This potentially effects the 60d as well (don't have one to test). I was having intermittent problems with the focus patterns feature and I finally traced it down to the screen being stored with the screen side facing inward. The problem does not seem to manifest itself otherwise and the focus patterns feature works flawlessly otherwise (and is very useful, thanks 8-).
With the screen stored inward (and not folded out or stored flush against the camera with the screen facing outward), the focus patterns fail to engage or stay engaged even if I toggle the DISP button to insure the display is off (of course in this orientation, it's impossible to visually verify the screen is off). If you start with the screen out and the patterns are engaged and working and then store the screen inward, the arrow buttons produce a strange (non-deterministic) pattern before changing to a standard single focus point (with subsequent arrow inputs). Also, pressing the focus pattern button (zoom in) first (after having patterns properly working and then turning in the screen) shows the last pattern (but using the arrows shows a non-deterministic pattern (or at least, it doesn't follow any obvious pattern) before switching to a single (traditional) focus point. Apparently, storing the screen inward on these cameras produces a third state (not screen on or screen off); or so it would seem.
Now that I understand this quirk I can avoid it though I tend to store the screen inward to protect it and keep my cheek prints off of it 8-).