The centroid of the lit areas weakens, and may lose focus a bit, but doesn't slow down much.

Studies of some other earthquakes, like the 1992 Landers and the 1999 Hector Mine earthquake, do show such hesitation. We speculate a lot about what causes the delays, but it is not certain.

Just anecdotally, sometimes it seems to take a while for the rock to rupture after the stress has been applied, this is also seen in simulations in the lab. A technical term that has been used is "barriers" absorb energy and slow down rupture, while "asperities" are places that release a lot of energy and speed up the rupture.

Work in progress.