Hello Rik,
I am going to try this technique myself this weekend. My research so far suggests flashing the print with light "three quarters of the way through the developing", or "as the print starts to appear" so I will use these two as starting points for experimentation.
I've also read that this technique works best with a Metol only developer so I'm not sure whether my standard Ilford Multigrade developer will produce the desired effect.
As with all experimentation record what you do and only change one variable each time.
Hopefully some of the knowledgeable people will tell us all the above is nonesense and/or point us in the right direction
Crispin