Closed esinyavuz closed 5 years ago
STDP in combination with delayed spike propagation has now been tested, and is confirmed working.
Introducing backpropagation delay is possible. Perhaps we should also take into consideration that the time constant of the dendrite probably is not equal to the time constant of the soma? i.e. membrane voltage may be significantly lower at postsynapse terminal afer backpropagating along the dendrites.
should we have a feature test on this?
new post_spike_time_in_sim
and pre_spike_time_in_post_learn
feature tests should address this
I believe we need to discuss where delays apply. Fromthe neurobiology side of things I would think what we implemented so far ar axonal delays and we assume that "backpropagation" of post-synaptic spikes for STDP etc does not incur a delay.