can we simulate neurobiology using CA?

[CARuler]

i was thinking about neurobiology and thought "hun those signals going across neurons look a lot like a lagdon's ant kind of CA" can we do something like this? here are some ideas and thoughts:

1. since we can already play with CA on any node structure, why can't we have the CA create and destroy nodes and connections? i will call these "node editing CA" (NECA). We will need these because the brain creates a destroys brain cells (which can be represented as nodes)
2. since these "ants" don't have a good way to store direction, these "ants" will probably spilt (they will also lose some energy as they move) or somehow encode directions in the nodes themselves
3. the energy of these "ants" will be continuous and the values held in the nodes will probably be too
4. there must be ways for the NECA to take in energy and inputs to effectively work (maybe add nodes for things like eyes? also, maybe make such nodes and connections indestructible?)

please help me with this idea
notes: the ants represent the signals passing along brain cells; the nodes represent brain cells