You're right about the journal paper, but the pop-sci title implies causation. So, for those who didn't click all the way through: There is zero proof of causation. It may well be that people getting dementia have increased problems sleeping. Or both may be caused by some third factor.