Human communication is more powerful and more flexible than any communication system we know. Yet intuitively we often compare our communicative abilities to the coding-decoding scheme of a computer, where all information is captured in the transmitted signal. Is this right, wonders Arjen Stolk in a new paper in PNAS. Other theories point out that one signal is not everything, and we interpret communication in the context of our conceptual understanding of the world, the person we interact with, and the history of the interaction. This understanding would be much slower than the signal itself. We’ve put two people in two MRI scanners while they solved a series of communicative problems together. We analysed how the activity of their brains related to each other, and to the communicative task. In line with the “computer” model of human communication large parts of their brain (the motor-system) were in sync with the signal they exchanged. Interestingly, although these brain regions were involved in transmitting a signal, they were not modulated by generating or understanding the meaning of that signal. Another region was, close the the auditory and semantic system in the temporal lobe. But this region wasn’t coupled to the signal; it was coupled across communicators at a very slow frequency, going up and down with the slower understanding of a communicative meaning at a conceptual level.