Since there are some good simulations of the neural circuits using this classic model, I'm curious about any reference or guideline that would be useful for this topic.
The attached article is one of them (written by researchers of the MIT etc.)
So from m experience, it becomes difficult to model Hodgkin Huxely Neurons in Matlab due to the solution of the multitude of differential equations that need to be numerically solved. You could attempt to do this, but if I may, I'd like to offer another alternative. There is a nice program called Neuron by Michael Hines and Ted Carnevale from Yale(open source) that is optimized for the solution of these equations. You can also control the program in Matlab, allowing for multiple runs with varying input stimuli. I found this approach quite a bit easier. While not directly answering your question, I hope this provides a helpful avenue.
We are currently working on building a platform for implementing HH and HH-like models in matlab. Matlab is rather slow in solwing differential equations as Brandon Coventry mentioned, so we use C# for solving. We are mainly interested in peripheral afferent fibers, so we do not have a model for networks. Neuron is probably the best platform for networks.
Attached is a simulation I wrote in C++ of a network of Hodgkin Huxley neurons. The paper that the simulation was for is https://www.researchgate.net/publication/234099572_NMDA_and_GABAB_(KIR)_Conductances_The_Perfect_Couple_for_Bistability. The excitatory Hodgkin Huxley neurons are described in http://www.nature.com/neuro/journal/v1/n4/abs/nn0898_273.html. The inhibitory neurons are described in http://www.jneurosci.org/content/16/20/6402.full.pdf.
Article NMDA and GABAB (KIR) conductances: the "perfect couple" for ...
Modeling connected HH neurons in Matlab is surely feasible. I've attached a PDF tutorial on HH programming that one of my professors borrowed from a class taught at Duke. I've also attached my implementation of a single neuron HH neuron in Matlab.
Edit: It is feasible to model ~1 million connected HH neurons these days in Matlab if your code is properly vectorized and sparse where necessary.