Much Ado about Mirror Neurons: Empathy, Autism, and Bias (Part 1)

by Anne Corwin

Mirror neurons are theorized to be, according to some of the more heavily popularized literature these days, neurons which activate in the primate brain upon observation of another individual performing an action.

The most frequently cited experiments suggesting the presence and functionality of mirror neurons have involved macaque monkeys—these monkeys were fitted with electrodes which allowed scientists to observe particular activity patterns in neurons in the premotor cortex.  The experiments sought to show a correlation between imitation, intention, and action that might shed light on the means by which primates (possibly including humans) may internally simulate the mental states of others.  But did it succeed?  Some might answer “yes”, and many since have attempted to correlate the mirror neuron experiment results with theories in autism research, since autism is commonly (though not necessarily accurately) associated with deficits in imitation, modeling, and empathy.

This discussion will attempt to explain the actual findings in monkey and human research associated with mirror neurons, and pose the argument that the mirror neuron studies being performed on human subjects do not necessarily imply for autism what much of contemporary popular science literature claims.  That is, while mirror neuron studies may indeed offer valuable information about brain differences and the various ways in which minds process and react to certain stimuli, these studies do not actually prove that the observed differences between autistic and nonautistic behavior and cognition are explained by a “dysfunctional” mirror neuron system.

Additionally, many recent publications linking autism and mirror neurons exhibit much in the way of biased language, faulty initial assumptions, and poor experimental design, which could in the long run end up serving as a detriment to both the autistic population (e.g., through the perpetuation of ideas suggesting that autistics lack some essential element of personhood) and the field of neuroscience (since good science must work tirelessly to avoid the kinds of bias that stem from prejudice and ignorance, such as that which resulted in “scientific” racism, which is now widely considered to be pseudoscientific and therefore a blemish on the history of science).

The neurons focused on in the macaque experiments are located in what is called the F5 area of the brain, and are thought to be associated with certain hand and mouth movements.  The paper Mirror neurons and the simulation theory of mind-reading, by Vittorio Gallese and Alvin Goldman, Trends In Cognitive Sciences – Vol. 2, No. 12, December 1998 explains the implications of the macaque experiment as follows:

These experiments showed that the activity of F5 neurons is correlated with specific hand and mouth motor acts and not with the execution of individual movements like contractions of individual muscle groups.  What makes a movement into a motor act is the presence of a goal. This distinction is very important since it allows one to interpret the role of the motor system not just in terms of the control of the dynamic variables of movement (like joint torques, etc.), but rather as a possible candidate for the instantiation of mental states such as purpose or intention.

The most profound implication here, according to researchers, is that the neurons in question might provide valuable clues about goal-setting and intention in response to observed movements, and by extension, the ability of primates to model the mental states of others.  While being aware that the macaques do not actually exhibit the sorts of behaviors generally associated with evidence of “theory of mind” in humans, nor tend toward imitating their peers, researchers nonetheless suspect that mirror neurons might still play a key role in human social communication. With regard to the possibility of mirror neurons existing in humans in the first place, Gallese and Goldman write:

Two lines of evidence strongly suggest that an action/observation matching system similar to that discovered in monkeys also exists in humans. The first refers to an elegant study by Fadiga et al.16 in which the excitability of the motor cortex of normal human subjects was tested by using Transcranic Magnetic Stimulation (TMS). The basic assumption underlying this experiment was the following. If the observation of actions activates the premotor cortex in humans, as it does in monkeys, this mirror effect should elicit an enhancement of the motor evoked potentials (MEPs) induced by TMS of the motor cortex, given its strong anatomical links to premotor areas. TMS was performed during four different conditions: observation of an experimenter grasping objects; observation of an experimenter doing aimless movements in the air with his arm; observation of objects; detection of the dimming of a small spot of light. The results of this study showed that during grasping observation MEPs recorded from the hand muscles markedly increased with respect to the other conditions, including the attention-demanding dimming detection task. Even more intriguing was the finding that the increase of excitability was present only in those muscles that subjects would use when actively performing the observed movements. This study provided for the first time evidence that humans have a mirror system similar to that in monkeys. Every time we are looking at someone performing an action, the same motor circuits that are recruited when we ourselves perform that action are concurrently activated.

In short: when human subjects in the TMS study observed other individuals moving, readings taken from their muscles seemed to reflect the sorts of muscular precursors to movements that would actually be required in order to perform the task being observed.  Note that actual neurons in the human brain were not electrode-probed as they were in monkeys; hence, while the TMS experiments do seem to indicate that observing someone performing a task can result in pre-motor movement potential in the muscles, they do not “prove” the existence of mirror neurons.  Nevertheless, it is not entirely farfetched to think that perhaps humans do have mirror neurons—the question is whether the performed experiments actually tell us anything about internal modeling of others’ minds or mental state (“mind reading”).  Gallese and Goldman define mind reading as follows:

Mind-reading is the activity of representing specific mental states of others, for example, their perceptions, goals, beliefs, expectations, and the like. It is now agreed that all normal humans develop the capacity to represent mental states in others, a system of representation often called folk psychology. Whether non-human primates also deploy folk psychology is more controversial (see last section of this article), but it certainly has not been precluded. The hypothesis explored here is that MNs are part of – albeit perhaps a rudimentary part of – the folk psychologizing mechanism.

The leap from detected premotor muscular activity to the presumption of simulated mental states seems to be quite a large one indeed (and not necessarily an appropriate one).  Continuing onward, Gallese and Goldman discuss the two basic theorized types of “human mind reading”, known as theory theory and simulation theory:

There is a large literature concerned with the nature of (human) mind-reading. Two types of approaches have dominated recent discussion: theory theory (TT) and simulation theory (ST) (Refs 20–22). The fundamental idea of TT is that ordinary people accomplish mind-reading by acquiring and deploying a commonsense theory of the mind, something akin to a scientific theory. Mental states attributed to other people are conceived of as unobservable, theoretical posits, invoked to explain and predict behavior in the same fashion that physicists appeal to electrons and quarks to predict and explain observable phenomena. On the standard presentation, the theory of mind possessed by ordinary people consists of a set of causal/explanatory laws that relate external stimuli to certain inner states (e.g. perceptions), certain inner states (e.g. desires and beliefs) to other inner states (e.g. decisions), and certain inner states (e.g. decisions) to behavior. This picture has been articulated by functionalist philosophers of mind23–26 as well as by developmental psychologists27,28. According to TT, attributing particular mental states to others arises from theoretical reasoning involving tacitly known causal laws. Much on this subject has been done by developmentalists, eager to determine how the mind-reading capacity is acquired in childhood29. Many interpret children’s changes in mind-reading skills as evidence in favor of TT because the skill changes are construed as manifestations of changes in theory30,31. Theory theorists differ among themselves as to whether theory of mind is acquired by a general-purpose scientizing algorithm32 or by the maturation of a domain specific module or set of modules33,34. This debate will not concern us here. ST arose partly from doubts about whether folk psychologizers really represent, even tacitly, the sorts of causal/explanatory laws that TT typically posits. ST suggests that attributors use their own mental mechanisms to calculate and predict the mental processes of others.

In other words, “theory theory” suggests a learning situation in which humans gradually come to understand how other people think through interacting, recording data about interactions, processing it, and coming up with various generalized and specialized notions of how others are likely to think and react in particular situations.  Gallese and Goldman, however, suggest that theory theory might not be enough to fully explain apparent “mind reading” abilities in humans, and posit that “simulation theory” might actually be a more likely explanation for the apparent capacity people seem to have allowing them to accurately intuit what others think and feel.  That is, rather than applying “scientific” algorithms to evaluate others’ behavior, people might have a more direct route to understanding the minds of others—a route that exists by virtue of “simulation hardware” such as mirror neurons.  While it does not seem that “mirror neurons” or their activity would actually preclude theory theory from being true, the notion that perhaps people “simulate” the mental processes of others at least seems worth exploring.

However, when exploring the simulation theory hypothesis in conjunction with autism research, the common implication is that nonautistic people have the ability to internally represent the mental states of others, whereas autistic people do not—and that the mirror neuron experiments might constitute something approaching empirical evidence for this.  Seeking to test this idea, researchers at UC San Diego performed the following experiment, as described in the April 2005 Science Daily article, Autism Linked To Mirror Neuron Dysfunction:

The UC San Diego team collected EEG data in 10 males with autism spectrum disorders who were considered “high-functioning” (defined as having age-appropriate verbal comprehension and production and IQs above 80) and 10 age- and gender-matched control subjects.

The EEG data was analyzed for mu rhythm suppression. Mu rhythm, a human brain-wave pattern, is suppressed or blocked when the brain is engaged in doing, seeing or imagining action, and correlates with the activity of the mirror neuron system. In most people, the mu wave is suppressed both in response to their own movement and to observing the movement of others.

Subjects were tested while they moved their own hands and while they watched videos of visual white noise (baseline), of bouncing balls (non-biologic motion) and of a moving hand.

As expected, mu wave suppression was recorded in the control subjects both when they moved and when they watched another human move. In other words, their mirror neuron systems acted normally. The mirror neurons of the subjects with autism spectrum disorders, however, responded anomalously—only to their own movement.

Upon first reading, the description of this experiment probably sounds fairly straightforward.  But examining it for a moment from a more critical perspective, all this experiment really tells us is that autistic brains seem to respond differently to a particular kind of stimulus (in a particular environment).  It does not actually prove that autistic brains are actually “dysfunctional” to begin with; it merely demonstrates an observation of difference in functionality between autistic and nonautistic cognition.  Nor is any attempt made to explain why mu waves might not be suppressed in autistics when those subjects viewed the moving hand—it could be that the autistic subjects were simply processing the data in their environment differently.  Of course, there is certainly nothing unscientific (or unethical) about noting these differences in reaction pattern and functionality between autistic and nonautistic brains, however, there is plenty to scrutinize when conclusions such as the following (from the same Science Daily article as quoted above) are drawn:

“The findings provide evidence that individuals with autism have a dysfunctional mirror neuron system, which may contribute to many of their impairments—especially those that involve comprehending and responding appropriately to others’ behavior,” said Lindsay Oberman, first author of the paper and UCSD doctoral student working in the labs of senior authors V.S. Ramachandran, director of the Center for Brain and Cognition, and Jaime Pineda, director of the Cognitive Neuroscience Laboratory.

While it is certainly not ridiculous to theorize that the mu wave experiment might carry implications that could explain autistic differences in processing, it must be acknowledged that the experiment does not actually “prove” any sort of correlation between mirror neurons and a person’s ability to comprehend others’ behavior or “respond appropriately”.  There might be any number of reasons why autistics might not seem to comprehend others’ behavior or respond appropriately to it, none of which necessarily have anything to do with mirror neuron activity or lack thereof.  Not to mention the fact that what comprises an “appropriate response” is subjective and culturally variable; in many Asian cultures, for instance, it is considered rude for people to make direct eye contact with their workplace superiors, and for children to make eye contact with adults.

And if one applies “theory theory” here (as opposed to simulation theory), the apparent disconnect between autistic and nonautistic patterns of relating and interacting comes into much sharper focus, since over time a typically-developing person will learn that other people have minds “like them” (because nonautistic people are more numerous), whereas an autistic person might be more likely to continue to see the minds of others as mysterious.  Perhaps the lack of mu wave suppression in autistics observing a hand motion reflects the fact that autistics do not automatically make the kinds of assumptions that nonautistic people do about others in their vicinity.  Perhaps it reflects the fact that autistic brains are “using” mu wave activity to accomplish cognitive functions not being tested for by the experiment described above.  Or perhaps it reflects something else entirely—we really won’t know until more experiments are performed.

But what about the “bottom line” here?  After all, autistics do seem to demonstrated reduced comprehension of the behavior of others, in addition to difficulties in imitation and an observed lack of “appropriate” (i.e., typical) responses to particular behaviors on the part of others, cultural variability notwithstanding.  Isn’t it just “grasping at straws” (politically correct straws, no less) to suggest that nothing concrete can be drawn from mirror neuron studies?  Not hardly.  In all these studies claiming an association between “broken mirror neurons” and autism, there is a distinct lack of any serious exploration of the ability (or lack thereof) of nonautistic people to understand autistic people.  Clearly, there is a communication breakdown here (between autistics and nonautistic), but not one that can be neatly and succinctly explained by suggesting that autistics are “broken”.  If mirror neurons can theoretically be said to indicate an ability to simulate others’ mental states and understand their feelings and motivations, then why are autistics so commonly described as “mysterious” and/or “otherworldly”?  Why don’t nonautistic people seem able to internally simulate the mental states of autistics?  Have any experiments been performed in this regard at all?  What seems to be happening here is that people are having tremendous difficulty accepting the heterogeneity of humanity and proposing instead that “human-ness” is dependent upon one’s possession of certain typical, majority traits—meaning that people who lack these traits are not merely different, but less than, limited, and perhaps not even people to begin with.

About the Author: Anne Corwin, an autistic woman who blogs at Existence is Wonderful, describes herself as “an engineer, science geek, sf/fantasy fan, amateur artist, cat appreciator, hyperlexic infovore, and maker of various and random quasi-functional objects.” This piece first appeared on the website of the Institute for Emerging Ethics and Technologies and is reprinted by permission.