"...but as people mastered the task, the systems [of the brain] stopped communicating. It turned out that people who could disengage these systems the fastest were also the best learners."

This may just be an artifact of the presentation, or my misreading of it, but from then on it seems to be assumed that fast learning is a consequence of fast disengagement, rather than the other way around, yet no evidence is offered here that explicitly supports this interpretation over the alternative.