Wednesday, August 6, 2008

Eight Problems for the Mirror Neuron Theory of Action Understanding

Regular readers (and perhaps even occasional readers) of Talking Brains are well aware that I have been rather critical of the interpretation of mirror neurons that dominates the literature, namely, that they are the basis of action understanding. I've finally synthesized all of these critical comments into a critical review titled "Eight Problems for the Mirror Neuron Theory of Action Understanding in Monkeys and Humans." The paper has recently been submitted for publication in the Journal of Cognitive Neuroscience. The review process should be interesting; I'll post updates on the paper's progress.

In the meantime, by way of preview, here are the Eight Problems. If anyone is interested getting a discussion going on any of these issues, I'd be happy to participate. Did I miss any problems? Am I wrong about the problems I listed? Just click on the "comments" link at the end of this entry and let me know!

1. There is no evidence in monkeys that mirror neurons support action understanding.

2. Action understanding can be achieved via non-mirror neuron mechanisms.

3. M1 contains mirror neurons

4. The relation between macaque mirror neurons and the “mirror system” in humans is either non-parallel or undetermined

5. Action understanding in humans dissociates from neurophysiological indices of the human “mirror system”

6. Action understanding and action production dissociate

7. Damage to the inferior frontal gyrus is not correlated with action understanding deficits

8. Generalization of the mirror system to speech recognition fails on empirical grounds


Anonymous said...

Why is the fact that M1 contains mirror neurons a problem? I'm sure it is articulated in the paper, but it's not obvious.

Greg Hickok said...

Because if M1 cells show "mirror" effects one could interpret all "mirror" responses, including those in F5 as unimplemented motor activations. Rizzolatti and colleagues realized this possible explanation and carried out the critical control, reported in their early Brain paper on mirror neurons. According to that study, M1 cells did not show mirror effects. They also measured motor evoked responses in the distal limbs of monkeys to make sure that covert motor plans were not driving the F5 responses. Again, no low level motor responses were found when monkeys viewed actions. This is what made mirror neurons so interesting. But now a recent study has found that M1 cells show mirror responses. The critical control does not hold up. Notice too, that the earliest human evidence for "mirror neurons" -- the MEP studies -- are open to the same criticism.

Drew H. said...

Great distillation of the problems. Looking forward to the paper.

meenakshi said...

do you mean to say then action understanding is not done through mirror neurons ?

Zarinah Agnew said...


I've made this point somewhere else on your blog. If mirror neurones are involved in action understanding then we can only be talking about highly congruent mirror neurones. The presence of broadly congruent mirror neurones is a problem for this theory; these cells respond during execution of a specific action but during observation of a range of actions. How can this be reconciled with a theory of accurate action understanding?

(PS. I should say that whilst I'm sure many have had similar thoughts, I believe G. Csibra should get the credit for publicly discussing these ideas in the form of an online conference)

Greg Hickok said...

Hi Zarinah,
It is true that the congruence properties of mirror neurons aren't discussed much, so that the general consumer of this research isn't aware that many mirror neurons don't have a one-to-one correspondence between executed and perceived actions that drive them. This is an important point that deserves attention in any serious attempt to explain action understanding in terms of mirror neurons. I don't think that it is necessarily a fatal problem for the general approach, though, because one might appeal to population coding for actions that are not uniquely coded by single neurons. Thanks for your comment!

Zarinah Agnew said...

Hi there,

I thought I'd mention this paper by Dinstein and colleagues (Brain Areas Selective for Both Observed and Executed Movements, 2006). They use repetition suppression to try to identify unique patterns of activity that are action specific in both the executed and observed domain. The paper is long and difficult to read (we covered it in journal club a few weeks ago) but it has an interesting approach to this question. Namely that cells specific to a certain stimulus wil show an adaptive response (repetition suppression).

They report the same areas as have been previously reported to show mirror responses. Unfortunately they did not find what would have been the most persuasive evidence for mirror responses; multimodal repetition suppression (i.e. when observing a specific action 'primes' the response to executing the same specific action). This may well be a power issue as they didn't have a great number of subjects.

This type of approach may be useful to address the issue of congruency.

Andrew Wilson said...

I'm pleased to see this critique - I've been grumpy about the mirror neuron concept for a long time but I'm not a neuroscientist nor do I have the time to really engage the literature in detail, so I didn't really feel I could quite get away with this.

Where I am more expert is in perception/action research. What do you know of biological motion perception? My feeling is that this literature provides a much more robust mechanism for action perception; the best work I know of comes from Nikolaus Troje's lab but the field's been alive and well for 50 years.

My main concern has always been that you only need mirror neurons to reconnect perception and action if perception and action are somehow separated to begin with. Given the ready availability of perceptual information for action, who needs mirror neurons?

Good luck with the paper, these are good questions that need to be asked.