I have not read the full text of the article, so I cannot comment on it beyond what is presented in the abstract. However, the increase in gray matter volume is the observation for which the authors claim the lowest P-value. For those who are not familiar with it, the P-value is a commonly abused test for statistical significance. If we assume the null hypothesis to be that there really is no difference between the brains of the twins, the low P-value would help to eliminate the null hypothesis as true. According to the numbers the authors provided, there appears to be a real difference in their gray matter volumes. The idea that exercise has an effect on the brain is not a new one and, if done correctly, perhaps this study helps to support that. That physical activity appears to be good for us probably is not a big surprise, either.
Medical and statistical peeps may have things to add here.