Hi Donald,
Let me confirm this, but I'm pretty sure it is dBV/sqrt(Hz). This is in contrast to the dBm/Hz that people have traditionally worked in.
This is because the dBV/sqrt(Hz) is basically just 20*log10() of the nV/sqrt(Hz) noise spectral density number. Whereas the dBm/Hz number is 10*log10() of the mW/Hz noise spectral density.
Thanks,
Joel