I am building a novel model of neural tissue for the purposes of Machine Learning and am currently trying to unpick the functions of the astroglia.
The literature suggests that astrocytes ensheath synapses, providing them with structural support, given that the synapse is an oily droplet between two neuron 'tentacles' this makes sense, performs caretaker functions at synapses it ensheaths with regard to neurotransmitters and, via its perivascular feet, provides resources from nearby blood vessels.
The literature goes on to suggest that the astrocyte may play a role in some neurological disorders if the ionic 'effects' from one set of perivascular feet, namely one set of synapses, 'leaks' across an internal chemical barrier within said astrocyte, to spread the 'effect' to another set of perhaps unrelated synapses. I can see how this could influence and cause unrelated learning, which may be a hallmark of some neurological disorders, one of which the paper suggests is schizophrenia and its patterns of unearned and irregular learning.
I can appreciate how some form of localized learning or whole- or partial-pathway learning, as opposed to pure individual neuron-level learning, would certainly be beneficial to a learning network in speeding up learning.
Do astrocytes connect to astrocytes and if so how and what chemical messaging is present there?
I apologize that I am not a neurochemist only a software developer with a background in electronics and while I can follow some of the chemistry, I am interested in the functional purpose of these biological components. Nature only had chemistry to work with when these mechanisms evolved but, as an aside, if it had had access to electronics and software, one supposes the neural tissue would have been far easier for us to comprehend.