Musical interactions between babies and their primary caregivers are very frequent during the early years of life and their impact on dyadic interaction and infants’ development has garnered significant attention in recent literature. However, the difficulties that natural observations entail have meant that research often carries out methodological manipulations that have a significant impact on the phenomenon studied. In order to clarify how to investigate best natural musical interactions and the information that these can provide, we have carried out a systematic review to analyze the proposed scenarios and the variables analyzed in the studies published on such interactions between main caregivers and babies under three years old. We have screened 971 articles and yielded 27. We have found a higher prevalence in the literature of studies on singing interactions, between mothers and babies under 12 months of age. We have also been able to identify two extremes in terms of methodological structuring of natural interactions. Regarding the analysis variables, a few behaviors are repeated throughout the studies, being emotions, rhythmic behaviors and characterizations of the vocal emissions common between parents and babies. Synchrony is the dyadic variable with the most weight and also one of the preferred focuses of interest in the most recent literature that has undergone a shift of focus from characterization of musical interactions to the search for the mechanisms that underlie and make them specific.