cnrl / CoNeX

Cortical Network for everything!
MIT License
3 stars 0 forks source link

Standard Structure #4

Closed saeedark closed 1 year ago

saeedark commented 1 year ago

Structure by @atenagm1375 for cortical column and neocortex.

This PR enables the ability to build a new network with provided structures automatically.

saeedark commented 1 year ago

We will need a min_delay to be used differently based on source and destination (inter-column vs intra-column). Basically, max_delay will only indicate the buffer size and this min_delay will represent the default (initial) delay index to read from. Also, if we assume that distal dendrites receive input from layers in the column and apical dendrites receive input from other columns, we can force a 1-unit difference between both max_delay and min_delay of the two dendrite types.

There is a technical problem here. as far as I know, we consider delays from one population to another one. In the current implementation of the delay mechanism, delays have been broken into pre-synaptic(Axon) and post-synaptic(Dendrite) parts. Calculating absolute delay is possible and I think they should be used as a measure for comparing minimum delay between synapses.

That being said, does each kind of synapse need its own minimum delay(Proximal, Apical and Distal)? should we define minimum for pre-synaptic as well as post-synaptic delay?

atenagm1375 commented 1 year ago

We will need a min_delay to be used differently based on source and destination (inter-column vs intra-column). Basically, max_delay will only indicate the buffer size and this min_delay will represent the default (initial) delay index to read from. Also, if we assume that distal dendrites receive input from layers in the column and apical dendrites receive input from other columns, we can force a 1-unit difference between both max_delay and min_delay of the two dendrite types.

There is a technical problem here. as far as I know, we consider delays from one population to another one. In the current implementation of the delay mechanism, delays have been broken into pre-synaptic(Axon) and post-synaptic(Dendrite) parts. Calculating absolute delay is possible and I think they should be used as a measure for comparing minimum delay between synapses.

That being said, does each kind of synapse need its own minimum delay(Proximal, Apical and Distal)? should we define minimum for pre-synaptic as well as post-synaptic delay?

I think we'd better define the min value for pre-synaptic only to handle the time difference between inter-column and intra-column connections more easily and assume a zero/one as min value for axons by default. Otherwise, we need means to distinguish if source and destination are in the same column (which is already there and usable but I say it will not be neat).