Is it conceptualised as the signal distribution moving rightward, or as the noise distribution moving leftward?
- Anybody can ask a question
- Anybody can answer
- The best answers are voted up and rise to the top
Technically it can go either way, and both situations would be conceptually and statistically equivalent. But it is conventional to fix the mean of the noise distribution at 0, so that increased discriminability corresponds graphically with the signal+noise distribution moving rightward along the latent axis.