I use NEST (Version 2.16.0) by using PyNN.
While using I came across the problem that NEST throws an error, when the rate of a
SpikeSourcePoisson (in PyNN) is updated from any positive value to 0 and to any positive
value again.
In other words, once the rate is set to zero, it cannot be changed anymore. It seems like
this issue is independant of the temporal location.
The exception thrown is the following:
python2.7:
/pathToNestSimulator/nest-simulator-2.16.0/nestkernel/event_delivery_manager.h:564:
nest::delay nest::EventDeliveryManager::get_slice_modulo(nest::delay): Assertion
`static_cast< std::vector< delay >::size_type >( d ) <
slice_moduli_.size()' failed.
Further, any rate change of a specific neuron is not applied before the next spike at the
previous rate. This is problematic when dealing with low rates as spikes occur only
rarely.
For instance, with a rate of 1 it is not unlikely that there is not a spike for a few
seconds.
As a result, a rate change can and will be applied extremely delayed.
I initially reported this at PyNN, but it seems to be a NEST issue:
https://github.com/NeuralEnsemble/PyNN/issues/650
What would you advise?