Action understanding: the role of expertise in adulthood and its development in infancy

The ability to understand others' intentions through observation of their action is crucial in social interactions. Several studies involving behavioural and neurophysiological methodologies suggest the existence of mechanisms linking action execution and perception in monkeys and humans that seem to be involved in action understanding. When we observe an action performed by another individual, our motor system internally simulates it, through a matching mechanism which maps the observed action onto the observer’s motor repertoire. However the question of how such mechanisms emerge and tune their features during development is still under debate. Developmental studies suggest that these mechanisms may be present in the first year of life, and may be modulated by infants' sensorimotor experience; so far there is no direct evidence in infancy that action observation induces in the observer the recruitment of the same motor program of the observed action. After a review of the current literature in Chapter 1, in Chapter 2 electromyography (EMG) is used for the first time in infancy to assess whether the observed action is directly mapped onto the infant’s motor system and internally simulated, and whether the properties of such simulation change during development, in 3-, 6-, 9-month-old infants. This first line of evidence shows that in 3-month-old infants the EMG activity is not modulated by the goal of an observed action which is not yet part of infants' motor repertoire; at 6 months of age, the observed action is simulated on-line, while in 9-month-olds the motor simulation is active at action onset, anticipating the final goal of the observed action. These results suggest that mirror mechanisms develop gradually, possibly according to infants’ greater experience and familiarity with the observed actions. In adulthood, the human mirror system seems to encode both goal-directed actions and intransitive (i.e. non goal directed) movements, thus suggesting that actions are coded both in terms of their goals and means to achieve them. In Chapter 3, the question whether in action simulation there is a predominance of coding goals or means of the observed action is investigated by showing highly familiar actions (e.g. grasping) executed by unusual effectors (e.g. foot). By means of transcranial magnetic stimulation (TMS) and EMG recordings, it is shown that observing a familiar action performed by an unusual effector activates, in the observer, not only the effector-specific motor program, but also the motor program of the effector usually involved in the observed action, suggesting that the action is remapped with respect to the observer’s typical manner of reaching the same goal. Studies in infancy show that infants' visual and motor familiarity with the observed action may influence their ability to understand its goal. In Chapter 4, by means of eye tracking technique, it is shown that 6-month-olds infants are able to discriminate between a familiar action, such as grasping, and a similar one executed in a biomechanically impossible manner (i.e. violating the constraints of human anatomy). Both biomechanically possible and impossible actions are coded as goal-directed, suggesting that biomechanical plausibility does not impair infants’ ability to ascribe goals to the observed actions. However, the familiarity with the possible action, when presented first, exerts an influence in coding the action as goal-directed more in the possible than in the impossible condition, suggesting that information about biomechanical properties of motion is relevant for 6-month-olds’ ability to anticipate the goal of the observed action. Given that infants in the first months of life appear to be sensitive to biological motion and biomechanical constraints of human movements, in Chapter 5 it is assessed whether the ability to discriminate between possible and impossible movements is already present shortly after birth. Two-days-old newborns are able to discriminate between biomechanical possible and impossible intransitive (i.e. non-goal-directed) hand movements, but not between static gestures. Overall, the present studies suggest that mechanisms linking motor and visual representations of movements are already present at birth, probably thanks to sensorimotor experience in the intrauterine life, and they develops in accordance with the observer’s sensorimotor experience with the observed actions or movements.