Given a conservative estimate of a quarter of a billion years for the Milky Way to rotate once, and let's imagine two or three supernovae, you'd be looking for

Given a conservative estimate of a quarter of a billion years for the Milky Way to rotate once, and let's imagine two or three supernovae, you'd be looking for a pattern of intersecting ripples in some already scarce interstellar matter smeared out over forty or so galactic revolutions. My guess is that you'd need a pretty good survey of local stellar fragments (black holes, neutron stars, or white dwarfs) and a very, very good survey of the density of local interstellar media. Modeling that "rewinding" would be interesting. I have no idea if we have the data now to accomplish it but that is interesting to think about.