I recently read The Autistic Brain by Temple Grandin. It is a fascinating book that really draws your attention to what is known and what is unknown about autism — and some of the problems with autism research that are caused by the fact that most researchers into autism are not themselves autistic, and the fact that there is a strong bias against self-reports, meaning there is too much focus on external expressions and not enough on what someone with autism is experiencing. It seems odd that autism research is, in this sense, a final holdout of behaviorism.
It should not be surprising that Grandin, being autistic herself, focuses on how autistics experience the world. But she also points out that autistics think about the world in a different way. Specifically, she notes that neurotypicals think in a top-down fashion (big picture before details), while autistics think in a bottom-up fashion (details before big picture). This has an interesting result. This means that neurotypicals tend to develop ideas from fewer sources, then take that theory and go back to the facts, where as autistics collect far more details and develop a theory from the details, from the facts. This also, coincidentally, is what allows autistics to see and understand patterns better and in more detail.
Also, this bottom-up thinking might help autistics to understand bottom-up processes better. By bottom-up processes, I am talking about self-organizing scale-free network processes like economies and other social systems, developmental biology, ecosystems, etc. They are created by having more and more things interacting with each other. Which is much the same way that bottom-up thinking works. Which brings us to issues of epistemology, specifically how concepts are formed..
With top-down thinkers, one needs very few examples of something to create a concept. If you see two or three cats, that can be enough to develop the concept of cat. However, bottom-up thinkers require many more examples before the concept is fully developed. Ten, fifty, or a hundred cats may be required. But once that concept is formed, it is well-formed. This can of course create problems with language acquisition, insofar as there is a correlation between words and things, actions, and qualities in the world. But it can also prevent one from developing new ideas before the situation is fully understood — a quality that could come in quite helpful if you are a scientist, for example.
It is also likely true that a more bottom-up, pattern-based thinker is going to see and understand bottom-up processes with complex patterns better than neurotypicals, meaning they will tend to understand spontaneous orders in general better than neurotypicals.This is because similar complex systems better model each other. Top-down thinking views the social world as hierarchical organizations; bottom-up thinking views the social world as spontaneous orders; the former thus make for better business people, while the latter make for better economists.