In a fit of recursion, I am going to begin my discussion of the scientific understanding of the mind by bringing up a piece of psychology research into how people perceive neuro-imaging. This not only gives a taste of what different types of research can be like, but reveals something rather disturbing: merely adding a brain scan image or two makes people more likely to rate an article as scientifically sound. This gets us into questions of what is and isn’t a good reason to believe any particular research conclusion, which is ultimately what I want to talk about in this series of articles.

At the present time there are basically two technologies that can give us some idea of the activity of a working brain: positron-emission topography (PET) and functional magnetic-resonance imaging (fMRI). They both have important limitations in terms of resolution, what they actually measure, and many other things besides, but they’re also pretty amazing technologies. They produce detailed 3D maps of the “activity” of a whole brain, which are often represented like this:

This is a 2D cross section of a living brain, where more “active” regions relative to a baseline scan are represented in red, while less “active” regions are blue. For fMRI, “active” means greater oxygen consumption, which is a measure of… well, neurons doing something. We don’t really know precisely what oxygen consumption means, and this is part of the problem with fMRI. Resolution is also a serious issue. A single fMRI pixel is represents a cube of tissue a few millimeters across, containing perhaps five million neurons, and the maximum imaging rate of one frame every few seconds is far too slow to investigate anything that happens quickly. There is a recent paper in the journal Nature which discusses all of this in painful depth, but one might imaginatively compare an fMRI image to an aerial photograph of a city: you can see some broad patterns, but basically no details. Actually, it’s worse that that: you may be able to make a traffic map, but you have no idea why the people are driving. Even so, it’s an extremely interesting technology, because it allows the first ever 3D “movies” of what parts of the brain might be involved in whatever that brain is doing at the time.

Actually, it’s a little too sexy. In the paper Seeing is believing: The effect of brain images on judgments of scientific reasoning, David P. McCabe and Alan D. Castel describe an experiment performed on 156 undergraduates at Colorado State University (everyone tests their theory first on the undergraduates; psychologically speaking, we now know a truly shocking amount about them.) First they wrote three bogus scientific articles, of the kind you might find in an online news report. These were titled “Meditation Enhances Creative Thought”, “Playing Video Games Benefits Attention”, and “Watching TV is Related to Math Ability”. Each claimed that neural imaging data showed a link between two different activities.

How can this be argued from fMRI results? The general form goes something like this:

1. Put volunteers in an fMRI machine and have them do some activity, like meditating.
2. Scan them and watch what regions of the brain show increased blood flow when they start meditating. Neural anatomy is complex, so there’s almost always a cool-sounding name for the region in question, such as the “right temporal lobe.”
3. Take some other people and put them in the fMRI machine and have them do something else, like solving logic puzzles or composing a sonnet.
4. See which bits of the brain seem to be using more oxygen this time around.
5. If some of the same bits of the brain are involved in both activities, suggest that there is some relationship between them.

The critical bit of reasoning of reasoning is step 5. Sure, meditation and creative thought may both use e.g. the right temporal lobe, but what have we actually learned in terms of cause and effect? We don’t get to decide upon the casual relationships of the universe; they’re already there, and the job of a scientist — or any sort of truth-seeker, really — is to try to figure out what reality has already laid down. (I’m going to ignore the solipsistic school of thought that says we create reality with our minds. I mean, I create reality with my mind too, but somewhere in between I use my hands, and those are physical objects.)

The (better) neuroscientists are aware of this, and in the (better) literature, you will only find statements of the sort “the right temporal lobe is associated with creative activity.” That’s it. Associated. If mediation is also “associated” with the right temporal lobe, does this mean meditation is “associated” with creative activity in the brain? Well, no. That’s like saying that cleaning your windshield and filling a test-tube are similar activities because both involve something made of glass.

And yet, these types of arguments are often convincing, when it comes to the study of the mind. Accordingly, the reasoning in all three of the bogus articles used in McCabe and Castel’s research was in fact wrong. Nonetheless, they asked each of their 156 subjects to read the articles and rate how much they agreed with the statement “the scientific reasoning in the article made sense.” They used a four four-point scale: “strongly disagree”, “disagree”, “agree”, and “strongly agree”, which gives numbers from 1-4.

They found that those who read articles with brain images gave an average rating of 2.90, approximately “agree,” and those who read articles without gave an average rating of 2.73, somewhat less. This was statistically significant to p<0.05, meaning that there was thought to be a 95% chance that this wasn’t simply a random fluke of measurement, like throwing down ten dice and having them all come up six. They also got all their control group and randomization procedures right. In a later article I’ll explain exactly what p values mean and why one has controls and randomizations at all, but for the moment the point is that, unlike the faulty reasoning far too common in discussions of fMRI results, this particular piece of psychology research might actually have teased apart a real casual relationship that exists inside people’s minds: adding an image of a brain scan to an article makes otherwise intelligent people far more likely to accept its fallacious reasoning.

Here’s a graph to convince you:

Why are people so easily mislead? Figuring out what is true turns out to be sort of difficult, of course, but there’s more too this whole question, reasons not simply academic. In a similar piece of research which tested the effect of neuroscience language on perceived credibility, Weisberg and colleagues write

Research on nonneural cognitive psychology does not seem to pique the public’s
interest in the same way [as nueroscience research], even though the two fields are concerned with similar questions.
…

Because articles in both the popular press and scientific journals often focus on how neuroscientific findings can help to explain human behavior, people’s fascination with cognitive neuroscience can be redescribed as people’s fascination with explanations involving a neuropsychological component.

In other words ultimately we all just want that feeling of understanding why. But feelings are not the same thing as reality, and that’s an important lesson about the mind too.