Autism has never been short of controversial theories. Hype about “cold” mothers in the 1970s has given way to groundless scares about MMR vaccinations and notions of extreme male brains. Are we anywhere near a consensus? Vilayanur Ramachandran and Lindsay Oberman think their group has some compelling evidence when it comes to explaining autism’s characteristically bizarre symptoms
“I KNOW Steven is trapped in there somewhere. If only you could find a way to tell our son how dearly we love him, perhaps you could bring him out, Dr Ramachandran.” How often have doctors heard that heartfelt cry from parents describing their autistic child? Sadly, though, medical science has quite a way to go before we can talk about ways to bring Steven out.
But why does Steven need “bringing out”? In the 1940s, two doctors, Leo Kanner in Baltimore and Hans Asperger in Vienna, independently described this devastating developmental disorder as autism, from the Greek “autos” or “self”. It is a perfect description, because autism’s most striking feature is complete withdrawal from the social world and a marked reluctance or inability to interact with people.
Advertisement
Watching six-year-old Steven, you are unlikely to notice anything wrong as he gets on with his drawing. His pictures of animals are beautiful. There’s one of a galloping horse that is so wonderfully animated it seems to leap out of the page. Physically, too, he seems perfectly healthy. But talk to him and you soon realise that there’s a sense in which Steven, the person, simply isn’t there. He is incapable of anything remotely resembling the two-way exchange of normal conversation. He refuses to make eye contact, and keeps fidgeting and rocking his body to and fro. All attempts at meaningful communication with him have been, and will be, in vain.
Although Steven’s symptoms seem disparate, they fall into two major groups: social-cognitive and sensorimotor. In the social-cognitive group is the single most important symptom: mental “aloneness” and a lack of contact with the world, especially the social world. Hand in hand with this comes an absence of emotional empathy, and a profound inability to engage in normal conversation. Even more surprising in a species known for its playfulness, he has no sense of play, no “pretend” games.
Though he is socially withdrawn, Steven has a heightened interest in his inanimate surroundings, such as his drawings, almost to the point of obsession. Often this leads to odd, narrow preoccupations and fascination with things that seem utterly trivial to most of us, such as memorising phone numbers. Autistic children also – crucially – often have difficulty miming and imitating other people’s actions.
There are dozens of theories of autism, which broadly divide into psychosocial explanations and physiological explanations. One ingenious psychological explanation by neuroscientists Uta Frith of University College London and Simon Baron-Cohen at the University of Cambridge is that autistic children are unable to form adequate “theories of other minds”. This is supported by indirect evidence that “normal” children do not use their general intelligence to make inferences but use specialised brain mechanisms that create internal models of the inner workings of other minds so they can predict the behaviour of others – and manipulate them if need be.
Since children with autism have such profound deficits in social interactions, it follows they may lack these brain mechanisms – whatever they are. Are there any mechanisms with functions that match those that are damaged in autism? Is there an anatomical explanation for the symptoms that are unique to autism? In the late 1990s, it occurred to our group (which at the time included Eric Altschuler; Lindsay Oberman joined in 2003) at the University of California, San Diego (UCSD) that a set of neurons in the premotor cortex in the frontal lobes fitted the bill exactly. These neurons had been identified in the brains of macaque monkeys by Giacomo Rizzolatti, Vittorio Gallese, Marco Iacoboni and their colleagues at the University of Parma, Italy.
The existence of command neurons that can control voluntary movements has been known for decades, but what surprised the researchers was an extra property: some of these neurons fired not only when the monkey reached for a peanut, but also when it watched the researcher reach for a peanut. The Italian group dubbed them “mirror neurons” or “monkey-see, monkey-do” neurons.
This observation had extraordinary implications. The neuron – or more accurately, the network it is part of – was not only generating a highly specific command – “reach for the peanut” – but also allowed the monkey to put itself in another monkey’s shoes. Primates, including humans, are highly social creatures, and being able to create an internal simulation of what is going on in the minds of other group members is crucial for survival. It may even be that the emergence and subsequent sophistication of mirror neurons in hominids played a crucial role in the development of such essentially human abilities as empathy, language, learning through imitation rather than trial and error, and perhaps even the rapid transmission of what we call “culture”.
“Creating an internal simulation of the minds of other group members is crucial for survival”
We were struck by the fact that it is precisely these properties of mirror neurons that are not functioning in autism. There is a perfect match between the deficits in autism and the functions of mirror neurons. Andrew Whitten’s group at the University of St Andrews, UK, was thinking the same thoughts around the same time, but the first, and much of the subsequent experimental evidence came from our research group, which included Altschuler, and also Jamie Pineda.
To find evidence of mirror neuron dysfunction, we had to probe the brains of autistic children naturally, without inserting electrodes. Fortunately, one component of an electroencephalogram (EEG), the mu wave, is blocked whenever someone makes the simplest voluntary movement, such as opening their fingers. Remarkably, this is also blocked when someone watches another person perform the same movement. So watching for mu wave suppression in an EEG turned out to be a simple way of monitoring human mirror neuron activity.
We were wary of generalising from limited data, so we carried out more experiments using 10 high-functioning people with autism spectrum disorders and 10 controls. We measured mu-wave suppression while they watched videos of a moving hand, a bouncing ball, visual noise such as “snow” on a detuned TV, or when they moved their own hands. Mu suppression was evident in the non-autistic people both when they observed hand movement and when they moved their own hands. As before, the EEGs of the people with autism also showed mu suppression when they moved their own hands, but not when they watched other people’s hand movements. Since our experiments in 2000, other groups have explored mirror neurons and autism. For example, a group of researchers led by Riitta Hari at Helsinki University of Technology used magneto-encephalopathy, an ultrasensitive technique for measuring the magnetic fields produced by the brain’s electrical activity, and initially reported no mirror neuron deficits, but subsequent work confirmed our experiments. More recently, Mirella Dapretto, Iacoboni and their colleagues at the University of California, Los Angeles, used fMRI to find a reduction in mirror neurons in autistic children, confirming our hunch.
All these results seem to provide conclusive evidence that deficient mirror neurons really do account for some of autism’s most important symptoms. Assuming we are right, do mirror neurons offer some possibilities for helping relatives and carers deal with a disorder that is still difficult to treat? Certainly the lack of mu-wave suppression could be used to screen babies for autism so that timely behavioural therapies can be started before other more obvious symptoms appear.
Pineda is currently investigating a second, more intriguing possibility. He is using biofeedback, which involves monitoring a child’s mu waves and displaying them on a screen to see if the child can somehow learn to suppress the waves. This technique might revive the mirror neurons should they be dormant rather than completely lost.
So much for the social-cognitive symptoms: what of the sensorimotor symptoms? While they are not the cardinal signs of autism, they need to be explained – especially since they cause most distress to autistic people. In the early nineties, we came up with the “salience landscape” theory in collaboration with William Hirstein at Elmhurst College, Illinois, and we still think it is the best candidate to account for the bizarre rockings to and fro, hypersensitivity or aversion to certain sounds, the avoidance of eye contact, and so on.
When we look at the world, we confront a bewildering sensory overload. After processing in the sensory areas, this information is relayed to the amygdala, which is the gateway to the limbic system – the emotional core of the brain. With input from stored knowledge, the amygdala gauges the emotional significance of the visual input: is it prey, predator, mate, boss, or utterly trivial? The messages cascade from the amygdala to the rest of the limbic system into the hypothalamic nuclei, and eventually into the autonomic nervous system, preparing us for action – feeding, fighting, fleeing, or sex.
Part of this preparation is behavioural, but part is also autonomic-sympathetic arousal, making us sweat in order to dissipate the heat from muscular exertion, for example. This arousal feeds into the brain in a positive feedback loop. So the emotional arousal amplifies itself, creating an autonomic storm every time we see a burglar, predator or pin-up. This allows the amygdala to create a “salience landscape”, in which hills and valleys correspond to high and low importance, and to carry out “emotional surveillance” of the sensory world.
Autistic children seem to have a distorted salience landscape. This may be partially due to indiscriminately increased – or reduced – connections between sensory cortices and amygdala and, possibly, between limbic structures and frontal lobes. When these connections are exaggerated, every trivial event or object sets off an autonomic storm. If the emotional arousal is less pronounced, the child may attach abnormally high significance to certain unusual stimuli, accounting for the strange preoccupations – including “savant” skills.
Conversely, if some of the connections between the sensory cortices and the amygdala are partially cancelled out by the distortions in the salience landscape, the child may ignore things – such as eye contact – that grab the attention of most normal children. To test this, we measured the reduction in skin resistance caused by sweating that accompanies autonomic arousal in 37 children with autism spectrum disorders compared with 25 children developing normally. The non-autistic group showed arousal for certain expected categories of stimuli but not for others. The autistic children showed a generally heightened autonomic arousal that was increased by the most trivial objects and events, while they ignored highly relevant stimuli.
“Heightened autonomic arousal was caused by the most trivial of events or objects in children with autism”
Whatever the causes, our embodied sense of self seems to depend crucially on a back-and-forth interaction between the emotional core of the brain and the sensory signals arriving from the world. If the connections that lead to such interactions are damaged, or fail to develop in childhood, it could result in a disturbing loss of normal development of an integrated self anchored in the body.
Of course, our theories about the salience landscape and mirror neurons are not necessarily mutually exclusive. If there are connections between the mirror neuron system and the limbic system, it is conceivable that distortions in limbic-sensory connections are what lead, ultimately, to a deranged mirror neuron system.
Whatever the underlying mechanisms turn out to be, we believe that children with autism have a dysfunctional mirror neuron system that may help explain many important features of autistic spectrum disorders, and that they have defects in their salience landscape that explain still others. It remains to be seen whether this is caused by genes connected to brain development, genes that predispose to certain viruses which, in turn, might predispose to seizures, or due to something else entirely. The final story of autism is still to be written; we just think we are at a very useful starting place. Someday we may find a way to “bring Steven out” – or at the least provide his family and doctors with a proper explanation of why this may not be possible.
Profile
Vilayanur S. Ramachandran is director of the Center for Brain and Cognition at the University of California, San Diego (UCSD) and adjunct professor at the Salk Institute, La Jolla. After training as a doctor, he gained a PhD in neuroscience from the University of Cambridge.Ramachandran became famous for his work on brain abnormalities, such as phantom limbs and synthaesthesia. In 2003, he gave the BBC Reith lectures, entitled The Emerging Mind, and in 2005 he received the Henry Dale prize and an honorary life fellowship from The Royal Institution of Great Britain. Among his books are Phantoms in the Brain (HarperCollins) and A Brief Tour of Human Consciousness (Pi Press).
Lindsay Oberman obtained her undergraduate degree at Emory University in Atlanta, Georgia, and is currently a PhD student at UCSD.
