Well you could imagine seeing a process that causes entanglement and just trying to halt this process somehow. Before a certain time you would simply see no entanglement at all and after some time you wouldn't see any effect on the entanglement. In between something might happen, but what exactly is a bit hard to tell.

None of this is beyond current models though, so I imagine you could predict the results quite well if the calculations were feasible. And if that can be done then that's probably exactly what they did.