nemo::Simulation Class Reference

Simulation of a single network. More...

Inherits nemo::ReadableNetwork.

Public Member Functions
virtual const firing_output &	step ()=0
virtual const firing_output &	step (const firing_stimulus &fstim)=0
virtual const firing_output &	step (const current_stimulus &istim)=0
virtual const firing_output &	step (const firing_stimulus &fstim, const current_stimulus &istim)=0
Modifying the network
Neuron parameters and state variables can be modified during simulation. However, synapses can not be modified during simulation in the current version of NeMo
virtual void	setNeuron (unsigned idx, unsigned nargs, const float args[])=0
void	setNeuron (unsigned idx, float a, float b, float c, float d, float u, float v, float sigma)
virtual void	setNeuronState (unsigned neuron, unsigned var, float value)=0
virtual void	setNeuronParameter (unsigned neuron, unsigned parameter, float value)=0
virtual void	applyStdp (float reward)=0
Queries
Neuron and synapse state is availble at run-time. The synapse state can be read back at run-time by specifiying a list of synpase ids (see addSynapse). The weights may change at run-time, while the other synapse data is static.
virtual float	getNeuronState (unsigned neuron, unsigned var) const =0
virtual float	getNeuronParameter (unsigned neuron, unsigned parameter) const =0
virtual float	getMembranePotential (unsigned neuron) const =0
virtual const std::vector < synapse_id > &	getSynapsesFrom (unsigned neuron)=0
Simulation (timing)
The simulation has two internal timers which keep track of the elapsed simulated time and wallclock time. Both timers measure from the first simulation step, or from the last timer reset, whichever comes last.
virtual unsigned long	elapsedWallclock () const =0
virtual unsigned long	elapsedSimulation () const =0
virtual void	resetTimer ()=0

---

Detailed Description

Simulation of a single network.

Concrete instances are created using the nemo::simulation factory function.

Internal errors are signaled by exceptions. Thrown exceptions are all of the type nemo::exception which in turn subclass std::exception.

---

Member Function Documentation

virtual const firing_output& nemo::Simulation::step

(

)

[pure virtual]

Run simulation for a single cycle (1ms) without external stimulus

virtual const firing_output& nemo::Simulation::step

(

const firing_stimulus &

fstim

)

[pure virtual]

Run simulation for a single cycle (1ms) with firing stimulus

Parameters:

fstim

An list of neurons, which will be forced to fire this cycle.

Returns:

List of neurons which fired this cycle. The referenced data is valid until the next call to step.

virtual const firing_output& nemo::Simulation::step

(

const current_stimulus &

istim

)

[pure virtual]

Run simulation for a single cycle (1ms) with current stimulus

Parameters:

istim

Optional per-neuron vector specifying externally provided input current for this cycle.

Returns:

List of neurons which fired this cycle. The referenced data is valid until the next call to step.

virtual const firing_output& nemo::Simulation::step	(	const firing_stimulus &	fstim,
		const current_stimulus &	istim
	)			[pure virtual]

Run simulation for a single cycle (1ms) with both firing stimulus and current stimulus

Parameters:

	fstim	An list of neurons, which will be forced to fire this cycle.
	istim	Optional per-neuron vector specifying externally provided input current for this cycle.

Returns:

List of neurons which fired this cycle. The referenced data is valid until the next call to step.

virtual void nemo::Simulation::setNeuron	(	unsigned	idx,
		unsigned	nargs,
		const float	args[]
	)			[pure virtual]

Set an existing neuron

Parameters:

	idx	neuron index
	nargs	number of parameters and state variables
	args	parameters and state variables (in that order) of the neuron

Precondition:

The parameter and state array must have the dimensions matching the neuron type specified when the neuron was first added.

void nemo::Simulation::setNeuron	(	unsigned	idx,
		float	a,
		float	b,
		float	c,
		float	d,
		float	u,
		float	v,
		float	sigma
	)

Change an existing Izhikevich neuron.

See also:

nemo::Network::addNeuron for parameters

virtual void nemo::Simulation::setNeuronState	(	unsigned	neuron,
		unsigned	var,
		float	value
	)			[pure virtual]

Change a single state variable for an existing neuron

Parameters:

	neuron	neuron index
	var	state variable index
	value	new value of the state variable

For the Izhikevich model variable indices 0=u, 1=v.

virtual void nemo::Simulation::setNeuronParameter	(	unsigned	neuron,
		unsigned	parameter,
		float	value
	)			[pure virtual]

Change a single parameter for an existing neuron

Parameters:

	neuron	neuron index
	parameter	parameter index
	value	new value of the state variable

For the Izhikevich model the parameter indices are 0=a, 1=b, 2=c, 3=d, 4=sigma.

virtual void nemo::Simulation::applyStdp

(

float

reward

)

[pure virtual]

Update synapse weights using the accumulated STDP statistics

Parameters:

reward

Multiplier for the accumulated weight change

virtual float nemo::Simulation::getNeuronState	(	unsigned	neuron,
		unsigned	var
	)			const [pure virtual]

Get a single state variable for a single neuron

Parameters:

	neuron	neuron index
	var	variable index

Returns:

state variable with index var.

For the Izhikevich model the variable indices are 0=u, 1=v.

virtual float nemo::Simulation::getNeuronParameter	(	unsigned	neuron,
		unsigned	parameter
	)			const [pure virtual]

Get a single parameter for a single neuron

Parameters:

	neuron	neuron index
	parameter	parameter index

Returns:

parameter with index parameter.

For the Izhikevich model the parameter indices are 0=a, 1=b, 2=c, 3=d, 4=sigma.

virtual float nemo::Simulation::getMembranePotential

(

unsigned

neuron

)

const [pure virtual]

Returns:

membrane potential of the specified neuron

virtual const std::vector<synapse_id>& nemo::Simulation::getSynapsesFrom

(

unsigned

neuron

)

[pure virtual]

Returns:

vector of synapse ids for all synapses with the given source neuron

The vector reference is valid until the next call to this function.

Implements nemo::ReadableNetwork.

virtual unsigned long nemo::Simulation::elapsedWallclock

(

)

const [pure virtual]

Returns:

number of milliseconds of wall-clock time elapsed since first simulation step (or last timer reset).

virtual unsigned long nemo::Simulation::elapsedSimulation

(

)

const [pure virtual]

Returns:

number of milliseconds of simulated time elapsed since first simulation step (or last timer reset)

virtual void nemo::Simulation::resetTimer

(

)

[pure virtual]

Reset both wall-clock and simulation timer