The studies tested maximum speed at 35m, not time to 35m. It was thus a test of maxV. Moreover, the 2006 study had an uphill-downhill, downhill only, uphill only and horizontal only group, and found the greatest improvement in the uphill-downhill group (i.e. contrast training) followed by the downhill only (i.e. overspeed training) group. Since the uphill group did not improve, the improvements were not due to the uphill portion of the combined group.
Forum members such as lkh and UKcheetah have seen substantial improvements with contrast workouts involving overspeed. Charlie has mentioned that, while not generally a safe training method, it may be a good option for someone who has to catch up on missed speed work. He has also described how athletes often achieve a certain wind assisted time earlier in a season, which then sets them up neurologically for running the same time under legal conditions a little later.
I’m not advocating the regular use of overspeed training by the way, just pointing out that the statement that overspeed training has no scientific findings behind it is incorrect.