Spiking Chemical Synapse Model

Watch the video tutorial on this subject!

Free C++ source code for the spiking chemical synapse model is available.

Figure 1. Spiking chemical synapse property dialog.

This dialog box is used to specify the properties of different types of spiking chemical synapse models. A spiking chemical synapse is a class of synapse in which transmitter release, and hence the post-synaptic conductance change, is only triggered by a spike in the pre-synaptic neuron (see also Non-Spiking Chemical Synaptic Types). See Synaptic Models for more details of how the different classes of synapse are modeled.

All the synapses in this module are derived from the Link object and share the properties of that item. For a description of these properties please see the text that discusses the Link properties.

Hebbian Properties

Allow forgetting
If this value is true, the synapse “forgets” its training unless it is periodically reinforced. If this value is false, training is a one-way process; a trained synapse maintains its augmented conductance state permanently.

Consolidation factor
If this factor is greater than 1, then well-trained synapses forget more slowly than poorly trained synapses. The consolidation factor extends the forgetting time window by an amount set by the product of the consolidation factor itself and the relative degree of augmentation of the synapse. Thus if the synapse is fully augmented, then the effective forgetting time window is the baseline forgetting time window multiplied by the consolidation factor.
Acceptable range: 1 to 1000.

See Synaptic Models for more details of how voltage dependent and Hebbian synapses are modeled.

Forgetting time window
This sets the baseline of the time window during which reinforcement of training must occur. After a Hebbian synaptic augmentation, the synaptic conductance declines linearly back to its baseline value during the course of the forgetting time window (but see Consolidation below).
Acceptable range: 0 to 10000000 msec.

Hebbian
Set this value to true to make the synapse Hebbian. This means that the conductance of the synapse increases when the post-synaptic neuron spikes within a certain time-window of the pre-synaptic input.

Learning increment (%)
Enter the relative amount by which the synaptic strength is augmented when the post-synaptic neuron spikes immediately after a Hebbian pre-synaptic input.
Acceptable range: 0.001 to 100 %.

Learning time window
Enter the time-window for which a post-synaptic neuron “remembers” a Hebbian pre-synaptic input. If the post-synaptic neuron spikes within this time-window the synaptic strength is augmented. The degree of augmentation depends upon how far into the time-window the input occurred. The closer in time the post-synaptic spike occurs relative to the pre-synaptic input, the greater the augmentation (to a maximum level set by the Increment parameter above). Augmentation varies linearly within the time window.
Acceptable range: 1 to 1000 msec.

Maximum augmented conductance
Enter the maximum synaptic conductance of the fully augmented (trained) synapse.
Acceptable range: Baseline conductance to 1000 uS/size.

Synapse Properties

Decay rate
Enter the time constant for the decay rate of the synaptic conductance. When a synapse is activated there is an initial step-increase in conductance (to the amplitude defined in the Maximum synaptic conductance parameter above), followed by an exponential decline back to zero with a time constant set by this Decay rate parameter..
Acceptable range: 0.01 to 1000 ms.

Equilibrium potential
Enter the equilibrium (reversal) potential for this synaptic type.
Acceptable range: -100 to 300 mV.

Facilitation decay
Enter the time constant of the rate of decay of facilitation. A long time constant means that the synapse shows facilitation or decrement (according to the value of the Relative facilitation parameter) at relatively low frequencies of activation, while a short time constant means that the synapse only facilitates or decrements with high frequency activation.
Acceptable range: 0.01 to 1000 ms.

Name
This is a text description of this synapse.

Relative facilitation
Enter the relative facilitation rate of the synapse. A rate greater than 1 means that the synapse post-synaptic conductance tends to increase with high-frequency repetitive activation (facilitation), while a facilitation rate less than 1 means that the post-synaptic conductance tends to decrease with repetitive activation (decrement).
Acceptable range: 0 to 10

Synapse Type
This is the type of synapse. It can be spiking-chemical, non-spiking chemical, or electrical. The is a read-only value.

Synaptic conductance
Enter the amplitude of the post-synaptic conductance change which this synapse mediates. This value is the initial conductance (time zero) of the unfacilitated synapse. A value of 0 means that the synapse is switched off. Note: amplitude is defined per unit size (effectively per unit membrane area). This means that conductance scales proportionally with changes in the relative electrical size of the neuron, so that changes in neuron size make no difference to the amplitude of the synaptic response (for comparison see Electrical Synaptic Types).
Acceptable range: 0 to 100 uS/size.

Voltage Dependent Properties

Maximum (unblocked) relative conductance
Enter the amplitude of the maximum synaptic conductance as a multiple of the Baseline conductance.
Acceptable range: 1 to 1000.

Saturate post-synaptic potential
Enter the post-synaptic membrane potential at which all voltage-dependent post-synaptic conductance block is fully removed, so that the conductance becomes equal to the Baseline conductance multiplied by the Maximum relative conductance.
Acceptable range: -100 to 100 mV.

Threshold post-synaptic potential
Enter the post-synaptic membrane potential at which the voltage-dependent post-synaptic conductance block starts to be removed.
Acceptable range: -100 to 100 mV.

Voltage dependent
Set this value to true to make the synaptic type voltage dependent (i.e. the conductance depends on the post-synaptic membrane potential)