The R/S naming system for stereocenters relies on the Cahn–Ingold–Prelog priority rules to rank the four substituents of a stereocenter; the R or S name is then attributed depending on the spatial orientation of the ranked substituents. While this is mostly used for asymmetric carbon atoms, it holds true of other stereocenters with four substituents.
However, nitrogen and phosphorus atoms can also act as stereocenters if the inversion of their configuration is blocked (e.g. because they're in a cyclic molecule) or slowed enough (e.g. because of bulky substituents). Apparently, these stereocenters are also labelled in R/S fashion. Examples I found include Tröger's base:
(above is the (5S,11S)-enantiomer, below is the (5R,11R)-enantiomer), and DIPAMP:
Having said that, I can't find any reference as to how the R/S naming scheme is applied to these tetrahedral stereocenters which don't have four substituents, but a lone electron pair and three substituents instead. I assume that the same scheme can be followed if one considers the lone pair as the lowest-ranked substituent, but I can't find any canonical reference as to that.
So, how are such stereocenters named? What authoritative reference can I find on this topic?