The article discusses the role of the mirror neuron system (MNS) in understanding others' intentions through action observation. The authors propose that the MNS functions within a predictive coding framework, which involves minimizing prediction errors at all levels of the cortical hierarchy engaged during action observation. This framework is equivalent to empirical Bayesian inference, where prior expectations emerge naturally from the hierarchical models. The MNS is composed of areas such as the ventral premotor cortex, inferior parietal lobule, and superior temporal sulcus (STS), which are reciprocally connected. The authors argue that the MNS can infer the most likely cause of an observed action by comparing predicted and actual motor commands, goals, and intentions, and adjusting these predictions based on context and prior information. This mechanism allows the MNS to distinguish between different intentions even when the observed movements are identical, as demonstrated in a thought experiment involving Dr. Jekyll and Mr. Hyde. The predictive coding account provides a detailed computational explanation for how the MNS infers intentions and connects generative models used in perceptual inference with forward models used in motor control.The article discusses the role of the mirror neuron system (MNS) in understanding others' intentions through action observation. The authors propose that the MNS functions within a predictive coding framework, which involves minimizing prediction errors at all levels of the cortical hierarchy engaged during action observation. This framework is equivalent to empirical Bayesian inference, where prior expectations emerge naturally from the hierarchical models. The MNS is composed of areas such as the ventral premotor cortex, inferior parietal lobule, and superior temporal sulcus (STS), which are reciprocally connected. The authors argue that the MNS can infer the most likely cause of an observed action by comparing predicted and actual motor commands, goals, and intentions, and adjusting these predictions based on context and prior information. This mechanism allows the MNS to distinguish between different intentions even when the observed movements are identical, as demonstrated in a thought experiment involving Dr. Jekyll and Mr. Hyde. The predictive coding account provides a detailed computational explanation for how the MNS infers intentions and connects generative models used in perceptual inference with forward models used in motor control.