Background Impaired imitation has been found to be an important factor contributing to social communication deficits in individuals with autism spectrum disorder (ASD). It has been hypothesized that the neural correlate of imitation, the mirror neuron system (MNS), is dysfunctional in ASD, resulting in imitation impairment as one of the key behavioral manifestations in ASD. Previous MNS studies produced inconsistent results, leaving the debate of whether "broken" mirror neurons in ASD are unresolved. Methods This meta-analysis aimed to explore the differences in MNS activation patterns between typically developing (TD) and ASD individuals when they observe biological motions with or without social-emotional components. Effect size signed differential mapping (ES-SDM) was adopted to synthesize the available fMRI data. Results ES-SDM analysis revealed hyperactivation in the right inferior frontal gyrus and left supplementary motor area in ASD during observation of biological motions. Subgroup analysis of experiments involving the observation of stimuli with or without emotional component revealed hyperactivation in the left inferior parietal lobule and left supplementary motor during action observation without emotional components, whereas hyperactivation of the right inferior frontal gyrus was found during action observation with emotional components in ASD. Subgroup analyses of age showed hyperactivation of the bilateral inferior frontal gyrus in ASD adolescents, while hyperactivation in the right inferior frontal gyrus was noted in ASD adults. Meta-regression within ASD individuals indicated that the right cerebellum crus I activation increased with age, while the left inferior temporal gyrus activation decreased with age. Limitations This meta-analysis is limited in its generalization of the findings to individuals with ASD by the restricted age range, heterogeneous study sample, and the large within-group variation in MNS activation patterns during object observation. Furthermore, we only included action observation studies which might limit the generalization of our results to the imitation deficits in ASD. In addition, the relatively small sample size for individual studies might also potentially overestimate the effect sizes. Conclusion The MNS is impaired in ASD. The abnormal activation patterns were found to be modulated by the nature of stimuli and age, which might explain the contradictory results from earlier studies on the "broken mirror neuron" debate. Keywords: Meta-analysis, Autism, Action observation, Mirror neuron, Emotion, fMRI, ES-SDM