It has often been claimed that dysfunction of the human mirror neuron system is a key cause of difficulties in imitation and social interaction in individuals with autism spectrum condition. However, this claim is very controversial. A variety of neuroscientific methods have been used to study the integrity of the mirror system in autism, but an overview of these results does not yet exist.
Objectives:
This study aims to evaluate the broken mirror hypothesis of autism, by conducting a systematic review of all neuroscientific studies of mirror neuron function in children and adults with autism spectrum disorder.
Methods:
25 studies were identified which use neurophysiological methods (EEG / MEG / TMS / fMRI / EMG / eyetracking) to examine mirror neuron systems in autism. These were evaluated in terms of whether or not they provide statistically robust evidence for abnormal mirror neuron responses in autism which cannot be attributed to other effects.
Results:
Considering studies of the MNS which use weakly localised methods (EEG / MEG / TMS / EMG), only the TMS studies (n=2) consistently report differences in autism. EEG (n=8), MEG (n=3) and EMG (n=2) all report mixed results with little clear evidence for abnormal mirror systems. 4 eyetracking studies all report normal behaviour in autism. 8 fMRI studies have directly measured MNS engagement. Studies using emotional stimuli (n=4) report some group differences, while those using non-emotional stimuli (n=4) mostly do not. Structural MRI studies do not report consistent abnormalities of the MNS.
Conclusions:
These data demonstrate that over 10 years after it was first suggested, there is little evidence to support idea that mirror systems are broken in autism or that this failure is a primary cause of poor social skills. I suggest that it is more useful to consider how social cues exert top-down control over imitation responses in ASC (Wang & Hamilton, 2012). In this model, basic visuomotor processing in the mirror neuron system is intact in ASC, but top-down modulation of this system by social signals is abnormal. This top-down model can account for current data and also for new results showing differential imitation of goal-directed and social actions in children with ASC (Marsh & Hamilton, IMFAR poster). Implications of this model for both research and therapies will be considered.
See more of: Brain Imaging
See more of: Brain Structure & Function