There is a posting on Talking Brains (here) by Greg Hickok discussing the change in the views about mirror neurons. You may remember how magical mirror neurons were supposed to be when they were discovered. They were the source of a Theory of Mind facility and of empathy. Their malfunction was the cause of autism. They brought the rabbit out of the hat without cognition or learning. I must admit that I feel a bit smug and also admit that I feel a little ashamed at how smug I feel. The post traces some of the slow retreat from mirror neuron hype.

The discoverers of the mirror neurons in macaques have re-thought some or their original ideas:

Exciting discoveries in neuroscience over the last several years have revolved around a mechanism that unifies action perception and action execution. The essence of this mechanism — the mirror mechanism — is that each time individuals observe an action done by others, a set of neurons that code that action are activated in the motor system. Because the observers are aware of the outcome of their motor acts, they also understand what the others are doing without the necessity of an intermediate cognitive mediation. In his talk, Rizzolatti will first present some new discoveries on the mirror mechanism in monkeys. He will then present evidence that humans possess the mirror mechanism and that the anatomical location of parieto-frontal mirror networks of monkeys and humans closely coincide. Then, he will discuss the limits of the mirror mechanism in understanding other people. He will stress that the parieto-frontal mirror mechanism is, however, the only mechanism that allows a person to understand others’ actions from the inside, giving the observing individual a first-person grasp of other individuals’ motor goals and intentions.

Hickok believes this does not go far enough:

Now the mirror system has a much more restricted role in which the system allows understanding “from the inside”. It’s still not at all clear to me that this concept actually does any work, but even granting this point, it is worth noting that mirror neurons are only responsible for understanding actions that the observer knows how to perform. This is a highly restricted domain of function when considering the range of actions that one needs to understand yet has no experience executing. For example, I’ve never actually punched someone in the face, but I need to be able to recognize and understand such an action should I see it (and I believe that I can do so).

And he is not too happy with the approach to intention and goals:

There is a shift away from the idea that mirror neurons code particular movements and toward the idea that they code motor goals or intentions. So mirror neurons don’t do their magic via motor simulation, but by activating the goal or intention directly. This sounds like a profound insight, but in fact it pushes mirror neurons right out of the motor system and into the dreaded cognitive system that Rizzolatti and colleagues so wish to avoid: “Because the observers are aware of the outcome of their motor acts, they also understand what the others are doing without the necessity of an intermediate cognitive mediation“… The goal or intention is not the movement itself, it is the consequences of the movement. It’s no wonder one doesn’t need a motor system to understand the goals or intentions of actions: the goals and intentions are not motor!

In a recent paper Hickok (with Hauser) gave some reasons why cognition was avoided in the first place. Here is the abstract:

It is hard to imagine a class of neurons that has generated more excitement than mirror neurons, cells discovered by Rizzolatti and colleagues in macaque area F5 that fire both during action execution and action observation. We suggest, however, that the interpretation of mirror neurons as supporting action understanding was a wrong turn at the start, and that a more appropriate interpretation was lying in wait with respect to sensorimotor learning. We make a number of arguments, as follows. Given their previous work, it would have been natural for Rizzolatti’s group to interpret mirror neurons as involved in action selection rather than action understanding. They did not make this assumption because, at the time, the data suggested that monkey behavior did not support such an interpretation. Recent evidence shows that monkeys do, in fact, exhibit behaviors that support this alternative interpretation. Thus, the original basis for claiming that mirror neurons mediate action understanding is no longer compelling. There are independent arguments against the action understanding claim and in support of a sensorimotor learning origin for mirror neurons. Therefore, the action understanding theory of mirror neuron function requires serious reconsideration, if not abandonment.

Personally I find the biggest part of the problem to be the question of understanding. Understanding is most likely to be the concepts we build and the relationship we build between them. The network of concept-cells that we have built does not mediate understanding, or allow understanding – it IS understanding. The mirror neuron in this model is just a special type of concept. A particular mirror neuron and the nature of its connections to all the other concepts of various kinds that it is related to becomes a small piece of our understanding. We do not understand by magic – we have to learn, create concepts, build models and connect dots. Of course, it may turn out that I have to change how I picture understanding, it happens, but it will not be by replacing it with magic.