Neuropsychological Deficits: Fact And Artifact About Marijuana Tests
A new study claims to show small deficits on neuropsychological tests in college students who started smoking marijuana early in life. It might get a lot of press. Prohibitionists love to bang the drum of marijuana-related cognitive deficits, so I’d like NORMLites to know how to make sense of this sort of research. The recurring themes in this literature involve several alternative explanations that never seem to dawn on journalists. These results often arise from artifacts of the study rather than physiological effects of the plant. I’d like to focus on a few: other drug use, dozens of statistical tests, the incentives for performance, and the demands communicated by the experimenters.
The latest paper of this type is actually pretty good. Researchers studied over 30 people aged 18-20 who started using before age 17 (their average starting age was around 15) and who smoked at least 5 days per week for at least a year. They compared them to a comparable bunch of non-users. I hate to see 15-year-olds using anything psychoactive, even caffeine. Spending full days in high school with less than optimal memory functioning is no way to lay the groundwork for a superb life. I admit that I want these same people to grow up and be the next generation of activists, so feel free to call me selfish when I emphasize NORML’s consistent message: THE PLANT IS NOT FOR KIDS WHO LACK MEDICAL NECESSITY.
OTHER DRUG USE?
First, we have to keep other drug use in mind. Unfortunately, the marijuana group in this study got drunk more than 4 times as much in the last six months as the controls. Given what we know about binge drinking and neuropsychological functioning, it’s going to be hard to attribute any differences between these groups to the plant. It’s just as likely that any deficits stem from pounding beers. Studying cannabis users who aren’t so involved with alcohol would help address neuropsychological functioning much better.
HOW MANY TESTS?
In addition, we should always consider the number of measures in any study. Many of these neuropsychological tasks have multiple trials that can be scored multiple ways. The more statistical tests you run, the more likely it is that you’ll find a statistically significant difference by chance. It’s kind of like flipping coins. It’s rare to flip four heads in a row. But if you flip a coin a thousand times, odds are high that somewhere in the list of a thousand results will be four heads in a row. These investigators got 48 different test scores out of the participants. You’d expect at least 2 of them to be significant just by chance. They found differences on 14 different scores, suggesting that something’s going on, but we’re not sure which results are the “real” differences and which ones arose by accident. (That’s why we replicate studies like this.) And, as I mentioned, it might all be because of the booze.
WHY WOULD ANYONE DO ALL THESE TESTS?
We also have to consider incentives for performance. Most researchers bring participants to the lab for a fixed fee and ask them to crank out a bunch of crazy puzzles and memory assessments. It’s unclear why people would feel compelled to strain their brains. The authors of this study were kind enough to mention some relevant work by my friend (and former student) Dr. Rayna Macher. Dr. Macher showed that cannabis users respond best when you make the effort worth their while. She focused on people who used the plant at least four times per week for a year or more. She read one group some standard instructions for a memory test. The other group got the regular instructions plus an additional sentence: “It is important that you try your very best on these tasks, because this research will be used to support legislation on marijuana policy.”
As you’d guess, this simple sentence fired them up. Compared to cannabis users who didn’t hear that sentence, they performed better on 3 out of 10 measures. (You’d expect less than one difference by chance.) And compared to the non-users, the folks who got the incentive sentence did just as well on all the tests. For those who didn’t hear the incentive sentence, users did less well than non-users on 1 of the 10.
I know that prohibitionists are going to try to call this amotivation. (See my rant on that when you get a chance) I call it putting effort where it pays. But given what we know about how these studies can hamper the reform of marijuana laws, users everywhere should do their best on all tests whenever they get the chance.
WE OFTEN DO WHAT EXPERIMENTERS EXPECT OF US
Last but not least, we have to consider the demands communicated by the experimenter. Decades of data now support the idea that people often do what others expect them to do, especially if they believe the expectation, too. Another friend and former student, Dr. Alison Looby De Young, showed that these expectations are critical in studies of neuropsychological performance and cannabis. She gave a neuropsychological battery to men who had used cannabis at least three times per week for the last two years. One group of men read instructions that said that cannabis had no impact on their performance on these tests. Another group read instructions that said that cannabis was going to make them perform poorly. You guessed it, those men who heard they were going to flub the tests performed worse on 2 of the 4 tests. (You’d expect less than one difference by chance). As you might imagine, some laboratories communicate their expectations about cannabis and cognitive function subtly or not so subtly. Some participants are bound to behave accordingly. So what looks like a cognitive deficit is just an artifact of the laboratory environment where experimenters stare daggers at cannabis users.
In the end, I’m glad that researchers do this work, but these effects are too small and fleeting to justify prohibition. We already know that cannabis isn’t for healthy kids. People who get heavily involved with the plant early in life might not perform as well as those who never touch cannabis even if investigators control for other drug use, AND use a sensible number of tests, AND provide appropriate incentives, AND communicate a reasonable expectation.
But how many people should go to jail for that?
If you said, “None,” you’ve done an excellent job on an important cognitive test.