We agree with Cook et al. that sensorimotor associative learning takes place pervasively in the brain, as well as among mirror neurons (MNs) in particular. However, many associative learning theories stress the importance of cognitive representations in Pavlovian and instrumental conditioning (Balleine & Dickinson Reference Balleine and Dickinson1998; Berridge & Robinson Reference Berridge and Robinson2003). Without consideration of the roles of cognitive representations, MNs end up being the action-executing motor neurons (the primary motor cortex), as opposed to the neurons responsible for action-planning (Schubotz & Von Cramon Reference Schubotz and Von Cramon2003) an issue recognized by Cook et al.

We recommend that the “cardinal feature” of MNs – sensorimotor matching properties – should be best understood in a way no different from that to understand how the unity of conscious perception is brought about by integrating distributed activities in the brain, known as the “binding problem” (Revonsuo & Newman Reference Revonsuo and Newman1999). This issue applies to a range of brain functions including perception (Treisman & Gelade Reference Treisman and Gelade1980), volitional emotion regulation (Phan & Sripada Reference Phan, Sripada, Armony and Vuilleumer2013; Swain et al. Reference Swain, Konrath, Brown, Finegood, Akce, Dayton and Ho2012), and social cognitive emotional interaction (Ho et al. Reference Ho, Gonzalez, Abelson and Liberzon2012). For example, the Theory of Event Coding (TEC) by Hommel et al. (Reference Hommel, Müsseler, Aschersleben and Prinz2001) suggested that a domain-general binding process can generate “event files,” that is, cognitive representations (memory) of events, which can be any to-be-perceived or to-be-generated incident in the environment. Interestingly, there is evidence that such “event files” can be formed after a single sensory-motor encounter and represented by a network of loosely linked nodes rather than a master file (Hommel Reference Hommel1998). Because TEC has been linked to MNs (Hommel Reference Hommel2004), a direct comparison and contrast of TEC and the account presented in the target article should be performed.

One specific example of MN activity is in the parent–infant dyad. Neuroimaging studies are beginning to explore this in mothers (Mayes et al. Reference Mayes, Swain and Leckman2005; Swain Reference Swain2011). Recent findings (Atzil et al. Reference Atzil, Hendler and Feldman2011; Reference Atzil, Hendler, Zagoory-Sharon, Winetraub and Feldman2012) have demonstrated brain activations in response to baby-videos that constitute a social network involved in mentalization, action representation, simulation, mirroring, and attention functions (Dodell-Feder et al. Reference Dodell-Feder, Koster-Hale, Bedny and Saxe2011; Iacoboni & Dapretto Reference Iacoboni and Dapretto2006; Keysers & Fadiga Reference Keysers and Fadiga2008). These neuro-hormonal networks underpin social interactions among kin and non-kin members of society (Decety Reference Decety2011; Swain et al. Reference Swain, Mayes and Leckman2004; Reference Swain, Kim and Ho2011; Reference Swain, Konrath, Brown, Finegood, Akce, Dayton and Ho2012), as they have been demonstrated to respond according to minute-by-minute interactive synchrony between mother and infant (Atzil et al. Reference Atzil, Hendler and Feldman2011). Such social brain responses involve the integration of higher-order cognitive functions implicated in mentalization and empathy, as well as more basic functions of perception (lingual gyrus) and action representation (motor areas). Possibly, when perception–action regions are co-activated with mentalization regions, they have a social-attentive function that enhances the salience of the social context and the planning of adequate action – such as in parent's ability to understand the intentions and desires of her infant and to respond with a synchronized conduct (Atzil et al. Reference Atzil, Hendler and Feldman2011).

These parental social-brain functions have been shown to be under close regulation of oxytocin (OT) (Gordon et al. Reference Gordon, Zagoory-Sharon, Leckman and Feldman2010). The brain–OT correlations may provide additional support to the notion that mothering is guided by greater motivational–emotional focus. Recent pharmaco-imaging study suggests that plasma OT levels significantly correlated with limbic–emotional brain areas among mothers and fathers (Atzil et al. Reference Atzil, Hendler, Zagoory-Sharon, Winetraub and Feldman2012). Oxytocin is central for the formation of social bonds in general and parenting in particular, and is critical for maternal behavior (Shahrokh et al. Reference Shahrokh, Zhang, Diorio, Gratton and Meaney2010) in animal models. In human studies, oxytocin has been established as important for many social competencies, including trust, “mind-reading,” and empathy (Bartz et al. Reference Bartz, Zaki, Bolger and Ochsner2011). In parenting, oxytocin was differentially related to limbic and cortical activations in mothers and fathers for whom arginine vasopressin may play a related role in modulating social brain circuits (Atzil et al. Reference Atzil, Hendler and Feldman2011) through motivation enhancement or cognitive modulation.

Perhaps, then, the “laboratory” of the parent–infant dyad will be useful in exploring the importance of MNs for parents, as well as the development of related systems. One example that has received no attention so far is the development of MNs in humans who either have no experience of motion or have the experience of impaired motion. For example, for babies where motor function is impaired, would MNs not develop? In this case, their leg-related sensorimotor cortex would neither be capable of action “understanding,” and so forth, nor significant neural activity in areas commonly accepted as MNs when witnessing leg movements, because they would have never been able to associate the visual experience with the motor experience and thereby never have developed MNs. A less extreme example may be to study MN development in children of parents with depression – in which psychomotor retardation may be a significant mediating factor in the detrimental effects of postpartum depression on infant development (Tronick & Reck Reference Tronick and Reck2009).