Since their discovery in the 1990s, mirror neurons have experienced a degree of fanfare uncommon to findings in the field of neuroscience. Mirror neurons are so named because they are activated both when a primate participates in a task, and while watching another complete the same task, thus “mirroring” the behavior of the other animal. This unique activation pattern has led some to suggest that mirror neurons are integral not only to imitation, but also to understanding that others have their own mental states (theory of mind). By extension, it has been hypothesized that mirror neurons are necessary for language acquisition and social interaction. Dysfunctions in mirror neurons have even been offered as a possible cause of autism.
Thus, they have come to be viewed as a very special kind of neuron, with a versatility and importance to brain function that is unrivaled by other types of brain cells. But, do mirror neurons deserve the exalted status that some have ascribed to them? In short: probably not.
Mirror neurons do seem to play an interesting role in cognition. Primate studies have found mirror neurons to be activated in correlation with focusing on a particular goal or intention of movement. They are also activated in a selective fashion, with specific groups corresponding to different goals of an action, e.g. grasping to move vs. grasping to eat. Of additional interest, they have been found to respond to sounds associated with an observed action.
fMRI studies in humans have revealed specific activity in areas where mirror neurons are thought to be located, such as the ventral premotor (vPM) and anterior intraparietal sulcus (alPS), during observation and imitation of movement.
But, while these findings in humans and non-human primates are intriguing, they don’t support the rampant speculation that has followed about the role of mirror neurons in overall cognitive function. Experiments with monkeys to date haven’t assessed the ability to imitate, experience empathy, display theory of mind, or use language. Of course, it is debatable to what extent some of these attributes even exist in non-human primates, or if they can be studied if they do.
As for humans, neuroimaging experiments have allowed scientists to determine which regions of the brain are active during imitation or observation of an action. The specific neurons that are utilized, however, and any physiological characteristics that make them unique, cannot be assessed with current imaging technology.
Thus, the roles attributed to mirror neurons in the last decade since their discovery may have been an overly ambitious attempt to describe their function. By extension, implying that their malfunction is critical in autism could really be jumping the gun.
In an essay in last week’s Nature, Antonio Damasio and Kaspar Meyer discuss the exaggerated claims about mirror neurons, and suggest a rational hypothesis for how they may work. Twenty years ago Damasio proposed a theory known as “time-locked multimodal activation” to explain the development of complex memories. The theory is based on the proposed existence of groups of neurons that, during the encoding of memories, receive input from a number of different sites. Damasio termed these neuronal groups convergence-divergence zones (CDZ). He suggested there are two types of CDZs: local CDZs, which collect information from areas close to a sensory cortex (e.g. the visual cortex), and non-local CDZs, which are higher-order structures of the brain where the information from local CDZs converges.
According to this theory, when a memory is formed—Damasio and Meyer use the example of a monkey opening a peanut shell—all the information about the event converges on a non-local CDZ. Then, if the monkey hears a peanut shell opened in the future, this would activate a local auditory CDZ, as well as the non-local CDZ where memories associated with the noise are stored. Signals are sent out from the non-local CDZ to all local CDZs that were involved in the original experience of the event, activating these sites and resulting in a sort of recreation of the original peanut-cracking.
Mirror neurons are represented by the non-local CDZs. In this scenario, however, mirror neurons are not physiologically unique. They are normal neurons involved in a network that has less to do with “mirroring” than with integrating and syncing the various aspects of elaborate memories. This does not take away from the role and function of this network, but should detract a little from the aggrandized status attributed to individual mirror neurons, in favor of an appreciation of the holistic complexity of the brain.
The CDZ hypothesis, however, has not been tested, although research does indicate that networks involved in observing and imitating behavior spread beyond purported mirror neuron sites. Regardless of whether the specifics of the CDZ hypothesis come to be supported by future studies, I feel it represents a more sensible approach to mirror neurons. To credit mirror neurons alone with a function that carries such importance, like the ability to infer the mental states of others, seems to oppose much of what has been learned thus far about neuroscience. We have never found language neurons, love neurons, or fear neurons. Instead we have found networks that spread throughout brain regions that correlate with the ability to experience these aspects of cognition. I suspect we will soon say the same about mirror neuron networks and their involvement in social interaction.
Damasio, A., Meyer, K. (2008). Behind the looking-glass. Nature, 454 (7201), 167-168. DOI:10.1038/454167a
Dinstein, I., Thomas, C., Behrmann, M., & Heeger, D.J. (2008). A mirror up to nature. Current Biology, 18 (1), 13-17.