Actually, there IS a way that the ball could graze the rim and the rotation of the ball not change … and that would be if the force from the impact of the ball hitting the rim was in a direction that was roughly the same as that which the ball was already rotating....
The change in rotation that I saw was in the same direction as the ball was already rotating. It just made the ball rotate in that direction slightly more quickly.
There's also a magnitude of force to consider … but we've already established that the contact with the rim was either non-existent (zero magnitude) or very slight - so the speed up of the ball's rotation was also very slight, and was made to look even more slight due to the super slow motion of the instant replays.
Always knew that my Speed School education would come in handy....