That's a damn weird correlation. I wonder what causes it, given that so many factors might affect the post-stroke neurologic outcome, potentially unknown and thus not adjusted by the researchers.
I'm not too informed on USA and Mexicans, but my first bet would be non-linguistic in nature: food. I wonder if language change isn't correlated to dietary changes (both change as the individual shifts culture), and that those dietary changes might increase the likelihood of a favourable outcome.