Alessandro Garofalo / Reuters

“I can’t even remember what happened that night” is a common joke/cry for help among people who recently drank to the point of blacking out. But there’s also evidence that drinking even a little bit can seriously impair learning and memory.

Sleep, especially the REM phase when dreams occur, is when memories get cemented into our minds. Alcohol blocks REM sleep, and as a result, says University of California, Berkeley, professor Matthew Walker, drinking can make you forget new information—even if the drinking happens days after the learning took place.

The most striking evidence of this phenomenon, which Walker describes in his recent book, Why We Sleep, is from a 2003 study by Carlyle Smith, a professor at Trent University in Canada. He invited 15 students to his lab and got them kind of drunk.

First, the participants were all (soberly) taught a new type of grammar system. They then were put into three groups: One drank a cocktail of orange juice with three ounces of vodka on the same night they learned the task; another group drank the cocktail two days later, but they had only orange juice the night of the learning and the following night; and a control group had no booze at all on any of the nights.

They were retested a few days later, when they were all sober. The group that drank alcohol on the same day of the learning session, perhaps expectedly, performed worse than the booze-free control group. The same-night drinkers forgot 50 percent of what they had learned. But surprisingly, the group that just had alcohol on the third night performed worse than the control group, too. Even though they had two restful, alcohol-free nights of sleep between the day they learned the task and the night they got slightly sloshed, they still forgot 40 percent of what they learned, Walker explains. That means if you’re in college, or in any environment where you’re learning new material, you can’t be angelic all week and get hammered on weekends without suffering the consequences. “The memories remain vulnerable and fragile to the impact of alcohol for days,” Walker told me.

Now, there are many caveats to this study. It only had 13 subjects, which would never fly in today’s trials, and it didn’t account for other things that could be affecting the participants’ memory. And Smith, though he’s still an emeritus professor of psychology, has gone on to study dream telepathy, a pretty fringe topic.

Smith acknowledges that the sample size was small, and that the dream-telepathy stuff is unorthodox. (“I don’t know what my colleagues think about that now,” he told me recently. “Some maybe avoid me in the halls.”)

But he stands by his findings on alcohol’s detrimental effects on memory. Alcohol can reduce the number of eye movements that take place during REM—spelled out, it’s called “rapid eye movement” sleep. “It’s a very sensitive state of sleep,” Smith said.

He also did another experiment that seemed to show that simply drinking right after you learned something didn’t impair memory. But drinking right before bed did. “You can drink alcohol, just be sure it has time to metabolize and get out of your system before you go to sleep,” he said. “Maybe you should drink in the afternoon.”

Phew, at least they’re not coming for our day drinking!

But wait.

Recently, Anya Topiwala, a senior clinical researcher in psychiatry at the University of Oxford, had been seeing some studies showing that light alcohol consumption supposedly protects against cognitive decline. But in the course of her practice, she was also seeing some older people with memory issues who drank moderately—say by sharing a bottle of wine at night.

She wanted to see whether the drinking was helping or hurting their brains. So she and her colleagues looked at 550 people who had been part of the famous Whitehall II study of the health of British civil servants. Most of the participants had been drinking moderately over the course of 30 years. Throughout that time, they had been questioned about how much they drink and given cognitive-functioning tests. Topiwala and her colleagues gave them some additional cognitive tests and scanned their brains.

They found that the more the participants drank, the smaller their hippocampus, an area of the brain responsible for memory, and the worse their neural “insulation” in another memory region called the corpus callosum. It didn’t help to drink “lightly,” or just a couple drinks a week. And those who drank the most—more than 21 alcohol units a week, or a pint of beer a day—had a 16 percent decline in their “lexical fluency,” or the number of words they could come up with off the top of their heads. And light drinkers weren't better than abstainers at lexical fluency, the study showed.

Topiwala and her colleagues found these effects even in men who drank a “moderate” amount, or between about five and seven pints of beer a week. That’s considered “14 to 21 units per week.” (Thought a “unit” was a single drink? That might be because wine glasses have gotten several times larger in recent years.)

Those studies that were finding light drinking was a good thing for the brain? They might have neglected to account for the fact that drinkers tend to be richer and better-educated than teetotalers, Topiwala said, or they might have been asking participants to recall how much they drank years ago—a somewhat unreliable measure.

Besides these two, other meta-analyses have found that young people who drink heavily are worse at remembering things like locking the door or mailing a letter than are abstainers or light drinkers. Another paper found that students who had about nine drinks a week, on average, already had a harder time with a word-related memory task. Another study has seemingly affirmed the sleep link, showing that college students who drink more sleep later, are sleepier in the daytime, and do worse on tests.

This is far from the sum total of alcohol’s negative effects—even just on sleep. Because alcohol blocks REM sleep so much, some alcoholics have their REM sleep “bleed over into their waking life,” Walker says. They start to have dreamlike delusions and hallucinations.

The problem is, it’s holiday-party season. It’s stressful, festive, and social—sometimes all at once. Although, of course, you don’t have to drink to socialize, celebrate, or control your stress. I recently put a call out on Twitter for stress-management advice from teetotalers, and they had lots of ideas. I also asked the researchers I interviewed how they’ve adjusted their behavior based on their findings.

Topiwala used to drink a glass of wine a day, but now she only has three or four a week, she said. Walker doesn’t drink in the evenings—but then again, he doesn’t drink at all. He also gives himself a “nonnegotiable sleep opportunity” of eight hours a night, even if it means sacrificing other things.

“Sleep is in a desperate state right now,” he said. But, “sleep is the best health-insurance policy you could imagine.”

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.