STDP Ordering Options

[mway]

For synapse inputs that arrive after a spiking neuron, STDP effects dictate that if within a close enough temporal proximity the future influence of the synapse would be lessened. What I would like to know is which of these two senarios is more correct.

Scenario A:
1. Post synaptic potential arrives.
2. Neuron depresses synaptic influence.
3. The membrane potential is updated based on this new influence value.

Scenario B:
1. Post synaptic potential arrives.
2. The membrane potential is updated per usual.
3. Neuron depresses synaptic influence so that any future arrival exhibit less effect.

I'm not if any studies have been done to this detail yet, but if someone could shed any light on this query i'd be very grateful.

Matt

[Paul King]

It may be a mistake to imagine that there is a specific order to these events because the events are not happening that quickly.

In your scenario A and B, steps 2 and 3 are reversed. This assumes that STDP synaptic weight updates happen immediately, within milliseconds, but this is not the case.

STDP, which is a combination of LTP and LTD, is a slow process involving the manufacture, trafficking, and insertion or deletion of neuroreceptors. It happens on the order of minutes to hours. Whereas spikes last 1 ms and happen a few per second.

It is understandable that the update might seem immediate, because that is how one might implement it in a computer simulation. However the experiments that produce the STDP effect do not measure it immediately after a single spike. The STDP plots that one sees are measured after applying hundreds of coincident spikes over many seconds or minutes. For a synaptic strength to change substantially could require days, as gene switching in the cell nucleus is involved.

A better way to think about STDP from the point of view of simulation is that potential weight changes are accumulated over time and are applied incrementally and slowly to the strength of the synapse.

There are other faster (and temporary) synaptic effects, such as paired-pulse facilitation and paired-pulse depression which occur within a few milliseconds, but this is by a different mechanism from STDP.