I think there are 3 feedback mechanisms in different levels:

1. Micro-scale : ephaptic coupling
2. Meso-scale : feedback loops within neural networks
3. Macro-scale : emotions

Based on the Schachter & Singer’s theory of emotions, I believe we can consider them as feedback mechanisms about the environment. So, perception is kind of feed-forward processing, emotions are a feedback mechanism and awareness (both about the physical world and about the emotions) is kind of backpropagation. But kind of, not exactly..

So, my idea about brain as 3 networks is also meaningful in this context:

• One network is about information processing (probably synaptic), one network provides a feedback mechanism (probably ephaptic), and the other network could be the one that controls the backpropagation (probably epigenetic) in the micro scale.
• Above that, we can talk about synaptic feedback loops, stable electrical fields (yeah, related with ephaptic coupling / ephaptic network) and neuroplasticity (also related with epigenetics / epigenetic network).
• In the macro scale, I think we can talk about perception, emotion and awareness, relatively.

I think meso-scale processes can be thought of as sum of the micro-scale processes, but in the macro-scale, I think we are observing emergent properties. If we can formulate these networks and interactions, Artificial General Intelligence (AGI) may not be so far away.