Yes... now we can do spherical neural topologies!

New SpikeFun update (v0.40) adds initial support for the spherical neural network topology (in a very early and buggy form - i.e. neuron picking is totally flawed!). Spherical topology is very useful, as it easily allows simulating of the "reentry" effect, which is also observed in the vertebrate central nervous system (so-called "thalamocortical loops"). Original idea for spherical topology for simulating reentry effects came from the Izhikevich et. al..

Unfortunately, with SpikeFun v0.40 it is still not possible to export the spiking data and identify self-organized neuronal groups, but this is something that is planned to be implemented in the future. Also, SpikeFun still does not support modeling of non-myelinated axonal collaterals (another "to be done" feature).

However, even in this early form it might be interesting to watch how STDP and reentry shape the behavior of neurons in the network.

Following picture shows SpikeFun 0.40 running a simulation with spherical network topology: