In an attempt to increase the “sexy” factor of brain imaging studies (what–you don’t think deoxygenated hemoglobin sounds sexy?), some press articles omit important details or subtleties that can distort, exaggerate, or even misrepresent the original results. To continue our series on brain imaging in the press (see Alex’s post on the lack of press articles featuring EEG), I came up with a list of four things you should know about press articles that feature functional magnetic resonance imaging (fMRI) research.
Much of this list is based off of a recent article written by my advisor, Diane Beck, in which she warns of the negative consequences of allowing the press to take liberties with fMRI results (Beck, 2010). She expands on many of the ideas I discuss here, but her message is primarily aimed at researchers using fMRI who may not always communicate with the press as clearly as they should. Meanwhile, the press is still happily buzzing about brain imaging (see “Part of Brain That Suppresses Instinct Identified”), so here are a few tips to help you get to the bottom of these fMRI press articles.
Yes, they’re pretty. But people are often so impressed with brain images featuring neural activation “hot spots” that they largely ignore the scientific reasoning behind the actual research (McCabe & Castel, 2008). In fact, you’re probably thinking I must be very smart right now because I have a picture of a brain in my post. It’s like those beer ads that show scantily clad women drinking the featured beer: they get your attention and maybe even convince you to buy their brand. But do the ads themselves tell you anything about the beer’s texture or flavoring? Definitely not. In the case of brain imaging, brain pictures speak louder than words, and that’s a bad thing if the science isn’t sound. Besides, showing a picture of where activity occurs in the brain probably isn’t going to help you understand the point of the study anyway. So ignore the shiny brain images and put on your critical thinking caps.
2) All fMRI studies are NOT created equal.
Just because a press article features brain imaging doesn’t mean that the study is inherently interesting. For example, Beck (2010) cites an article from The Guardian titled “Brain scans pinpoint how chocoholics are hooked,” featuring a study that found that people who find chocolate especially rewarding (a.k.a., chocoholics) have greater activity in their reward areas of the brain when seeing or tasting chocolate than those who aren’t as enthused by chocolate. News Flash! As it turns out, this study did not discover how chocoholics get hooked on chocolate; it only confirmed that reward responses to chocolate are associated with a person’s self-reported preference for chocolate.
Another category of press articles to watch out for are those that showcase fMRI studies that discover the neural correlates of some complex task without an overarching theory or framework–so called “fishing experiments.” For instance, we might ask, “Which brain regions are involved in social networking?” without any initial psychological theories of social networking. These types of studies typically result in a laundry list of brain regions that are more active for the task of interest (e.g., using Facebook) compared to a control task (e.g., online shopping). Let’s say our social networking experiment resulted in greater activity in several regions throughout the brain for Facebooking versus online shopping. We would then have to come up with possible explanations for the implications of each of these regions for social networking. For instance, we might find greater activity in areas associated with reward, social interaction and face processing; so our hypothetical experiment might confirm that social networking is more rewarding, social, and involves faces to a greater extent than online shopping, which means…what? It wouldn’t tell us anything interesting about what drives us to use Facebook. Even so, I’d bet that a headline like “Social Networking Activates Reward Circuits in the Brain” would catch your eye.
The bottom line is, fMRI studies are only as good as the theories or concepts that are being investigated. So always ask yourself two questions:
1) Does this article tell me more than what I probably already know?
2) Do the findings help explain a behavior or cognitive process?
3) fMRI images are not real-time snapshots.
I once saw an episode of “House” where the doctors were scanning the brain of a patient while she thought about emotional topics. As the patient lay in the scanner, the TV doctors were shocked to see that the patient’s language areas lit up while her emotion areas lay dormant–indicating that she could talk about emotional subjects without feeling strong emotions, i.e. symptoms of a psychopath. At first, I laughed at this unrealistic portrayal of fMRI; but then I thought, “Wait–is this really how people think brain imaging works?”
Well, in case you didn’t know, fMRI does not work by showing brain regions lighting up in real-time. On the contrary, fMRI data is typically only acquired every few seconds, and the analysis can take days due to the sheer amount of brain data that gets processed. Also, brain areas don’t just light up during a particular task; these activation “hot spots” are actually the result of a comparison of regional blood flow during a task of interest versus a control task. The basic idea is that the more active a brain area is, the more oxygen it requires, which is carried by the protein hemoglobin in blood. As a region uses up oxygen, hemoglobin becomes deoxygenated, and the MRI scanner measures the ratio of deoxygenated to oxygenated hemoglobin (also referred to as blood-oxygenation-level-dependent or BOLD signal).
By observing this BOLD signal over time, we can then compare which regions are more active for one task versus another. For example, take the hypothetical patterns of brain activity from our social networking study (i.e., greater activity in reward, social interaction and face processing areas). It wouldn’t necessarily be the case that these regions were only active while using Facebook, only that they were more active during Facebooking versus online shopping. If the press article actually mentions the comparison that was used to find an activated region, you can then decide for yourself if the researchers used an appropriate baseline task. For instance, how would we interpret our results if we had compared neural activity while using Facebook versus Twitter? What if we compared using Facebook to actually interacting with real live people? Each of these comparisons would answer a different question about the underlying cognitive processes that are associated with social networking. Thus, a comparison of two tasks is always necessary for getting a relative measure of activity in any brain region, and an appropriate comparison is needed to substantiate any claims about how the mind works.
4) Be wary of claims that neuroscientists can “read your mind.”
Many press articles use this sort of phrase when they feature studies that analyze fMRI data using pattern recognition. One of the latest trends in fMRI research, pattern recognition is generally used to analyze large-scale patterns of neural activity during various tasks. For instance, this technique was recently used by our lab to show that specific patterns of activity in visual cortex are associated with distinct natural scene categories such as mountains and beaches (Bernhardt-Walther et al., 2009). By training a computer to learn the activation patterns that are elicited by a particular task or condition (e.g., scene category), you can then test the computer by inputting new neural data and asking, “Which condition (scene category) was the person viewing based on this never-before-seen neural data?” The output is then the computer’s best guess based on how well it learned to associate a given pattern of activity with a specific condition.
While this technique is pretty cool, it is not really mind-reading. To actually read your mind, neuroscientists would have to guess what you’re thinking about by first determining the specific patterns of activity that are associated with every possible thing you could think of. As you might have guessed, the science just isn’t at that point yet.
Well, that turned out to be four lengthy things to keep in mind about fMRI press articles. Hopefully with the combined efforts of responsible researchers, careful journalism and educated readers, we can improve the way that fMRI results are communicated to the general public. Happy reading!
Diane M. Beck (2010). The appeal of the brain in the popular press. Perspectives on Psychological Science : 10.1177/1745691610388779
Walther DB, Caddigan E, Fei-Fei L, & Beck DM (2009). Natural scene categories revealed in distributed patterns of activity in the human brain. The Journal of neuroscience : the official journal of the Society for Neuroscience, 29 (34), 10573-81 PMID: 19710310
McCabe DP, & Castel AD (2008). Seeing is believing: the effect of brain images on judgments of scientific reasoning. Cognition, 107 (1), 343-52 PMID: 17803985