I have come across an article about a language learning theory, named Recursive Hierarchical Recognition and thought it is worth to be shared here. Though it is intentionally bounded to explain language learning rather than whole perception, it’s built upon the same foundations about brain such as hierarchy, temporal (sequence) prediction, Hebbian learning as HTM theory is. I guess it has some successful practical applications as well, which shows how those concepts, even at a very abstract level, are powerful.
I think, in some sense, the cortical circuitry that Numenta discovers may get changed or updated over time but the fundamental ideas of HTM theory are so true that it must/does work any field that is somehow related with perception.