Note that the article uses "verbal memory and semantic fluency" as a proxy for general cognition, so readers will almost automatically score better than viewers. They also seem to have modeled all factors as linear, and "cannot rule out (...) the possibility of reverse causality".

So to answer your question: it depends on whether you consider all assumptions of the article relevant. One thing the article certainly doesn't contradict is the old adage "use it or lose it".