This isn't really true. When Human Action was written there already was a fair bit of literature on behavioural psychology and other fields, but Mises made a conscious choice to not base his theories on them. In the first part of Human Action Mises does bring this point up but makes a clear distinction between investigations into human behaviour that are rooted in the natural and empirical sciences (i.e. biology) and his own flavour which is rooted in a Kantian-esque introspection. This concept is referred to in Mises' work as well as in the works of his successors as methodological dualism, and is a core tenet of Austrian social analysis.
Social science isn't "experimentable" as biology or physics... And if you try hard to force it you will end up with what's called social engineering instead...
W. Brian Arthur is an economist I like. Complexity theory is also very exciting for me. I do not have any experience in those, but they and other things inspired how I perceive the world.
I probably did not deliver what I think clearly. The link above might be helpful to deliver what I mean.