There are at least two advantages to scanning a B&W negative in RGB and then converting to B&W afterwards, in PS. They are usually relatively minor advantages, but in some situations, this can be a useful technique.
The first advantage is that if there is water, fire or any other damage to the negative, it almost always is of a color different from that of the image on the negative. This difference can allow one to artificially crank up the saturation and semi-automatically select the damaged spots / areas (and then, appropriately deal with them) rather than needing to identify each damage spot/area by hand. This can save you an enormous amount of time.
Scanning B&W's in RGB for this purpose is a technique strongly recommended by Ctein in what I consider the bible of digital restoration:
"Digital Restoration from Start to Finish: How to repair old and damaged photographs"
http://www.amazon.com/Digital-Restoration-Start-Finish-photographs/dp/0240812085
The second advantage arises because B&W negatives are never purely gray - they actually all have a slight color cast. Say you do an 8 bit per channel scan. If you scan in B&W, you will have a maximum of 256 different levels available. In contrast, if you scan in 8 RGB, the color cast is slightly different from point to point, and so you effectively have nearly 3x the number of different color levels, so you can set PS to 16 bpc, and then, after desaturation, effectively wind up with a 9 bpc image out of an 8 bpc scan. Of course, if your scanner a has suitably large dynamic range and can output 16 bpc files the above is moot, but many low end scanners don't have this capability. Also, if your negatives are perfectly exposed and processed, then you won't have to dig information out of nearly opaque or nearly transparent areas of the negative, and so having more available quantization levels isn't necessary.
Cheers,
Tom M