I am not a computer scientist and I don't know much about AI, but I want to make a suggestion on how to make Neuro more human like. I don't know whether you've ever heard about Niklas Luhmann. He was a famous sociologist and he specialized in systems theory. I think his theory of meaning can help you make Neuro process meaning in a more human-like way.
Luhmann basically theorized that meaning is processed in three dimensions: factual, time, and social dimensions. Factual dimensions process object distinctions: this-and-not-that. Time process future and past distinctions. Social dimension processes ego-alter distinctions. Meaning is a surplus of reference with a focal point--actuality--and a peripheral horizon of possibilities. It's a super complicated theory and, if you are interested, you should read his books such as Introduction to Systems Theory (which is a collection of his lectures) or Social Systems (which is actually a book about general systems theory, not about social systems).
I don't know much about AI, but I imagine how human process meaning can be replicated. It would require three special algorithms: one for processing the factual dimension of meaning, one for time dimension, and one for social dimension. And then you need one algorithm for synthesizing the three dimensions. Maybe it will require too much computational power. But hopefully you find this helpul