Overview of models

loaded_spiketrains

class networkunit.models.loaded_spiketrains(name=None, backend='storage', attrs=None, **params)

Abstract model class for spiking data. It has an example loading routine for hdf files, is able to display the corresponding rasterplot with self.show_rasterplot(), and if the self.params contains: align_to_0=True, the spiketrains all start from 0s, max_subsamplesize=x, only the x first spike trains are used.

stochastic_activity

class networkunit.models.stochastic_activity(name=None, backend='storage', attrs=None, **params)

Model class which is able to generate stochastic spiking data

Parameters:

• size (int) – Number of spike trains

• t_start (quantity) – starting time

• t_stop (quantity) – ending time

• rate (quantity) – average firing rate

• statistic ('poisson', 'gamma'(to be implemented))

• correlation_method ('CPP', 'spatio-temporal', 'pairwise_equivalent', None) –

  CPP - compound Poisson process generating correlated activity of

  size ‘assembly_sizes’ with mean correlation according to ‘correlations’.

  spatio-temporal - generates CPP correlated groups and shifts the

  spike trains randomly within +- 0.5*’max_pattern_length’ to create spatio-temporal patterns.

  pairwise_equivalent - generates pairs of correlated spike trains

  so that the amount of correlation is equivalent to a correlated group with parameters ‘assembly_sizes’ and ‘correlations’

• expected_bin_size (quantity) – bin_size with which correlations are calculated to be able to generate the pairwise equivalent.

• correlations (float, list of floats) – Average correlation for the correlated group(s). Pass a list of floats if there are multiple groups with different correlations. If 0, it generates homogenous Poisson activity.

• assembly_sizes (list of ints) – Size(s) of correlated group(s). Empty list for no correlations.

• bkgr_correlation (float) – Background correlation (to be implemented)

• max_pattern_length (quantity) – Maximum pattern length for spatio-temporal patterns.

• shuffle (bool) – Shuffle the spike trains to separate the correlated groups

• shuffle_seed (int)

Example

nest_simuation

class networkunit.models.nest_simulation(name, nest_instance=None, attrs=None, model_params=None)