Much of the brain's visual processing can change dynamically with changes in environment.
For example, a common experiment in college psych courses is to give a student glasses that flip the world upside-down. It takes a few days for the student's brain to adapt to the new inputs, and then they see the world normally (and revert after a few days w/o the glasses). Patients with macular degeneration can wear glasses that stretch-map the visual input around areas of missing vision (in the manner of a cylindrical mirror). After some time, they report seeing the world normally - their visual system has adapted and remapped the input.
I wonder if the effect simply measures the amount of reading the subject does; in other terms, perhaps it's just measuring the amount of fine-focus eye training? What does the test show for people who play a lot of arcade games (shooters, especially ones that throw a lot of targets at you)? Or people who use a lot of visual perception in their daily lives?
The article stated that the authors "tested for other possible explanations". Also, the correlation was at most 71%, note that flipping a coin is expected to correlate to around 50%. Their data seems to be awfully well clustered and the slope seems to be due to the outliers. The first study used 12 subjects, and the second only 53.
I'm unconvinced. It could be promising, but I would like to see correlations from more data.