If you wanted to measure performance variability, you could do it with reaction tests. I would say that a lot of the reaction tests are static interactions. So it's, "When you see the yellow triangle, press the button as soon as possible." They're not dynamic interactions. Most of what we do on a daily basis are dynamic interactions. Hearing somebody speaking is a dynamic cadent interaction. We have to keep up with the words. It's not just one word sitting there and you say, "Okay, what's that word?" We don't do static interactions; we do very dynamic interactions. You've got to predict in those interactions. So that's one thing. The eye tracking is really a continuous dynamic test of attention. The other thing is you do collect-- because of this camera and the eye position, you can collect many, many data points within a second-- hundreds, maybe thousands of data points within a second of eye tracking depending upon how fast the camera is. In the reaction time test you only collect a few data points, and so you have to go to 20, 30 minutes. The other problem is that reaction times are effort-related. So if I don't feel like performing well on a neurocognitive test, I just delay my reaction times, whereas the eye tracking you're either eye tracking or you're not. There's no effort issues. You're either following the red ball going around or you're not. So when we do test-retest reliability, we look at-- for instance, we've done this in soldiers; we test them and then we bring them back 2 weeks later and test them again-- there's a very high test-retest reliability. So it says basically that you can take the test and if there's any real change, it's really because of the neurology and not so much because of the testing itself. So I think the test-retest reliability, I think it's a very quick test. The actual test itself is 30 seconds. Because you get a lot of data points within a short period of time, you're looking at continuous performance tasks for attention, I think that makes it-- Now, you could say that there are other parts of attention. People have a problem knowing what a yellow triangle is. People may have trouble with memory. They may be blind, they may have motor problems. All those things have to be assessed as well. And I would say that these are just measuring certain parameters. We're measuring a continuous visual attention test. If you want to measure something else, then you use some other test. I think the idea that there's going to be one test that's going to be used solely for coming up with a diagnosis of concussion or mTBI is ludicrous. We don't do that in medicine. If you look at diagnoses, they're based upon the history, they're based upon imaging, they're based upon symptoms, quantitative testing, and so on. You bring all that together and you come up with a diagnosis. So I think we're looking at--and the military, the Defense Department, is certainly looking at--from this perspective is that there's not going to be just one test we're going to be using. We're going to be looking at multiple different parameters and looking at a basket of tests and from that, over time, see how they relate to each other in terms of producing diagnostic criteria that both have immediate clinical utility but also in terms of prognosis as well.