For Mirror Neurons, Picture Grows Cloudier


by Aalok Mehta

July 20, 2009

In the early 1990s, neuroscientists chanced upon what many would come to consider one of the greatest brain discoveries of the century: the existence of mirror neurons, cells that activate both when an animal performs a specific movement and when it observes that action being performed by another.

Ever since Giacomo Rizzolatti and his Italian colleagues published their results, based on single-cell recordings in the motor cortex of monkeys, in the journal Brain, these unique cells have lit up the imaginations not just of scientists but also of the public. Theorists have speculated that the neurons help us understand and anticipate the actions of others and that they may even underlie high-level concepts such as empathy, language and art. Some experts have also linked mirror-neuron defects to autism and related disorders.

But a controversial new paper asserts that mirror neurons may not exist in humans, at least not in the same sense. Psychology professor Alfonso Caramazza of Harvard University and his colleagues used functional magnetic resonance imaging (fMRI) to test for the presence of the cells in healthy adults. The team employed the theory of adaptation—that the brain response to a given action decreases for a repeat occurrence, at least during a short window of time.

The results showed that adaptation occurred when people performed the same random action several times or when they saw the same action repeatedly. It also occurred when they saw an action and then performed it. But adaptation did not occur when people first performed an action and then saw it, demonstrating a mismatch in how action and observation affect the brain.

“What we found is that there is an asymmetry in adaptation. The signal decreases when you go from seeing to acting, but not when you go from acting to seeing,” he says. “I would argue that this is evidence against mirror neurons.” The paper was published online June 2 in the Proceedings of the National Academy of Sciences.

But other researchers in the field say the paper goes too far and criticize both the conclusions and the experimental setup. “In life science in general and in brain imaging in particular, a negative result can't possibly prove the nonexistence of a phenomenon. It is astonishing that Caramazza's paper makes those claims and that a high-profile journal like PNAS publishes it,” Marco Iacoboni, a professor of psychiatry and biobehavioral sciences and a mirror neuron expert at the University of California, Los Angeles, wrote in an e-mail.

“The Caramazza paper is seriously flawed,” he added, because of limitations in using fMRI to study adaptation. In particular, he said, the technique does not actually allow scientists to circumvent the spatial resolution limits of fMRI and determine the activity of small populations of neurons, like the mirror neurons mentioned in this study.

Direct observation of mirror neurons in humans has long been, since it’s impossible to do the kind of single-neuron electron recording common in monkey studies, says Ilan Dinstein, a postdoctoral researcher at the Weizmann Institute of Science in Israel and lead researcher on a similar study that found none of the three cases of adaptation reported in Caramazza’s work.

A few scientists have watched for neural responses that might be expected from mirror neurons during surgery with patients receiving therapeutic electrode implants. Otherwise, the standard methods are either to look at adaptation or analyze the patterns of brain activity using fMRI. Both techniques are limited and assess relatively large areas of the brain instead of anything close to individual cells. For instance, Dinstein adds, Caramazza and his colleagues might have missed the small window of timing in which adaptation occurs, or movement may have a longer delay than observation, throwing off the results.

In fact, Dinstein says, the data could support the theory of mirror neurons as easily as detract from it. “You can spin this either way,” he says. “They found three of the four cases they were looking for. They’re only missing the final criterion, and we feel strongly that this is not the critical condition. … They say the cup is mostly empty; we say the cup is three-quarters full. I’m surprised they emphasized the part that is missing.”

Still, at least some scientists accept that the work might suggest an altered role or reduced prominence for mirror neurons. “The fact that they didn’t find adaptation doesn’t mean it didn’t happen. It’s hard to say,” says Lindsay Oberman, a postdoctoral research fellow at the Beth Israel Deaconess Medical Center who uses transcranial magnetic stimulation techniques to study mirror neurons. “But I’m willing to be open to the possibility that they might not work in the exact [same] sense in humans as in monkeys.”

That’s the kind of healthy skepticism Caramazza is hoping for, given the scope of claims made about mirror neurons. “I have no doubt that there are cells that respond both to executing a movement and observing a movement, but the mirror neurons may be activating by recognition. They may not be playing a causal role,” he says. “We need to study in great detail what these neurons do, because the stakes are so high.”