
w_max_inh = -5. # lower bound on inh weights

A_ltd = 14.e-5 # (mV-1)
A_ltp = 8.e-5  # (mV-2)

vth_m = -20. # (mV)
vth_p = 7.5 # (mV)

# -----