As they repeated throughout our statistics classes, correlation does not prove causation.
The lead researcher of that study had previously found that middle-aged people working more than 55 hours a week tend to be disproportionately slow-witted, and to be more at risk for dementia.
So the question is, which causes which?
If you want something to be true hard enough, it comes true. Jeez, Chuck, what's all this statistics crap? Don't they teach you anything about cognition?