Is mirror neuron activity just a mirage?
By David Robson They’ve been used to explain autism, empathy and why porn turns us on. Now some of the past findings that mirror neurons have been said to explain have been called into question by new research which suggests that these past investigations have not looked closely enough. Mirror neurons are brain cells that fire both when performing an action, and when the brain is observing that action being performed by someone else. Because of this, they have been a focus of keen interest among scientists looking for the neurological roots of things like empathy or imitation, and even morality. They have been directly observed in monkeys by measuring the firing of single neurons in the brain. In humans, mirror neurons can’t be observed directly, so most studies have instead looked for them by taking fMRI scans of subjects performing and observing various activities, and then finding the regions of the brain that light up in both situations. Now Ilan Dinstein from New York University has cast doubt on these studies. He claims that they haven’t examined these regions in fine enough detail: the neurons responsible for the increased activity when observing may be different from those that are active when performing, he says. Assigning them a role in autism and morality is premature, he concludes. To try to pick apart which neurons are active when observing and which when performing, Dinstein and a team from New York University and the University of Washington in Seattle recruited five volunteers to play the rock, paper, scissors game while an fMRI scanner recorded images of their brain activity. While the volunteers played 360 times against a video recording of an opponent, the researchers trained a computer program to recognise the characteristic patterns of brain activity associated either with the observation of each hand gesture, or the production of each hand gesture. The team studied activity in three regions: the motor cortex, the visual cortex and the anterior intraparietal sulcus (aiPS) – a region previously assumed to be full of mirror neurons. More: See an interactive graphic of the human brain Dinstein’s team then tested whether the software could successfully predict from brain activity alone whether a subject had seen or produced a hand gesture. As expected, using brain activity in the visual cortex it could predict an observed gesture with high accuracy (72%). Similarly, using brain activity in the motor cortex it predicted which gesture had been produced by the subject. If the aiPS was full of mirror neurons, the activity in this region would have been similar during both observation and execution, and the “mind reading” software should have been able to use activity in this region seen during observations to predict executed gestures, and vice versa. Yet it was able to do this no better than would be expected by chance. This shows the area is largely populated by a mix of motor and visual neurons, Dinstein claims. “Previous studies have ignored these other neural populations, which in fact form a staggering majority,” he says. “Our study puts a huge question mark on their conclusions because we have shown they were not measuring the activity of mirror neurons.” Christian Keysers from the BCN Neuroimaging Center in Groningen, the Netherlands, says he doesn’t find the scarcity of mirror neurons surprising, since experiments on primates have already suggested the same result. He says he has used a similar technique to detect mirror neuron activity from some fMRI scans. He agrees Dinstein’s research does suggest that some conclusions drawn from previous experiments may be wrong. Marco Iacoboni from the University of California, Los Angeles, who had previously reported evidence for mirror neurons in the human brain, is not convinced by the research. He points out that the method Dinstein used to analyse the brain activity could only correctly identify the gestures 70-80% of the time, even when analysing the visual or motor cortices alone – so it is not surprising that it failed to notice the more subtle changes within the aiPS. Journal reference: The Journal of Neuroscience, DOI:10.1523/JNEUROSCI.3585-08.2008 More on these topics: