I generally side with those that say that increased cycling infrastructure does not significantly increase number of cyclists, in general and in the big picture, i.e. national statistics.
Although, as in real estate, it's all about location, location, location. Here in Florida we don't have much in the way of cycling infrastructure, although I'm seeing more bike lanes built here and there. However, my family is from Maryland just outside D.C. and I've seen the cycling infrastructure built up there, including a MUP placed on the new Woodrow Wilson bridge, which is part of the interstate system.
I do see more cyclists now in that area, but many of these cyclists I believe you must be very careful in counting them as an increase; simply because many of them transit via car to a point to ride their bikes. However, you do see an increase in the number of commuters, but this is where location, location, location comes into play. Even if I were not a cyclists I could imagine myself wanting to ride a bike into D.C. (or any big city) as opposed to dealing with traffic and parking, which for many is a major problem, both in finding a spot (which many times is a good walk from your destination) and also costs, either in parking meters or parking lots; this cost can be significant.
So in these cases I definitely see how increased cycling infrastructure can attract more cyclists. However, beyond that I really don't think much of the projects justify the costs, but this must be looked at on a case-by-case basis.
I'm not against increased cycling infrastructe; I'm only saying you can't assume that if you build it they will come.