My take on this has always been that our brain (with the loverly theory of mind) is normally very good a recognising
intent in other people (as a by product of recognising that they have intent).
In our prehistory this would have been a wonderful adaptation:
Rustling in the long grass, what's that?
Twig snaps in the darkness, what's that?
If you are prone to concluding that the above sinister sounds were the effects of a purposeful actor (say a stalking leopard or sneaking member of the tribe from across the valley) that has you as the focus of it's attention (it's tryin' to get ya) then you would be more inclined to take evasive action and (hopefully) survive.
If you conclude that the sound is simply the wind and it does not flag any threat awareness reaction you will most likey screen it out.
I remember speaking to a young man with Asperger's syndrome who I used to support and he said that sometimes he would not notice people in the room with him untill they spoke (thus confirming they were worth paying attention to).
I wonder if this is because he did not flag them as intenful actors in his evironment. He would say they were part of the furniture untill they spoke up or moved.
So perhaps seeing the intent of an imaginary actor in the environment is a survival mechanism that went a bit too far.