Neonatal imitation

I'm telling you, the baby started it!

What are these cheeky babies up to? They're demonstrating neonatal imitation, as newborns can from the moment they're born.

Researchers in 1977 got adults to demonstrate certain facial gestures - tongue protrusion, lip protrusion, or open mouth - for 15 seconds, to newborns. After this, the newborns produced the same facial gesture more often than other facial gestures, see Figure below. The same researchers in a later study found the same effect, in babies less than an hour old. They also showed that newborns imitate an adult poking their tongue to one side (a gesture which is less likely to be made spontaneously, or as a reflex).

There's still some controversy over whether this phenomenon is real. An alternative explanation is that the sight of the adult's face causes general arousal in the newborn, with increased actions of all sorts.

Assuming the phenomenon is real, how do newborns do it? How could a newborn translate their visual perception of an adult poking their tongue out, into their own facial motor action?

This is an example of an explanation proposed within an embodied cognition framework, which is quite different to those proposed within traditional frameworks. The debate partly turns on different ideas about the role which symbolic 'representation' plays in human thought and action.

Representation is central to the classical/computational model of brain function. A computer processes information by manipulating the physical patterns within its central processing unit; those physical patterns are themselves symbols/representations of external objects, actions, or concepts. By this model, neural function works the same - the physical pattern of a group of neurons and their connections (or perhaps even a single neuron) becomes a symbol for an external object, which the brain manipulates. This manipulation is what 'thought' or 'cognition' is.

An example of a cognitive representation is a mental map. If you can imagine life before Google Maps, we used to rely on street directories when driving somewhere new. We'd look at the street map, printed on paper, and try to commit the map to memory. It does seem like some sort of neural pattern must work as a representation of the paper map, in our brains. 

But other brain functions which might seem to involve this sort of representation, probably don't at all. Like in the outfielder problem, described in a previous post. It seems unlikely that a fielder catches a ball by mentally simulating the ball's trajectory, within a 3D mental map of the playing field.

Much of the scepticism about neonatal imitation has stemmed from an assumption that, for a newborn to imitate an adult's facial gesture, they'd have to use a mental representation of the gesture. Newborns "can visually observe a stimulus, store an abstract representation of that stimulus, and compare it shortly thereafter to a proprioceptive representation (i.e., their own movements)."

Could a newborn's brain really be capable of a such a computational feat?

My reading of the research is that neonatal imitation is real enough. But it might not involve this sort of representation.

An alternative explanation is that newborns might be born with functioning 'intermodal translation' - connection between the sensory channels of vision and proprioception (the sense of one's own body part positions in space). A close connection is needed between these channels, for an infant to master basic sensorimotor tasks such as reaching and grasping. But it's not been clear just how early this sensorimotor coordination might come online. Looking at a baby under 6 weeks age, doesn't seem like they have a lot of coordination going on.

But perhaps a skill like reaching and grasping isn't the first use which intermodal translation is put to. A tight and rich connection between visual perception and proprioception could also serve as a link between an infant seeing an adult's facial gesture, and the infant reproducing the same facial gesture. "In the case of the neonate, we do not have to posit some prior internal cognition that the infant is attempting to express or externalize through imitation. The infant has seen something, a facial gesture, and is expressing what she/he has seen, by taking up the capacity the infant has for that expressive movement on her/his own face."

That is, the infant doesn't need to use an abstract mental representation of the facial gesture she/he has just seen. The infant can skip that and go straight to a physical representation of it instead (on her/his own face).

If this intermodal translation idea is right, then it's likely that mirror neurons play a role. Mirror neurons, found in monkeys, are neurons which are active both when a monkey performs a particular motor action, and when the monkey sees another monkey perform that action. It's likely that mirror neurons operate in the human brain also (though their purported role has been overhyped, de rigeur for neuroscience findings).

Interesting stuff. But what then is the purpose of neonatal imitation? An obvious possibility is that it might have the "function of maintaining adult/infant social interaction until infants are capable of intentionally influencing such interactions themselves."

Ho hum. A more appealing explanation (to me anyway) is that it might play a foundational role in an infant's early sense of self, of others, and of intersubjectivity. Mind you, current evolutionary theory apparently contends that few if any evolutionary innovations serve only one purpose - so maybe it's a bit of both. (This principle would then also apply to intermodal translation - necessary for both basic sensorimotor performance, and for foundational social cognition.)

Further reading:

Gallagher S, "How the Body Shapes the Mind" (2005) Oxford University Press.


Related papers:

Meltzoff AN and Moore MK, "Imitation of facial and manual gestures by human neonates" (1977) 198 Science 75.

Meltzoff AN and Moore MK, "Imitation, memory, and the representation of persons" (1994) 17 Infant Behavior and Development 83.

Jones SS, "The development of imitation in infancy" (2009) 364 Philosophical Transactions of the Royal Society B 2325.

Bjorklund DF, "A note on neonatal imitation" (1987) 7 Developmental Review 86.