Even though I've been spouting off for the last few months about the problems with mirror neurons, having read this review paper carefully, I have to admit there was much I didn't know about their properties or the theories behind their function. Unfortunately, learning more only solidified my concerns and lead to even more doubts and questions about the mirror neuron theory of action understanding, imitation, and speech processing.
There is a LOT to discuss in the Rizzolatti & Craighero paper. It took us a couple of hours in class this week to get only about halfway through. Quite an interesting read, but I have to say, it's not exactly the tightest paper I've read. Plenty of speculation, hints of circularity, over-generalization, etc. We all agree that mirror neurons are very interesting neural creatures, but the idea that they are the basis for action understanding is darn near incoherent, and by the authors' own admission effectively untestable.
Since there's so much to discuss, I'll breakdown the summary discussion of the Rizzolatti & Craighero paper (hencefore RC) into a few posts.
A couple of initial observations:
1. There are two types of "visuomotor" neurons in macaque F5, "canonical neurons, which respond to the presentation of an object, and mirror neurons, which respond when the monkey sees object-directed action" p. 170. Having not read all the literature on F5 physiology, I wondered if anyone has ever claimed that canonical neurons are the neural basis for object understanding. If not, why? Maybe this is a naive question, but if a cell that responds to the perception of an object-directed action is the basis of action understanding, why isn't a cell that responds to an object the basis of object understanding? Somebody correct me.
2. It is claimed that there are two types of mouth-related mirror neurons, "ingestive mirror neurons" and "communicative mirror neurons." (p. 171). However, according to RC's definitions, in fact communicative mirror neurons do not exist:
a) "Mirror neurons in which the effective observed and effective executed actions correspond in terms of the goal... and means for reaching the goal... have been classed as 'strictly congruent' ... Mirror neurons that do not require the observation of exactly the same action that they code motorically have been classed as 'broadly congruent.' (p. 170)
b) "The most effective observed action for [communicative mirror neurons] is a communicative gesture... However, from a motor point of view they behave as the ingestive mirror neurons, strongly discharging when the monkey actively performs an ingestive action" (p. 171).
So basically, what RC are calling "communicative mirror neurons" are just ingestive mirror neurons that are "broadly congruent." I wonder if monkeys confuse the perception of communicative gestures with the perception of ingestive actions?
3. Cells in the monkey superior temporal sulcus respond to the perception of actions. "STS appears to code a much larger number of movements than F5... [and] STS neurons do not appear to be endowed with motor properties." (p. 171). Hmm. So STS codes for a much larger number of actions than F5. Sounds to me like STS, not F5, is the neural basis of action understanding... If STS is not the basis for action understanding, what IS it doing? RC's position on the relation between the STS and mirror system is downright puzzling: "STS is strictly related to it [the mirror neuron circuit] but, lacking motor properties, cannot be considered part of it." (p. 172). What does it mean to be "strictly related" but "not part of"???