A Calm Look at the Most Hyped Concept in Neuroscience – Mirror Neurons (Wired)

Mirror neurons, if you believe the hype, are the keys to learning (both procedural and affective), empathy, altruism, intersubjectivity, and a host of other complex skills. In the article below, Christian Jarrett offers a summary for a recent scientific paper on mirror neurons and what we currently know (and don't know) - beyond the hype.

Here is the abstract by J.M. Kilner and R.N. Lemon for the study being reviewed:

Mirror neurons were discovered over twenty years ago in the ventral premotor region F5 of the macaque monkey. Since their discovery much has been written about these neurons, both in the scientific literature and in the popular press. They have been proposed to be the neuronal substrate underlying a vast array of different functions. Indeed so much has been written about mirror neurons that last year they were referred to, rightly or wrongly, as ‘‘The most hyped concept in neuroscience’’. Here we try to cut through some of this hyperbole and review what is currently known (and not known) about mirror neurons.

Good stuff - and a little needed clarity. From Wired.

Full Citation:
Kilner, JM and Lemon, RN. (2013, Dec 2). What We Know Currently about Mirror Neurons. Current Biology, Volume 23, Issue 23, R1057-R1062.

A Calm Look at the Most Hyped Concept in Neuroscience – Mirror Neurons

By Christian Jarrett
12.13.13


Image: Mark Dumont/Flickr
Last year I suggested that mirror neurons are the most hyped concept in neuroscience. Discovered in the 90s by neuroscientists in Italy studying monkeys, these are motor cells in the brain (involved in the control of movement) that are also activated – mirror-like – by the sight of the same movement by others. Thankfully a new open access review has just been published that provides us with a calm update on what we know so far about these fascinating cells.

First, here’s some background on the hype. Neuroscientist V.S. Ramachandran says these cells shaped our civilisation; in fact he says they underlie what it is to be human – being responsible for our powers of empathy, language and the emergence of human culture, including the widespread use of tools and fire. When mirror neurons don’t work properly, Ramachandran believes the result is autism.

For the record, a detailed investigation earlier this year found little evidence to support his theory about autism. Other experts have debunked Ramachandran’s claims linking mirror neurons to the birth of human culture. The activity of mirror neurons can be altered by simple and brief training tasks showing that these cells are just as likely to have been shaped by culture as the shaper of it.

The exaggerated and oversimplified story about mirror neurons has been swallowed whole by the media and much of the public. For a blast of this neuro-bunk try searching for “mirror neurons” on the Daily Mail website. For instance, the paper ran an article earlier this year that claimed the most popular romantic films are distinguished by the fact they activate our mirror neurons. Another claimed that it’s thanks to mirror neurons that hospital patients benefit from having visitors. In fact, there is no scientific research that directly backs either of these claims, both of which represent reductionism gone mad.

A brief search on Twitter also shows how far the concept of powerful empathy-giving mirror neurons has spread into popular consciousness. “‘Mirror neurons’ are responsible for us cringing whenever we see someone get seriously hurt,” the @WoWFactz feed announced to its 398,000 followers with misleading confidence earlier this month. “Mirror neurons are so powerful that we are even able to mirror or echo each other’s intentions,” claimed self-help author Dr Caroline Leaf in a tweet sent a few weeks ago.

In fact we do not yet have the research to show that mirror neurons are vital for human empathy, and there are reasons to believe that empathy is possible without them. For starters, we are able to comprehend the intentions behind the actions of other people or animals even if we’ve never performed, or are incapable of performing, their actions ourselves. Many brain damaged patients who can no longer produce speech are still able understand it. There are other patients who have lost the ability to express emotion yet can still understand the emotion of others.

Now a pair of neuroscientists in London have published a welcome review in the respected journal Current Biology entitled “What we know currently about mirror neurons.” In contrast to the hype that usually surrounds these cells, James Kilner and Roger Lemon at UCL have taken a calm, objective look at the literature.

They acknowledge that it is difficult to interpret mirror neuron activity in humans (using brain imaging) and so they focus on the 25 papers that have involved the direct recording of individual brain cells in monkeys. This research reveals that motor cells with mirror-like properties are found in parts of the front of the brain involved in motor control (so-called premotor regions and in the primary motor cortex) and also in the parietal lobe near the crown of the head.

Reading their paper it soon becomes clear that the term “mirror neurons” conceals a complex mix of cell types. Some motor cells only show mirror-like responses when a monkey sees a live performer in front of them; other cells are also responsive to movements seen on video. Some mirror neurons appear to be fussy – they only respond to a very specific type of action; others are less specific and respond to a far broader range of observed movements. There are even some mirror neurons that are activated by the sound of a particular movement. Others show mirror suppression – that is their activity is reduced during action observation. Another study found evidence in monkeys of touch-sensitive neurons that respond to the sight of another animal being touched in the same location (Ramachandran calls these “Gandhi cells” because he says they dissolve the barriers between human beings).

Importantly, Kilner and Lemon also highlight findings from monkeys showing how the activity of mirror neurons is modulated by such factors as the angle of view, the reward value of the observed movement, and the overall goal of a movement, such as whether it is intended to grasp an object or place it in the mouth. These findings are significant because they show how mirror neurons are not merely activated by incoming sensory information, but also by formulations developed elsewhere in the brain about the meaning of what is being observed. This is not to detract from the fascination of mirror neurons. It does show they are not the beginning of a causal path. Rather they are embedded in a complex network of brain activity.

Finally, it’s worth highlighting Kilner and Lemon’s useful summary of where we are at regards identifying mirror neuron function in humans. While Ramachandran and others have been quick to find the roots of humanity in these cells, the reality is that we’re only at the early stages of establishing whether mirror neurons exist in humans. Single-cell recording of the kind used in monkeys is too invasive to be performed in people, other than in exceptional circumstances (such as during required brain surgery). The single study of this kind published to date did find evidence for mirror neurons in the human frontal cortex and temporal lobe.

Brain imaging studies with humans have also reported what looks like mirror neuron activity in many of the same brain regions identified in monkeys. However, many of these papers only looked at the observation of actions, so they can’t determine if the same brain regions are involved in both action and observation. Other brain imaging studies have exploited the principal of adaptation – neurons get less responsive the more they’re activated. If a brain region has mirror properties there should be signs of this fatigue after action performance and observation – in fact the results are mixed with two of five adaptation studies failing to find evidence of mirror-like properties. This could be because mirror neurons don’t show adaptation, but we’ll have to see.

James Kilner and Roger Lemon are to be applauded for providing this much needed overview of the field. No doubt about it – mirror neurons are an exciting, intriguing discovery – but when you see them mentioned in the media, remember that most of the research on these cells has been conducted in monkeys. Remember too that there are many different types of mirror neuron. And that we’re still trying to establish for sure whether they exist in humans, and how they compare with the monkey versions. As for understanding the functional significance of these cells … don’t be fooled: that journey has only just begun.



~ Christian Jarrett is a cognitive neuroscientist turned science writer. He’s editor of The British Psychological Society’s Research Digest blog, staff writer on their magazine The Psychologist, and a columnist for 99U. He’s also author of The Rough Guide to Psychology, editor of 30-Second Psychology, and co-author of This Book Has Issues. His next book due in 2014 is Great Myths of the Brain.

Read more by Christian Jarrett
Follow @Psych_Writer on Twitter.

Gary Marcus - Hyping Artificial Intelligence, Yet Again

Over at The New Yorker, psychologist and cognitive scientist Gary Marcus (author of Kluge: The Haphazard Evolution of the Human Mind [2008] and The Birth of the Mind: How a Tiny Number of Genes Creates The Complexities of Human Thought [2004]) does a nice job of stripping away the hype from artificial intelligence promotion. I am grateful for Marcus.

Hyping Artificial Intelligence, Yet Again

Posted by Gary Marcus
January 1, 2014

According to the Times, true artificial intelligence is just around the corner. A year ago, the paper ran a front-page story about the wonders of new technologies, including deep learning, a neurally-inspired A.I. technique for statistical analysis. Then, among others, came an article about how I.B.M.’s Watson had been repurposed into a chef, followed by an upbeat post about quantum computation. On Sunday, the paper ran a front-page story about “biologically inspired processors,” “brainlike computers” that learn from experience.

This past Sunday’s story, by John Markoff, announced that “computers have entered the age when they are able to learn from their own mistakes, a development that is about to turn the digital world on its head.” The deep-learning story, from a year ago, also by Markoff, told us of “advances in an artificial intelligence technology that can recognize patterns offer the possibility of machines that perform human activities like seeing, listening and thinking.” For fans of “Battlestar Galactica,” it sounds like exciting stuff.

But, examined carefully, the articles seem more enthusiastic than substantive. As I wrote before, the story about Watson was off the mark factually. The deep-learning piece had problems, too. Sunday’s story is confused at best; there is nothing new in teaching computers to learn from their mistakes. Instead, the article seems to be about building computer chips that use “brainlike” algorithms, but the algorithms themselves aren’t new, either. As the author notes in passing, “the new computing approach” is “already in use by some large technology companies.” Mostly, the article seems to be about neuromorphic processors—computer processors that are organized to be somewhat brainlike—though, as the piece points out, they have been around since the nineteen-eighties. In fact, the core idea of Sunday’s article—nets based “on large groups of neuron-like elements … that learn from experience”—goes back over fifty years, to the well-known Perceptron, built by Frank Rosenblatt in 1957. (If you check the archives, the Times billed it as a revolution, with the headline “NEW NAVY DEVICE LEARNS BY DOING.” The New Yorker similarly gushed about the advancement.) The only new thing mentioned is a computer chip, as yet unproven but scheduled to be released this year, along with the claim that it can “potentially [make] the term ‘computer crash’ obsolete.” Steven Pinker wrote me an e-mail after reading the Times story, saying “We’re back in 1985!”—the last time there was huge hype in the mainstream media about neural networks.

What’s the harm? As Yann LeCun, the N.Y.U. researcher who was just appointed to run Facebook’s new A.I. lab, put it a few months ago in a Google+ post, a kind of open letter to the media, “AI [has] ‘died’ about four times in five decades because of hype: people made wild claims (often to impress potential investors or funding agencies) and could not deliver. Backlash ensued. It happened twice with neural nets already: once in the late 60’s and again in the mid-90’s.”

A.I. is, to be sure, in much better shape now than it was then. Google, Apple, I.B.M., Facebook, and Microsoft have all made large commercial investments. There have been real innovations, like driverless cars, that may soon become commercially available. Neuromorphic engineering and deep learning are genuinely exciting, but whether they will really produce human-level A.I. is unclear—especially, as I have written before, when it comes to challenging problems like understanding natural language.

The brainlike I.B.M. system that the Times mentioned on Sunday has never, to my knowledge, been applied to language, or any other complex form of learning. Deep learning has been applied to language understanding, but the results are feeble so far. Among publicly available systems, the best is probably a Stanford project, called Deeply Moving, that applies deep learning to the task of understanding movie reviews. The cool part is that you can try it for yourself, cutting and pasting text from a movie review and immediately seeing the program’s analysis; you even teach it to improve. The less cool thing is that the deep-learning system doesn’t really understand anything.

It can’t, say, paraphrase a review or mention something the reviewer liked, things you’d expect of an intelligent sixth-grader. About the only thing the system can do is so-called sentiment analysis, reducing a review to a thumbs-up or thumbs-down judgment. And even there it falls short; after typing in “better than ‘Cats!’ ” (which the system correctly interpreted as positive), the first thing I tested was a Rotten Tomatoes excerpt of a review of the last movie I saw, “American Hustle”: “A sloppy, miscast, hammed up, overlong, overloud story that still sends you out of the theater on a cloud of rapture.” The deep-learning system couldn’t tell me that the review was ironic, or that the reviewer thought the whole was more than the sum of the parts. It told me only, inaccurately, that the review was very negative. When I sent the demo to my collaborator, Ernest Davis, his luck was no better than mine. Ernie tried “This is not a book to be ignored” and “No one interested in the subject can afford to ignore this book.” The first came out as negative, the second neutral. If Deeply Moving is the best A.I. has to offer, true A.I.—of the sort that can read a newspaper as well as a human can—is a long way away.

Overhyped stories about new technologies create short-term enthusiasm, but they also often lead to long-term disappointment. As LeCun put it in his Google+ post, “Whenever a startup claims ‘90% accuracy’ on some random task, do not consider this newsworthy. If the company also makes claims like ‘we are developing machine learning software based on the computational principles of the human brain’ be even more suspicious.”

As I noted in a recent essay, some of the biggest challenges in A.I. have to do with common-sense reasoning. Trendy new techniques like deep learning and neuromorphic engineering give A.I. programmers purchase on a particular kind of problem that involves categorizing familiar stimuli, but say little about how to cope with things we haven’t seen before. As machines get better at categorizing things they can recognize, some tasks, like speech recognition, improve markedly, but others, like comprehending what a speaker actually means, advance more slowly. Neuromorphic engineering will probably lead to interesting advances, but perhaps not right away. As a more balanced article on the same topic in Technology Review recently reported, some neuroscientists, including Henry Markram, the director of a European project to simulate the human brain, are quite skeptical of the currently implemented neuromorphic systems on the grounds that their representations of the brain are too simplistic and abstract.

As a cognitive scientist, I agree with Markram. Old-school behaviorist psychologists, and now many A.I. programmers, seem focused on finding a single powerful mechanism—deep learning, neuromorphic engineering, quantum computation, or whatever—to induce everything from statistical data. This is much like what the psychologist B. F. Skinner imagined in the early nineteen-fifties, when he concluded all human thought could be explained by mechanisms of association; the whole field of cognitive psychology grew out of the ashes of that oversimplified assumption.

At times like these, I find it useful to remember a basic truth: the human brain is the most complicated organ in the known universe, and we still have almost no idea how it works. Who said that copying its awesome power was going to be easy?

Gary Marcus is a professor of psychology at N.Y.U. and a visiting cognitive scientist at the new Allen Institute for Artificial Intelligence. This essay was written in memory of his late friend Michael Dorfman—friend of science, enemy of hype.

Photograph: Chris Ratcliffe/Bloomberg/Getty