Home
CV & Biography
Publications
Collaborators
All Souls College
Latest

Where do mirror neurons come from?

Mirror neurons are thought to play important roles in imitation and action understanding.  In a recent paper (Heyes, 2010), I argued that mirror neurons are forged by associative learning - the same kind of learning that produces Pavlovian conditioning.  The theory is outlined in the following figure.  If this theory is correct, we should expect to find mirror neurons in a wide range of animals, and it should be possible to 'repair' the mirror neuron system using sensorimotor training. 

 

Fig. 1. Mirror neurons from associative learning. The associative hypothesis (or 'Associative Sequence Learning' hypothesis; Heyes, 2001, 2005; Heyes and Ray, 2000) proposes that mirror neurons are acquired in the following way. Before learning (A), sensory neurons in the superior temporal sulcus, which are responsive to different high level visual properties of an observed action (S1, S2, Sn; Oram and Perrett, 1994, 1996) are weakly and unsystematically connected (dashed arrows) to some motor neurons in the premotor (Rizzolatti et al., 1988) and parietal cortices (M1, M2, Mn; Gallese et al., 2002), which discharge during the execution of actions with different high-level properties. For example, S1, which fires during observation of a precision grip, is weakly connected to both M1 and M2, which discharge during execution of a precision grip and a power grip, respectively. The kind of learning that produces mirror neurons (B) occurs when there is correlated (i.e. contiguous and contingent, Table 2) activation of sensory neurons and motor neurons that are each responsive to similar actions. For example, when an adult imitates an infant's facial movements (a), there might be correlated activation of neurons that are responsive to the observation (Sn) and execution (Mn) of lip protrusion. Correlated activation of Sn and Mn increases the strength of the connection between them, so that activation of Sn is propagated toMn. Therefore, after learning (C),Mn is active, not only during execution of lip protrusion, but also, via its connection with Sn, during observation of lip protrusion, i.e.Mn has become a lip protrusion mirror neuron. Correlated activation of sensory and motor neurons encoding the same property of action occurs not only when we are imitated (a), but also when we use optical mirrors (b), watch our own actions (c), and observe others during the kind of synchronous activities involved in sports and dance training (d) (Ray and Heyes, in press). Correlated activation of visual and motor neurons can also be produced indirectly by accompanying sounds. When the same sound (e.g. smacking) has been heard during observation and execution of an action (e.g. lipsmacking), hearing the sound will activate both visual and motor neurons encoding that action (Heyes and Ray, 2000; Keysers et al., 2003).