Neural simulations¶
Overview¶
A simulation of a network is like an experiment with the difference that it takes place inside the computer’s memory rather than in the physical world.
Like in a real experiment, you need a system which you want to investigate. Moreover, you need a clear idea of what you want to learn from the experiment. In the context of a network simulation this means that you have to know which input you want to give to your network and which output you expect.
The next section will illustrate the main concepts of NEST simulations, using a simple simulation. The following sections will then give a step-by-step introduction to the main concepts of NEST simulations. Finally, we will discuss a complex example.
A simple simulation¶
The simplest simulation in NEST is that of a network which contains just one neuron:
SLI ] /iaf_psc_alpha Create /neuron Set
We are going to simulate a standard integrate and fire model with
resting potential at -70 mV and spike threshold at -55 mV. In this line,
we use the model iaf_psc_alpha
to create a neuron. The command
Create
returns a handle to the created neuron, which we store in the
variable neuron
.
Next, we would like to add a stimulation and a recording device to the neuron, so that we will see something during the simulation.
In our little example, we want to inject a current into the neuron and record its membrane potential. Thus, we need to create the corresponding devices and connect them to the neuron:
SLI ] /voltmeter Create /vm Set
This command creates a voltmeter node. The handle to the new voltmeter
is stored in the variable vm
.
By default, the voltmeter will only record membrane potential values, so we configure it to show us the time stamp of each value as well. We also set it to print the time and potential to the screen and we set the recording interval to 0.1 ms. The default is 1.0 ms.
SLI ] vm << /record_to /screen /interval 0.1 >> SetStatus
The double angled brackets <<
and >>
delimit a dictionary
definition which consists of successive /key value
pairs.
After setting the parameters of the voltmeter, we connect it to the neuron, using the handles we have created above:
SLI ] vm neuron Connect
Now we create a DC generator which will supply a constant current to our neuron:
SLI ] /dc_generator Create /stim Set
Next, we want to set the amplitude of the DC generator such that it delivers enough current to elicit a spike in the neuron:
SLI ] stim << /amplitude 600.0 >> SetStatus
Here, we only adjust the strength of the input, we set its new value to
600.0
, which is the amplitude in pA.
We have to connect the DC generator to the neuron:
SLI ] stim neuron Connect
We can now run the simulation and expect to see some results:
SLI ] 15.0 Simulate
The command Simulate
runs the simulation for the specified number of
milliseconds. Below, you see a transcript of the simulation:
Nov 10 08:57:51 Simulate [Info]:
Simulating 15 ms.
Nov 10 08:57:51 Scheduler::prepare_nodes [Info]:
Please wait. Preparing elements.
Nov 10 08:57:51 Scheduler::prepare_nodes [Info]:
Simulating 3 local nodes.
Nov 10 08:57:51 Scheduler::update [Info]:
Simulating using OpenMP.
1 0.1 -70
1 0.2 -70
1 0.3 -70
1 0.4 -70
1 0.5 -70
1 0.6 -70
1 0.7 -70
1 0.8 -70
1 0.9 -70
1 1 -70
1 1.1 -70
1 1.2 -69.7612
1 1.3 -69.5248
1 1.4 -69.2907
1 1.5 -69.0589
:
1 10.5 -55.3751
1 10.6 -55.2818
1 10.7 -55.1894
1 10.8 -55.098
1 10.9 -55.0075
1 11 -70
1 11.1 -70
1 11.2 -70
1 11.3 -70
:
1 12.7 -70
1 12.8 -70
1 12.9 -70
1 13 -70
1 13.1 -69.7612
1 13.2 -69.5248
1 13.3 -69.2907
1 13.4 -69.0589
1 13.5 -68.8295
1 13.6 -68.6023
1 13.7 -68.3775
1 13.8 -68.1548
1 13.9 -67.9343
1 14 -67.7161
Nov 10 08:57:51 Scheduler::resume [Info]:
Simulation finished.
Nov 10 08:57:51 Scheduler::finalize_nodes() [Info]:
using OpenMP.
After some initial messages from the simulation scheduler, we see the output from the voltmeter. The number in the left column represents the node ID of the model neuron and the center column the network time in milliseconds. The right column contains the values of the membrane potential at that time. The potential is given in mV.
By default, NEST uses a simulation stepsize of 0.1 ms. With a simulation time of 15.0 ms, we have 150 simulation steps.
The neuron that we have simulated was a standard integrate-and-fire
neuron [Tuckwell91] with a resting potential of -70 mV and a threshold
at -55.0 mV. We see the first effect of the DC input current at 1.2 ms.
This time delay is due to several reasons: The dc_generator
emits
the first current output at the end of the first time step, i.e., at 0.1
ms. Since the connection between generator and neuron was created with
the default delay of 1 ms, the current signal arrives at the neuron at
1.1 ms and thus affects the membrane potential for the first time during
the time step from 1.1 ms to 1.2 ms.
During the time step from 10.9 ms to 11.0 ms, the membrane potential crosses the threshold value -55.0 mV. Thus, the neuron emits an output spike at 11.0 ms and the membrane potential is then reset to -70.0 mV and clamped to the resting value for 2 ms, the refractory period of the neuron. After the refractory period, the membrane continues to depolarize due to the continuing input current.
Nodes and Models¶
A neuronal model in NEST is represented as a directed weighted graph. Nodes in the graph can either be neurons or devices, while the edges correspond to the synapses. On the level of their implementation, these elements are C++ classes.
An overview of all available node models can be retrieved from the
NEST kernel using GetKernelStatus /node_models get
, while passing
the key ‘synapse_models’ will return the list of available synapse
models.
You can find a list of all available neuron models in our model directory.
Creating nodes¶
Before continuing with the example, we reset NEST, to clear all nodes that we have created before.
SLI ] ResetKernel
Sep 21 10:13:39 Network::clear_models [Info]:
Models will be cleared and parameters reset.
Nodes are created from a model, using the command Create
.
SLI ] /iaf_psc_alpha Create ==
1
In the fist line, we create one integrate and fire neuron from the model
iaf_psc_alpha
.
The return value of Create
is an integer that identifies the last
node that was created in the network (note that this can be different
from 1 if you have not called ResetKernel before)
. This integer is
called the node’s node ID (the network as a whole owns the node ID
0
, therefore the ids of user-created nodes start with 1
). Often,
it is neccessary to have a large number of nodes of the same type. The
command Create can also be used for this purpose. The following line of
code create 10 integrate and fire neurons:
SLI ] /iaf_psc_alpha 10 Create ==
11
Status information¶
Nodes have a state which can be extracted and modified. In the follwing example, we display the status information of one the neurons in the layer we have created above:
SLI ] 1 ShowStatus
--------------------------------------------------
Name Type Value
--------------------------------------------------
archiver_length integertype 0
C_m doubletype 250
E_L doubletype -70
frozen booltype false
global_id integertype 1
I_e doubletype 0
local booltype true
local_id integertype 1
model literaltype iaf_psc_alpha
node_type literaltype neuron
parent integertype 0
recordables arraytype <arraytype>
state integertype 0
tau_m doubletype 10
tau_minus doubletype 20
tau_minus_triplet doubletype 110
tau_syn_ex doubletype 2
tau_syn_in doubletype 2
thread integertype 0
t_ref doubletype 2
t_spike doubletype -1
vp integertype 0
V_m doubletype -70
V_reset doubletype -70
V_th doubletype -55
--------------------------------------------------
Total number of entries: 24
Using the command SetStatus
, it is possible to change the entries of
this so called status dictionary. The following lines of code change
the threshold value V_th to -60 mV:
SLI ] 1 << /V_th -60.0 >> SetStatus
SLI ] 1 GetStatus /V_th get =
-60
Please note, that SetStatus
checks if a property really exists in a
node and will issue an error if it doesn’t. This behavior can be changed
by the following command:
<< /dict_miss_is_error false >> SetKernelStatus
Then, NEST is very tolerant with respect to the property that you are
trying to change: If it does not know the property, or if the property
cannot be changed, there will be no error, but only a warning. In any
case, SetStatus
does complain if the new value does not match in the
expected type:
SLI ] 1 << /V_th (60) >> SetStatus
Dec 01 15:33:54 SetStatus_ad [Error]: TypeMismatch
Expected datatype: doubletype
Provided datatype: stringtype
In order to find out, which properties of a given model can be changed an which not, you have to refer to the model’s documentation.
Connections¶
Connections between nodes define possible channels for interactions
between them. A connection between two nodes is established, using the
command Connect
.
Each connection has two basic parameters, weight and delay. The weight determines the strength of the connection, the delay determines how long an event needs to travel from the sending to the receiving node. The delay must be a positive number greater or equal to the simulation stepsize and is given in ms.
Example 1¶
SLI ] /iaf_psc_alpha Create /n1 Set
SLI ] /iaf_psc_alpha Create /n2 Set
SLI ] /iaf_psc_alpha Create /n3 Set
SLI ]
SLI ] n1 n2 Connect
SLI ] n1 n3 Connect
To inspect the parameters of a connection, one first needs to obtain a
handle to the connection. This is done using the command
GetConnections
. It takes a dictionary that at least contains the id
of the source node and will return a list of handles for all outgoing
connections. The search can be restricted by using the optional
parameters target and synapse_type.
Example 2¶
SLI ] << /source n1 >> GetConnections /c1 Set
SLI ] c1 length ==
2
SLI ] << /source n1 /target n2 >> GetConnections /c2 Set
SLI ] c2 length ==
1
To actually see the parameters of the connection, GetStatus
is used,
just like it is for nodes.
Example 3¶
SLI ] c1 0 get GetStatus info
--------------------------------------------------
Name Type Value
--------------------------------------------------
delay doubletype 1
receptor integertype 0
sizeof integertype 32
source integertype 1
synapse_model literaltype static_synapse
target integertype 2
weight doubletype 1
--------------------------------------------------
Total number of entries: 7
To change the parameters of a connection, SetStatus
is used, just
like it is for nodes.
Example 4¶
SLI ] c1 0 get << /weight 2.0 >> SetStatus
SLI ] c1 0 get GetStatus /weight get ==
2.000000e+00
Devices¶
Devices are network nodes which provide input to the network or record its output. They encapsulate the stimulation and measurement process. If you want to extract certain information from a simulation, you need a device which is able to deliver this information. Likewise, if you want to send specific input to the network, you need a device which delivers this input.
Devices have a built-in timer which controls the period they are active.
Outside this interval, a device will remain silent. The timer can be
configured using the command SetStatus
.
By definition, a device is active in the interval \((t_1,t_2)) if we can observe events \(E) with time stamps \(t_E) which obey \(t_1 <= t_E < t_2) for all \(E) . In other words, the interval during which the device is active corresponds to the range of time-stamps of the device’s events.
Note that it is not possible to generate/observe an event with time stamp 0.
Device parameters¶
The following entries of the status dictionary are the same for all stimulation and recording devices:
Property Type Description /start
double First time of activity,
relative to the value of origin
in ms. /stop
double First time
of inactivity, relative to the value of origin
in ms. /origin
double Origin of the device clock, relative to the network time in ms.
In general, the following must hold:
stop >= start
If stop = start, the device is inactive.
Stimulation devices¶
A range of devices is available for the stimulation of neurons. The most important ones are listed in the following table. For details, refer to the documentation of the respective decive.
Model name Description spike_generator
Device to generate spikes at
specific times. poisson_generator
Device to generate poisson
shotnoise. dc_generator
Device to generate a constant current.
ac_generator
Device to generate an alternating (sine) current.
step_current_generator
Device to generate a step current with
different amplitudes at different times.
Example 5¶
SLI ] /iaf_psc_alpha Create /n Set
SLI ] /poisson_generator Create /pg Set
SLI ] pg << /rate 220.0 Hz >> SetStatus
SLI ] pg n Connect
Recording devices¶
All devices which are used to observe the state of other network nodes
are called recording devices. Examples are multimeter and
spike_recorder
.
Recording devices have properties which control the amount, the
format, and the destination of their output. The latter is done by
setting their property record_to
to the name of the recording
backend to use. To dump recorded data to a file, set ascii
, to
print to the screen, use screen
and to hold the data in memory,
set memory
, which is also the default for all devices. Data
stored in memory can be retrieved after the simulation using
GetStatus
. To get a list of all available recording backends, run
SLI ] GetKernelStatus /recording_backends get ==
A list of node models including all available device models can be retrieved by calling ``GetKernelStatus /node_models get`. The most important devices are:
voltmeter
Device to observe membrane potentials.multimeter Device to observe arbitrary analog quantities.
spike_recorder
Device to observe spike times.
Please note that the connection direction for analog recorders (all
except spike_recorder
in above list) is inverted with respect to
other recorders, i.e. the recorder has to connected to the neurons in
this case.
Example 6¶
SLI ] /iaf_psc_alpha Create /n Set
SLI ] /voltmeter Create /vm Set
SLI ] /spike_recorder Create /sr Set
SLI ] vm n Connect
SLI ] n sr Connect
Simulation¶
NEST simulations are time driven. The simulation time proceeds in
discrete steps of size dt
, set using the property resolution
of
the root node. In each time slice, all nodes in the system are updated
and pending events are delivered.
The simulation is run by calling the command t Simulate
, where t
is the simulation time in milliseconds