Learning in the HTM algorithm

It might be interesting to set up a small experiment where a result from psychology was reproduced. I am thinking of this paper:

Very old study that showed consciousness/attention/awareness (pick one, not really significant here) is not required for learning.