Thursday, August 27, 2009

Marco Iacoboni on imitation and sociality

Thanks to a tip from Tony Barnhart, I learned this morning of a talk at ASU today relevant to my last post ("Imitation, isolation, and independence") by UCLA neuroscientist Marco Iacoboni. Although I wasn't able to stay for the Q&A session, I did get to hear his entire presentation, titled "Imitation and Sociality: The Role of Neural Mirroring." His talk covered the following points (from his initial agenda slide):
  • Imitation in human behavior
  • Potential neural precursors in primates
  • Neural mechanisms of human imitation
  • Neural circuitry for imitation and language
  • Imitation and empathy
Dr. Iacoboni was introduced by new ASU prof. Art Glenberg, who started right off by pointing out that the existence of mirror neurons is itself controversial, and some "don't think there's much of interest proved about mirror neuron systems." Dr. Iacoboni thanked Prof. Glenberg for beginning with the "elephant in the room," and said that the question has never been raised about the existence of mirror neurons in monkeys, and suggested that some people don't want there to be homologous systems in humans, e.g., for the sake of human exceptionalism or denial of evolution. (Has your blood pressure gone up yet, Tony?)

Imitation in human behavior
He started by briefly discussing the role of imitation in human behavior, citing Andrew Meltzoff's 1977 article in Science ("Imitation of Facial and Manual Gestures by Human Neonates," (PDF) 198:75-78), noting that Meltzoff is probably the only guy to publish a photograph of himself sticking out his tongue in Science. Imitation, the copying of the behavior of another, is pervasive by humans. People copy body positions and movements, and such imitation promotes liking. (As an aside, he said that he has been interviewed by Glamour (July 2003) about his work, and can have a second career as a consultant to Internet dating services if mirror neurons turn out not to exist.) Imitation facilitates communication and conversation, and people tend to even synchronize the way they talk. (I know I've heard multiple stories of people whose accents have been changed by being around people with different accents.)

Potential neural precursors in primates
Mirror neurons were first discovered in macaque monkeys, in the ventral premotor cortex. It was found that neurons in this area fired when monkeys engaged in grasping behavior, and also fired to a lesser extent when those monkeys observed other monkeys engaged in grasping behavior. (Here, Iacoboni cited Gallese et al., Brain, 1996.)

Neural mechanisms of human imitation
Iacoboni said that the term "mirror" may be good for marketing, but may also be misleading. Mirror neurons are defined physiologically rather than anatomically, by behavior rather than location in the brain. They have motor properties, and are specialized for actions, including sensory attributes of actions, but not mere peceptions. They are not simply "monkey-see, monkey-do" cells--while 1/3 tend to fire for very specific actions, 2/3 fire for other sorts of complementary actions. Mirror neurons have abstract codings for hidden actions, action sounds, and intentions, not just specific actions. Mirror neurons that fire in response to a grasping action of picking up a laser pointer would also fire if the details of the action were obscured by a screen. The sound of tearing paper can fire mirror neurons that fire when observing paper being ripped. And if there are variant actions that achieve the same purpose, such as bringing food to the mouth, the same mirror neurons can fire. Mirror neurons learn and have some degree of plasticity.

Iacoboni's model predicts that observing an action should have the lowest level of activation for mirror neurons, performing a motor task should have a medium level, and imitation--both seeing and doing an action--should have the strongest level of activation. And that is what his research has found.

At UCLA, they've done parallel work with monkeys and humans, and identified apparently homologous brain regions between the two. The specific region where mirror neurons were first discovered, the F5 region, appears to be homologous to the BA44 region of the human brain. The "BA" stands for Brodmann Area, a part of Broca's area associated with language--those with lesions to that area have Broca's aphasia, which reduces language fluency and makes speech slow and difficult. This raises the question of whether the mirroring is effectively covert verbalization in humans.

Experiments with transcranium magnetic stimulation (TMS), where a magnetic copper coil placed against the head creates an electrical flow in the brain, interfering with the underlying electrical activity in the brain--essentially adding noise and causing disruption--have enabled a way to demonstrate causation where functional magnetic resonance imaging (fMRI) could only show correlation. Iacoboni called this a shift "from brain mapping to brain zapping." If you zap an area and cause a deficit in a particular behavior or function, you show the causal involvement of that area in the production of that behavior. Doing experiments with TMS of Broca's area vs. a control site, using an imitation task and a control task, show the essential role of Broca's area in imitation. (Here, Iacoboni cited Heiser, et al., Eur. J. of Neuroscience, 2003.)

Iacoboni showed a diagram that he labeled the "core imitation circuit" which involved three locations of the brain--the superior temporal surface (STS), which manages visual input to the system via a visual or pictorial description of an action, which then feeds to the parietal mirror neuron system (MNS), which has the motor details of an action, which then feeds to the frontal MNS, which deals with the goal or intention of an action. (There were two-way arrows between STS and parietal MNS, and between parietal MNS and frontal MNS.)

Neural circuitry for imitation and language
Iacoboni said that an old theory of speech perception which had been abandoned has now been brought back by mirror neurons. That theory is the motor neuron theory, which says that to perceive speech sounds, you simulate the generation of the same speech. Speech perception involves speech simulation. In experiments that compared brain activation of speaking and listening, he suggested that he found evidence to support this. (This must be complicated by the fact that when you speak, you hear yourself. He cited Meister, et al., Current Biology, 2007.)

He discussed hemispheres of the brain and action sounds, where the right and left motor cortexes were subjected to TMS stimulation. I didn't quite get the details of this, but apparently a response was stronger for the left hemisphere, which is dominant for language. (He cited Azir-Zadeh et al., Eur. J. of Neuroscience, 2004.) He also referred to research of somatotopic maps, indicating that even when you read sentences about hand and foot actions (as opposed to seeing them), you get activation of the motor neurons for those areas.

He then spoke about how meaning is encoded in the brain, distinguishing a symbolic approach to "embodied semantics," favoring the latter view. In the embodied view, the meanings of words are grounded in sensorimotor experience and meaning is given by associations with sensorimotor activation.

He described an experiment in how mirror neurons code intentions, where subjects were shown short videos. There were first contexts, such as a set of cookies, a teapot, gnutella, etc., set up as though someone was going to have a snack; contrasted with this was the same items, with just cookie crumbs, and empty cup, and so forth, as though someone had already had a snack. There were contrasting actions--a hand grasping the edge of a cup (as though putting it down or picking it up to serve someone else), vs. a hand grasping the handle of a cup, for the action of drinking. And then there were intention conditions, with each combination of actions embedded in a context. The result was to find a difference in activation between the intention settings, as well as between action and intention; with the act of drinking generating more activation in the inferior frontal gyrus. (Here he cited Iacoboni, PLoS Biology, 2005, "Grasping the intentions of others with one's own mirror neuron system.")

He next showed a diagram of MNS interactions, showing imitative learning and social mirroring (or empathy, or "emotional contagion"). Imitative learning involves the MNS interacting with the pre-motor cortex, while social mirroring involves the MNS interacting with the insula and the limbic system.

Imitation and empathy
He spoke about "the chameleon effect"--some people are more imitative than others, and a tendency to imitate is correlated with a tendency to be more empathetic. He showed two photographs of President Jimmy Carter and his chief of staff, Hamilton Jordan, at two different times at the same event; in both cases the chief of staff was in the same physical position as Carter, standing next to or slightly behind him.

When feeling what others feel, the mirror neurons simulate facial expressions, which then feed through the insula to the limbic system, where you feel the emotion. He referred to research on imitating and observing facial expressions proposing a neural model of empathy in humans (Carr et al., PNAS, 2003).

We are "wired for empathy," he said, and notes that he used to quote a French phenomenologist on this point, but since that's not popular among U.S. philosophers he needed to find a champion of the analytic school of philosophy. He offered two quotes from Ludwig Wittgenstein, one which began "We see emotions. We do not see facial contortions and make the inference that he is feeling joy, grief, boredom. We describe a face immediately as sad, radiant, bored, even when we are unable to give any other description of the features." (From Remarks on the Philosophy of Psychology, vol. 2, p. 100.) The other began "'I see that the child wants to touch the dog but doesn't dare.' How can I see that? - Is this description of what is seen on the same level as a description of moving shapes and colors? Is an interpretation in question?Well, remember that you may also mimic a human being who would like to touch something, but doesn't dare. And what you mimic is after all a piece of behaviour." (From Remarks on the Philosophy of Psychology, vol. 1, p. 177.)

He then spoke of experiments with facial expression photos shown to kids and asked to imitate them, where they used fMRI and compared to measures of social competence, number of play dates, number of friends, etc., and found a high correlation between mirror neuron activation and social competence. (He cited Pfeifer et al., NeuroImage, 2008.)

This then led to the issue of autism, which he described with a slide heading titled, "Broken mirrors in autism?" He spoke of observation/imitation tasks with two groups of kids, those with autism spectrum disorder and a control set, which yielded differential activity in motor neurons. (He cited Dapretto et al., Nature Neuroscience, 2006.)

After a quote from Eric Hoffer ("When people are free to do as they please, they usually imitate each other"), he spoke about human single-neuron recordings done with depth electrode readings on epilepsy patients undergoing very invasive methods to identify the focal points of seizures for surgery to remove or destroy minimal amounts of brain tissue to stop the seizures. They have studied about 10 patients per year over the last three years, using modified depth electrodes that each have 9 microwires, extending from them into the brain, one ground, and eight which each record for a single cell. On these patients they've done experiments with observation and execution of a grasping task, and with observation and imitation of facial expressions. They've taken records from the temporal lobe, amygdala, hippocampus, and other parts of the brain, and found that about 8% of cells measured have mirroring properties.

He then described some differences between human and monkey mirror neurons, the key one of which is that in some cases where mirror neurons show an increase in firings from an execution or imitation, a decrease is seen when observing. For monkeys, by contrast, the activations always go up for both observation and execution. He suggested that this may be due to a human differentiation between self and other. Humans have cases where there are excitatory effects, inhibitory effects, and opposite effects between observation and execution. There are mirror responses in humans in areas where they are not found in monkeys, the results appear to be more flexible, and there can be more prolonged responses, perhaps due to greater complexity (e.g., the language and meaning aspect?).

He ended by saying he was proud to say that his work falls within the tradition and support of Darwinian evolution--that his book, Mirroring People: The New Science of How We Connect with Others (I think you should always be skeptical of any book with a subtitle that starts with the words "The New Science of ..."), argues that mirror neurons have been selected (naturally) to facilitate social interactions. He asserted that this solves the problem of other minds, and provokes a major revision of long-standing beliefs--that we need to change the idea that we've evolved for self-preservation, and instead we're "wired for involvement and care." He concluded that he is a believer in the importance of neuroscience to society, and that rather than being isolated in an ivory tower, scientists have a responsibility to go to society and communicate their work. (And his book is written for a popular audience.)


Magic Tony said...

Hey, Jim. I haven't had a chance to read your post completely (leaving for Poland in just a few hours), but I did want to note that Iacoboni completely mischaracterized the argument that's made regarding the difficulty pinpointing mirror neurons in humans. It's not about human exceptionalism, and it's not anti-evolutionary sentiment. People like Iacoboni hypothesize that mirror neurons are the source of things like empathy, language, and imitation. Last I checked, monkeys suck at all of these things. If monkeys' mirror neuron systems (which HAVE been observed) do not produce these sorts of traits, why should we believe that they do in humans?

On another small note, you know it's a dubious presentation when they discuss the Meltzoff infant imitation study as positive evidence of anything. It has been debunked on too many occasions to count.

I'm nauseous...and not because I just had a root canal.

Jim Lippard said...

I think the answer in his presentation to the question is based on the differences he described between the monkey and human results, as well as the TMS data.

When you get a chance, can you post further critique, in particular the debunking of Meltzoff you mention?

Magic Tony said...

Hi Jim. More to come, but while I'm thinking about it, I wanted to give you some information about the flaws in the Meltzoff work. The person who did most of the debunking (and who has received very little notoriety for it) is Susan S. Jones, now at Indiana University. Her relevant publications can be found here: Long story short, the only behavior that Meltzoff could get infants to imitate reliably (and which can be replicated) was the tongue protrusion. Unfortunately, infants protrude their tongues when given any sort of stimulation (including blinking lights and the overture from "The Barber of Seville"). Also, infants stop "imitating" the tongue protrusion a few months after birth...strange for something which is thought to be innate and hard-wired.

Brian H said...

Magic, I feel your pain! Or not.

Ever reared a kid? They ape parents' emotions enthusiastically. Sometimes.

And monkeys can learn by watching each other.

Brian H said...

Existence proof:

If that's not mirroring in a human infant, Magic, I'd be curious just what you think it is!

Magic Tony said...

Brian, you're misunderstanding the argument. The argument Meltzoff (and Iacoboni) make is that newborns imitate...that it's a hard-wired behavior. Strong evidence against this interpretation of the Meltzoff data has been provided by the researcher I linked to in my previous comment. I don't think the baby in your video would be labeled as a newborn.

I agree that monkeys can learn by observing (albeit, very rudimentary behaviors), but observational learning is not the same as imitation.

Brian H said...

The operation of mirror neurons is, at core, translation of observed behaviors into actions and skills. How is observational learning not imitation? A distinction without a difference.

Magic Tony said...

Wrong, Brian. There's a rather large difference.

Mirror neurons do not translate observed behaviors into actions. When a monkey's mirror neurons are firing due to observation, they never actively imitate the behavior.

You should read this: