Marcel Brass and Cecilia Heyes (2005)
Imitation: is cognitive neuroscience solving the correspondence problem?
Trends in Cognitive Sciences 9, 10, 489-495
Authors see imitation as copying body movement although at the same time claims that p489 ‘imitation provides a foundation for language acquisition, skill learning, socialization, and enculturation’
P493
More than a century of research on imitation has left us with a crucial functional problem: how are we able to transform a visual representation of an action into motor output. In the present review we have considered whether recent work in cognitive neuroscience has helped us to
solve this correspondence problem.
The correspondence problem: How does the imitator know what pattern of motor activation will make their action look like that of the model? P489 ‘ when we observe another person moving we do not see the muscle activation underlying their movement but rather the external consequences of that activation. So, how does the observer’s motor system ‘know’ which muscle activations will lead to the observed movement;’
Specialist theories – there are functional ( ? modules) and neurological mechanisms dedicated to controlling imitation.
eg of a specialist theory AIM (active intermodal matching)
P 490 ‘coding allows a visual representation of an observed action to be matched up with a pattern of motor activation that can produce the same action.’
Generalist theories – problem solved by general mechanisms of associative learning and action control. Neuroscience support – imitation is based on the automatic activation of motor representations by movement observation. ie motor representations depend on learned perceptual-motor links and are externally triggered.
eg IM (ideomotor theory) and ASL( associative sequence learning model– p490 ‘subsumes imitation under a general account of motor control and the associative sequence learning model (ASL)….. all actions are represented in the form of ‘images’ of sensory feedback….observing someone else executing an action leads to an activation of an internal motor representation in the observer because the observed action is similar to the content of the equivalent motor representation…… ASL complements IM by explaining imitation, not only of ‘transparent’ actions such as finger movements, but also ‘opaque’ actions such as facial expressions, where the observer’s image of the model is not normally similar to the sensory feedback received during performance of the same action ASL does this by postulating that each action guiding image is a compound of two action representations –one encoding visual information (what the action looks like), and the other containing somatosensory information and motor commands (what the action feels like and how it is initiated). The visual and motor components become linked through Hebbian learning. Like IM, ASL assumes that learning of this kind can occur whenever we look down and watch our own actions. However, it points out that self-observation of an opaque movement will normally give rise to a ‘nonmatching vertical association’, i.e. to a visuo-motor link that cannot support imitation because the visual component is not similar to the visual input received during observation of another person performing the same action. Performance of an opaque movement leads to the establishment of a ‘matching vertical association’, a visuo-motor link that can support imitation, only in the kinds of environment created by optical mirrors, imitative social partners and explicit training regimes.’
Evidence
Behavioral: RT of the copy of observed movements.(nb paradigm of congruent & incongruent movements, see Bird et al compatible, incompatible actions note Interference effects of this kind occur both when kinematic aspects of the movement are observed [17–19],
and when the terminal posture is presented alone 15.20)
Neuroimaging studies: participants need not be required to perform any action at all. P491 ‘Therefore, these studies provide yet clearer evidence that passive observation of action is sufficient to generate motor activation (see [21,22] for reviews). A large number of studies have now demonstrated that passive observation of action leads to activation of a set of brain regions known to be involved in movement execution [23–27] (see Box 2). Interestingly, although motor activation also occurs during observation of abstract stimuli, it seems to be strongest for the observation of human biological motion . Areas like the inferior parietal cortex seem to be strongly ‘tuned’ (innately or through learning) to biological motion, whereas frontal regions like Broca’s area seem to be concerned with more abstract aspects of the action like the goal [25,66,67]. This latter point is further supported by a recent demonstration that premotor areas involved in the processing of biological motion are also activated by sequences of abstract stimuli [68].’
P491 Imitation is typically effector-specific; to support this imaging studies suggest somatotopic organization for observed movements in motor-related areas [29–31], particularly in premotor cortex.
If we assume that special purpose mechanisms tend to be implemented in distinctive cortical areas, the failure to find ‘hot spots’ that are active during imitation ( KRO the assumption that it has to be a specific anatomical area) but not during passive movement observation is supportive of generalist solutions to the correspondence problem.
The role of learning in imitation and observation of biological motion.
ASL model, suggests that imitative performance can be either direct or linguistically mediated.
Another study relevant to learning found stronger automatic imitation effects when body movements were viewed from an ‘own person’ perspective (at the angle from which one views ones own movements) than from an ‘other person’ perspective [36]. This finding was recently confirmed with transcranial magnetic stimulation [37]. This is what one would expect if, as generalist theories suggest, imitation depends on experience of ones own actions’
‘The results of two neuroimaging studies indicate that activation of cortical areas involved in imitation and movement observation depends on learned expertise in performing the observed movements. Using an elegant experimental design, Calvo-Merino and colleagues [38] presented capoeira dancers, expert classical ballet dancers and non-dancer control participants with video-clips of closely matched capoeira and ballet movements (Figure 3).
The capoeira experts showed stronger activation in the premotor, parietal and posterior STS regions when observing capoeira movements than when observing ballet movements, and the ballet experts showed stronger activation in the same areas when observing ballet movements than when observing capoeira movements.’
P493
Problem for generalist theories : ‘If imitation depends on shared representations of perception and action, how can we distinguish between internally generated and externally triggered motor representation’ also why don’t we imitate all of the time?
Is there an inhibition process . Brass and colleagues [44,45] investigated the functional and neuronal mechanismsinvolved in inhibition of imitative response tendencies. These studies showed that the inhibition of imitative behaviour involves cortical areas which are known to be involved in distinguishing self from other, rather than response inhibition per se. These findings provide support for the view that action observation activates motor representations of the same kind that guide internally-generated action. If they did not, it would be unnecessary for inhibition of imitation to call on mechanisms that distinguish self and other [45]. Also ‘Research involving neurological patients indicates that this problem is not purely theoretical. Some patients withprefrontal lesions are echopractic; they tend to imitate observed behaviour, rather than to follow verbal or symbolic instructions [41]. Sometimes this even leads to compulsive ‘imitation behaviour’ [42,43].’
