“modest, but I would say somewhat important association of short sleep and dementia risk”

Link. We need animal models, there is no way to figure out the causality in humans.