Intermediate usage example

Here we show how to use NNMT to calculate the power spectra of the Potjans and Diesmann [2012] microcircuit model.

import numpy as np
import matplotlib.pyplot as plt
import nnmt

First we try loading previous results, which makes running this script for the second time much faster. If there are no stored results, we instantiate a microcircuit model.

try:
    network = nnmt.models.Network(file='microcircuit.h5')
except IOError:
    network = nnmt.models.Microcircuit('network_params.yaml',
                                       'analysis_params.yaml')

We compute the working point, which only requires the network parameters defined in network_params.yaml (see minimal example as well).

print('Compute working point')
nnmt.lif.exp.working_point(network)

Then we continue by computing the transfer function, the delay distribution matrix, the effective connectivity, and finally the power spectrum. This requires the definition of the analysis frequencies in analysis_params.yaml.

print('Compute transfer function')
nnmt.lif.exp.transfer_function(network)
print('Compute delay_dist_matrix')
nnmt.network_properties.delay_dist_matrix(network)
print('Compute effective connectivity')
nnmt.lif.exp.effective_connectivity(network)
print('Compute power spectra')
nnmt.lif.exp.power_spectra(network)

We retrieve the analysis frequencies and power spectra and plot the results

freqs = network.analysis_params['omegas'] / 2 / np.pi
spectra = network.results['lif.exp.power_spectra'].T

for i in range(8):
    plt.plot(freqs, spectra[i], label=network.network_params['populations'][i])
plt.yscale('log')
plt.xlabel('frequency (1/s)')
plt.ylabel('power')
plt.legend()
plt.show()

# save results to h5 file
network.save('microcircuit.h5')