Add publication

[SomeUserName1]

Hi Michael,

Is there a paper or a preprint or something similar on Jaxley?

Can you shortly elaborate on the neuron model and the difference between Jaxley and NEURON?

Regarding NEURON, the book says:

The axon and dendrites are simple cylinders, with uniform diameters and membrane properties along their lengths. The dendrites are passive, while the soma and axon have Hodgkin-Huxley sodium, potassium, and leak currents, and are capable of generating action potentials (Hodgkin and Huxley 1952).

However, that was written in 2004.
From what I understand, NMODL allows for extensions, however just for ion channels.

Jaxley also implements an extensible HH-model, correct?
Is Jaxley in theory capable of supporting GPCRs in some way?
What about dendritic non-linearities, e.g. synaptic clustering, shunting, Ca-spikes and the like?

Thanks in advance
Fabian

[michaeldeistler]

Dear Fabian,

thanks for reaching out!

We are currently preparing the preprint for Jaxley and will submit it this week (that’s why we just made the code public). I will let you know once it is out!

Yes, (like NEURON), Jaxley also allows to implement any kind of custom ion channels, see here.

And yes, (like NEURON), Jaxley allows to have channels in any kinds of neurite (e.g. only in the soma, or in the dendrite). See the Insert mechanisms subsection here.

Because Jaxley allows to insert any kind of channel in any kind of neurite (e.g., dendrite), it can also model dendritic nonlinearities, shunting inhibition, calcium spikes, etc. You can find a collection of currently implemented channels here.

What are GPCRs?

Hope this helps!

Michael

[michaeldeistler]

The big differences between Jaxley and NEURON are that:

1. Jaxley runs on CPU and GPU without any changes to the code
2. Jaxley allows to compute the gradient with respect to parameters

[huangziwei]

GPCRs

G protein-coupled receptors? In Jaxley, we use the concept of Synapse instead of Receptor, which deviates slightly from the convention used in NEURON, but usage-wise it's actually very similar.

https://jaxleyverse.github.io/jaxley/tutorial/05_channel_and_synapse_models/#your-own-synapse

https://github.com/jaxleyverse/jaxley-mech/tree/main/jaxley_mech/synapses

[michaeldeistler]

The preprint for Jaxley is on biorxiv now, see here.