Mirror Neurons

Abstract

Mirror neurons are cells that discharge both when an animal performs a goal-directed action and when it observes another individual performing a similar action. First reported in the ventral premotor cortex of the macaque monkey, they were quickly proposed as a neural mechanism for understanding the actions of others from the inside, by simulation rather than inference. This article traces the field from the original single-cell recordings in area F5, through the mapping of a parieto-frontal mirror system in the human brain, to the theories that link mirror neurons to action understanding, imitation, intention reading, and language. It also gives full weight to the vigorous debate the theory has provoked, including the associative-learning account and the argument that mirroring may reflect rather than cause understanding. A worked example and three interactive demonstrations make the core findings concrete.

Keywords: mirror neuron, action understanding, motor resonance, imitation, embodied simulation

What Mirror Neurons Are

A *mirror neuron* is a neuron that is active both when an individual executes a particular goal-directed action and when it watches someone else perform that same or a similar action. The defining property is this matching, or *mirroring*: a single cell participates in both doing and seeing, so the observed action is mapped onto the observer's own motor repertoire. Because the cell's discharge during observation echoes its discharge during execution, watching another person grasp a cup engages, in part, the very circuit the observer would use to grasp the cup themselves (di Pellegrino, Fadiga, Fogassi, Gallese, & Rizzolatti, 1992).

This simple finding carried a large claim. If perceiving an action automatically activates its motor representation, then an observer might grasp what another is doing not by reasoning about it but by covertly reproducing it, a *motor resonance* that gives immediate, first-person access to the meaning of the act (Rizzolatti & Craighero, 2004). That proposal made mirror neurons one of the most discussed discoveries in modern neuroscience, invoked to explain imitation, empathy, and the evolution of language, and it also drew sustained criticism that the field is still working through.

Discovery in Macaque Area F5

Mirror neurons were found by accident in the early 1990s in Parma, while Giacomo Rizzolatti's group recorded from single neurons in the ventral premotor cortex, area F5, of macaque monkeys performing hand actions such as grasping food. Some of these cells, they noticed, also fired when the monkey merely watched an experimenter make the same movement, even though the monkey itself stayed still. The neurons did not respond to the sight of the object alone, nor to a mimed movement with no target; they required an actual, goal-directed action (di Pellegrino et al., 1992). A fuller study confirmed that a substantial population of F5 neurons had this property and that it was tied to the goal of the act rather than to the specific movements that achieved it (Gallese, Fadiga, Fogassi, & Rizzolatti, 1996).

Rizzolatti and colleagues sorted these cells by how closely the observed action that drove a neuron matched the executed action it coded. *Strictly congruent* neurons required essentially the same action in both conditions; *broadly congruent* neurons responded to an observed action that was related to, but not identical with, their preferred motor act. This congruence was the crux of the interpretation: because the observed and executed responses corresponded, the visual response could be read as an activation of the motor representation of the same goal (Rizzolatti, Fadiga, Gallese, & Fogassi, 1996). The first demonstration lets the reader set a cell's execution and observation firing rates and read off a mirror index that captures this congruence, reproducing the Worked Example below.

The Human Mirror System

Single neurons cannot ordinarily be recorded in the human brain, so the human mirror system was first mapped indirectly. Functional imaging showed that watching and imitating hand actions engages a network centered on the inferior frontal gyrus and the inferior parietal lobule, the human counterparts of the monkey's F5 and parietal mirror areas, with imitation activating them more strongly than mere observation (Iacoboni et al., 1999). A meta-analysis of 125 human functional-imaging studies later delineated this network in detail, confirming a core parieto-frontal circuit with mirror-like properties while showing that regions beyond the classical areas also carry them (Molenberghs, Cunnington, & Mattingley, 2012).

Direct evidence in humans arrived when neurosurgical patients, implanted with depth electrodes for clinical monitoring, allowed recordings from single cells during both the execution and the observation of actions. A subset of neurons, notably in the medial frontal and medial temporal lobes, showed the mirroring signature at the single-cell level, the first hard confirmation that human mirror neurons exist (Mukamel, Ekstrom, Kaplan, Iacoboni, & Fried, 2010). Anatomical and physiological work has since widened the picture from a two-node circuit to an *extended* mirror network, distributed across parietal, premotor, and other regions and shaped by the observer's own motor experience (Bonini, 2017).

Functions and Theories

The boldest claim made for mirror neurons is that they underlie *action understanding*: that we grasp the meaning of an observed act directly, because observing it reactivates our own motor knowledge of it, rather than by drawing an inference about the actor's mind. Gallese and the philosopher Alvin Goldman argued that this makes mirror neurons a plausible neural basis for *simulation* accounts of mind-reading, in which we understand others by covertly running their states in our own system (Gallese & Goldman, 1998). On this view motor resonance is not a byproduct of perception but its interpretive engine.

Two findings sharpened the functional case. First, some F5 neurons proved to be *audiovisual*: they fired not only when a monkey saw or performed a noisy action, such as cracking a peanut, but also when it merely heard the action's sound in the dark, and they stayed quiet for control sounds, showing that the representation is of the action itself, abstracted from the sensory channel (Kohler et al., 2002). Second, mirror neurons in the inferior parietal lobule were found to discharge differently during the *same* grasp according to what the actor would do next, eating the object or placing it, so that the neuron's activity at the grasp already coded the agent's overarching intention (Fogassi et al., 2005). This *intention reading* extended the theory from what an action is to why it is being done. The second and third demonstrations reproduce these two results.

Rizzolatti and Sinigaglia later reframed the account to answer critics, arguing that the mirror mechanism supplies a specific kind of understanding, a first-person motor grasp of *how* an action is done and toward what proximate goal, which complements rather than replaces more cognitive routes to understanding others (Rizzolatti & Sinigaglia, 2010). In this more careful form the theory claims that mirroring contributes to action understanding, not that it is the whole of it.

Worked Example: Computing a Mirror Index

Suppose an experimenter records an F5 neuron across two conditions and wants a single number for how well it mirrors. During grasping execution the cell fires at 45 spikes per second; during observation of the same grasp it fires at 30 spikes per second; its spontaneous baseline is about 8 spikes per second. Both responses sit well above baseline, so the cell is active in both conditions, the first requirement for calling it a mirror neuron.

A convenient congruence measure is the *mirror index*, defined as one minus the absolute difference of the two rates divided by their sum. Here the difference is 45 minus 30, which is 15, and the sum is 45 plus 30, which is 75. The ratio is 15 divided by 75, or 0.20, so the mirror index is one minus 0.20, that is, 0.80. An index of 0.80 is high: the observation response is four-fifths as matched as a perfect echo would be, marking the cell as broadly-to-strictly congruent.

Contrast a second cell that fires 45 spikes per second on execution but only 10 on observation. Now the difference is 35 and the sum is 55, the ratio is 0.64, and the mirror index falls to 0.36, a weakly congruent cell whose visual response only faintly tracks its motor preference. The lesson is quantitative: mirroring is a graded property, not an all-or-none label, which is exactly why the original studies sorted neurons into strict and broad congruence classes rather than a single category. The first demonstration lets the reader vary both rates and watch the index and its classification change.

The Debate Over Function

No claim about mirror neurons has gone unchallenged. The sharpest critique, from Gregory Hickok, laid out eight specific problems for the proposal that mirror neurons are the basis of action understanding, among them that lesions to the human mirror regions do not reliably abolish the ability to understand actions, that monkeys understand actions they cannot perform, and that activity correlated with understanding need not be its cause (Hickok, 2009). The core worry is a confusion of correlation with mechanism: a motor system that resonates to observed actions may be reading out an understanding computed elsewhere rather than producing it.

A second line of criticism concerns where the mirroring comes from. Cecilia Heyes and Caroline Catmur have argued that mirror properties are largely *acquired* through sensorimotor associative learning, built up whenever we watch ourselves act or act in synchrony with others, rather than being an innate adaptation for understanding. On this account mirror neurons are real and interesting but are a *product* of ordinary learning mechanisms, and the grand functional claims for them outran the evidence (Heyes & Catmur, 2022). This associative view makes different predictions, for instance that mirror responses can be reversed by counter-training, and it reframes mirror neurons as a case study in how experience wires the brain rather than as a dedicated module for social cognition.

Discussion

Three decades on, a measured synthesis is possible. That mirror neurons exist is not in doubt: they have been recorded directly in the monkey and the human brain, and a parieto-frontal system with mirror properties is one of the better-established findings of cognitive neuroscience. What remains contested is the *interpretation*, how much of social cognition rests on this mechanism. The strongest early claims, that mirror neurons are the neural basis of empathy, language, and mind-reading, ran well ahead of the data, and the reaction against them was warranted.

For cognitive psychology the durable contribution is a concrete instance of *embodied cognition*: a demonstration that perceiving an action is not a purely visual event but engages the motor system, so that perception and action are more tightly coupled than a strict input-output model assumes. Whether or not mirror neurons cause action understanding, they show that the brain represents others' behavior in a partly motoric code, which bears on learning by imitation, on social cognition, and on how attention to others' actions is organized. The safest reading is that mirroring is one contributor among several to how we make sense of one another.

Current Directions

Recent work has moved away from asking whether mirror neurons cause understanding and toward characterizing what the extended mirror network actually computes and how experience shapes it. A thirtieth-anniversary review by Bonini and colleagues surveys how the field has broadened, from the original two-node circuit to a distributed system, and how mirror mechanisms are being applied in domains such as motor rehabilitation, where action observation is used to drive plasticity, and social neuroscience (Bonini, Rotunno, Arcuri, & Gallese, 2022). The emphasis has shifted from a single grand function to a mechanism with several graded, context-dependent roles.

At the same time the associative-learning program has forced sharper experiments. If mirror properties are learned, they should be manipulable by training and should vary with sensorimotor history, and testing these predictions has become a research front in its own right (Heyes & Catmur, 2022). Open questions include how mirror responses are read out by the rest of the brain, how they develop in infancy, and how far findings in the macaque transfer to the human system, all of which will decide which of the theory's original promises survive.

Common Misconceptions

Mirror neurons were discovered in humans.

They were first recorded in the ventral premotor cortex of the macaque monkey; direct single-cell evidence in humans came almost two decades later from neurosurgical patients (Mukamel et al., 2010).

A mirror neuron fires for any movement it sees.

The original cells required a goal-directed action on an object; they did not respond to the object alone or to a mimed movement with no target (di Pellegrino et al., 1992).

Mirror neurons are proven to be the basis of empathy and language.

These are hypotheses, and contested ones; lesion and behavioral evidence has been used to argue that mirroring may accompany rather than produce understanding (Hickok, 2009).

Mirror neurons are an innate, dedicated module.

An influential alternative holds that their mirror property is largely built by sensorimotor associative learning over development, not hardwired for social understanding (Heyes & Catmur, 2022).

Glossary

Action understanding.

Grasping the goal or meaning of an observed action; the function most often attributed to mirror neurons and the most debated.

Area F5.

A subregion of the macaque ventral premotor cortex where mirror neurons were first recorded.

Associative-learning account.

The view that mirror properties are acquired through correlated experience of seeing and doing actions, rather than being innate.

Audiovisual mirror neuron.

A mirror neuron that responds to the sound of an action as well as to seeing or performing it, coding the action across sensory channels.

Broadly congruent neuron.

A mirror neuron whose observed-action trigger is related to, but not identical with, the action it codes during execution.

Congruence.

The degree to which a mirror neuron's response during observation matches its response during execution of the same action.

Embodied cognition.

The broad view that cognition depends on the body's sensorimotor systems; mirror neurons are a frequently cited example.

Inferior parietal lobule.

A parietal region carrying mirror neurons, including those whose grasp response codes the actor's later intention.

Intention reading.

Inferring what an actor will do next; parietal mirror neurons discharge differently at a grasp according to the intended outcome.

Mirror index.

A graded measure of congruence used here, one minus the normalized difference between a cell's execution and observation firing rates.

Motor resonance.

The automatic activation of an observer's own motor representations by the sight of another's action.

Simulation theory.

The proposal that we understand others by covertly reproducing their states in our own systems, for which mirror neurons were offered as a substrate.

Strictly congruent neuron.

A mirror neuron that requires essentially the same action during observation and execution.

Ventral premotor cortex.

The frontal motor region, containing area F5 in the monkey and its inferior-frontal counterpart in humans, central to the mirror system.

Key Researchers

Luciano Fadiga

(b. 1961). A member of the Parma team who used transcranial magnetic stimulation to demonstrate that the human motor system is activated during action observation; Professor at the University of Ferrara and the Italian Institute of Technology.
ORCID - Wikipedia - Google Scholar - Faculty page

Leonardo Fogassi

(b. 1958). A member of the Parma team who showed that parietal mirror neurons code the goal of an action chain and can predict an actor's intention from the grasp; Professor at the University of Parma.
ORCID - Wikipedia - Faculty page

Vittorio Gallese

(b. 1959). A co-discoverer of mirror neurons and the originator, with Alvin Goldman, of the embodied-simulation account of how they support understanding others; Professor of Physiology at the University of Parma.
ORCID - Wikipedia - Google Scholar - Faculty page

Cecilia Heyes

(b. 1960). A psychologist at the University of Oxford and a leading proponent of the associative-learning account of mirror neurons, arguing that their mirror property is built by sensorimotor experience.
ORCID - Wikipedia - Google Scholar - Faculty page

Gregory Hickok

A cognitive scientist at the University of California, Irvine and a prominent critic of the mirror-neuron theory of action understanding, author of *The Myth of Mirror Neurons*.
Google Scholar - Faculty page

Marco Iacoboni

A neuroscientist at the University of California, Los Angeles who mapped the human mirror system with functional imaging and linked it to imitation and intention understanding; director of the UCLA Neuromodulation Lab.
ORCID - Wikidata - Faculty page

Giacomo Rizzolatti

(b. 1937). The neurophysiologist who led the Parma team that discovered mirror neurons in macaque area F5 and set the research program that followed; Emeritus Professor at the University of Parma and a Fellow of the Royal Society.
Wikipedia - Google Scholar

Frequently Asked Questions

What are mirror neurons?

Mirror neurons are cells that fire both when an individual performs a goal-directed action and when it observes another individual performing a similar action. This matching of observation to execution led to the proposal that they help us understand others' actions by mapping them onto our own motor system (di Pellegrino et al., 1992).

Where were mirror neurons discovered?

They were first recorded in the ventral premotor cortex, area F5, of the macaque monkey by Giacomo Rizzolatti's group in Parma in the early 1990s, and later found in the inferior parietal lobule (Gallese et al., 1996).

Do humans have mirror neurons?

Human imaging has long shown a parieto-frontal system with mirror-like properties, and direct single-cell recordings in neurosurgical patients confirmed neurons with the mirroring signature, so the answer is yes (Mukamel et al., 2010).

What is the mirror system supposed to do?

The main proposals are that it supports action understanding, imitation, and reading the intentions behind actions, by reactivating the observer's own motor representations of what is seen (Rizzolatti & Craighero, 2004).

Can mirror neurons tell what someone intends to do?

In the monkey, parietal mirror neurons discharge differently during the same grasp depending on whether the object will be eaten or placed, so their activity at the grasp already predicts the actor's intention (Fogassi et al., 2005).

Are mirror neurons really the basis of empathy and language?

These are hypotheses that outran the evidence. Critics have marshalled lesion and behavioral data to argue that mirroring may reflect understanding computed elsewhere rather than produce it (Hickok, 2009).

Are mirror neurons innate?

Not necessarily. An influential account holds that their mirror property is largely acquired through sensorimotor associative learning across development, which predicts that mirror responses can be changed by training (Heyes & Catmur, 2022).

What do mirror neurons respond to besides sight?

Some F5 neurons are audiovisual: they respond to the characteristic sound of an action, such as cracking a peanut, even in the dark, showing that they code the action itself and not just its visual appearance (Kohler et al., 2002).

References

Bonini, L. (2017). The extended mirror neuron network: Anatomy, origin, and functions. The Neuroscientist, 23(1), 56-67. https://doi.org/10.1177/1073858415626400

Bonini, L., Rotunno, C., Arcuri, E., & Gallese, V. (2022). Mirror neurons 30 years later: Implications and applications. Trends in Cognitive Sciences, 26(9), 767-781. https://doi.org/10.1016/j.tics.2022.06.003

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. https://doi.org/10.1007/BF00230027

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. https://doi.org/10.1126/science.1106138

Gallese, V., Fadiga, L., Fogassi, L., & Rizzolatti, G. (1996). Action recognition in the premotor cortex. Brain, 119(2), 593-609. https://doi.org/10.1093/brain/119.2.593

Gallese, V., & Goldman, A. (1998). Mirror neurons and the simulation theory of mind-reading. Trends in Cognitive Sciences, 2(12), 493-501. https://doi.org/10.1016/S1364-6613(98)01262-5

Heyes, C., & Catmur, C. (2022). What happened to mirror neurons? Perspectives on Psychological Science, 17(1), 153-168. https://doi.org/10.1177/1745691621990638

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. https://doi.org/10.1162/jocn.2009.21189

Iacoboni, M., Woods, R. P., Brass, M., Bekkering, H., Mazziotta, J. C., & Rizzolatti, G. (1999). Cortical mechanisms of human imitation. Science, 286(5449), 2526-2528. https://doi.org/10.1126/science.286.5449.2526

Kohler, E., Keysers, C., Umiltà, M. A., Fogassi, L., Gallese, V., & Rizzolatti, G. (2002). Hearing sounds, understanding actions: Action representation in mirror neurons. Science, 297(5582), 846-848. https://doi.org/10.1126/science.1070311

Molenberghs, P., Cunnington, R., & Mattingley, J. B. (2012). Brain regions with mirror properties: A meta-analysis of 125 human fMRI studies. Neuroscience & Biobehavioral Reviews, 36(1), 341-349. https://doi.org/10.1016/j.neubiorev.2011.07.004

Mukamel, R., Ekstrom, A. D., Kaplan, J., Iacoboni, M., & Fried, I. (2010). Single-neuron responses in humans during execution and observation of actions. Current Biology, 20(8), 750-756. https://doi.org/10.1016/j.cub.2010.02.045

Rizzolatti, G., Fadiga, L., Gallese, V., & Fogassi, L. (1996). Premotor cortex and the recognition of motor actions. Cognitive Brain Research, 3(2), 131-141. https://doi.org/10.1016/0926-6410(95)00038-0

Rizzolatti, G., & Craighero, L. (2004). The mirror-neuron system. Annual Review of Neuroscience, 27, 169-192. https://doi.org/10.1146/annurev.neuro.27.070203.144230

Rizzolatti, G., & Sinigaglia, C. (2010). The functional role of the parieto-frontal mirror circuit: Interpretations and misinterpretations. Nature Reviews Neuroscience, 11(4), 264-274. https://doi.org/10.1038/nrn2805