Abstract
Rather than demonstrating a simple overlap between action observation and execution fMRI responses, cross-modal fMRI adaptation has been proposed as a more suitable technique to confirm the presence or absence of mirror neurons in the human brain (Dinstein et al., 2007). Due to their multimodal properties and their proposed role in action simulation, it is suggested that regions containing mirror neurons should yield decreased responses when observing a particular action that was previously executed (or vice-versa), as opposed to executing and observing (or vice-versa) two different actions. However, the few studies that have employed this technique to study the possible presence of mirror neurons in the human brain have yielded conflicting results (Chong et al. 2008; Kilner et al., 2009; Lingnau et al., 2009; de la Rosa et al., 2016). Here we tested this proposal for the first time in monkeys using a motor-to-visual fMRI adaptation experiment. Two rhesus monkeys (male, 6–7 kg) were trained to perform in the scanner (Siemens 3 Tesla) either a number of consecutive grasping or reach-only movements, followed by observation of a video depicting either a similar grasping or reach-only motor act. Whole brain fMRI analysis did not reveal any voxels that showed a significant response reduction for same compared to different trials, even at a very low threshold. Moreover, detailed region-of-interest analysis of brain regions previously shown to house mirror neurons (premotor area F5, parietal areas AIP and PFG) also did not yield a significant reduction in fMRI signal for same compared to different trials. In line with a previous single cell study that failed to find repetition suppression effects in macaque F5 during action observation (Caggiano et al., 2013), our results suggest that care should be taken when using fMRI adaptation technique to confirm presence (or absence) of mirror neurons in the brain.