Predicting imitative performance through cortico-cerebellar circuits: A multivariate and effective connectivity study

Abstract

The ability to accurately imitate actions requires the contribution of the Mirror Neuron System (MNS) and of prefrontal and cerebellar regions. The present study aimed at investigating whether functional interaction between cortical areas and the cerebellum during the observation of complex bimanual actions can predict individual ability to imitate the same actions. Nineteen healthy participants underwent an fMRI task in which they observed complex bimanual action sequences (paper folding) and subsequently imitated the same sequences. Control conditions included passive observation of bimanual actions, observation of reaching movements, observation of actions without intent to imitate, and observation of natural landscapes. Participants’ imitation performance was video-recorded and scored for accuracy. Univariate whole-brain regression, multivariate pattern recognition, and generalized psychophysiological interaction analyses were used to assess whether activation patterns during the observation phase could predict subsequent imitation performance. The results showed that: (i) observing actions during the imitation condition activated parietal, premotor, prefrontal cortex, and lateral cerebellum; (ii) activation levels in the left anterior intraparietal sulcus (aIPS), ventral premotor cortex (PMv), dorsolateral prefrontal cortex (DLPFC), and right lateral cerebellum (CB VI) predicted imitation accuracy; (iii) a bilateral distribution pattern involving aIPS, PMv, DLPFC, and CB VI better predicted imitation performance than a whole-brain approach; (iv) increased effective connectivity between the right CB VI, left aIPS, and left DLPFC during observation-to-imitate condition correlated with higher imitation accuracy. These findings underscore the role of the cerebellum within the MNS in simulating observed actions and enabling their accurate reproduction.