miv_simulator.network

Network initialization routines.

Functions

connect_cell_selection(env)	Loads NeuroH5 connectivity file, instantiates the corresponding synapse and network connection mechanisms for the selected postsynaptic cells.
connect_cells(env)	Loads NeuroH5 connectivity file, instantiates the corresponding synapse and network connection mechanisms for each postsynaptic cell.
connect_gjs(env)	Loads NeuroH5 connectivity file, instantiates the corresponding half-gap mechanisms on the pre- and post-junction cells.
find_gid_pop(celltypes, gid)	Given a celltypes structure and a gid, find the population to which the gid belongs.
init(env)	Initializes the network by calling make_cells, init_input_cells, connect_cells, connect_gjs.
init_input_cells(env)	Initializes cells with predefined spike patterns.
ld_bal(env)	For given cxvec on each rank, calculates the fractional load balance.
lpt_bal(env)	Load-balancing based on the LPT algorithm.
make_cell_selection(env)	Instantiates cell templates for the selected cells according to population ranges and NeuroH5 morphology if present.
make_cells(env)	Instantiates cell templates according to population ranges and NeuroH5 morphology if present.
make_input_cell_selection(env)	Creates cells with predefined spike patterns when only a subset of the network is instantiated.
merge_spiketrain_trials(spiketrain, ...)	rtype: ndarray
run(env[, output, output_syn_spike_count])	Runs network simulation.
set_union(a, b, datatype)
shutdown(env)	Forces NEURON to make it delete its MPI communicator and shut down properly.

miv_simulator.network.ld_bal(env)

For given cxvec on each rank, calculates the fractional load balance.

Parameters:

env – an instance of the miv_simulator.Env class.

miv_simulator.network.lpt_bal(env)

Load-balancing based on the LPT algorithm. Each rank has gidvec, cxvec: gather everything to rank 0, do lpt algorithm and write to a balance file.

Parameters:

env – an instance of the miv_simulator.Env class.

miv_simulator.network.connect_cells(env)

Loads NeuroH5 connectivity file, instantiates the corresponding synapse and network connection mechanisms for each postsynaptic cell.

Parameters:

env (Env) – an instance of the miv_simulator.Env class

Return type:

None

miv_simulator.network.find_gid_pop(celltypes, gid)

Given a celltypes structure and a gid, find the population to which the gid belongs.

Return type:

str

miv_simulator.network.connect_cell_selection(env)

Loads NeuroH5 connectivity file, instantiates the corresponding synapse and network connection mechanisms for the selected postsynaptic cells.

Parameters:

env – an instance of the miv_simulator.Env class

miv_simulator.network.connect_gjs(env)

Loads NeuroH5 connectivity file, instantiates the corresponding half-gap mechanisms on the pre- and post-junction cells.

Parameters:

env (Env) – an instance of the miv_simulator.Env class

Return type:

None

miv_simulator.network.make_cells(env)

Instantiates cell templates according to population ranges and NeuroH5 morphology if present.

Parameters:

env (Env) – an instance of the miv_simulator.Env class

Return type:

None

miv_simulator.network.make_cell_selection(env)

Instantiates cell templates for the selected cells according to population ranges and NeuroH5 morphology if present.

Parameters:

env – an instance of the miv_simulator.Env class

miv_simulator.network.make_input_cell_selection(env)

Creates cells with predefined spike patterns when only a subset of the network is instantiated.

Parameters:

env – an instance of the miv_simulator.Env class

miv_simulator.network.init_input_cells(env)

Initializes cells with predefined spike patterns.

Parameters:

env (Env) – an instance of the miv_simulator.Env class

Return type:

None

miv_simulator.network.init(env)

Initializes the network by calling make_cells, init_input_cells, connect_cells, connect_gjs. If env.optldbal or env.optlptbal are specified, performs load balancing.

Parameters:

env (Env) – an instance of the miv_simulator.Env class

Return type:

None

miv_simulator.network.shutdown(env)

Forces NEURON to make it delete its MPI communicator and shut down properly.

TODO: This may no longer be required in more recent versions of neurons

miv_simulator.network.run(env, output=True, output_syn_spike_count=False)

Runs network simulation. Assumes that procedure init has been called with the network configuration provided by the env argument.

Parameters:

• env (Env) – an instance of the miv_simulator.Env class

• output (bool) – if True, output spike and cell voltage trace data

• output_syn_spike_count (bool) – if True, output spike counts per pre-synaptic source for each gid