Hello NEST community,
I am at the outset of a large project and I would like to start off with the right simulator. Perhaps someone here can advise on the viability of NEST for my particular needs. I will be modeling neuromorphic hardware with some unique design constraints, such as:
- A complex dendritic arbor - Dendrite potentials may be coupled with their neighbors - Several dendrites may feed into other dendrites (a several-layer-deep dendritic tree) - The arbor eventually feeds into the soma which has typical LIF behavior, with a membrane potential determined by the sum of all directly connected dendrites - The solution for the leaky-integrator term of dendrites must be drawn from a precalculated lookup table
The question here is how best to facilitate these dendritic nuances within NEST? Any counsel on whether it makes sense to pursue modeling such a system in NEST would be highly appreciated. We would like to eventually scale to the maximum number of neurons our hardware/clusters and your simulator can facilitate.
Thank you, Ryan
Dear Ryan,
Thanks for writing in! All of the things you mention are probably in principle possible with NEST Simulator, but we would have to look in detail at some things, in particular the dendrite-to-dendrite coupling. It is possible this could be handled by the "gap junction" type connection in NEST. You would define as many input (or "receive") ports ("rports") to your neuron as there are connection sites between the two dendrites. Each of these would be connected pairwise with a gap junction connection. See https://nest-simulator.readthedocs.io/en/latest/models/gap_junction.html and reference [1] there. Gap junctions in NEST use the waveform relaxation method, so please double-check that that meets your needs.
Support for complex dendritic arbors was recently added; please see https://nest-simulator.readthedocs.io/en/latest/models/cm_default.html for details. You could probably customize this model with the specific set of ion channels and dynamics that you need.
It would be interesting to tackle the remainder with NESTML (https://nestml.readthedocs.org/). We would be happy to follow up on implementing a lookup table in an efficient manner. Compartmental models support is also coming soon for NESTML (you will be able to define your ion channels and dynamics in the NESTML syntax, rather than in C++).
Hope these provide some starting points, please let us know how you fare!
With kind regards, Charl
On Thu, Oct 13, 2022, at 17:37, Ryan O'Loughlin wrote:
Hello NEST community,
I am at the outset of a large project and I would like to start off with the right simulator. Perhaps someone here can advise on the viability of NEST for my particular needs. I will be modeling neuromorphic hardware with some unique design constraints, such as:
- A complex dendritic arbor
- Dendrite potentials may be coupled with their neighbors
- Several dendrites may feed into other dendrites (a several-layer-deep dendritic tree)
- The arbor eventually feeds into the soma which has typical LIF behavior, with a membrane potential determined by the sum of all directly connected dendrites
- The solution for the leaky-integrator term of dendrites must be drawn from a precalculated lookup table
The question here is how best to facilitate these dendritic nuances within NEST? Any counsel on whether it makes sense to pursue modeling such a system in NEST would be highly appreciated. We would like to eventually scale to the maximum number of neurons our hardware/clusters and your simulator can facilitate.
Thank you, Ryan
NEST Users mailing list -- users@nest-simulator.org To unsubscribe send an email to users-leave@nest-simulator.org
-
Charl Linssen -
Ryan O'Loughlin