Pretty sure you cant use "solid like that, and since you cant get x,y coordinates from it anyways, I'd suggest placing all the "solid" objects into a family, and using overlapping at offset. +n y will get the bottom, -n y will get the top etc.
Its that or making up some collision dummies you place at image points.
That's odd, cause we can have one big event "collision with solid", but we can't ask which solid it was (except for manually putting all solids and shit in separate groups or even making stand-alone events for each solid). Guys, do something about it.