Mirror Neurons and the On-going Debate

This essay discusses mirror neurons—a cutting edge topic in cognitive neuroscience. I argue that mirror neurons are strictly action-predicting in that they cannot yet be linked to understanding the intention in others.

For the past twenty years, the field of neuroscience has been intrigued by research on a specific type of neuron that has shown interesting physiological similarities between performing a motor task and observing the performance of a motor task. These ‘mirror neurons’ were discovered by di Pellegrino, Fadiga, Fogassi, Gallesse, and Rizzolatti (1992) who used single cell recordings in macaques' f5 region of the ventral premotor cortex to show that some neurons fired both while a monkey performed an action and while it observed another monkey performing the same action. 

Since then, debate has sprung up around whether these neurons are indeed what they are claimed to be by some: the fundamental building blocks of human empathy and social interaction. Others argue that this phenomenon is merely a reaction of the motor cortex in training for future movement and that it cannot yet be claimed to contribute to such social behavior. While drawing on conflicting evidence of monkey and human studies, the current essay argues that mirror neurons are strictly predicting action in the motor cortex and that they cannot yet be linked to understanding high-level intention in others. The main argument put forth is that empathy and goal-understanding, which require emotional, sympathetic, and motivational comprehension, cannot be derived from research thus far on mirror neurons.

Mirror Neurons in the Macaque

To begin, the first finding on mirror neurons in 1992 was limited in its validity because of the total amount of neurons that were reported on (Casile, 2012). Research was not done on this topic again until 1996, when more specific questions elicited a closer look into the phenomenon. Gallese, Fadiga, Fogassi, and Rizollati (1996) found similar neuronal activation patterns within dense groups of neurons in the pre-motor cortex of both acting and observing monkeys. More importantly, they found that these neurons only responded to an action being performed with an object and not when monkeys solely observed an object. Almost 20% of the neurons that were tested responded to an observed monkey’s action and half of these neurons fired in preference of hand movements in specific directions.

Many papers published around this time started to discuss possible explanations as to why neurons would be involved in cortical motor processes during observation of an action. The original claim was merely that of imitation of the motor cortex in a response to visual information. However, some began to suggest that these findings were evidence for evolutionary advances in monkeys and that they could explain a functional role in our social development as a species (or at least, in the social development of the Macaque monkey). Others commented that these neurons were able to “understand the goal of action” and discussion arose from ideas like these (Rizzolatti & Luppino, 2001). Authors like Gallesse and Leannorad claimed that these findings were developments of the motor cortex in representing the actions of others, internally, in order for the organism to engage in future behavior.

Much of the debate that continues to surface from studies on mirror neurons is centered on the philosophical issue of what is meant by claims of mirror neuron function. Although much of the field understands that ‘mirroring’ motor activation points to motor mimicry in an observing animal, discussion on how this activation relates to higher-level processing is at the heart of the debate. Terms such as ‘goals’, ‘intention’, and ‘understanding’ are often not thoroughly defined and lead to an inconsistency of research questions and varying interpretations of results. It can be rightfully argued that motor mimicry, to some extent, can contribute to ‘action understanding’ in an observing monkey. That is, the monkey’s brain is representing motor features of the action that it is seeing and therefore leads to the monkey’s understanding of the behavior (in the motor sense). However, understanding ‘intention’, ‘goals’, ‘motivation’ and other higher-level behaviors cannot be summed up with terms like ‘action understanding’. So although mimicry does suggest a certain degree of ‘understanding an action’, this essay argues that it fails in support of theories of mind-reading and empathy, both of which are attributed to more complex representations in the brain.

To further investigate the role of mirror neuron activity, research began to focus on changing the environment in which areas of the cortex could become excited.  It also investigated how vision and movement could be manipulated in order to isolate when and where mirror neurons were becoming the most active. Umiltà, Kohler, Gallese, Fogassi, Fadiga, Keysers, and Rizzolatti (2001) compared recordings in f5 between a monkey able to see the full extent of a reaching and grabbing task and a monkey able to see only a reaching task. They found similar activation of mirror neurons when vision of the grasping task was both occluded and not occluded. The level of activation in the occluded condition, however, was much lower than that of the non-occluded condition. Although this effect was small, it provided evidence to the authors that these neurons were responding as a way for monkeys to understand the intended goal of the grasper. This claim was argued because mirror neurons still fired even when a monkey could not see the result of a motor task (Casile, 2012). Kohler, Keysers, Umiltà, Fogassi, Gallese, and Rizzolatti (2002) also found these results to be true in a similar task where monkeys were able hear the sounds of visually occluded actions (cracking a nut; dropping a stick, etc.). This suggested that even though the visual system was not able to “see” the action (because the actual movement was occluded), mirror neurons were able to express understanding of the goal of the experimenter (e.g. crack the nut).

Authors including Iocobanni and Galesse argued that these sounds caused mirror neurons to fire which, in turn, caused the monkey to ‘understand’ the intention of the action even without visual representation of that action. A different interpretation of these results is that hearing the crack of a nut is most likely associated with the act of breaking a nut, therefore causing neurons in f5 to fire based on spreading activation within the pre-motor cortex and auditory processing areas (Hickok, 2009). Another interpretation of these findings, provided by Southgate, Csibra, and Gergely (2007), explained that babies are able guess the future action of another, given the right amount of visual information. Based on this evidence, the authors argued that action anticipation can be done without making full inference of another’s intentions. In regard to the current mirror neuron debate, this study suggests that visual information may provide enough data for the brain to anticipate the action of the grasper based on what it can ‘see’. Counterevidence does not necessarily support the idea that the observer ‘understands’ the mind of the grasper, but that they are merely able to recognize the kinesthetic information required to carry out the same action themselves (Gerrans, 2010).

Fogassi, Ferrari, Gesierich, Rozzi, Chersi, and Rizzolatti (2005) compared mirror neuron activation of monkeys when observing a researcher place an object into a cup or place an object into their mouth. Results showed that there were separate subsets of neurons active for each task within the inferior parietal lobules, suggesting that these mirror neurons were encoding the goal of the action rather than just the action itself.  A counter argument could say that different actions elicit different activation of related sensory modalities. Moreover, related cortical areas are expecting to receive certain information based on what can be assumed by the motor action that is present (when preparing to eat: opening mouth, salivation, tactile information, etc; when preparing to place an item into a cup: the sound of the object dropping; the feel of the container, etc.). This association of cortical areas between the motor cortex and other sensory areas could be linked by mirror neurons in order to predict how a task were to be experienced if it were to happen to the observer and not to assume high-level cognitive function of another (Hickok, 2009).

Mirror Neuron Systems in Humans

Human studies looking at mirror neuron activation began to surface within the field to draw comparisons between monkey mirror neurons and possible mirror neuron systems within the human brain. Not only are human studies useful for comparing and contrasting physiological evidence for mirror neurons, they are also useful in exploring how behaviors of the two species can be related by mirror neuron activation.  The differences in methodology between human and monkey studies will be discussed later. 

Many functional imaging studies have shown that humans indeed have neurons that behave in a similar way to those within the ventral pre-motor and parietal areas of the macaque monkey. Buccino, Vogt, Ritzl, Fink, Zilles, Freund, and Rizzolatti (2004) found mirror neuron activation in humans who passively watched guitar chords being played. Human mirror neuron circuits showed to be firing in similar ways to those of the macaque monkey, but in more complex patterns (Carlson, 2012, p 281).

Prior research shows that the mirror neuron system in humans is excited when the kinesthetic features of an action are noticed. This activation in the motor cortex exists to prepare the brain for the same performance in the future (Cisek & Kalaska 2004). It seems to be a way for the motor cortex to practice movement (which is required for learning) and not as a way to understand the high-level intentions and motivations of others. More evidence for this idea comes from a study by Haslinger, Erhard, Altenmüller, Schroeder, Boecker, and Ceballos-Baumann (2005) who showed that there was activation of the mirror neuron system, the visual system, and the auditory system when people imagined playing a song that they had previously learned on a piano. Mirror neurons could belong in a junction between sensory areas that share activation when the brain is trying to predict events in the world.

Predicting events does not necessarily mean ‘mind-reading’ in this sense. Claims of this ‘mindreading’ nature come from those like Rizolatti, Fogassi, and Gallese (2001) who argue that “[A]n action is understood when its observation causes the motor system of the observed to resonate”. This is too bold a claim if ‘action-understanding’ is meant to encompass more than just the activation of motor and related sensory areas in the brain. Mirror neurons could be a part of the process of understanding the cortical features of action but cannot yet be linked to such complex tasks as ‘mindreading’ or inference. Areas in the cortex previously shown to be involved during mind-reading tasks have not been shown to correlate completely with human mirror neuron circuits (Pineda, 2009). Moreover, even though both humans and monkeys have mirror neurons, monkey mind-reading abilities are far below those of humans and there is question as to whether this ability even exists in the monkey. If this is so, there is no evidence explaining why both species have similar mirror neuron systems, but that monkeys fail in comparison to human mind-reading abilities (Hickok, 2009; Gerrans, 2010).

A few studies provide relevant evidence to suggest that cortical areas involved in action understanding are separate from areas that include mirror neuron systems in humans. One study by Catmur, Walsh, and Heyes (2007), had subjects view a video of hand movements (most specifically movement of the fingers). Unsurprisingly, mirror neurons responded normally when viewing the movement of the pinky, which activated pinky-moving areas within the cortex. Later, subjects were taught and trained to move their fingers in ways that were inconsistent with the hand movement on the screen. For example, when the hand on the video moved its pinky, the subject was instructed to move their thumb (and vice versa). Interestingly, mirror neuron activation seemed to be responding in the new learned fashion, where pinky areas would light up when viewing thumb movements. This evidence suggests that the mirror neuron system can be trained and ultimately change based on cues that are learned by the motor system. One could argue that this is because there is now a new understanding of intention that elicits a different pattern of motor activation. A more sound argument is that this phenomenon is closely related to associative learning in the cortex where neurons are firing in patterns based on previously learnt coding. If mirror neurons serve to understand the intent of action, why would training in an inconsistent way with the observed action cause an opposite pattern of activation? One would assume that even if I trained to move my hand in a different way than how I see it move, I could still understand the intention of the original action. This shows evidence that ‘understanding’ and mirror neurons may not be as closely linked as some believe.  

The argument for the dissociation between motor areas and ‘understanding-areas’ can be further addressed. Buccino, Lui, Canessa, Patteri, Lagravinese, Benuzzi, and Rizzolatti, (2004) had humans observe non-communicative gestures (like biting) and communicative gestures (lip moving, lip smacking, and barking) from three species (human, monkey, and dog). FMRI readings during biting sequences showed mirror neuron activation regardless of the species being observed. However, only communicative gestures of the human and monkey species elicited mirror neuron activation. Hickok (2009) interprets these results as inconsistent with the ‘understanding theory,’ because of how the intention of all three gestures should be equally understood given the correct context.

Human and Monkey Comparison

Interesting studies on both humans and monkeys show that mirror neuron activation seems to be much stronger when motor tasks are already familiar to the one observing the action. Calvo-Merino, Glaser, Grezes, Passingham, and Haggard (2005) had dancers in different fields of expertise watch a collection of videos that were either from their own field or from another field. All dancers showed a certain level of activation of mirror neuron circuits while watching all forms of ballet, but the most activation when watching videos that were consistent with their own fields of expertise. Similarly, Rochat, Caruana, Jezzini, Escola, Intskirvelli, Grammont, Gallese, Rizolatti, and Umilita (2010) had monkeys use different tools to perform different actions and recorded their pre-motor cortex activation. Monkeys seemed to have a stronger activation when observing the use of a familiar tool than when observing the use of a new tool. These findings suggest that there is savings on re-learning in the mirror neuron system. But why would ‘understanding the goal’ be the underlying motivation of this system when an observer could potentially understand action intention in scenarios where they did not know how to use the tool (Hickok, 2009)? In support of the current essay’s hypothesis, it could be argued that the mirror neuron circuit recognizes both actions that have been learned before and actions that are novel. It recognizes familiar actions more strongly because the neural basis of using a tool has already been established. Mirror neurons could be acting in combination with other brain areas (most specifically the visual and sensory systems) in order to predict this kind of movement.

It has also been found that, in tasks that activate mirror neurons, monkeys are unable to imitate action as efficiently as human subjects (Visalberghi & Fragaszy, 2001). This suggests that mirror neurons do not specifically encode imitation because monkeys show both mirror neuron activation and poor imitation skills. Humans however, show activation of mirror neuron systems and are equipped for action imitation. This is argued to be a fundamental part of the learning process in humans (Buccino et al. 2004). For these conflicting reasons, some researchers believe that because monkeys cannot imitate, their mirror neurons serve a different function: that of goal understanding. However, I am convinced that monkey deficit in imitation abilities are due to a lack of parallel processing within areas involved in learning in the pre-frontal cortex (Rizzolatti, 2005). Understanding the physical elements of an action is much lower-level than that of understanding the ‘goal’ of an action. Given this, it is hard to believe that monkeys are unable to imitate action but that they can understand the goal of an action. I state this because it seems odd that activity in the pre-frontal cortex, which would be required to infer motivation, is lacking in the monkey brain and yet monkeys are supposed to have skills necessary to understand another monkey’s mind.

Due to differences in information processing abilities of monkeys and humans, these studies are often difficult to compare. I believe that another difficulty in comparing mirror neuron function in humans and monkeys has to do with differences in practiced methodology. Studies with humans are able to record millions of cells at a time while studies with monkeys often report from individual cells within the cortex. This difference in methodology leads to incomparable results because of the lack of detail in research on humans. Although I do not discuss this much further, it needs to be understood that these limitations should not be the only grounds to explain differences in mirror neuron function within humans and monkeys.

Conclusion and Future Work

As mentioned before, a huge problem within this debate stems from the rather ill-defined descriptions of commonly used terms. To what extent are monkey and human mirror neuron systems actually a part of the process of ‘intention-understanding’? Vladimir Kosonogov (2012) advocates for the idea that if mirror neurons are indeed coding levels of meaning when observing action, there is much ambiguity as to what mirror neurons would consider as a ‘goal-driven’ behavior. Patricia Churchland (2001) argues that if the connection between mirror neurons and mindreading is to be made, the definition of ‘intention’ needs to be flushed out. She also explains that understanding intention is a highly context-dependent task that requires much more complex activation than what can be seen in the motor and parietal cortices. This kind of high-level processing, in her eyes, cannot be done by individual neurons no matter where they are located.

Future studies on the involvement of mirror neurons in predictive action should focus on asking more salient questions about the similarities and differences between human and monkey mirror neuron systems while also trying to get more representative samples of individual mirror neurons within humans. Researchers should, most importantly, work on defining areas of thought more clearly so that mirror neuron research can become more efficient and successful in answering its many diverse questions. The next decade of research within this fascinating field of neuroscience will give us a better understanding of our own behaviors as well as how our behaviors influence others. Compromise within this debate will only come if we remember to be conservative in the claims that we make and tread lightly in our interpretations.

References

Buccino, G., Lui, F., Canessa, N., Patteri, I., Lagravinese, G., Benuzzi, F., & Rizzolatti, G. (2004). Neural circuits involved in the recognition of actions performed by nonconspecifics: An fMRI study. Journal of cognitive neuroscience, 16(1), 114-126.

Buccino, G., Vogt, S., Ritzl, A., Fink, G. R., Zilles, K., Freund, H. J., & Rizzolatti, G. (2004). Neural circuits underlying imitation learning of hand actions: an event-related fMRI study. Neuron, 42(2), 323-334.

Calvo-Merino, B., Glaser, D. E., Grèzes, J., Passingham, R. E., & Haggard, P. (2005). Action observation and acquired motor skills: an FMRI study with expert dancers. Cerebral cortex, 15(8), 1243-1249.

Carlson, N. R. (2012). Physiology of Behavior 11th Edition. Pearson.

Casile, A. (2013). Mirror neurons (and beyond) in the macaque brain: an overview of 20 years of research. Neuroscience letters, 540, 3-14.

Catmur, C., Walsh, V., & Heyes, C. (2009). Associative sequence learning: the role of experience in the development of imitation and the mirror system. Philosophical Transactions of the Royal Society B: Biological Sciences,364(1528), 2369-2380.

Churchland, P. S. (2011). Braintrust: What neuroscience tells us about morality. Princeton University Press.

Cisek, P., & Kalaska, J. F. (2004). Neural correlates of mental rehearsal in dorsal premotor cortex. Nature, 431(7011), 993-996.

Di Pellegrino, G., Fadiga, L., Fogassi, L., Gallese, V., & Rizzolatti, G. (1992). Understanding motor events: a neurophysiological study.Experimental brain research, 91(1), 176-180.

Fogassi, L., Ferrari, P. F., Gesierich, B., Rozzi, S., Chersi, F., & Rizzolatti, G. (2005). Parietal lobe: from action organization to intention understanding.Science, 308(5722), 662-667.

Gerrans, P. (2010). Mirror neuron systems: The role of mirroring processes in social cognition. Cognitive Neuropsychiatry, 15(5), 501-504.

Haslinger, B., Erhard, P., Altenmüller, E., Schroeder, U., Boecker, H., & Ceballos-Baumann, A. O. (2005). Transmodal sensorimotor networks during action observation in professional pianists. Journal of cognitive neuroscience,17(2), 282-293.

Hickok, G. (2009). Eight problems for the mirror neuron theory of action understanding in monkeys and humans. Journal of cognitive neuroscience,21(7), 1229-1243.

Kohler, E., Keysers, C., Umilta, M. A., Fogassi, L., Gallese, V., & Rizzolatti, G. (2002). Hearing sounds, understanding actions: action representation in mirror neurons. Science, 297(5582), 846-848.

Kosonogov, V. (2012). Why the Mirror Neurons Cannot Support Action Understanding. Neurophysiology, 44(6), 499-502.

Pineda, J. A. (Ed.). (2009). Mirror neuron systems: The role of mirroring processes in social cognition. Springer.

Rizzolatti, G. (2005). The mirror neuron system and its function in humans.Anatomy and embryology, 210(5), 419-421.

Rizzolatti, G., & Luppino, G. (2001). The cortical motor system. Neuron, 31(6), 889-901.

Rizzolatti, G., Fadiga, L., Gallese, V., & Fogassi, L. (1996). Premotor cortex and the recognition of motor actions. Cognitive brain research, 3(2), 131-141.

Rizzolatti, G., Fogassi, L., & Gallese, V. (2001). Neurophysiological mechanisms underlying the understanding and imitation of action. Nature Reviews Neuroscience, 2(9), 661-670.

Rochat, M. J., Caruana, F., Jezzini, A., Intskirveli, I., Grammont, F., Gallese, V., & Umiltà, M. A. (2010). Responses of mirror neurons in area F5 to hand and tool grasping observation. Experimental brain research, 204(4), 605-616.

Southgate, V., Senju, A., & Csibra, G. (2007). Action anticipation through attribution of false belief by 2-year-olds. Psychological Science, 18(7), 587-592.

Umilta, M. A., Kohler, E., Gallese, V., Fogassi, L., Fadiga, L., Keysers, C., & Rizzolatti, G. (2001). I know what you are doing: a neurophysiological study.Neuron, 31(1), 155-165.

Visalberghi E, Fragaszy D (2001) Do monkeys ape? Ten years after. In: Dautenhahn K, Nehaniv C (eds) Imitation in animals and artifacts. MIT Press, Boston, MA