Um, this sounds like a couple of unrelated things.

First, there are a huge number of cues to depth, of which stereo is one, but not the only one (shown because pirates (Arr!) can see). You mention parallax (known in vision research by many names, but mostly some variation of structure from motion), which is actually very similar to stereo (If you think about it, the problem a monocular visual system solves combining an image at time N and time N+1 to infer the 3d structure of what it's looking at is very similar to the problem a 2-eyed system solves at any given time step (the big difference is that if both the world and the observer are moving, structure from motion can be more ambiguous, while if you know the geometric relationship between the two eyes, stereo's pretty well constrained)).

However, there are MANY other cues to depth, ranging from familiar size (imagine how much it would mess up your driving if scale-models of Hummers the size of bicycles became popular vehicles (ignore the serious engineering problems for now) - you'd have some real trouble knowing whether that was a big SUV far away, or a tiny car nearby (you'd be able to tell the difference using the distance between it and the visual horizon (meaning: the fact that it got way too low in your visual field), but that's just because you know your own observation height, as well (another cue)). The ability of a single eye to focus is also a cue to distance (Called "accomodation"), but it's useful range is limited to a few meters out. There are other cues as well (Palmer's late 90's textbook "Vision Science" has a nice, and pretty complete list of them (I had to enumerate them for part of my undergrad thesis)).

As I understand it, stereo isn't all that important for driving, anyhow (your eyes are pretty close together, so as objects get further away, the ability to use stereo to distinguish a given distance offset gets smaller pretty fast). Losing vision in one eye would probably have it's greatest effect on your field of view, at least for driving.

-Ross

RE: your first comment about eye strain, note that the comparison is between "action" video games requiring hand-eye coordination, and non-action video games. This way, the researchers controlled for much of the eye strain (unless doing active visual search for evil soldiers causes more strain than passively watching your little virtual people do little virtual things).

Vision-based puzzles and paining surely involve the visual system, and spatial reasoning, but neither requires the kind of serious on-line (meaning, immediate, time-sensitive) reasoning that shooting before you're shot does. The idea of this line of work is that the more immediate visuo-motor demands translate to improvements in visual sensitivity. A puzzle might require "visualizing" something in your head, as may painting, but neither requires something as immediate (which may well use different on-line mechanisms). As for violence, it's possible that the violent content lends extra immediacy or importance to the learning, if it's sufficiently realistic (The idea of a message from some other part of your brain to the parts to the on-line visuo-motor learning that your very survival depends on your learning this quickly), but I don't think that's been investigated yet (I talked to one of the grad students on this project, and he told me that the problem is controlling for how compelling the video game is (It's tough for a small research crew to make something less violent than, but otherwise identical to, a modern video game).

As for why contrast sensitivity might be important, it's not for tests that look like contrast sensitivity tests (shades of grey next to each other that you have to distinguish between). In the dark (for example) it's often very useful to distinguish one shade of almost-black from another. If you want to tell military camouflage from it's surroundings, you might be able to use it's very slightly different color from whatever it's hiding behind. If you wanted to tell if two pieces of clothing were the same, or almost the same tone, contrast sensitivty might help. There are plenty of other better examples, but I've gotta get back to work.

-Ross