Not sure about this exact study, but studies like this usually ignore confounding factors that cause people to abstain. The headline leads people to think that they should drink a bit to achieve the result, but that’s usually never the case.

For example there were a bunch of studies about how sleeping 10 hours per night has lower mortality rate than sleeping 12+ hours (oh no - are you sleeping too much?!?), but upon further review they realized that people who are really sick or alcoholics are more likely to sleep 12+ hours per night.

In this case, wouldn’t be surprised if there are reasons people fit into the zero drink category (health issues) that are associated with cognitive decline.