Robots, mirror neurons, virtual reality and autism
To address a growing public awareness and apparent increase in prevalence of autism (Prior, 2003), psychologists need to devise improved treatment options based on the latest scientific discoveries.
One popular treatment option is the use of ‘psychoeducational’ intervention strategies. The use of such strategies by educationists might be more fruitful in helping children with autism with the added application of scientific methods that seem to monitor what has been called a ‘mirror neuron’ (MN) dysfunction in autism. So why do MN system dysfunction and difficulties in imitation apply to children with autism, and how can educationists facilitate such discoveries for the benefit of those affected?
Recent research in cognitive neuroscience has advanced our knowledge of the brain areas underpinning imitation – a process often exploited in education that is said to provide a foundation for language acquisition, skill learning, socialisation and enculturation (Brass & Heyes, 2005; Iacoboni et al., 1999). In the 1990s, Giacomo Rizzolatti and co-workers from Parma University discovered a compelling property of some neurons located in an area towards the front of the monkey brain called F5, which discharge for both the observation and execution of similar actions, such as watching someone pick up a peanut and picking it up independently. More recently, brain imaging findings have revealed that these so-called ‘mirror neurons’ exist in the human equivalent of F5, known as area 44, or Broca’s region (Binkofski et al., 1999; Buccino et al., 2001).
These findings led the researchers from Italy to suggest that mirror neurons allow their owners to understand observed actions by mapping them on to their own brain areas used to execute similar actions (Rizzolatti & Craighero, 2004). But it doesn’t stop there; accumulating evidence suggests that these mirroring properties are likely to be the brain basis of more complex higher order social cognitive processes. These might include the ability to decode other people’s intentions, or the ability to ‘step into another persons shoes’ to appreciate their viewpoint on a matter of discussion.
Interestingly, because mirror neurons seem to control functions that are lacking in those with autism, Williams and co-workers (2001) suggested that a mirror neuron (MN) dysfunction is likely to exist in those with this developmental disorder. Furthermore, these researchers suggested that this MN dysfunction may result in not only a poor understanding of other people’s thoughts and intentions but also a lack of empathy and ability to imitate. Imitation is a common part of human interaction; consider the contagiousness of yawning or the spontaneity of shadowing another person drinking on the opposite side of the dinner table (Schürmann et al., 2005). Could mirror neurons underpin these behaviours and be deficient in the brains of those with autism?
Monitoring and mapping mirror neurons
Cognitive neuroscientists have used a variety of imaging techniques to monitor the mirror neuron system in the human brain. A relatively new technique is ‘functional magnetic resonance imaging’ (fMRI), which uses radio waves and a strong magnetic field to scan a person’s brain and highlight the areas being used during a research task. The technique achieves this by highlighting brain regions with elevated concentrations of oxygenated blood (compared to regions of deoxygenated blood) since this corresponds to more active areas of the brain. For example, Iacoboni and others (1999) scanned healthy participants who were asked to perform a finger movement while they watched another person carry out the exact same action or a slightly different one. Regardless of what was seen, the same brain areas were activated, but importantly, when observation corresponded exactly with the executed finger movement, this activation was significantly more intense, supporting the existence of a mirror neuron system used for observation/execution matching. Also, fMRI has revealed that children with autism show no MN system activity while imitating and observing emotional expressions, again implying that this neural dysfunction results in social problems (Dapretto et al., 2006).
In addition, an ingenious fMRI study has shown that more MN system activation exists in dancers who observe actions for which they have motor expertise, compared to dancers who are less familiar with the dance actions observed (Calvo-Merino et al., 2005). Importantly, this implies that this system can develop through learning.
But perhaps the most useful technique for the not-so-distant future is the use of an electroencephalogram (EEG) to record oscillations in a so-called ‘mu-wave’ in the brain. Mu-waves are electromagnetic oscillations arising from the synchronous and coherent electrical activity of large groups of neurons in the brain. Scientists from the University of California, San Diego have recently found that mu-waves are suppressed not only during the execution of action, but also during its observation. During observed actions the MN system is the only motor area active, so Oberman and her colleagues inferred that mirror neurons induce mu-wave blockage. Therefore, monitoring mu-waves during action observation seems to offer a safe non-invasive method for measuring MN system activity.
It is also interesting to note that this research group has highlighted that one can voluntarily manipulate ones’ own mu-waves, using neurofeedback training. Perhaps this process is the brain basis for how people can develop their own MN system. But how might the mu-wave behave in children with autism?
Mu-waves in autism
Interestingly, the group in San Diego found that children with autism block the mu-wave when performing an action, but not when merely observing one (Oberman et al., 2005). This is likely to characterise a dysfunction in observation/execution matching, the process that many consider mirror neurons to play a big role in. The MN system seems to control social cognitive processes deficient in autism, adding to reasons to believe that this developmental disorder is at least partly brought about by MN system dysfunction.
Research so far has shown that a reliable diagnosis of autism spectrum disorder can be made in children aged two to three years, although the process is lengthy because of a need to use several assessment methods including parent reports and observations (Moore & Goodson, 2003). The ability to monitor mu-wave function raises the possibility that the identification and diagnosis of autism could be a quicker process, allowing existing behavioural strategies for the support of families with a child with autism to start earlier, improving the chance that such attempts at intervention will work effectively (Ramachandran & Oberman, 2006). Educationists might then also need to design new age-appropriate strategies to encourage imitation to take effect after a more rapid diagnostic assessment process.
However, investigations into the patterning of mu-wave suppression have involved only high-functioning children with autism; findings may not apply to children with lower-functioning autism.
If work by the San Diego group is to benefit the education of children with autism, more extensive research with less able children with autism should be carried out in order to promote the inclusive educational environment viewed by many as a means to remove barriers and discrimination and improve outcomes (Lindsay, 2003).
Making mirrors move with robots
Press and co-workers (2005) used ‘electromyogram’ (EMG) recordings to show that the MN system can be stimulated to varying degrees by both human and robotic action. Taken together with the finding that the MN system can develop through learning, this implies that educational strategies using robots may help children with autism to remediate their underactive MN systems. Furthermore, ‘neurocomputing’ (the field that studies the functioning of a biological brain using computer technology) has shown that the presence or absence of EEG-measured mu-wave suppression can be used in order to test what aspects of robot movement are likely to induce MN system activity (Oberman et al., 2007).
Interestingly, when Oberman’s team combined their own findings with work by Tai and colleagues (2004), they concluded that the MN system is activated only when robot actions are perceived as autonomous owing to the robot’s controller being out of sight. These findings may have implications for education, in that they suggest that the perception of one-on-one robot–human interaction might be important for strategies designed to help activate MN systems in children with autism, to improve their ability to imitate.
It is known that children with autism enjoy predictable environments. More recently, Simon Baron-Cohen from Cambridge University has put forward a theory asserting that people on the autistic spectrum are often highly driven to attend to and create systems, a behaviour he has coined ‘systemising’ (Baron-Cohen, 2002; The Psychologist, February 2008). It could therefore be that interaction with robots (which can be made predictable and systematic in appearance) offers the perfect way for children with autism to be introduced to imitation.
Indeed, observational studies have shown that children with autism like to watch and eventually copy a robot’s actions (Robins et al., 2005). Importantly, monitoring mu-waves throughout ‘robot education’ with children with autism may reveal that their increased attention to such humanoid friends induces either increased MN system activity shown via voluntary mu-wave suppression or a compensatory strategy to allow new levels of imitative behaviour.
However, the challenge for this kind of research will be to see if the imitation of robot behaviour will eventually transfer to the imitation of people in the lives of children set back in development by autism. If this proves possible, the future imitation-education of children with autism may well rely on the use of humanoid robots controlled by peers or teachers in schools. When considering that imitation is a fundamental for more complex modes of communication, breakthroughs in robot–human interaction may eventually lead to those with autism displaying more motivation to initiate social behaviour and to engage in conversation.
In addition to research using robots, there are reports advocating the use of virtual-reality technology with children with autism. In light of the fact that some imitation takes the form of social mirroring – the area of imitation likely to be worse off in those with autism (Iacoboni, 2005) – the simulation of everyday situations using virtual reality might be the best way to promote both social mirroring and social understanding.
And indeed, British researchers (Mitchell et al., 2007) demonstrated gains in social understanding in six adolescents with autism, owing to experience in a virtual environment of a café. Specifically, there were several instances of significant improvement in social judgements made to videos of café and bus scenes, directly after exposure to the virtual environment. Therefore, it seems that contemporary evidence for a generalised improvement in social skills after exposure to a virtual-reality intervention exists for those with autism.
Intriguingly, a combination of virtual reality with voluntary mu-wave control may well offer an effective supplement to robot-imitation education. The Baron-Cohen group have used predictable objects (vehicles) in a series called ‘The Transporters’ in order to teach emotion-recognition to children with autism (see The Psychologist, February 2007). A future educational strategy might utilise the same characters in a virtual-reality setting. The voluntary control of an EEG-measured mu-wave during the presentation of an emotional expression found on a ‘transporters character’ could be used in imitation education. This educational paradigm would promote the mu-wave activity found in a healthy person’s MN system. It seems that a whole intervention package taking advantage of the latest technological know-how is just waiting to materialise!
To conclude, findings from neuroimaging and neurocomputing studies can be used to help develop educational strategies to improve basic ‘people skills’ in those who lack them. The San Diego research group appears at the forefront of these advancements and the use of their findings in the educational setting, although at first glance like something out of science fiction, is certainly no light-year leap away.
- Mark Alex Turner is an assistant psychologist in Hertfordshire Parntership NHS Foundations trust
Baron-Cohen, S. (2002). The extreme male brain theory of autism. Trends in Cognitive Sciences, 6, 248–254.
Binkofski, F., Buccino, G., Posse, S. et al. (1999). A fronto-parietal circuit for object manipulation in man: evidence from an fMRI- study. European Journal of Neuroscience, 11, 3276–3286.
Brass, M. & Heyes, C. (in press). Imitation: Is cognitive neuroscience solving the correspondence problem? Trends in Cognitive Sciences, 9, 489–495.
Buccino, G., Binkofski, F., Fink, G.R. et al. (2001). Action observation activates premotor and parietal areas in a somatotopic manner. European Journal of Neuroscience, 13, 400–404.
Calvo-Merino, B., Glaser, D.E., Grèzes, J. et al. (2005). Action observation and acquired motor skills: An fMRI study with expert dancers. Cerebral Cortex, 15, 1243–1249.
Dapretto, M., Davies, M.S., Pfeifer, J.H. et al. (2006). Understanding emotions in others: Mirror neuron dysfunction in children with autism spectrum disorders. Nature Neuroscience, 9, 28–30.
Iacoboni, M. (2005). Neural mechanisms of imitation. Current Opinion in Neurobiology, 15, 632–637.
Iacoboni, M., Woods, R.P., Brass, M. et al. (1999). Cortical mechanisms of human imitation. Science, 286, 2526–2528.
Lindsay, G. (2003). Inclusive education: A critical perspective. British Journal of Special Education, 30, 3–12.
Mitchell, P., Parsons, S. & Leonard, A. (2007). Using virtual environments for teaching social understanding to adolescents with autistic spectrum disorders. Journal of Autism and Developmental Disorders, 37, 589–600.
Moore, V. & Goodson, S. (2003). How well does early diagnosis of autism stand the test of time? Follow-up study of children assessed for autism at age 2 and development of an early diagnostic service. Autism, 7, 47–63.
Oberman, L.M., Hubbard, E.M., McCleery, J.P. et al. (2005). EEG evidence for mirror neuron dysfunction in autism spectrum disorders. Cognitive Brain Research, 24, 190–198.
Oberman, L.M., McCleery, J.P., Ramachandran, V.S. & Pineda, J.A. (2007). EEG Evidence for mirror neuron activity during the observation of human and robot actions. Neurocomputing, doi:10.1016/j.neucom.2006.02.024
Press, C., Bird, G., Flach, R. & Heyes, C. (2005). Robotic movement elicits automatic imitation. Cognitive Brain Research, 25, 632–640.
Prior, M. (2003). Is there an increase in the prevalence of autism spectrum disorders? Journal of Paediatrics and Child Health, 39, 81–82.
Ramachandran, V.S. & Oberman, L.M. (2006). Broken mirrors: A theory of autism. Scientific American, 295, 39–45.
Rizzolatti, G. & Craighero, L. (2004). The mirror-neuron system. Annual Review of Neuroscience, 27, 169–192.
Robins, B., Dautenhahn, Te Boekhorst, R. & Billard, A. (2005). Robotic assistants in therapy and education of children with autism: Can a small humanoid robot help encourage social interaction skills? Universal Access in the Information Society (UAIS), 4, 105–120.
Schürmann, M., Hesse, MD., Stephan, K.E. et al. (2005). Yearning to yawn: The neural basis of contagious yawning. NeuroImage, 24, 1260–1264.
Tai, Y.F., Scherfler, C., Brooks, D.J. et al. (2004). The human premotor cortex is mirror only for biological actions. Current Biology, 14, 117–120.
Williams, J.H.G., Whiten, A., Suddendorf, T. & Perret, D.I. (2001). Imitation, mirror neurons and autism. Neuroscience and Biobehavioural Reviews, 25, 287–295.
Picture credit: Chris Turner
BPS Members can discuss this article
Already a member? Or Create an account
Not a member? Find out about becoming a member or subscriber