The aim of the study was to investigate a method for analysing mother-infant communication considering its vocal, gestual, visual and affective aspects. Microanalyses of video and audio documents made possible a precise measure of time intervals in the different modalities studied. Thus, we were able to observe how these modalities are coordinating during interaction. Comparing mutual attunement in both early interaction and improvised jazz music leads us to consider rhythmicity in early communication, focalising on the research of time units in all modalities. Both macro analysis (looking at narrative episodes), and micro analysis (studying pulses) were performed. Our results highlight the existence of clearly bounded narrative episodes, constituted of time units, particular for each expressive modality, but always stable and around 1 second. However, we think that these multiple rhythmicities that formed communication are also constituted of a common stable unit. This hypothesis should explain why different rhythm succeed in joining the same beat, thus creating emotional climax such as those which were observed in our data.