Mirror neurons seemed like a revelation when they were first discovered—cells that light up both when you move to grasp an object and when you see someone else do it. Theories about the importance of mirror neurons began to proliferate, faster than the evidence accumulated, and now University of California Irvine cognitive scientist Gregory Hickok has taken it upon himself to throw cold water on all this overexuberance in his new book, The Myth of Mirror Neurons: The Real Neuroscience of Communication and Cognition. We chatted with Hickok recently about what the scientific evidence says—and doesn’t say—about these fascinating cells.
(Note: This interview has been lightly edited for clarity and length.)
WSF: Why do you think the mirror neuron theory became so widely accepted?
GH: Mirror neurons came along and seemed to offer a simple explanation for complex cognitive problems—stuff like language, empathy, and disorders like autism. We just simulate actions, and that’s how we understand them; and you can extend that to language, and other people’s mental processes.
WSF: But you were skeptical almost from the beginning, right?
GH: Yeah. I’d heard about them toward the end of the 1990s, and like everyone else I was really interested. But, being a speech researcher, I knew that damage to the motor speech system didn’t cause speech perception problems—There’s this famous syndrome called Broca’s aphasia, defined by patients who have difficulty generating speech but have little difficulty comprehending speech. So in the language domain, the theory just didn’t hold any water.
Nonetheless, the theory spread into the language domain, so I had to deal with it, and that pushed me to go back and look at the foundational reports in monkeys.
WSF: What sort of functions do we know that mirror neurons have in monkeys?
GH: Mirror neurons were initially discovered in research on motor control—and it was and remains beautiful work. The University of Parma group that found them was interested in whether information on object shape can get integrated with your grasping actions. So, if you’re reaching for a cell phone versus a cup on your desk, the shape and location are going to change the grasp you use. This group had found cells in the motor cortex of the macaque that would respond both during observation of objects and reaching for objects. They noticed, between trials, that as the experimenter reached into the display box the monkey was looking at—to place an object or pull one out—some of the cells they were recording were firing while the monkey was watching these actions.
WSF: Is there an evolutionary advantage to having your brain cells fire when you see another monkey (or person) grabbing something?
GH: The actions of other animals are clearly relevant to our own actions. Think about a monkey in the wild; they need to pay attention to what other monkeys are doing—if they’re being threatened, or grooming, or searching for food. Watching kind of determines your own action selection.
What was puzzling about mirror neurons was they showed a precise correspondence between the observed action and the motor preference of the same cells. But macaques tend not to imitate in that precise a way. So what behavior in the macaques could this class of cells support? That’s what led to the action understanding idea, because they didn’t have another explanation.
WSF: Has anyone directly observed this mirror neuron behavior in humans?
GH: Direct evidence is still pretty sparse, and that was a big research effort for awhile.
I have never questioned the existence of these cells or cells like these in humans, though. Humans have the ability to imitate actions that other people generate, even if they’re nonsense motions. So there must be a way of perceiving movements and then reproducing them. In order to do that, you need some sort of system that translates between the sensory representation and the motor representation.
WSF: Is there any evidence that would lead to you to revise your opinion of mirror neurons?
GH: The evidence that we haven’t seen is that damaging the mirror system could cause deficits in understanding. All the data we have from monkeys is essentially correlative: the monkey is observing an action, and the cell fires. But that doesn’t tell us if the cell is critical to that understanding. The simplest, most direct demonstration is to disrupt the cells’ function and see if the monkeys have deficits in understanding. Those sorts of experiments have not been done. The Parma group suggested they’re impossible to do, because the mirror system involves a wide number of brain areas, and if you damage that much, you induce cognitive deficits. I’m not convinced that that’s the case, but that’s their claim.
The evidence in humans has pretty much supported the view that it is quite possible to severely disrupt the motor system without necessarily causing deficits in understanding actions. In order to show that action understanding is correct, you’d want to find examples of individuals who have lost the ability to speak, and correspondingly, can’t understand speech. Sometimes those go together, but more often than not they dissociate. People with cerebral palsy can have severe speech production problems, but it’s been shown they can understand words just as well as anybody.
WSF: So, why a whole book on the myth of mirror neurons?
GH: The mirror neuron enterprise hasn’t seemed to lose all that much steam despite the fact that several of us have been pointing at evidence contradicting the claims. Since it’s working its way into other areas of research beyond my own, I thought it would be important to lay all the evidence out.
Autism, for example, has been claimed to be a result of “broken mirrors,” some kind of mirror neuron disruption. If theory that isn’t quite solid is being used to drive research, it’s useful to set the record straight.
Check out this excerpt from The Myth of Mirror Neurons, on the potential disadvantages of mirroring:
Mirror neuron theorists tout the virtue of “direct” understanding via motor stimulation without the need for complex cognitive inference. For example, if you observe a friend reaching toward a raisin, there is no need to run through the logical possibilities—Ally likes raisins, Ally is hungry, Ally is probably reaching for the raisin to eat it; by simulating her movement, you understand it “directly” and “automatically.” This might be considered a theoretical plus since explaining complex things using simple mechanisms is, after all, a good thing, but it also has potentially catastrophic consequences. If we had to simulate the actions of others in our own motor system in order to understand them, this could interfere with nonmirror actions that we may need to respond with.
Sports provide a good example. In many sports, athletes must respond to the actions of other players, typically in a nonmirrored fashion. A batter in a baseball game must prepare a swing in response to observing a pitch. A boxer must duck or block an incoming punch. A goalkeeper must lunge or jump upon seeing a kick. Examples like this abound and in most of them, reaction time is the key to success. Now, if the batter, the boxer, and the goalkeeper had to first simulate the actions he or she observed, this would tend to slow reaction times because the motor system would have to activate a simulated program that competes and interferes with the one required to successfully respond to the action. Mirroring is therefore potentially maladaptive in situations that require fast, nonmirror reactions in response to observed actions.
Excerpted from The Myth of Mirror Neurons: The Real Neuroscience of Communication and Cognition
by Gregory Hickok. Copyright © 2014 by Gregory Hickok. With permission of the publisher, W. W. Norton & Company, Inc. All rights reserved.