High Minded Science


By Alex Berardino


Words are important, especially those spoken by authorities on medical science. That’s why it was so disappointing to see so many loose and unfounded words being tossed around in the popular press in regards to a paper by Gilman et al. in the Journal of Neuroscience. The paper presents potential differences in the shape, size and density of specific brain areas, the Amygdala and Nucleus Accumbens, between people who use marijuana moderately, and those who don’t use it at all. These brain areas have been shown to be important for reward processing and are also implicated in addiction. The paper also showed a weak correlation between amount of marijuana use and the size of these differences. Notice that I specifically said differences between these groups, and correlation between these measures. The same care and effort was not taken in most reports about this paper. These results were reported as though they showed changes in the size of structures within the marijuana user’s brains caused by the amount of marijuana that they used, and subsequently that these early changes led to the forgetfulness and lack of focus common to long-term, dependent marijuana users. To be fair to the journalists, the paper and its authors are not exactly clear on which interpretation they subscribe to.


Perhaps the importance of the difference between these two interpretations is not clear, but the differences are vital to how we are to act based on this new information. In a time when our society is experimenting with the legalization of marijuana, it is important that we are accurately informed about the actual dangers and possible benefits of the drug and its effects on the brain. Before we break down the difference in interpretation, let’s first break down what the scientists actually did.


The authors collected structural MRI scans of 20 moderate marijuana users, and 20 controls matched for age, sex and educational attainment. Structural MRI scans show a snapshot of the tissue that makes up the brain, which allows you to see the shape and size of the structures that comprise it. You can think of them like a picture of the brain itself, not of its activity.   Despite many similarities, everyone’s brain is a little bit different. The sizes vary, the folds and protrusions don’t match up perfectly from person to person, and the borders between one area and another are not perfectly consistent. When we conduct studies like the one cited here, where we want to find variations across people’s brains, we have to first warp and squeeze each individual MRI scan so that they fit, as well as they can, inside a template brain scan. The template itself is generated by averaging across a large set of brain scans, so that what is left over is assumed to be a “representative” brain. That is exactly what the authors of this study did. They warped, or registered, each individual brain onto this same template. After doing so, they could compare any remaining variations that exist across the two groups.


The authors determined the average sizes, shapes and density of neurons of the Amygdala and Nucleus Accumbens of the moderate marijuana users and compared this to those of the controls. The strongest finding of the paper shows that the left Nucleus Accumbens has, on average, larger volume, and is denser, in marijuana users than in controls. The differences are reported as significant, but the standard errors are quite large and the distributions overlap by a large amount, suggesting that the differences are not entirely reliable.


It behooves us to stop and ask ourselves just exactly what it means that this area has larger volume and is more dense with neurons. Does it mean that the area should show increased performance, or decreased? Is the architecture of this area scrambled, or arranged in an orderly fashion like that of the controls? The truthful answer is that we don’t know.   There is no good accepted answer for what these differences mean, because the true machinery of the brain is built at a scale that is too fine for an MRI to resolve. That’s not to say that the finding isn’t important, just that we should be cautious in our interpretation of it.


Next, the authors binned the members of the marijuana group into subgroups by how many joints per day each member smoked. They report finding a correlation between number of joints smoked, and the size of the difference in the left Nucleus Accumbens. This is the tricky part though, because our algorithms for registering these brains to the same template brain, rely on identifying the very neural landmarks that vary from person to person, and can, by pure bad luck, be obscured or imprecisely located by poor resolution of the MRI. To account for any possible variations caused by errors in registration, or in misidentification of the boundaries of areas, its important to look at variations averaged across many subjects. This helps to ensure that the only variations that remain are those that are present across the whole set of brains, and not those due to these errors. Comparisons across small subsamples of these groups are generally unreliable.


This is important to keep in mind because this trend, between number of joints and size of change, is used as a justification for suggesting that marijuana is causally changing these areas. The fact remains that this study measured differences between groups. It was not a longitudinal study of one group of people who did not use marijuana and then began to use marijuana. Interpreting these data as showing that marijuana use is causing these changes, and not simply correlated with differences in this region, is a difficult interpretation to support. We hear it all the time, correlation is not causation, but it is an important point.


Leaving aside critiques of the science, unsubstantiated extrapolations of misinterpretations pepper many of the articles reporting on this paper. One of the authors of the study, Dr. Hans Breiter, was quoted in the Huffington Post saying, “We think [sic] we are seeing here is a very early indication of what becomes a problem later on with prolonged use, things like lack of focus and impaired judgment”. These are bold claims, unsubstantiated by anything presented in the paper. No attempt was made to show any behavioral evidence suggesting a tendency toward lack of focus or impaired judgment between the groups, or for the members of the marijuana group across time, nor were the results presented so impressive that they suggest major rewiring of the brain after exposure to marijuana.


It is important on a basic level to be wary of making claims we can’t support with evidence. Neuroskepticism is a widespread movement these days. Distrust for institutional authority doesn’t come from thin air. It comes from a sense that information is being withheld, or warped to fit an agenda, or handled by people who are incapable of handling it. Sometimes these claims are founded in truth, sometimes in conspiracy and misunderstanding. The former, at least, we can control by maintaining a strict relationship between evidence and statements of authority.