Benett I Bertenthal; Ty W. Boyer.
CogSci 2011
Previous studies had shown that 9 month old infants who were exposed to the passive experience of observing an experimenter search repeatedly for a toy tended to make the perseverative error and search incorrectly search for the toy in the last step of the A-not B error task. In this study the experimenters tried to assess whether observing a novel mechanical claw do the same also has a similar effect on the infants. When they realised that following similar demonstration by a mechanical claw the percentage of infants committing the error goes down, they further tried to assess if a brief familiarisation with the method of action or the goal of action of the mechanical claw before the actual experiment in any way affects the chances of the infants performing the error. Such studies ultimately led to the conclusion infants' brief experiences at observing certain actions facilitates the activation of a corresponding motor representation and hence covert imitation.
A lot of previous studies have tried to show the different stages of development of understanding and perception of goal-directed actions in infants through observing the perception of different tasks performed by experimenters or mechanical claws as goal directed by different age group of infants. All such studies suggested a final conclusion that there is a bidirectional relation between the perception and production of actions. James' ideomotor theory (1890), Prinz’s (1997) common coding theory and the recent discovery of mirror neurons are all theories that support the idea of perception and production of actions sharing common representational resources. However, although there is sufficient evidence of matching neuronal representations in case of observation or performance of actions in case of adulta, the same in case of infants is not very clear. Hence this study contributes a lot towards understanding the development of this linked representation.
Some preliminary evidence that supports the idea of such a matching system in infants is the mu rhythm desynchronisation of infants' EEG in response to observation of certain goal directed actions but not to other movements ((e.g., Lepage & Theoret, 2007). Similar desynchronisation was seen when subjects perform an action, thus forming a basis for a matching model between perception and production of action. In another relevant behavorial study, Longo and Bertenthal (2006) tested whether 9- month-old infants would show evidence of covert imitation following observation of action.
In this context it is necessary to describe the A not B error task. In this task typically an experimenter hides an attractive toy under box A within easy reach of an infant. The baby searches for the toy under box A and finds it and the process is repeated several times. Ultimately, in the critical trial, the experimenter moves the toy under box B, also within easy reach of the infant in his full view. However babies of 10 months or younger continue to look under box A although s/he saw the toy being moved to box B, thus committing which is comminly known as the perseverative error.
Now, according to the claim that infants simulate observed actions through corresponding motor representations, then they should demonstrate the perverative error not only after actively participating in the A not B task but also after observing the experimenter search for the toy under box A for multiple times before the critical trial. And unsurprisingly infants in both cases, i.e. who themselves recovered the toy from under box A in the initial trials and who observed an experimenter do the same, identically committed the perseverative error in the critical trial. However interestingly contralateral reaching (i.e. by crossing the midline of the body) was not as common as ipsilateral reaching (i.e. reaching on the same side of the body as the object) suggesting that the motor representation for contralateral reaching is not as well developed. Thus the results from these experiments suggested that infants demonstrated covert imitation by mapping the observed action to corresponding motor representation only in case the perceived action overlapped with a sufficiently developed mirror representation. In the experiments talked about in this particular paper, they only went ahead with this study to further explore the prerequisites for covert imitation by conducting two new experiments where infants were primed in different ways before being exposed to an A not B error task.
This was conducted in 3 parts, in each of which the infant was exposed to the search trials of the task performed by a mechanical claw instead of an experimenter.
Experiment 1a - the experimenter controlling the mechanical claw was completely out of view for the infants.
Experiments 1b - 1. only his arms were visibe. 2. The infants were familiarised briefly with the mechanical claw before the task when the experimenter played with the claws along with other objects infront of the infants fo 2 min.
This was done to dispel the unnatural curiosity or anxiousness associated with the claws due to their novelty
Experiment 1c - 1. Only his arms visible. 2. The infants were familiarised with the mechanical claws brief1ly before the commencement of the task when the experimenter demonstrated the operation of the claws to them by rolling a ball and retrieving it with the claws
This was an attempt to portray the mechanical claws as a means to an end instead of just another object.
Interestingly in 1a only 40% of infants reached incorrectly for box A as opposed to 70% who were reported to have done so when they observed the initial search trials being performed by a human experimenter. This might mean that either the perception-action mechanism was limited to biological models or that the infants could not sufficiently encode the action of the claws because of their unfamiliar nature.
However, even in experiment 1b the results werent much different with onlt 46.7% of infants searching incorrectlt under box A.
In experiment 1c 66.7% infants searched under box A. Such a result was statistically much significant to be attributed to random chance ( p >= 0.05)
Previous studies had suggested that non-biological motions do activate the motor system but to a much lesser degree as compared to biological motions. Now these current experiments showed that the matching mechanism can reach a considerable degree of abstraction based on sufficient experience. The familiarisation phase helped because the infants could consider the mechanical claws as effectors of goal directed actions rather than mere objects and could hence map these to a means-end motor representation more successfully.
This series of experiments were meant to assess if infants' motor representations of contralateral reaching can also be strengthened with familiarisation.
In experiment 2a the infants observed the experimenter reach out to box A exclusively with his contralateral hand for a familiaristaion period of two minutes
In experiment 2b only ipsilateral reaching was demonstrated during the familiarisation.
During the main task in both cases the experimenter was visible and always hid and retrieved the object with his contralateral hand.
Longo and Berenthal (2006) had earlier reported that only 50% of 9 month old infants searched incorrectly in the critical trial after observing the experimenter retrieve the objects with his coontralateral hand iin the initial trials. However contrastingly in this case 66.7% infants searched incorrectly in the A location in experiment 2a where they had been familiarised with contralateral reaching. This implied that infants not only visually encoded that the experimenter was reaching with his contralateral hand but also activated a suitable motor representation. In experiment 2b the number of infants reaching incorrectly was not significant enough to rule out chance and hence a conclusion to the effect that simply priming the infants with repeated reaching is sufficient to activate a motor representation for contralateral reach cannot be reached.
Thus it can be concluded from this study that infants' brief experiences of observing an action is encoded not only by their visual systems but also by their motor systems and that these representations become more prominent after a brief familiarisation phase which appears to prime the motor system. How such priming contributes to permanent learning or what is the need of such a direct connection between anobserved action and its motor representation is unclear and subject to ongoing research. Another concern regarding the conclusions of this experiment is the perseverative error being studied in these experiments is only a proxy for covert imitation and hence an indirect assessment of motor representation. More direct evidence like the EEG of infants while performing the A not B error task can serve as more concrete proof for these theories.
Wikipedia : http://en.wikipedia.org/wiki/A-not-B_error