September 16, 2013 | 1
On the heels of the decade of the brain and the development of neuroimaging, it is nearly impossible to open a science magazine or walk through a bookstore without encountering images of the human brain. As prominent neuroscientist, Martha Farah, remarked “Brain images are the scientific icon of our age, replacing Bohr’s planetary atom as the symbol of science”.
The rapid rise to prominence of cognitive neuroscience has been accompanied by an equally swift rise in practitioners and snake oil salesmen who make promises that neuroimaging cannot yet deliver. Critics inside and outside of the discipline have both been swift to condemn sloppy claims that MRI can tell us who we plan to vote for, if we love our iPhones, and why we believe in God. Yet, the constant parade of overtrumped results has lead to the rise of “The new neuro-skeptics” who argue that neuroscience is either unable to answer the interesting questions, or worse, that scientists have simply been seduced by the flickering lights of the brain.
The notion that MRI images have attained an undue influence over scientists, granting agencies, and the public gained traction in 2008 when psychologists David McCabe and Alan Castel published a paper showing that brain images could be used to deceive. In a series of experiments, they found that Colorado State University undergraduates rated descriptions of scientific studies higher in scientific reasoning if they were accompanied by a 3-D image of the brain (see Figure), rather than a mere bar graph or a topographic map of brain activity on the scalp (presumably from electroencephalography).
Critics of cognitive neuroscience have largely assumed that the pretty images which persuaded McCabe and Castel’s naïve participants have also seduced academics, journalists, and policy makers. Researchers in fields ranging from psychology to English literature were lured, so the argument goes, into using an extravagant research tool that has not advanced their disciplines in meaningful ways. These claims have hardly been limited to snide remarks over drinks at academic conferences, the McCabe and Castel paper has been cited several hundred times in scientific papers and been used to discount the scientific value of neuroimaging.
Some neuroscientists have started to push back. A recent critical review suggests that McCabe and Castel may have gotten it wrong—that brain images possess little-to-no special persuasive power. Most systematically, a series of 10 experiments—with over 2000 subjects—designed to replicate the original experiments found that brain images “exerted little to no influence”. Likewise, a pattern of failed replications in other labs suggests that the effect of brain images in persuasion may actually be trivial. A popular blogger, the Neuroskeptic, has argued that critics of brain imaging may be themselves the victims of another kind of seductive allure, “the allure of that which confirms what we already thought we knew.”
So what are we to believe: Are brain images persuasive or not? Although the idea seems intuitively appealing, the data is decidedly mixed. It seems like a puzzle that will continue to spark debate and research. However, we suspect it’s a puzzle that was actually solved in 1980—decades before cognitive neuroscientists began using MRI.
In the 1960 and 70’s, psychologists studying persuasion confronted a morass of conflicting findings. Some studies would find, for example, that peripheral aspects of a message such as the attractiveness of a speaker were more persuasive than the actual arguments. Use a sexy model, and it doesn’t matter what he or she says – an adage that explains 98% of beer advertising. But then other studies would find the opposite. Sometimes it was the central aspects of a message – the power of its arguments – that mattered, while people remained unfazed by attractiveness or other tangential cues. There was, in the words of one psychologist, “reigning confusion in the area”.
For many years, debate raged about which findings were “true” and which were “false”. But poring over the literature, two young psychologists, Rich Petty and John Cacioppo, realized that far more informative questions were instead, “when does each pattern hold, and why?” Asking the right questions instead of seeking the “correct” answer ultimately led Petty and Cacioppo to develop a theory—the Elaboration Likelihood Model—that resolved apparent contradictions in the literature and fundamentally changed the science of persuasion.
The Elaboration Likelihood Model (ELM) posits that when people are motivated and able to carefully evaluate messages they tend to be persuaded by central aspects of a message (e.g., the strength of its arguments). In contrast, when they are not motivated or able to elaborate, they tend to be persuaded by more peripheral aspects of a message (e.g., the attractiveness or professional credentials of a speaker). Groundbreaking work by psychologist Shelly Chaiken and others emerged around the same time making similar claims. These theories were a watershed moment in the field, and have been proven to be hugely powerful and influential frameworks for understanding persuasion across a variety of contexts and fields. These papers have been cited several thousand times over the past few decades.
It is, for this reason, remarkable that the recent articles on the persuasive power of brain images do not draw upon these highly influential models of persuasion. If they did, the question about whether or not brain images are persuasive would likely be discarded in favor of more informative questions. For instance, the persuasive value of a brain image very likely depends on whether or not the image is central or peripheral to the message, and whether or not the audience is motivated or able to elaborate on the message. These models would predict that scientists and other critical consumers would be the least likely to find brain images persuasive, unless they had a direct bearing on the research question at hand (e.g., when the images are being used to make an argument for spatial associations or dissociations). There are exceptions, of course. But even many of these exceptions have been well mapped out by persuasion researchers.
From a psychological perspective, we have much to gain from using these well-developed theories to understand responses to particular phenomenon, like brain images. More generally, we suggest that debates of this nature—does X influence Y?—are not as useful as theoretical work designed to understand how or why X influences Y. “Do brain images seduce?” is not the right question. “When and why might they seduce?” is. Armed with a theory about underlying psychological processes, one can generate nuanced hypotheses about when specific effects are likely to occur. People are more influenced by peripheral cues when they don’t or can’t think carefully about persuasive messages. For that reason it’s probably a good idea to use sexy models in ads that will air late at night or when people are already drunk. And your sexy picture of a brain is more likely to persuade an audience of non-experts or of hung-over experts on the last morning of the conference.
Rather than documenting the veracity or effect size of particular effects, this type of theoretical work is our central task as scientists of the mind and brain. If we understand how the mind works, then we can predict how it will behave in particular circumstances. And if the history of persuasion research is any guide, asking the right questions is likely to be bear more fruit than trying to prove that brain images are (or are not) seductive.
Images: Figure from the paper by McCabe & Castel (2008), Cognition. Warhol Brains © 2012 Jay Van Bavel, New York University – All Rights Reserved, created in collaboration with Crossroads Creative.