Why do imageLoad/imageStore take potentially negative coordinates?

This is likely not important, but when looking at functions like imageLoad and imageStore. They seem to take in integer values instead of what, I assume, should be unsigned integers. I know it's not valid to fetch a texel outside the bounds of an image. Is this just a product of some decision in the past, or is there a legitimate reason you would want to fetch a pixel from a negative pixel coordinate?