I guess a small nuance of Canon MTF charts should be clarified. I mentioned that they depict a lot of information, both at maximum aperture as well as f/8. It is a bit confusing, but at f/8, the maximum resolution is limited by diffraction at that aperture. At f/8, while maximum resolution is limited, it produces an 86% accurate reproduction of the test chart, where as at f/1.4, where maximum resolution is incredibly high, it produces only a 50% accurate reproduction of the test chart. The reason for the lower reproduction accuracy at f/1.4 is due to the presence of significant, uncorrected optical aberrations. As such, you are effectively getting LESS than the 86 lp/mm maximum resolution possible at f/8, and far less than the 494 lp/mm maximum resolution possible at f/1.4.
Additionally, for most lenses, it usually only takes a couple full stops down on the aperture to eliminate optical aberrations as the most significant resolution-diminishing factor, leaving diffraction as the only remaining detractor. That would be why f/4 exhibits more resolution than f/8, since the maximum diffraction-limited resolution at f/4 is 173 lp/mm, where as the maximum diffraction-limited resolution at f/8 is 86 lp/mm.
In light of this, I may be abusing the Canon MTF chart to depict "approximate resolution" at any given aperture, and I'm not really sure thats possible. At f/4 for the 50/1.4, its apparent that you get the best scene replication possible at that aperture, as that seems to be where the effects of diffraction and aberration normalize. At best, you could resolve the maximum amount of detail theoretically possible by whatever given sensor you are using (i.e. 116 lp/mm for the Canon 7D), however in reality you will see some diminished accuracy due to the low-pass filter and the nature of bayer sensors. In all honesty, I cannot accurately claim any resolution above what you could get at f/4, as at any wider or narrower aperture, resolution will fall off only semi-predictably.