A possible role for mirror neurons in human speech has been raised by others (e.g., Hickok Reference Hickok2010). Here, I propose that mirror neurons (MNs) may play a role in other forms of vocal communication, specifically wherever imitative or responsive vocalization is found. In humans, examples would include contagious laughing or crying, and synchronized singing in duets or choral groups. In more general terms, I wish to raise the possibility that mirror-like neurons exist to foster vocal communication between individuals, where rapid exchange of information is required. Communication of this type exists in nature in duetting birds and mammals, and in rapid vocal exchanges between two individuals, where the appropriate response to a particular vocalization is a vocal response given after a brief interval. I further propose that neurons that serve this function are part of inherited neural circuitry that mediates a communication system, but become active only through experience.
There are numerous species of birds and mammals that engage in rapid and frequent vocal exchanges. These may be referred to as “duetting,” “antiphonal calling,” or simply vocal exchanges. In birds and some primates, duetting seems to serve to maintain or strengthen the pair-bond. In some cases, the duetting pair call in such rapid synchrony that it is often difficult to determine which partner is performing which part (e.g., Geissmann Reference Geissmann2002; Müller & Anzenberger Reference Müller and Anzenberger2002).
Such synchronization obviously requires a neural system that can manage this behavior, and, presumably, also mediate the learning processes that lead up to the performance of this behavior (Brenowitz et al. Reference Brenowitz, Arnold and Levin1985. I propose that a “mirror-neuron–like” system may be central to the performance of this behavior. Such an ensemble of neurons would, in each partner, monitor both the immediately ensuing component and trigger the expected output. How such an ensemble becomes activated in an individual may depend on associative learning, but there also must be a genetically derived program that puts together the components of the ensemble in the first place.
While Cook et al. propose that experimentation in the lab is needed to demonstrate the essential role of associative learning in MN formation, most of the behaviors I am addressing are probably not likely to occur in lab settings. However, some primates widely used in laboratory settings engage in vocal exchanges that can be observed while they are in captivity, and for which an established history in behavioral neuroscience exists. In squirrel monkeys (Saimiri), females who are affiliative partners (as determined by their close association during foraging and rest periods) produce “chuck” vocalizations in rapid exchanges of no more than a few hundred msec (Newman & Bernhards Reference Newman and Bernhards1997). Each affiliative partnership exchanges chucks with each other, and each individual makes chucks that are sufficiently distinct acoustically so that their partner can recognize them on the basis of their vocalizations alone. Experimentation has shown that one particular part of the chuck (referred to as the “flag”) is essential for accurate recognition and response (Soltis et al. Reference Soltis, Bernhards, Donkin and Newman2002). This behavior starts out during development as a more general “contact call,” in which young females respond indiscriminately to the chucks of other females. Gradually, over a year or two, a female learns to distinguish, and differentially respond to, the chucks of its mother and other affiliating females (McCowan & Newman Reference McCowan and Newman2000). The genetic component of this behavior is in the tendency to make chucks in the first place. The associative learning component comes from a young female gradually learning to restrict her chuck responses to the chucks of affiliative partners. While MN-like populations of neurons need not be necessary for this behavior, I propose that ensembles that regulate the production, monitoring, and subsequent response (and hence “MN-like”) would be more efficient (and likely favorably selected for) over populations of neurons that might be engaged in a variety of behavioral activities.
Another primate vocal communication system that might be mediated by a MN-like system is antiphonal calling in the common marmoset (Callithrix jacchus). Here, members of a mated pair exchange loud “phee” calls with other mated pairs within an interval of 5 sec or less (Miller & Wang Reference Miller and Wang2006). This behavioral system requires that a vocalizer recognizes when its partner calls, and when a member of another pair calls, and responds only to the latter with less than a 5-sec delay. No one knows how the brain mediates this behavior, but a study (Miller et al. Reference Miller, DiMauro, Pistorio, Hendry and Wang2010) using immunocytochemistry of c-fos gene expression has identified areas of the cortex that show neural activity to hearing a phee, to producing a phee, and to the production and hearing of a respondent's phee. No one, as yet, has recorded from single units in the identified frontal cortical areas in the marmoset, so it would be interesting to learn if MN-like activity was found there. Such specialized activity would likely occur both when the vocalizer called and when there were responses.
To summarize, ensembles of neurons that possess mirror-neuron–like properties are likely to exist in a wide range of birds and mammals, so that a correspondingly broad approach to identifying and learning more about these systems is needed. Cook et al. suggest that experiments in the lab would be needed to reveal the MNs emerging during associative learning. My suggestion that MN-like systems exist widely in nature implies that behavioral tests performed in a lab setting would be insufficient to explore the full extent of their role in communication. Some species are too vulnerable in nature or otherwise protected from invasive experimentation, but, with the improving technology to record neural activity in freely moving animals, it might be possible to study these systems. Some work being done on European Starlings (Sturnus vulgaris) has potential here (George et al. Reference George, Alcaix, Henry, Richard, Cousillas and Hausberger2010).
A possible role for mirror neurons in human speech has been raised by others (e.g., Hickok Reference Hickok2010). Here, I propose that mirror neurons (MNs) may play a role in other forms of vocal communication, specifically wherever imitative or responsive vocalization is found. In humans, examples would include contagious laughing or crying, and synchronized singing in duets or choral groups. In more general terms, I wish to raise the possibility that mirror-like neurons exist to foster vocal communication between individuals, where rapid exchange of information is required. Communication of this type exists in nature in duetting birds and mammals, and in rapid vocal exchanges between two individuals, where the appropriate response to a particular vocalization is a vocal response given after a brief interval. I further propose that neurons that serve this function are part of inherited neural circuitry that mediates a communication system, but become active only through experience.
There are numerous species of birds and mammals that engage in rapid and frequent vocal exchanges. These may be referred to as “duetting,” “antiphonal calling,” or simply vocal exchanges. In birds and some primates, duetting seems to serve to maintain or strengthen the pair-bond. In some cases, the duetting pair call in such rapid synchrony that it is often difficult to determine which partner is performing which part (e.g., Geissmann Reference Geissmann2002; Müller & Anzenberger Reference Müller and Anzenberger2002).
Such synchronization obviously requires a neural system that can manage this behavior, and, presumably, also mediate the learning processes that lead up to the performance of this behavior (Brenowitz et al. Reference Brenowitz, Arnold and Levin1985. I propose that a “mirror-neuron–like” system may be central to the performance of this behavior. Such an ensemble of neurons would, in each partner, monitor both the immediately ensuing component and trigger the expected output. How such an ensemble becomes activated in an individual may depend on associative learning, but there also must be a genetically derived program that puts together the components of the ensemble in the first place.
While Cook et al. propose that experimentation in the lab is needed to demonstrate the essential role of associative learning in MN formation, most of the behaviors I am addressing are probably not likely to occur in lab settings. However, some primates widely used in laboratory settings engage in vocal exchanges that can be observed while they are in captivity, and for which an established history in behavioral neuroscience exists. In squirrel monkeys (Saimiri), females who are affiliative partners (as determined by their close association during foraging and rest periods) produce “chuck” vocalizations in rapid exchanges of no more than a few hundred msec (Newman & Bernhards Reference Newman and Bernhards1997). Each affiliative partnership exchanges chucks with each other, and each individual makes chucks that are sufficiently distinct acoustically so that their partner can recognize them on the basis of their vocalizations alone. Experimentation has shown that one particular part of the chuck (referred to as the “flag”) is essential for accurate recognition and response (Soltis et al. Reference Soltis, Bernhards, Donkin and Newman2002). This behavior starts out during development as a more general “contact call,” in which young females respond indiscriminately to the chucks of other females. Gradually, over a year or two, a female learns to distinguish, and differentially respond to, the chucks of its mother and other affiliating females (McCowan & Newman Reference McCowan and Newman2000). The genetic component of this behavior is in the tendency to make chucks in the first place. The associative learning component comes from a young female gradually learning to restrict her chuck responses to the chucks of affiliative partners. While MN-like populations of neurons need not be necessary for this behavior, I propose that ensembles that regulate the production, monitoring, and subsequent response (and hence “MN-like”) would be more efficient (and likely favorably selected for) over populations of neurons that might be engaged in a variety of behavioral activities.
Another primate vocal communication system that might be mediated by a MN-like system is antiphonal calling in the common marmoset (Callithrix jacchus). Here, members of a mated pair exchange loud “phee” calls with other mated pairs within an interval of 5 sec or less (Miller & Wang Reference Miller and Wang2006). This behavioral system requires that a vocalizer recognizes when its partner calls, and when a member of another pair calls, and responds only to the latter with less than a 5-sec delay. No one knows how the brain mediates this behavior, but a study (Miller et al. Reference Miller, DiMauro, Pistorio, Hendry and Wang2010) using immunocytochemistry of c-fos gene expression has identified areas of the cortex that show neural activity to hearing a phee, to producing a phee, and to the production and hearing of a respondent's phee. No one, as yet, has recorded from single units in the identified frontal cortical areas in the marmoset, so it would be interesting to learn if MN-like activity was found there. Such specialized activity would likely occur both when the vocalizer called and when there were responses.
To summarize, ensembles of neurons that possess mirror-neuron–like properties are likely to exist in a wide range of birds and mammals, so that a correspondingly broad approach to identifying and learning more about these systems is needed. Cook et al. suggest that experiments in the lab would be needed to reveal the MNs emerging during associative learning. My suggestion that MN-like systems exist widely in nature implies that behavioral tests performed in a lab setting would be insufficient to explore the full extent of their role in communication. Some species are too vulnerable in nature or otherwise protected from invasive experimentation, but, with the improving technology to record neural activity in freely moving animals, it might be possible to study these systems. Some work being done on European Starlings (Sturnus vulgaris) has potential here (George et al. Reference George, Alcaix, Henry, Richard, Cousillas and Hausberger2010).
ACKNOWLEDGMENT
This research was supported (in part) by the Intramural Research Program of the NIH, NICHD.