What’s so different about human speech?

Upon leaving the island, Odysseus is warned that storms lie ahead. His route home to Ithaca passes the sirens; monsters whose beautiful, haunting voices lure sailors to their deaths.  

He sets a course and explains his intention to the crew. At his command they fasten him to the mast, seal their ears with wax, and prepare for their encounter.

They reach treacherous waters. The siren song reaches into Odysseus’ mind, resonating with his deepest longings. The storm rages within. He struggles, but his bindings, the result of his clear intention, secure him tightly to the mast.

Forced to stand still and listen, he finds that he starts to hear the voices for what they really are; the empty fears of his own soul. He relinquishes his fight and hears the voice at his own still centre. The storm calms.

The crew notice that he has returned to his senses. They cut him loose.

He is indeed a wise and worthy captain.


This ancient footprint, first made in soft mud, is an index which shows us the passing of an three-toed Theropod dinosaur.  Denver, Colorado (Image: Wikimedia Commons)

This ancient footprint, first made in soft mud, is an index which shows us the passing of an three-toed Theropod dinosaur. Denver, Colorado (Image: Wikimedia Commons)

Speaking involves transmitting and interpreting intentional signs, some of which are also used in the instinctual communications of animals. These signals are of three kinds: 1. An index physically shows the presence of something, e.g. wolves tracking their prey by scent.

2. An ‘icon’ resembles the thing it stands for, like a photograph or a painting. Dolphins, apes and elephants recognise their own reflection; we assume that they interpret this two-dimensional image as representing their three-dimensional physical selves.

3. A symbol associates an unrelated form with a meaning. Our words are symbols, linking an idea with unique sound-and-movement sequences. They do not resemble the things they represent.

The ‘Union Jack’, a symbol of Great Britain since the union of Great Britain and Ireland in 1801.  It is made up of three other flag symbols; the Cross of St George for England (insert, top), St Andrew’s Saltaire for Scotland (centre), and for Northern Ireland, St Patricks Saltire (below) (Image: Wikimedia Commons)

The ‘Union Jack’, a symbol of Great Britain since the union of Great Britain and Ireland in 1801. It is made up of three other flag symbols; the Cross of St George for England (insert, top), St Andrew’s Saltaire f... moreor Scotland (centre), and for Northern Ireland, St Patricks Saltire (below) (Image: Wikimedia Commons)

Symbolism is almost unknown amongst animals, with a few rare exceptions. A stereotypical form of symbol is the ‘waggle dance’ of honeybees.

Although chimpanzees can be taught to use some sign gestures, they do not naturally communicate using symbols. In contrast, we use our symbolic language intentionally.

Our uses of speech are unique. We revisit our memories, order our thoughts and make future plans. With a destination in mind, we can listen for our ‘inner voice’, map out our route, take a stand against the storm of inner and outer distractions, and find our way home.

How is our speech unique?

Normal speech is already multi-channel; our words are accompanied by the musicality of our speaking, and our facial expressions and other physical gestures transmit many layers and levels of complex meaning.  Writing is another mode of communicating our language.  Social media transmits our language into virtual worlds.  The online social networking service Facebook commissioned ‘Facebook Man’ to commemorate their 150 millionth user   (Image: Wikimedia Commons)

Normal speech is already multi-channel; our words are accompanied by the musicality of our speaking, and our facial expressions and other physical gestures transmit many layers and levels of complex meaning. Writing is ... moreanother mode of communicating our language. Social media transmits our language into virtual worlds. The online social networking service Facebook commissioned ‘Facebook Man’ to commemorate their 150 millionth user(Image: Wikimedia Commons)

Aspects of our language ability are found in other animals, but the way we have combined and developed these traits is uniquely human.

1.  We use any available channel.

Most human languages use vocal speech. Under circumstances where speaking is not possible, we find other ways, e.g. sign languages and Morse code.

2.  We build our words from parts that gain meaning as they are combined

Most of the syllables we use to build words lack meaning on their own. Combining them together (as in English) or adding tonal shifts (as in Chinese) creates words.

3.  We code our words with meanings, making them into symbols

The chimpanzee (Pan troglodytes) known as Washoe (1965-2007) was the first non-human animal to be taught American sign language.  She lived from birth with a human family, and was taught around 350 sign words.  It was reported that upon seeing a swan, Washoe signed "water" and "bird".  Chimpanzees are capable of learning simple symbols.  However Washoe did not make the transition to combining these symbols together into new meanings  (Image: Wikimedia commons)

The chimpanzee (Pan troglodytes) known as Washoe (1965-2007) was the first non-human animal to be taught American sign language. She lived from birth with a human family, and was taught around 350 sign words. It was rep... moreorted that upon seeing a swan, Washoe signed “water” and “bird”. Chimpanzees are capable of learning simple symbols. However Washoe did not make the transition to combining these symbols together into new meanings (Image: Wikimedia commons)

Symbols are ‘displaced’, i.e. they do not need to resemble the thing they represent. Our words symbolise ideas, experiences and things.

4.  We combine these symbols to make new meanings.

We build words into phrases and stories, use these to revisit and share our memories, combine them into new forms, and communicate this information to others in various ways. Combining different symbols brings us a new understanding, which changes how we respond.

Look at this painting. As you do, consider what feelings it provokes.

'Wheatfield with crows' by Vincent Van Gogh, 1890.  (Image; wikimedia commons)

‘Wheatfield with crows’ by Vincent Van Gogh, 1890. (Image; wikimedia commons)

It is, of course, by Vincent Van Gogh. As you may know, his choices of colour and subject material were a personal symbolic code. He often used vibrant yellows, considering this colour to represent happiness.

His doctor noted that during his many attacks of epilepsy, anxiety and depression, Van Gogh tried to poison himself by swallowing paint and other substances.

As a consequence, he may have ingested significant amounts of toxic ‘chrome yellow’, which contains lead(II) chromate (PbCrO4).

Now consider this statement.

“This is the last picture that Van Gogh painted before he killed himself” (John Berger 1972, p28)

Look again at the picture.

What do you feel this time?

'Wheatfield with crows' by Vincent Van Gogh, 1890.  (Image; wikimedia commons)

‘Wheatfield with crows’ by Vincent Van Gogh, 1890. (Image; wikimedia commons)

Certainly our response has changed, though it is difficult to articulate precisely what is different. The image now seems to illustrate this sentence. Its symbolic content has altered for us. This example shows how combining two types of information –an image and text- can change the meaning it symbolises.

Some animals can be trained to recognise simple symbols. The psychologist Irene Pepperberg taught her African Grey parrot ‘Alex’ to count; he learned to use numbers as symbols, and could identify quantities of up to 6 items.

An African Grey Parrot (Psittacus erithacus).  Irene Pepperberg’s parrot, Alex, learned very basic grammar, could identify objects by name, and could count  (Image: Wikimedia Commons)

An African Grey Parrot (Psittacus erithacus). Professor Irene Pepperberg’s parrot, Alex, learned basic grammar, could identify objects by name, and could count (Image: Wikimedia Commons)

5.  The order in which we combine symbols defines their meaning

We put word symbols together into phrases, sentences, descriptions, sayings, stories, poems, documents, manuals, plays, oaths, promises, parodies, plays, pantomimes….

The ordering of words follow rules (grammar and syntax). Animals such as dogs and dolphins show some form of syntactical ability, but there is no evidence that they are on the verge of using what we understand as language. The order of words shows us their relationship, allowing us to understand how they are interacting. We change the order of our words and phrases to change the meaning we wish to communicate.

For instance; this makes sense.

‘Jane asked Simon to give these flowers to you.’

This doesn’t quite fit our normal understanding of reality…

‘These flowers asked Simon to give Jane to you.’

This works, but the meaning has changed.

‘Simon asked you to give these flowers to Jane’

However grammar is not enough . The words in combination need to ‘make sense’ for us to understand the meaning the speaker wishes to communicate.

What does this enable us to say?

‘The treachery of images’ by Belgian surrealist painter, Rene Magritte (1928-9).   Much of Magritte’s work explored the combination of words and images, and the way that this challenges the meaning that we understand from the components on their own.  This combination of words and image have been deliberately chosen so that they contradict each other.   What the artist says is true. However, it isn’t a pipe! It is a two-dimensional representation of a pipe (Image: Wikimedia Commons)

‘The treachery of images’ by Belgian surrealist painter, Rene Magritte (1928-9).Much of Magritte’s work explored the combination of words and images, and the way that this challenges the meaning that we understand... more from the components on their own. This combination of words and image have been deliberately chosen so that they contradict each other.What the artist says is true. However, it isn’t a pipe! It is a two-dimensional representation of a pipe (Image: Wikimedia Commons)

When we make new combinations of words, or add words to a visual signal such as a gesture, we create a new meaning.

We can add adjectives to a description, add qualifiers, combine phrases into a sentence, and make statements one after the other so that our listener associates these ideas. This process is known as ‘recursion’, a linguistic term borrowed from mathematics.

Our ideas about time vary between cultures, but we all mentally ‘time travel’ by revisiting our memories. For instance, the scent of something can evoke a memory that transports us back into an earlier event; suddenly we experience again the emotions and sensations we felt at that time. Putting our current selves into the past memory, or imagining a future scenario and inserting ourselves into that story, is a form of recursion.

Memory allows us to link speaking and listening with the meanings of our words. Our language is well structured to easily express recursive ideas. This shows us that our thinking uses recursion.

Why are we able to do this?

An illustration by Randolph Caldecott (1887) for ‘the House that Jack Built’.  This traditional British nursery rhyme uses recursion to build up a cumulative tale.  The sentence is expanded by adding to one end (end recursion).  Each addition adds an increasingly emphatic meaning to the final item of the sentence (i.e. the house that Jack built) (Image: Wikimedia Commons).  One final version of combined phrases ends like this; This is the horse and the hound and the horn That belonged to the farmer sowing his corn That kept the cock that crowed in the morn That woke the priest all shaven and shorn That married the man all tattered and torn That kissed the maiden all forlorn That milked the cow with the crumpled horn That tossed the dog that worried the cat That killed the rat that ate the malt That lay in the house that Jack built.

An illustration by Randolph Caldecott (1887) for ‘the House that Jack Built’. This traditional British nursery rhyme uses recursion to build up a cumulative tale. The sentence is expanded by adding to one end (end r... moreecursion). Each addition adds an increasingly emphatic meaning to the final item of the sentence (i.e. the house that Jack built) (Image: Wikimedia Commons). One final version of combined phrases ends like this;This is the horse and the hound and the hornThat belonged to the farmer sowing his cornThat kept the cock that crowed in the mornThat woke the priest all shaven and shornThat married the man all tattered and tornThat kissed the maiden all forlornThat milked the cow with the crumpled hornThat tossed the dog that worried the catThat killed the rat that ate the maltThat lay in the house that Jack built.

Our thinking capacity, through which we learn and remember, means that we can copy and learn to use language. Although some brain regions appear specialised for roles in memory and language, our ‘language function’ uses our entire brain, and cannot be dissociated from our minds.

Our ‘language brain’ includes the ‘basal ganglia’; these are neurons which connect the outer cortex and thalamus with lower brain regions.

We need this connectedness to coordinate movements in our fingers, to understand the relationships between words that are inferred by their order in our phrases, and to solve abstracted (theoretical) problems. This network interacts with ‘mirror neurons’ which allow us to relate to and decode the posture, speech and emotional cues of others.

The  basal ganglia that influence our speech also regulate the muscles controlling our posture. Standing is therefore more than just balancing on two legs; it is a whole body activity and requires much finer muscle control than walking on all fours. It also frees the hands, which allows us to manipulate tools. Lieberman suggests that it is the fine motor control required to maintain our upright posture which pre-adapted our ancestors for manipulating hand tools as well as the tongue, lips and other structures that make speech possible.  This upright posture is linked with a remodelling of our breathing apparatus, giving us more control over our larynx.

Philip Lieberman’s work with people suffering from Parkinson’s’ disease suggests that it is the ability to remember that makes speaking possible. Parkinson’s patients have degraded nerve circuits in their basal ganglia, so these patients have short term memory problems and difficulties with balancing and making precise finger movements. They also struggle with understanding and using metaphors and longer word sequences. This suggests that when we speak we are using the circuitry for sorting and remembering movement sequences, irrespective of whether these are producing words or actions.

Our posture has remodelled the evolution of our entire physiology from breathing to childbirth.  It frees the hands, allowing us to perform delicate and precise sequences of tasks.  Selection for the ability to precisely sequence our manual motor skills may have provided our ancestors the means to better sequence their thoughts (Image: Wikimedia Commons)

Our posture has remodelled the evolution of our entire physiology from breathing to childbirth. It frees the hands, allowing us to perform delicate and precise sequences of tasks. Selection for the ability to precisely ... moresequence our manual motor skills may have provided our ancestors the means to better sequence their thoughts (Image: Wikimedia Commons)

 

The basal ganglia that influence our speech also regulate the muscles controlling our posture. Standing is therefore more than just balancing on two legs; it is a whole body activity and requires finer muscle control that walking on all fours.  It also frees the hands, which allows us to manipulate tools.

Lieberman suggests that it is the fine motor control required to maintain our upright posture which pre-adapted our ancestors for manipulating hand tools as well as the tongue, lips and other structures that make speech possible.  This upright posture is linked with a remodelling of our breathing apparatus, giving us more control over our larynx.

The nerve networks that control our limbs and voices are linked across all vertebrates. Our basic ‘walking instinct’ initially activates Central Pattern Generator circuits driving movement in all four limbs. These are the same neural outputs that control our lips, tongue and throat.

Conclusions: What does this say about our language?

Captain Odysseus stands upright against the mast. This posture is distinct to our species, and has many implications for our speech, language and other actions  (Image: Wikimedia commons)

Captain Odysseus stands upright against the mast. This posture is distinct to our species, and has many implications for our speech, language and other actions
(Image: Wikimedia commons)

  • Our hominin ancestors evolved to use symbolic words and stories as a code to store and share memories, develop new skills and ideas, and coordinate their intentions and actions with their tribe.
  • When we revisit our memories or ‘reword’ our experiences into new sequences, we remodel the past, and project our thoughts into the future.
  • The control we have over our vocal sounds is linked with our neural circuits for movement. The ability to balance ideas and manipulate our tongues is linked to our ability to stand upright, balance on two feet and manipulate tools with our hands.
  • Language, then, is a cultural tool that allows us to order our thoughts, go beyond our instincts, share our intentions, and choose our own story.

Text copyright © 2015 Mags Leighton. All rights reserved.

References
Berger J (1972) ‘Ways of Seeing’ Penguin books Ltd, London, UK
BickertonD and Szathmáry E (2011) ‘Confrontational scavenging as a possible source for language and cooperation’ BMC Evolutionary Biology 11:261  doi:10.1186/1471-2148-11-261
Corballis MC (2007) ‘The uniqueness of human recursive thinking’ American Scientist Volume 95 (3), May 2007, Pages 240-248
Corballis, M.C.(2007) ‘Recursion, language, and starlings’ Cognitive Science 31(4) 697-704
Everett D (2008) ‘Don’t sleep, there are snakes: Life and language in the Amazonian jungle’ Pantheon Books, New York, NY (2008)
Everett, D (2012) ‘Language: the cultural tool’ Profile Books Ltd, London, UK
Gentner TQ et al (2006) ‘Recursive syntactic pattern learning by song birds’ Nature, 440;1204–1207

Riddles in code; is there a gene for language?

‘I have…’

Words are like genes; on their own they are not very powerful.  But apply them with others in the right phrase, at the right time and with the right emphasis, and they can change everything.

‘I have a dream…’

Genes are coded information.  They are like the words of a language, and can be combined into a story which tells us who we are.

The stories we choose to tell are powerful; they can change who we become, and also change the people with whom we share them.

‘I have a dream today!’


Language is a means for coding and passing on information, but it is cultural, and definitely non-genetic.  Nevertheless, for our speech capacity to have evolved, our ancestors must have had a body equipped to make speech sounds, along with the mental capacity to generate and process this language ‘behaviour’.  Our body’s development is orchestrated through the actions of relevant genes.  If the physical aspects of language ultimately have a genetic basis, this implies that speech must derive, at least in part, from the actions of our genes.

The hunt for genes involved with language led researchers at the University of Oxford to investigate an extended family (known as family KE).  Some family members had problems with their speech.  The pattern of their symptoms suggested that they inherited these difficulties as a ‘dominant’ character, and through a single gene locus.

The FoxP2 gene encodes for the ‘Forkhead-Box Protein-2’; a transcription factor.  This is a type of protein that interacts with DNA (shown here as a pair of brown spiral ladders), and influences which genes are turned on in the cell, and which remain silent.   This diagram shows two Forkhead box proteins, which associate with each other when active.  This bends the DNA strand and makes critical areas of the genetic code more accessible (Image: Wikimedia Commons)

The FOXP2 gene encodes for the ‘Forkhead-Box Protein-2’; a transcription factor. This is a type of protein that interacts with DNA (shown here as a pair of brown spiral ladders), and influences which genes are turne... mored on in the cell, and which remain silent. This diagram shows two Forkhead box proteins, which associate with each other when active. This bends the DNA strand and makes critical areas of the genetic code more accessible (Image: Wikimedia Commons)

Discovery of another unrelated patient with the same symptoms confirmed that the condition was linked to a gene known as FOXP2  (short for ‘Forkhead Box Protein-2’).  This locus encodes a ‘transcription factor’; a protein that influences the activation of many other genes.  FOXP2 was subsequently dubbed ‘the gene for language’.  Is that correct?

Not really.  FOXP2 affects a range of processes, not just speech.  The mutation which inactivates the gene causes difficulties in controlling muscles of the face and tongue, problems with compiling words into sentences, and a reduced understanding of language.  Neuroimaging studies showed that these patients have reduced nerve activity in the basal ganglia  region of the brain.  Their symptoms are similar to some of the problems seen in patients with debilitating diseases such as Parkinson’s and Broca’s Aphasia; these conditions also show impairment of the basal ganglia.

Genes code for proteins by using a 3-letter alphabet of adenine, thymine, guanine and cytosine (abbreviated to A, T, G and C).  These nucletodes are knwn as ‘bases’ (are alkaline in solution) and make matched pairs which form the ‘rungs of the ladder’ of the DNA helix. Substituting one base for another (as happens in many mutations) can change the amino acid sequence of the protein a gene encodes.  Changes may make no impact on survival, allowing the DNA sequence to alter over time.  Changes that affect critical sections of the protein (e.g. an enzyme’s active site), or critical proteins like FoxP2, are rare (Image: Wikimedia Commons)

Genes code for proteins by using a 3-letter alphabet of adenine, thymine, guanine and cytosine (abbreviated to A, T, G and C). These nucletodes are knwn as ‘bases’ (are alkaline in solution) and make matched pairs w... morehich form the ‘rungs of the ladder’ of the DNA helix. Substituting one base for another (as happens in many mutations) can change the amino acid sequence of the protein a gene encodes. Changes may make no impact on survival, allowing the DNA sequence to alter over time. Changes that affect critical sections of the protein (e.g. an enzyme’s active site), or critical proteins like FOXP2, are rare (Image: Wikimedia Commons)

Genes provide the code to build proteins.  Proteins are assembled from this coding template (the famous triplets) as a sequence of amino acids, strung together initially like the carriages of a train and then folded into their finished form.  The amino acid sequences of the FOXP2 protein show very few differences across all vertebrate groups.  This strong conservation of sequence suggests that this protein fulfils critical roles for these organisms.  In mice, chimpanzees and birds, FOXP2 has been shown to be required for the healthy development of the brain and lungs.  Reduced levels of the protein affect motor skills learning in mice and vocal imitation in song birds.

The human and chimpanzee forms of FOXP2 protein differ by only two amino acids. We also share one of these changes with bats.  Not only that, but there is only one amino acid difference between FOXP2 from chimpanzees and mice.  These differences might look trivial but they are probably significant.  FOXP2 has evolved faster in bats than any other mammal, hinting at a possible role for this protein in echolocation.

Mouse brain slice, showing neurons from the somatosensory cortex (20X magnification) producing green fluorescent protein (GFP).  Projections (dendrites) extend upwards towards the pial surface from the teardrop-shaped cell bodies. Humanised Foxp2 in mice causes longer dendrites to form on specific brain nerve cells, lengthens the recovery time needed by some neurons after firing, and increases the readiness of these neurons to make new connections with other nerves (synaptic plasticity).  The degree of synaptic plasticity indicates how efficiently neurons code and process information (Image: Wikimedia Commons)

Mouse brain slice, showing neurons from the somatosensory cortex (20X magnification) producing green fluorescent protein (GFP). Projections (dendrites) extend upwards towards the pial surface from the teardrop-shaped ce... morell bodies. Humanised Foxp2 in mice causes longer dendrites to form on specific brain nerve cells, lengthens the recovery time needed by some neurons after firing, and increases the readiness of these neurons to make new connections with other nerves (synaptic plasticity). The degree of synaptic plasticity indicates how efficiently neurons code and process information (Image: Wikimedia Commons)

Changing the form of mouse FOXP2 to include these two human-associated amino acids alters the pitch of these animals’ ultrasonic calls, and affects their degree of inquisitive behaviour.  Differences also appear in their neural anatomy.  Altering the number of working copies (the genetic ‘dose’) of FOXP2 in mice and birds affects the development of their basal ganglia.

Mice with ‘humanised’ FOXP2 protein show changes in their cortico-basal ganglia circuits along with altered exploratory behaviour and reduced levels of dopamine (a neurotransmitter  that affects our emotional responses).  So too, human patients with damage to the basal ganglia show reduced levels of initiative and motivation for tasks.

This suggests that FOXP2 is part of a general mechanism that affects our thinking, particularly around our initiative and mental flexibility.  These are critical components of human creativity, and are as it happens, essential for our speech.

Basal ganglia circuits process and organise signals from other parts of the brain into sequences.  Speaking involves coordinating a complex sequence of muscle actions in the mouth and throat, and synchronising these with the out-breath.  We use these same muscles and anatomical structures to breathe, chew and swallow;  our ability to coordinate them affects our speech, although this is not their primary role.

Family KE’s condition, caused by a dominant mutation in the FoxP2 gene, follows an autosomal (not sex-linked) pattern of inheritance, as shown here.   Dominant mutations are visible when only one gene copy is present.  In contrast a recessive trait is not seen in the organism unless both chromosomes of the pair carry the mutant form of the gene.   The FoxP2 transcription factor protein is required in precise amounts for normal function of the brain.  The loss of one working FoxP2 gene copy reduces this ‘dose’ which is enough to cause the problems that emerged as family KE’s symptoms (Image: Annotated from Wikimedia Commons)

Family KE’s condition, caused by a dominant mutation in the FOXP2 gene, follows an autosomal (not sex-linked) pattern of inheritance, as shown here.Dominant mutations are visible when only one gene copy is present. In... more contrast a recessive trait is not seen in the organism unless both chromosomes of the pair carry the mutant form of the gene. The FOXP2 transcription factor protein is required in precise amounts for normal function of the brain. The loss of one working FOXP2 gene copy reduces this ‘dose’ which is enough to cause the problems that emerged as family KE’s symptoms (Image: Annotated from Wikimedia Commons)

In practice, very few of our 25,000 genes are individually responsible for noticeable characteristics.  Most genetically inherited diseases result from the effects of multiple gene loci.  FOXP2 is unusual because of its ‘dominant’ genetic character.  It does not give us our language abilities, but it is involved in the neural basis of our mental flexibility and agility at controlling the muscles of our mouths, throats and fingers.

In addition, genes are only part of the story of our development.  The way we think and subsequently behave alters our emotional state.  Feeling stressed or calm affects which circuits are active in our brain.  This alters the biochemical state of body organs and tissues, particularly of the immune system, modifying which genes they are using.

The dance between the code stored in our genes and the consequences of our thoughts builds us into what we are mentally, physically and socially.  This story is ours to tell.  By our experience, and with this genetic vocabulary, we create what we become.

Text copyright © 2015 Mags Leighton. All rights reserved.

References
Chial H (2008)  ‘Rare genetic disorders: Learning about genetic disease through gene mapping, SNPs, and microarray data’ Nature Education 1(1):192  http://www.nature.com/scitable/topicpage/rare-genetic-disorders-learning-about-genetic-disease-979
Clovis YM et al. (2012) ‘Convergent repression of Foxp2 3′UTR by miR-9 and miR-132 in embryonic mouse neocortex: implications for radial migration of neurons’  Development 139, 3332-3342.
Enard, W (2011) ‘FOXP2 and the role of cortico-basal ganglia circuits in speech and language evolution’  Current Opinion in Neurobiology  21; 415–424
Enard, W et al (2009)  A Humanized Version of Foxp2 Affects Cortico-Basal Ganglia Circuits in Mice  Cell 137 (5); 961–971  http://www.sciencedirect.com/science/article/pii/S009286740900378X
Feuk L et at., Absence of a Paternally Inherited FOXP2 Gene in Developmental Verbal Dyspraxia, in The American Journal of Human Genetics, Vol. 79 November 2006, p.965-72.
Fisher SE and Scharff C (2009) ‘FOXP2 as a molecular window into speech and language’  Trends in Genetics 25 (4); 166-177
Lieberman P  (2009)  ‘FOXP2 and Human Cognition’  Cell 137; 800-803
Marcus GF & Fisher SE (2003) ‘FOXp2 in focus; what can genes tell us about speech and language?’  Trends in Cognitive Sciences 7(6); 257-262
Reimers-Kipping S et al. (2011) ‘Humanised Foxp2 specifically affects cortico-basal ganglia circuits’ Neuroscience 175; 75-84
Scharff C & Haesler S (2005) ‘An evolutionary perspective on Foxp2; strictly for the birds?’ Current opinion in Neurobiology 15:694-703
Vargha-Khadem F et al. (2005) ‘FOXP2 and the neuroanatomy of speech and language’  Nature Reviews Neuroscience 6, 131-138 http://www.nature.com/nrn/journal/v6/n2/full/nrn1605.html
Wapshott N (2013)  ‘Martin Luther King's 'I Have A Dream' Speech Changed The World’ Huffington post, 28th August 2013  http://www.huffingtonpost.com/2013/08/28/i-have-a-dream-speech-world_n_3830409.html
Webb DM & Zhang J (2005) ‘Foxp2 in song learning birds and vocal learning mammals’  Journal of Heredity 96(3);212-216

Upon reflection; what can we really see in mirror neurons?

“Mirror, mirror on the wall,
who is the fairest of them all?”

“Fair as pretty, right or true;
what means this word ‘fair’ to you?
Fair in manner, moods and ways,
fair as beauty ‘neath a gaze…
Meaning is a given thing.
I cannot my opinion bring
to validate your plain reflection!
You must make your own inspection.”


Mirrors shift our perspective, enabling us to see ourselves directly, and reflect ideas back to us symbolically.  But how do we really see ourselves?

Quite recently, neuroscientists discovered a new type of nerve cell in the brains of macaques which form a network across the primary motor cortex, the brain region controlling body movements.  These nerves are intriguing; they become active not only when the monkey makes purposeful movements such as grasping food, but also when watching others do the same.  As a result, these cells were named ‘mirror neurons’.

Comparison of human (above) and chimpanzee brain sizes (Image: Wikimedia Commons)

Comparison of human (above) and chimpanzee brain sizes (Image: Wikimedia Commons)

In monkeys and other non-human primates, these nerves fire only in response to movements with an obvious ‘goal’ such as grabbing food.  In contrast, our mirror network is active when we observe any human movement, whether it is purposeful or not.  Our brains ‘mirror’ the actions of ourselves and others’ actions, from speaking to dancing.

However it is to interpret what function these cells are performing.  Different researchers suggest that mirror neurons enable us to:

– assign meaning to actions;

– copy and store information in our short term memory (allowing us to learn gestures including speech);

– read other people’s emotions (empathy);

– be aware of ourselves relative to others (giving us a ‘theory of mind’, i.e. we have a mind, and the contents of other people’s minds are similar to our own).

Whilst these opinions are not necessarily exclusive, they do seem to reflect the different priorities of these experts.  Their varied interpretations highlight how difficult it is to be aware of how our beliefs and assumptions affect what our observations can and cannot tell us.

What we can say, is that the behaviour of these nerve cells shows that our mirror neuron responses are very different from those of our closest relatives, the primates.

What do we know about mirror neurons from animals?

A tribe of stump tail macaques (Macaca arctoides) watch their alpha male eating.   Macaques making meaningful gestures, such as grabbing for food, triggers the same mirror neurone network in the animal performing the action and the observers. (Image: Wikimedia commons)

A tribe of stump tail macaques (Macaca arctoides) watch their alpha male eating.When these macaques observe a meaningful gesture, such as grabbing for food, this triggers a shift in the electrical status of the same mot... moreor neurons in their brain as the observed animal is using to perform the action.  This ‘mirroring’ is found in other primates, including humans.(Image: Wikimedia commons)

Researchers at the University of Parma first discovered mirror neurons in an area of the macaque brain which is equivalent to Broca’s area in humans.  This brain region assembles actions into ordered sequences, e.g. operating a tool or arranging our words into a phrase.  Later studies show these neurons connect right across the monkey motor cortex, and respond to many intentional movements including facial gestures.

The macaque mirror system is activated when they watch other monkeys seize and crack open some nuts, grab some nuts for themselves, or even if they just hear the sound of this happening.  Their neurons make no responses to ‘pantomime’ (i.e. a grabbing action made without food present), casual movements, or vocal calls.

Song-learning birds also have mirror-like neurons in the motor control areas in their brains.  Male swamp sparrows’ mirror network becomes active when they hear and repeat their mating call.  Their complex song is learned by imitating other calls, suggesting a possible role for mirror neurons in learning.  This is tantalising, as we do not yet know the extent to which mirror neurons are present in other animals.

Our visual cortex receives information from the eyes, which is then relayed around the mirror network and mapped onto the motor output to the muscles.   First, [purple] the upper temporal cortex (1) receives visual information and assembles a visual description which is sent to parietal mirror neurons (2).  These compile an in-body description of the movement, and relay it to the lower frontal cortex (3) which associates the movement with a goal.   With these observations complete, inner imitation (red) of the movement is now possible. Information is sent back to the temporal cortex (4) and mapped onto centres in the motor cortex which control body movement (5) (Image: Annotated from Wikimedia Commons)

Our visual cortex receives information from the eyes, which is then relayed around the mirror network to the motor output to the muscles. First, [purple] the upper temporal cortex (1) receives visual information and ass... moreembles a visual description go the action.  This is sent to parietal mirror neurons (2) which compile an in-body description of the movement, and relay it to the lower frontal cortex (3) where the movement becomes associated with a goal. With these observations complete, an inner imitation (red) of the movement is now possible. Information is sent back to the temporal cortex (4) and mapped onto centres in the motor cortex which control body movement (5) (Image: Annotated from Wikimedia Commons)

 

The primate research team at Parma suggest the mirror system’s role is in action recognition, i.e. tagging ‘meaning’ to deliberate and purposeful gestures by activating an ‘in-body’ experience of the observed gesture.  The mirror network runs across the sensori-motor cortex of the brain, ‘mapping’ the gesture movement onto the brain areas that would operate the muscles needed to make the same movement.

An alternative interpretation is that mirror neurons allow us to understand the intention of another’s action.  However as monkey mirror neurons are not triggered by mimed gestures, the intention of the observed action presumably must be assessed at a higher brain centre before activating the mirror network.

How is the human mirror system different?

Watching another human or animal grabbing some food creates a similar active neural circuit in our mirror network.

The difference is that our nerves are activated by us observing any kind of movement.  Unlike monkeys, when we see a mimed movement, we can infer what this gesture means.  Even when we stay still we cannot avoid communicating; the emotional content of our posture is readable by others.  In particular we readily imitate other’s facial expressions.

As we return a smile, our face ‘gestures’.  Marco Iacoboni and co-workers have shown that as this happens, our mirror system activates along with our insula and amygdala.  This shows that our mirror neurons connect with the limbic system which handles our emotional responses and memories.  This suggests that emotion (empathy) is part of our reading of others’ actions.  As we see someone smile and smile back, we feel what they feel.

Spoken words deliver more articulated information than can be resolved by hearing  alone.  Our ability to read and copy the movements of others as they speak may be how we really distinguish and understand these sounds.  This ‘motor theory of speech perception’ is an old idea.  The discovery of mirror-like responses provides physical evidence of our ability to relate to other people’s movements, suggesting a possible mechanism for this hypothesis.

Sound recording traces of the words ‘nutshell’, ‘chew’ and ‘adagio’.  Our speech typically produces over 15 phonemes a second.  Our vowels and consonants ‘overlap in time’, and blur together into composite sounds.  This means that simply hearing spoken sounds does not provide us with enough information to distinguish words and syllables.  In practice we decode words from this sound stream, along with emotional information transmitted through the tone and timbre of phrases, facial expressions and posture (Images: Wikimedia Commons)

Sound recording traces of the words ‘nutshell’, ‘chew’ and ‘adagio’. Our speech typically produces over 15 phonemes a second. Our vowels and consonants ‘overlap in time’, and blur together into composite... more sounds. This means that simply hearing spoken sounds does not provide us with enough information to distinguish words and syllables. In practice we decode words from this sound stream, along with emotional information transmitted through the tone and timbre of phrases, facial expressions and posture (Images: Wikimedia Commons)

Further studies suggest that these mirror neurons are part of a brain-wide network made of various cell types.  Alongside the mirror cells are so-called ‘canonical neurons’, which fire only when we move.  In addition, ‘anti-mirrors’ activate only when observing others’ movements.  Brain imaging techniques show that frontal and parietal brain regions (beyond the ‘classic’ mirror network) are also active during action imitation.  It is not clear how the system operates, but in combination we relate to others’ actions through the same nerve and muscle circuits we would use to make the observed movements.  We relate in this way to what is happening in someone else’s mind.

Are mirror neurons our mechanism of language in the brain?

Macaques (Macaca fuscata) grooming in the Jigokudani HotSpring in Nagano Prefecture, Japan.   Human and monkey vocal sounds arise in different regions of the brain.  Primate calls are mostly involuntary, and express emotion.  They are processed by inner brain structures, whereas the human speech circuits are located on the outer cortex (Image: Wikimedia Commons)

Japanese macaques (Macaca fuscata) grooming in the Jigokudani Hotspring in Nagano Prefecture, Japan. Human and monkey vocal sounds arise from different regions of the brain. Primate calls are mostly involuntary, and exp... moreress emotion. They are processed by inner brain structures, whereas the human speech circuits are located on the outer cortex (Image: Wikimedia Commons)

Mirror-like neurons activate whether we are dancing or speaking.  Patients with brain damage that disrupts these circuits have difficulties understanding all types of observed movements, including speech.  This suggests that we use our extended mirror network to understand complex social cues.

Our mirror neuron responses to words map onto the same brain circuits as other primates use for gestures.  However signals producing our speech and monkey vocal calls arise from different brain areas.  This suggests that our speech sounds are coded in the brain not as ‘calls’ but as ‘vocal gestures’.  This highlights the possible origins of speaking as a form of ‘vocal grooming’, which socially bonded the tribe.

When we think of or hear words, our mirror network activates the sensory, motor and emotional areas of the brain.  We thus embody what we think and say.  Michael Corballis and others consider that mirror neurons are part of the means by which we have evolved to understand words and melodic sounds as ‘gestures’.

Woman Grasping Fruit’ by Abraham Brueghel, 1669; Louvre, Paris.  The precision control of the grasping gesture she uses to pluck a fig from the fruit bowl is unique to humans.  The intensity of her expression implies many layers of meaning to what we understand from this picture (Image: Wikimedia Commons)

Woman Grasping Fruit’ by Abraham Brueghel, 1669; Louvre, Paris. The precision control of the grasping gesture she uses to pluck a fig from the fruit bowl is unique to humans. The intensity of her expression implies ma... moreny layers of meaning to what we understand from this picture (Image: Wikimedia Commons)

What is unclear is how we put meaning into these words.  Some researchers have suggested that mirror neurons anchor our understanding of a word into sensory information and emotions related to our physical experience of its meaning.  This would predict that our ‘grasp’ of the meaning of our experiences arises from our bodily interactions with the world.

Vocal gestures would have provided our ancestors with an expanded repertoire of movements to encode with this embodied understanding.  Selection could then have elaborated these gestures to include visual, melodic, rhythmical and emotional information, giving us a route to the symbolic coding of our modern multi-modal speech.

We produce different patterns of mirror neuron activity in relation to different vowel and consonant sounds, as well as to different sound combinations.  Also, the same mirror neuron patterns appear when we watch someone moving their hands, feet and mouth, or when we read word phrases that mention these movements.

Wilder Penfield used the ‘homunculus’ or ‘little man’ of European folklore to produce his classic diagram of the body as being mapped onto the brain.  A version of this is shown here; mirror neurons map incoming information onto the somatosensory cortex (shown left) and outputs to the muscles from the motor cortex (right).  These brain regions lie adjacent to each other (Image: Wikimedia Commons)

Wilder Penfield used the ‘homunculus’ or ‘little man’ of European folklore to produce his classic diagram of the body as being mapped onto the brain. A version of this is shown here; mirror neurons map incoming ... moreinformation onto the somatosensory cortex (shown left) and outputs to the muscles from the motor cortex (right). These brain regions lie adjacent to each other (Image: Wikimedia Commons)

We process word sequences in higher brain centres at the same time as lower brain circuits coordinate the movements required for speech production and non-verbal cues.  Greg Hicock suggests that our speech function operates by integrating these different levels of thinking into the same multi-modal gesture.

Mirror neurons connecting the brain cortex and limbic system may allow us to synchronously process our understanding of an experience with our emotional responses to it.  This allows us to consciously control our behaviour, adapt flexibly to our world, and communicate our understanding to others and to ourselves.

Smoke and mirrors; what do these nerves really show and tell?

People floating in the Dead Sea.   Our ability to read emotional information from postures means that we can intuit information about people’s emotional state even when they are not visibly moving (Image: Wikimedia Commons)

People floating in the Dead Sea. Our ability to read emotional information from postures means that we can intuit information about people’s emotional state even when they are not visibly moving (Image: Wikimedia Comm... moreons)

The word ‘mirror’ conjures up strong images in our minds.  This choice of name may have influenced what we are looking for in our data on mirror neurons.  However they appear crucial for language.  This and other evidence suggests that our ability to speak and to read meaning into movement is a property of our whole brain and body.

Single nerve measurements show that the mirror neuron network is a population of individual cells with distinct firing thresholds.  Different subsets of these neurons are active when we see similar movements made for different purposes.  This suggests that the network responds flexibly to our experience.

Cecilia Heyes’ research shows that our mirror network is a dynamic population of cells, modified by the sensory stimulus our brain receives throughout life.  She suggests that these mirror cells are ‘normal’ neurons that have been ‘recruited’ to mirroring, i.e. adopted for a specialised role; to correlate our experience of observing and performing the same action.

This gives us a possible evolutionary route for the appearance of these mirror neurons.  Recruitment of brain motor cortex cells to networks used for learning by imitation would create a population of mirror cells.  This predicts that;

i. Mirror-like networks will be found in animals which learn complex behaviour patterns, such as whales. (They are already known in songbirds.)

One of the many dogs Ivan Pavlov used in his experiments (possibly Baikal); Pavlov Museum, Ryazan, Russia. Note the saliva catching container and tube surgically implanted in the dog's muzzle.   These dogs were regularly fed straight after hearing a bell ring.  In time, the sound of the bell alone made them salivate in anticipation of food.  This experience had trained them to code the bell sound with a symbolic meaning, i.e. to indicate the imminent arrival of food (Image: Wikimedia Commons)

One of the many dogs Ivan Pavlov used in his experiments (possibly Baikal); Pavlov Museum, Ryazan, Russia. Note the saliva catching container and tube surgically implanted in the dog’s muzzle. These dogs were regu... morelarly fed straight after hearing a bell ring. In time, the sound of the bell alone made them salivate in anticipation of food. This experience had trained them to code the bell sound with a symbolic meaning, i.e. to indicate the imminent arrival of food (Image: Wikimedia Commons)

ii. It should be possible to generate a mirror-like network in other animals by training them to associate a stimulus with a meaning, perhaps a symbolic meaning as in Pavlov’s famous ‘conditioned reflex’ experiments with dogs.

Mirror neurons then, show us that something unusual is going on in our brain.  They reveal that we use all of our senses to relate physically to movement and emotion in others, and to understand our world.  They are part of the system we use to learn and imitate words and actions, communicate through language, and interact with our world as an embodied activity.

However beyond this, we cannot yet see what else they reveal.  Until we do, our conclusions about these neurons must remain ‘as dim reflections in a mirror’.

Conclusions

  • Monkey mirror neurons relate the observations of intentional movements to a sense of meaning.
  • The human mirror network activates in response to all types of human movements, including the largely ‘hidden’ movements of our vocal apparatus when we speak.

    Double rainbow. The second rainbow results from a double reflection of sunlight inside the raindrops; the colours of this extra bow are in reverse order to the primary bow, and the unlit sky between the bows is called Alexander's band, after Alexander of Aphrodisias who first described it (Image: Wikimedia commons)

    Double rainbow. The second rainbow results from a double reflection of sunlight inside the raindrops; the raindrops act like a mirror as well as a prism.  The colours of this extra bow are in reverse order to the prima... morery bow, and the unlit sky between the bows is called Alexander’s band, after Alexander of Aphrodisias who first described it. (Image: Wikimedia commons)

  • These neurons are a component of the neural network that allows us to internally code meaning into our words, and ‘embody’ our memory of the idea they symbolise.
  • The mirror network neurons seem to be part of an expanded empathy mechanism that connects higher and lower brain areas, allowing us to understand our diverse experiences from objects to ideas.
  • These cells are recruited into the mechanism by which we learn symbolic associations between items (such as words and their meaning).  This shows that it is our thinking process, rather than the cells of our brain, that makes us uniquely human.

Text copyright © 2015 Mags Leighton. All rights reserved.

References
Aboitiz, F & García V R (1997) The evolutionary origin of the language areas in the human brain. A neuroanatomical perspective   Brain Research Reviews 25(3);381-396. doi: 10.1016/S0165-0173(97)00053-2
Aboitiz, F et al.  (2005) Imitation and memory in language origins’ Neural Networks 18(10);1357.. doi: 10.1016/j.neunet.2005.04.009
Arbib, M (2005) ‘The mirror system hypothesis; how did protolanguage evolve?’  Ch 2 (p21-47 ) in Language Origins  -  Tallerman M (ed), Oxford University Press, Oxford.
Arbib, M A (2005) ‘From monkey-like action recognition to human language; an evolutionary framework for neurolinguistics.’  The behavioural and Brain Sciences 2: 105-124
Aziz-Zadeh L et al (2006) Congruent Embodied Representations for Visually Presented Actions and Linguistic Phrases Describing Actions’  Current Biology 16(18); 1818-1823
Braadbaart, L (2014) ‘The shared neural basis of empathy and facial imitation accuracy’  NeuroImage 84; 367 – 375
Bradbury J (2005) ‘Molecular Insights into Human Brain Evolution’. PLoS Biology 3/3/2005, e50  doi:10.1371/journal.pbio.003005
Carr, L.et al. (2003) ‘Neural mechanisms of empathy in humans: A relay from neural systems for imitation to limbic areas’    Proceedings of the National Academy of Sciences of the United States of America  100(9); 5497-5502
Catmur, C. et al (2011) ‘Making mirrors: Premotor cortex stimulation enhances mirror and counter-mirror motor facilitation’  (2011) Journal of Cognitive Neuroscience, 23 (9), pp. 2352-2362.  doi: 10.1162/jocn.2010.21590  http://www.mitpressjournals.org/doi/pdf/10.1162/jocn.2010.21590
Catmur C et al. (2007)  ‘Sensorimotor Learning Configures the Human Mirror System’  Current Biology 17(17) 1527-1531  http://www.sciencedirect.com/science/journal/09609822
Corballis MC (2002)  ‘From Hand to Mouth: The Origins of Language’ Princeton University Press, Princeton, NJ, USA
Corballis MC (2003)  ‘From mouth to hand: Gesture, speech, and the evolution of right-handedness’   Behavioral and Brain Sciences 26(2); 199-208
Corballis, M (2010) ‘Mirror neurons and the evolution of language’ Brain and Language 112(1); 25-35  doi: 10.1016/j.bandl.2009.02.002
Corballis, M.C. (2012) ‘How language evolved from manual gestures’ Gesture 12(2); PP. 200 – 226
Ferrari PF et al. (2003) ‘Mirror neurons responding to the observation of ingestive and communicative mouth actions in the monkey ventral premotor cortex’ European Journal of Neuroscience 17 (8); 1703–1714
Ferrari, P.F. et al (2006)  ‘Neonatal imitation in rhesus macaques’  PLoS Biology, 4 (9), pp. 1501-1508  doi: 10.1371/journal.pbio.0040302  http://biology.plosjournals.org/archive/1545-7885/4/9/pdf/10.1371_journal.pbio.0040302-L.pdf Galantucci, B et al (2006) ‘The motor theory of speech perception reviewed’  Psychon Bull Rev. 2006 June; 13(3): 361–377. PMCID: PMC2746041 NIHMSID: NIHMS136489  http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2746041/
Gallesse V et al (1996)  ‘Action recognition in the premotor cortex’  Brain 119:593–609
Gentilucci M & Corballis MC (2006) ‘From manual gesture to speech: A gradual transition’  Neuroscience & Biobehavioral Reviews 30(7); 949-960
Hage, S.R., Jürgens, U. (2006) Localization of a vocal pattern generator in the pontine brainstem of the squirrel monkey  European Journal of Neuroscience 23(3); 840 – 844  doi: 10.1111/j.1460-9568.2006.04595.x
Heyes C (2010) ‘Where do mirror neurons come from?’  Neuroscience and behavioural Reviews 34(4); 575-583  http://www.sciencedirect.com/science/article/pii/S0149763409001730
Heyes CM (2001) ‘Causes and consequences of imitation’ Trends in Cognitive Sciences 5; 245–261
Hickok G (2012)  ‘Computational neuroanatomy of speech production’ Nature Reviews Neuroscience 13, 135-145  doi:10.1038/nrn3158
Jürgens, U (2003) From mouth to mouth and hand to hand: On language evolution Behavioral and Brain Sciences 26(2); 229-230
Kemmerer D and Gonzalezs-Castillo J (2008) ‘The Two-Level Theory of verb meaning: An approach to integrating the semantics of action with the mirror neuron system’  Brain Lang. 112(1);54-76  doi: 10.1016/j.bandl.2008.09.010. Epub 2008 Nov 8.
Keysers C & Gazzola V (2009)  ‘Expanding the mirror: vicarious activity for actions, emotions, and sensations’ Curr Opin Neurobiol. 2009 Dec;19(6):666-71. doi: 10.1016/j.conb.2009.10.006. Epub 2009 Oct 31. Review.
Kohler E et al (2002) Hearing Sounds, Understanding Actions: Action Representation in Mirror Neurons Science297(5582);. 846-848 DOI: 10.1126/science.1070311 
Molenberghs P et al (2012) ‘Activation patterns during action observation are modulated by context in mirror system areas’  NeuroImage59(1); 608–615
Molenberghs P et al. (2012)  ‘Brain regions with mirror properties: A meta-analysis of 125 human fMRI studies’  Neuroscience & Biobehavioral Reviews 36(1); 341-349  http://dx.doi.org/10.1016/j.neubiorev.2011.07.004
Molenberghs P, et al. (2009)  Is the mirror neuron system involved in imitation? A short review and meta-analysis.  Neurosci Biobehav Rev. 2009 Jul;33(7):975-80. doi: 10.1016/j.neubiorev.2009.03.010. Epub 2009 Apr 1.
Mukamel R et al (2010) Single-neuron responses in humans during execution and observation of actions Current biology 20(8); 750–756  http://dx.doi.org/10.1016/j.cub.2010.02.045
Pohl, A et al.  (2013) ‘Positive Facial Affect – An fMRI Study on the Involvement of Insula and Amygdala’  PLoS One 8(8): e69886. doi:10.1371/journal.pone.0069886  PMCID: PMC3749202
Prather, J. F., Peters, S., Nowicki, S., Mooney, R. (2008). "Precise auditory-vocal mirroring in neurons for learned vocal communication." Nature 451: 305-310.
Pulvermüller F (2005) ‘Brain mechansims linking language and action’  Nature Reviews Neuroscience 6:576-582
Pulvermüller F et al (2005) ‘Brain signatures of meaning access in action wordrecognition’  Journal of Cognitive Neuroscience 17;884-892
Pulvermüller F et al (2006)  ‘Motor cortex maps articulatory features of speech sounds’ PNAS 103 (20); 7865–7870  doi: 10.1073/pnas.0509989103
Rizzolatti G & Luppino G  (2001)  ‘The cortical motor system’ Neuron 31:889–901
Rizzolatti G (1996)  ‘Premotor cortex and the recognition of motor actions. Cogn. Brain Res. 3:131–41
Pavlov IP (1927)  ‘Conditioned Reflexes; an investigation of the physiological activity of the cerebral cortex’  OUP, London (republished 2003 as ‘Conditioned Reflexes’ Dover Publications Ltd, NY, USA)
Rizzolatti G, et al (1996)  ‘Localization of grasp representation in humans by PET: 1. Observation versus execution. Exp. Brain Res. 111:246–52
Rizzolatti G, Fogassi L, Gallese V. 2002.  ‘Motor and cognitive functions of the ventral premotor Cortex’  Curr. Opin. Neurobiol. 12:149–54
Rizzolatti G. et al (2001) ‘Neurophysiological mechanisms underlying the understanding and imitation of action’  Nature Reviews Neuroscience,2(9);661-670. doi: 10.1038/35090060
Umilta et al (2001) ‘I know what you are doing.  A neurophysiological study’  Neuron 31:155-165