1 Introductory Remarks

The mechanism represented by mirror neurons (mirror mechanism) unifies action perception and action execution (Di Pellegrino et al. 1992; Gallese et al. 1996; Rizzolatti et al. 1996a; Gallese et al. 2002; Fogassi et al. 2005; Rozzi et al. 2008). The essence of this mechanism is the following: each time an individual observes another individual performing an action, a set of neurons that encode that action is activated in the observer’s own cortical motor system.

In this chapter we will first introduce the basic function of goal coding in the motor system, and describe the properties of the parieto-frontal action observation/action execution (mirror) circuit in monkeys and humans. Then we will show how, based on first-person knowledge, the mirror system encodes the intention of other individuals and how this function can be impaired in autistic patients. Finally we will show some evidence of plasticity within the mirror system.

2 Goal Coding in the Monkey Cerebral Cortex

A traditional view on information processing in the cerebral cortex maintained that its posterior (parietal and temporal) sector is devoted to perception (high order elaboration of sensory input), while its anterior (frontal) sector plays a crucial role in movement programming and execution, on the basis of information provided by the “perceptual” part of the cortex. This basically serial view was challenged by the neuroanatomical and neurophysiological data accumulated in the last three decades. Briefly, neuroanatomical data showed that most of the connections between posterior and anterior cortical areas are reciprocal, thus indicating that the flow of information runs in parallel, leading to a strict reciprocal influence between action and perception (Rizzolatti et al. 1998; Rizzolatti and Luppino 2001). Neurophysiological data showed that the motor cortex, far from being a purely executive cortical sector, contains stored representations of the goals of motor acts (Rizzolatti et al. 1988). Through the above-mentioned neuroanatomical connections, the role of these motor signals is that of providing a meaning to the incoming sensory information provided by the posterior cortical areas. For example, when I see an object in the external space, besides visual recognition, its physical properties are immediately transformed in a motor format, that is, in the goal-related motor act appropriate for interacting with that object (see Jeannerod et al. 1995). However, if the context does not allow the execution of this motor act, the activation of the motor system remains in the state of a potential motor act. Thus, our understanding of the external world is, at least partly, based on the automatic activation of the motor system.

Evidence for goal coding in the motor system has been given by single neurons recording experiments carried out on ventral premotor cortex (area F5, see Fig. 9.1) showing that most of its motor neurons discharge during the execution of goal-related motor acts such as grasping, manipulating, breaking, etc., rather than during execution of simple movements, i.e., body-parts displacements without a specific goal (e.g., finger flexion) (Rizzolatti et al. 1988; Kakei et al. 2001). Compelling evidence that this is the case was recently provided by Umiltà et al. (2008). They recorded single neurons in monkeys trained to grasp objects using two different types of pliers: “normal pliers,” which require typical grasping movements of the hand, and “reverse pliers,” which require hand movements executed in the opposite order (i.e., closing first and then opening the fingers). The results showed that F5 neurons discharged during the same phase of grasping in both conditions, regardless of whether this involved opening or closing of the fingers.

Fig. 9.1
figure 00091

Lateral view of the monkey brain showing the parcellation of the agranular frontal and posterior parietal cortices. Agranular frontal areas are defined according to Matelli et al. (1985, 1991). All posterior parietal areas are defined according to Pandya and Seltzer (1982) and Gregoriou et al. (2006). AI inferior arcuate sulcus; AS superior arcuate sulcus; C central sulcus; IP inferior parietal sulcus; L lateral fissure; P principal sulcus; STS superior temporal sulcus

Full size image

Area F5 belongs to the above-mentioned set of circuits connecting parietal and frontal cortex. Specifically, it is connected with a sector of the inferior parietal lobule (IPL), namely areas PFG (see Fig. 9.1) and AIP (an area buried inside the rostral part of the inferior parietal sulcus). Interestingly, the functional properties of IPL motor neurons seem to be similar to those of F5 neurons, that is, they are active during the execution of goal-directed motor acts rather than the single movements constituting them (Hyvärinen 1982; Sakata et al. 1995; Fogassi et al. 2005; Rozzi et al. 2008).

3 The Parieto-Frontal Mirror Circuit

3.1 The Monkey Parieto-Frontal Network

The mirror mechanism was originally discovered in the ventral premotor cortex of the macaque monkey (area F5) (Di Pellegrino et al. 1992; Gallese et al. 1996; Rizzolatti et al. 1996a). Single neuron recordings showed that in this area there are neurons that fire both when a monkey executes a specific motor act and when it observes another individual (either a conspecific or an experimenter) performing the same motor act (mirror neurons, Fig. 9.2). Mirror neurons do not respond to the simple object presentation and do not respond, or respond only weakly, to the observation of the experimenter performing a hand motor act (e.g., grasping) without a target object.

Fig. 9.2
figure 00092

Mirror neuron responding during observation and execution of a hand grasping motor act. The neuron shows a visual response when an experimenter grasps a piece of food in front of the monkey and when the monkey grasps the same piece of food from the experimenter’s hand. The silence between the visual and the motor responses corresponds to the time in which the experimenter approaches the plate with food to the monkey, before it grasps it. Rasters and histograms are aligned with the moment in which the experimenter’s hand touches the food. Abscissae: time; ordinates: spikes per bin; bin width: 20 ms (modified from Rizzolatti et al. 1996a)

Full size image

Although the response of most mirror neurons is not influenced by many visual details of the observed act, some of them show specificity for the direction or the space sector in which the act is performed or the hand (left or right) used by the observed agent.

Since their first discovery it has been suggested that mirror neurons have a prominent role in the understanding of the goal of observed motor acts. However, one could have argued that their response, due to the visual presentation of a motor act, could in principle express a simple visual recognition of a biological movement, without allowing to assign it a motor meaning. This criticism was overcome by two studies in which it was demonstrated that the vision of the motor act is not a necessary requisite for activating mirror neurons. In a first study (Umiltà et al. 2001) it has been demonstrated that mirror neurons responded both when the monkey could fully observe a grasping act and when it could see only part of it because the hand–target interaction was hidden behind a screen. Interestingly, the response was absent when the monkey knew that no object was present behind the screen (mimicked hidden motor act), suggesting that mirror neurons have access to prior contextual information in order to retrieve the motor representation corresponding to the observed motor act, despite the absence of a full visual description of the motor act.

In a second study (Kohler et al. 2002) it has been demonstrated the capacity of mirror neurons to respond to the sound of motor acts. These neurons have been called audio-visual mirror neurons and constitute a subcategory of F5 mirror neurons. They activate when a monkey not only observes, but also hears the sound of a motor act. Their response is specific for the type of motor act seen and heard. For example, these neurons respond to peanuts breaking when the act is only observed, only heard, or both heard and observed, and do not respond to the vision and sound of another act, or to nonspecific sounds. The presence of audio-visual mirror neurons demonstrates that, beyond the visual input, also the acoustic input related to biological actions can have access to the representation of the goal of motor acts.

The most important property of mirror neurons is the congruence they show between the visual and the motor response, that is, the matching between the goal of the observed motor act and that of the executed motor act. This property is crucial, because it enables the observer to understand what another individual is doing. In other words, during observation of a motor act, the corresponding motor representation is automatically retrieved in the motor cortex of the observer. Note that, during observation, observers normally do not mimic the observed motor acts. This means that an inhibitory mechanism is at work, so that the “motor resonance” elicited in the observer does not become an overt motor output. Interestingly, very recently Kraskov et al. (2009) demonstrated that half of F5 mirror neurons that activated during grasping execution were inhibited during grasping observation. This inhibition could, at least in part, explain why the observed motor act is not automatically converted in its execution.

Up to now we described, as main function of mirror neurons, that of understanding the goal of motor acts, without entering in the issue of which could be their role in the behavioral reactions consequent to the observation of other individuals’ actions. A more recent study (Caggiano et al. 2009) allows to propose some answer in this direction. The main aim of the study was that of assessing whether the discharge of mirror neurons can be modulated by the distance at which the observed act is performed. The same motor act was performed by the experimenter inside the monkey reaching space (peripersonal space) or outside it (extra-personal space). It has been found that 50 % of mirror neurons were differently active in the two conditions. Of them, half discharged stronger when the experimenter grasped a piece of food within the monkey peripersonal working space, while the other half responded better when the same motor act was performed in the extra-personal space. Interestingly, when the monkey working space was shortened by the presence of a barrier, extra-personal neurons started to discharge strongly also within the peripersonal space, as if it were become far. Taken together, these data suggest that mirror neurons could code other’s action within different spaces, and that this property could be related to the possibility to socially interact (cooperate, compete) with others.

Mirror neurons are also present in the rostral part of the IPL, particularly in area PFG (Gallese et al. 2002; Fogassi et al. 2005; Rozzi et al. 2008) and AIP (the anterior intraparietal area) (Rizzolatti et al. 2009). The properties of parietal mirror neurons are quite similar to those of F5. Both areas PFG and AIP are heavily connected with F5 (Borra et al. 2008; Rozzi et al. 2006; Gerbella et al. 2011). These two areas receive higher order visual information from the cortex located inside the superior temporal sulcus (STS) (Rozzi et al. 2006; Borra et al. 2008, see also Fig. 9.1). STS areas encode, as mirror areas, biological actions, but they lack motor properties. AIP receives also connections from the inferior temporal gyrus (Borra et al. 2008). This input could provide the mirror areas with information concerning object identity.

3.2 Evidence for New Types of Mirror Neurons

LIP mirror neurons. An interesting function that involves an interaction between two individuals is shared attention. When an individual, for example, is looking in a given direction, an observer located in front of him tends to gaze to the same direction (Gaze following). This behavior can be functional to share the same target at which the first individual’s gaze is directed. Neurophysiologically, observation of the eye position of another monkey is known to activate neurons in the STS (Perrett et al. 1992). Only recently, however, it has been demonstrated the presence of mirror neurons for eye movements in the lateral intraparietal area (LIP). This area, located inside the intraparietal sulcus (IPS), is part of a circuit involving the frontal eye field and plays a crucial role in organizing intended eye movements. Most of its neurons discharge when the monkey looks in a specific direction (Barash et al. 1991). Interestingly, a subset of them has been found to discharge also when a monkey observed another monkey looking in the neuron motor preferred direction (Shepherd et al. 2009). This finding suggests that the motor system involved in the control of eye movements towards targets is endowed with a mirror-like mechanism. In sharing attention, the automatic social reaction to another individual’s gaze might rely on this mirror-like mechanism.

VIP mirror neurons. Previous studies showed that VIP neurons encode tactile and visual stimuli delivered in the peripersonal space of the monkey (Colby et al. 1993; Duhamel et al. 1998). Ishida et al. (2009) demonstrated that some of these neurons also respond to stimuli presented in the peripersonal space of an individual located at about one meter far from the monkey and facing it. It is important to keep in mind that area VIP is strictly connected with area F4, which represents peripersonal space and whose neurons discharge during reaching movements. It is plausible, therefore, that neuronal responses that seem to be induced by visual stimuli actually represent potential motor acts directed towards specific body parts (Fogassi et al. 1996). The study on VIP neurons is of great interest because it shows that the mirror mechanism of this area encodes body-directed rather than object-directed motor acts, thus opening fascinating possibilities for individuals to encode the body of others.

Altogether, the described studies on LIP and VIP indicate that the function of mirror neurons is related to the motor properties of the areas in which they are located.

4 The Human Parieto-Frontal Mirror System

Sensory, motor, and cognitive functions can be studied in humans by means of electrophysiological (EEG; MEG; TMS) and neuroimaging (PET, fMRI) techniques. These techniques have been successfully employed in the last 15 years to demonstrate that an action observation/action execution mirror circuit also exists in humans.

Brain imaging studies have shown that, as in the monkey, this action observation/action execution mirror circuit is formed by two main regions (Fig. 9.3): (1) the inferior sector of the precentral gyrus plus the posterior part of the inferior frontal gyrus (IFG); (2) the IPL including the cortex located inside the IPS (see Rizzolatti and Craighero 2004; Rizzolatti et al. 2009). Additional cortical areas (such as the dorsal premotor cortex and the superior parietal lobule) have been also occasionally found to be active during action observation. These areas are active when volunteers are asked to observe proximal arm movements directed to a particular location in space (Filimon et al. 2007).

Fig. 9.3
figure 00093

Cortical areas belonging to the parieto-frontal mirror system. Gray-shaded regions indicate cortical sectors activated during action observation, that become also active during execution of the same actions. Note that in some studies additional cortical areas (e.g. dorsal premotor cortex and superior parietal lobule) can activate during observation of reaching or body movements. A rostral sector of the superior temporal sulcus also activate during action observation, but not during action execution. IFG, inferior frontal gyrus; ifs, inferior frontal sulcus; IPL, inferior parietal lobule; IPS, intraparietal sulcus; L, lateral sulcus; M1, primary motor cortex; PMD, dorsal premotor cortex; PMV, ventral premotor cortex; sfs, superior frontal sulcus; SPL, superior parietal lobule; STS, superior temporal sulcus (from Cattaneo and Rizzolatti 2009)

Full size image

By using single-subject fMRI analyses, evidence has been recently provided that other cortical areas (e.g., SI, SII, middle temporal cortex) become active during action observation and action execution (Gazzola and Keysers 2009). It has been suggested that these activations outside the “classical” mirror areas reflect additional mechanisms (e.g., internal models) that are triggered by the mirror mechanism. These activations would enrich the information about other individuals’ actions that the mirror mechanism provides.

In agreement with early findings (Rizzolatti et al. 1996b; Buccino et al. 2001; Decety et al. 2002), a series of new fMRI studies provided strong evidence that the human mirror parieto-frontal circuit encodes the goal of observed motor acts. Gazzola et al. (2007a, b) instructed volunteers to observe video clips in which either a human or a robot arm grasped objects. In spite of differences in shape and kinematics between the human and robot arms, the parieto-frontal mirror circuit was activated in both conditions. These results were extended by Peeters et al. (2009), who investigated the cortical activations in response to the observation of motor acts performed by a human hand, a robot hand, or a tool. They found bilateral activation of a mirror network formed by intraparietal and ventral premotor cortex regardless of the acting effector. In addition, the observation of tool actions produced a specific activation of a rostral sector of the left anterior supramarginal gyrus, suggesting that this sector specifically evolved for tool use.

Unlike monkeys, the parieto-frontal mirror circuit of humans becomes also active during the observation of individual movements (Rizzolatti et al. 1999; Lui et al. 2008). The initial evidence for this mechanism was based on TMS experiments that indicate that the observation of others’ movements results in an activation of the muscles involved in the execution of those movements (Fadiga et al. 1995; Strafella and Paus 2000; Gangitano et al. 2001). Additional support comes from EEG and MEG studies showing that the observation of movements without a goal desynchronizes the electroencephalographic rhythms recorded from motor areas (Hari et al. 1998; Cochin et al. 1998; Kilner et al. 2009). These data suggest that in humans both observation of goal-directed actions and of simple movements can activate the motor system. These two types of activation are very likely used for different purposes.

5 Understanding Actions Based on First-Person Knowledge

Most of the data reviewed up to now indicate that action understanding is based on a first-person motor knowledge. However, it has been proposed that action understanding could occur by analyzing the different visual elements of the observed actions and applying to them some form of inferential reasoning (see Wood and Hauser 2008). Actually, in some cases, motor behavior might require a mechanism different from mirroring in order to be understood. The capacity of humans to recognize animals’ actions that do not belong to the human motor repertoire and cannot be captured by motor generalization is a typical example in this regard. Evidence for the existence of both a mirror and a non-mirror mechanism in non-conspecifics action recognition has been provided by Buccino et al. (2004b). In their fMRI study volunteers were presented with video clips showing motor acts that did or did not belong to the human motor repertoire. The former consisted of ingestive actions performed by a conspecific or by animals (dog and monkey). The latter consisted of communicative gestures (silent speech, dog barking, and monkey lip-smacking). Although all volunteers recognized the observed motor acts regardless of whether or not they belonged to their own motor repertoire, the parieto-frontal mirror system was activated during observation of all ingestive actions and during observation of silent speech. Instead, no activation of parieto-frontal mirror areas was found in the case of those acts that did not belong to it (e.g., dog barking). The areas that became active in the last case were occipital visual and STS areas. These data indicate that the recognition of others’ motor behavior can rely on the mere processing of its visual aspects, but it does not provide the observer with information necessary for a real understanding of the message (e.g., the communicative intent of the barking dog). By contrast, when the observed motor act activates the motor system through the mirror mechanism, that action is not only visually recognized but also understood, because there is a sharing of motor goal by the observer and the agent. In other terms, the observed action is understood from the inside in motor terms and not from the outside as a mere visual description.

6 Understanding the Motor Intentions of Others

6.1 A Matching Mechanism Based on Action Organization

When we perform a complex action we intend to achieve a given behavioral goal. Thus, our intention does not correspond to a general preparation to act, but specifies an ultimate goal. In this sense, the term intention is used with a meaning different from that used by other authors in neuroscience. For example, according to some proposals, intention represents a sort of readiness to start a movement, according to others, a preparation of a precise movement or motor act, including programming of some motor parameters (for example, direction of an impending reaching movement). On the contrary, according to our proposal, the agent’s intention includes the selection of a final goal—on the basis of his motivation and of the context—and the organization of the sequence of motor acts necessary to achieve this goal. Interestingly, each motor act belonging to an intentional action has its subgoal, the achievement of which is instrumental for the unfolding of the action sequence, because it prepares the following motor act. The questions are how intentional actions are coded in the parieto-premotor cortical circuits and whether the neurons coding the goal of motor acts are influenced by the ultimate action goal. In order to provide a first answer to this question, grasping neurons were recorded from areas PFG and F5 while the monkey executed a motor task and observed the same task, performed by an experimenter, in which the same motor act (grasping) was embedded into two different actions (grasping to eat and grasping to place) (Fogassi et al. 2005; Bonini et al. 2010). The results showed that a high percentage of parietal and premotor neurons discharged differently when the monkey performed the grasping act, depending on the final goal of the action (either eating or placing) in which the act was embedded. This finding implies that areas F5 and PFG are constituted of chains of neurons in which each neuron encodes a given motor act and is linked to another one selective for the next motor act in the sequence (Fogassi et al. 2005; Rizzolatti et al. 2006). Together they encode a specific action intention (e.g., eating or placing).

Similarly to the motor task, during the visual task it has been found that most mirror neurons discharged differently during observation of grasping, when this act was embedded into different actions. Because in this case the grasping act was performed by the observed agent, it was suggested that the neuronal selectivity for the action goal during grasping observation represents a prediction of the action outcome. Thus, in agreement with the “chain” interpretation of the results of the motor task, the observation of a motor act embedded in an action would activate a chain corresponding to a specific intention. This activation would allow one to understand automatically the motor intentions of others.

These data underline two important concepts. First, the intention to achieve a given motor goal is directly represented in the motor system by a dedicated “chained” neuronal organization. Thus, the motor system does not only encode the goals of motor acts, but also the ultimate action goals. Second, in spite of the mentalistic interpretation of the strategies we use to decode others’ intentions, the motor system offers a very simple, automatic mechanism to decode others’ intention in most of the contexts of our daily life. Once again, this mechanism provides first-person knowledge of others’ behavior.

Evidence that in humans the parieto-frontal mirror circuit is also involved in intention encoding was first provided by an fMRI experiment by Iacoboni et al. (2005). The experiment consisted of three conditions. In the first (“context condition”) the volunteers saw video clips showing scenes arranged as to represent an ongoing breakfast or arranged as if the breakfast had just finished (“context” condition); in the second, the volunteers saw video clips showing a hand grasping a mug on an empty background (“action” condition); in the third, they saw videos showing the same hand motor act within the two contexts (“intention” condition). In this latter condition, the context provided clues for understanding the motor act intention. The results showed that the intention condition induced a stronger activation, relative to the other two conditions, in the caudal IFG of the right hemisphere.

The presence of a chain mechanism underlying intentional actions has been indirectly shown in humans with a behavioral experiment very similar to that used in monkeys and described above. Cattaneo et al. (2007) asked children to grasp a piece of food for eating or for placing it in a container, or to observe an experimenter performing the same actions. During both execution and observation conditions, the EMG activity of the mylohyoid (MH) muscle—a muscle involved in mouth opening—was recorded. Both the execution and the observation of the eating action determined an increase of MH activity during the reaching phase, before object contact, whereas no MH activity was recorded during the execution and the observation of the placing action. This indicates that, as soon as the action starts, the entire motor “chain” involved in action execution is activated. On the observation side, the activation of the same chain would allow the observer to predict what action the agent is going to execute and thus to understand the agent’s motor intention.

A mirror mechanism, located in the fronto-mesial areas, can also play a role in representing the motor behavior of others in advance. It has been shown that the “Bereitschaftspotentials,” an electrophysiological marker of the readiness to act (Deecke et al. 1969), occurs not only when an individual executes a motor act, but also when the nature and the onset time of an upcoming action performed by another individual is predictable on the basis of a visual cue (Kilner et al. 2004).

6.2 Mirroring Intentions and Inferring Reasons

Intention understanding is a multilayered process involving different levels of action representation, from the motor intention that drives a given chain of motor acts to the propositional attitudes (beliefs, desires, etc.) that—at least in humans—can be assumed to explain the observed behavior in terms of its plausible reasons. Thus, while in our daily life we are usually able to understand others’ intention through a fast, automatic process, very likely relying on the mirror mechanism, there are cases in which additional inferential processes (Rizzolatti and Sinigaglia 2007; Gallese 2007) are needed. In line with these considerations, recent empirical data showed that, although the parieto-frontal mirror mechanism is active in all conditions in which a motor task has to be directly understood, when volunteers were required to judge the reasons behind the observed actions, there was an activation of a sector of the anterior cingulate cortex and of other areas of the so-called mentalizing network (de Lange et al. 2008). Activation of the same network was also shown in a study (Brass et al. 2007) that investigated unusual actions performed in implausible vs. plausible contexts, as well as in a study (Liepelt et al. 2008) that studied the neural basis of reason inference in non-stereotypic actions.

The areas belonging to this network have as yet not been demonstrated to have mirror properties. There have been several proposals trying to integrate these two ways of understanding other’s intentions (Kilner and Frith 2007; Keysers and Gazzola 2007). However, differently from the mirror system, there are currently no neurophysiological data that can explain how the “mentalizing network” might work.

6.3 Intention Understanding in Autistic Patients

Autistic spectrum disorder (ASD) is a syndrome characterized by impairment in social skills, communicative abilities, emotional responses, and motor behavior (see Frith 2003). Although a number of electrophysiological and brain imaging experiments (Oberman et al. 2005; Théoret et al. 2005; Dapretto et al. 2006; Martineau et al. 2008) showed that individuals with ASD have an impairment of the mirror mechanism, some recent behavioral studies have challenged this view (Hamilton et al. 2007; Leighton et al. 2008). Cattaneo et al. (2007) provided an answer to this discrepancy. They asked children with ASD to grasp a piece of food either for eating or for placing it and, in another condition, to observe an experimenter performing these actions. The activity of the mylohyoid (MH) muscle, a muscle involved in mouth opening, was recorded. Unlike typically developing children (see above), in whom MH activation was already present during the “reaching” and “grasping” phases of the grasping-for-eating action, children with ASD showed MH activation only during the “bringing-to-the-mouth” phase. Furthermore, while typically developing children exhibited MH activation when observing a grasping-for-eating action, ASD children did not. These data indicate that children with ASD have a severe impairment in motor organization that includes a deficit in chaining motor acts into intentional actions and, as a consequence, a lack of activation of intentional motor chains during action observation. ASD children, in order to understand others’ actions do not use their internal motor knowledge, but another cognitive strategy. This interpretation is supported by a recent study showing that, although ASD children can understand the meaning of a motor act (e.g., grasping) performed by another agent, they are not able to understand the intention of the whole action. In fact, in order to understand intention, they must rely not on the observed motor behavior, but on the semantics of the object that is being manipulated or on the context in which the motor act takes place (Boria et al. 2009).

7 Plasticity of the Mirror System

A very important issue strictly linked to the properties of the mirror system is whether mirror neurons activity can be modified by experience and learning. Although in monkeys this issue will be probably best addressed in the future by chronic recording experiments, some recent monkey data show that mirror neurons can achieve new properties during visuomotor learning. Rochat et al. (2010), in a study in which F5 grasping neurons were recorded in monkeys trained to grasp objects using tools, reported the presence of F5 mirror neurons responding to the observation of grasping motor acts performed by an experimenter with the hand or with a tool, although the response during observation of grasping made with the tool was weaker than that during observation of hand grasping. This study illustrates that when a novel motor act is incorporated in the own motor repertoire, this allows to establish a new motor resonance during the observation of this act, provided that its goal is similar to that achieved with the hand.

Among the studies showing the presence of the mirror system in humans, a couple of them addressed the issue of whether a different motor experience could determine a different activation of this system. In a first fMRI study (Calvo-Merino et al. 2005) participants, who included classical dancers, dancers of Capoeira (a South American dance), and people naïve in professional dance, observed video clips showing steps of classical dance or Capoeira. Although all groups had an activation of the mirror system, the observation of Capoeira with respect to classical dance caused a greater activation in the Capoeira dancers, while the opposite was observed in classical dancers. Naïve subjects did not show differential activation between the two conditions.

In a second study, similar experience-dependent changes in the mirror system have been described in expert (Bangert et al. 2006) and in naïve piano players, but subjected to training (Lahav et al. 2007), that were required to listen to music after motor training.

The plastic change of the mirror system with motor experience was observed in the course of learning by Cross et al. (2006). In their study, expert dancers had to learn and rehearse novel, complex whole-body dance sequences for 5 weeks. Functional MRI was performed every week while the dancers observed and imagined performing movement sequences, half of which were rehearsed and half unpractised. The results showed that the activation of the mirror system was modulated by the dancers’ motor experience, with an increase of activity in PMv and IPL during observation of the rehearsed sequences.

The mirror system reveals its plasticity also in situations in which individuals lack an effector or are blind. In a study by Gazzola et al. (2007b) two aplasic individuals, born without arms or hands, participated in an fMRI study in which they had to observe goal-related hand motor acts. Typically developed subjects, observing the same videos, were used as control. In a second part of the study, both aplasic and normal subjects executed mouth and foot motor acts, while only control subjects performed hand motor acts, in order to map the effectors motor representation. This study achieved two important results. First, during observation, aplasic subjects presented a mirror system activation similar to that of controls. Second, during hand motor acts observation, in the frontal cortex they had an activation of the mouth and foot representation. This means that there was a recruitment, from the motor repertoire, of cortical representations involved in the execution of motor acts that achieve similar goals, i.e., taking possession of an object, using different effectors. Thus, the mirror system is not only modified by motor experience, but also undergoes plastic changes similar to those already demonstrated in sensory systems after deprivation of the afferent input.

In another study, Ricciardi et al. (2009) showed that when congenitally blind patients listened to the sound of actions there is an activation of a fronto-parieto-temporal system, corresponding to the regions activated in the normally sighted controls, during observation of and listening to the same actions. Furthermore, the sound of familiar actions caused a greater activation of the mirror system in both blind and normally sighted subjects. Thus, regions that in normal developing individuals are devoted to visuomotor integration during observation/execution of actions, accomplish the same functions in congenitally blind individuals, by exploiting a different sensory channel.

8 Conclusions

The discovery of mirror neurons has opened a wide spectrum of investigations in the motor cognitive domain and beyond, because it constitutes a basic mechanism matching action execution and action observation that allows the understanding of other’s actions from inside. Interestingly, this mechanism seems to be a very basic way of understanding, since its presence has been demonstrated not only in humans and monkeys, but also in singing birds, like swamp sparrows (Prather et al. 2008) and zebra finches (Keller and Hahnloser 2009). Furthermore, this mechanism appears to constitute a deep link between individuals that is fundamental for establishing interindividual relationships. The evidence on people with autism suggests a strong role of motor knowledge and of the mirror mechanism, based on this knowledge, in mediating the capacity to understand others’ behavior and to entertain interindividual interactions.