[frankinstien]

Well, I got this epiphany, I wondered since neural synapses can transmit multiple transmitter types simultaneously, how could that be a coding advantage?

When neurons have multiple receptor types as well, you easily see how those transmitter channels could all be independently paralleled information channels soliciting independent processing.  So a synapse then could represent multiple wires as opposed to what was thought to be only one.

But then I thought it might be an error correction scheme to ensure the quality of the signal to receiving neurons by sending some redundancy?

But then I thought what if when a neuron is in a repetitive mode each firing had some shift in the transmitters' strength ratios?
That makes for a more sophisticated neural code than what ANNs are doing today.

Here's a good link about the subject matter.

[MagnusWootton]

Cool idea.

Me personally, I dont think the real brain is going to help me finish A.I.   I'm of the opinion the neuron can be anything that suffices to transfer input to output,  maybe the real neuron is actually a really good one, but its useless to emulate it because u dont know what its doing exactly and you may as well just use Coll Pitt's,  does the job fine.