We are enthusiastic about the synthesis and expansion of the predictive coding of music (PCM) hypothesis in the recently published Review by Vuust et al. (Vuust, P., Heggli, O. A., Friston, K. J. & Kringelbach, M. L. Music in the brain. Nat. Rev. Neurosci. 23, 287–305 (2022))1. This latest version of the PCM hypothesis represents a substantial generalization from the domain of time (rhythm) to the domain of frequency (melody and harmony). Here, we consider limitations to the framing of the hypothesis as it currently stands, and propose generalizations that speak to a cross-cultural understanding of music beyond the Western tradition2.

Vuust et al. (see their Fig. 1) define three ‘constituents of music’: melody, harmony and rhythm. While melody and rhythm are indeed constituents traditionally found in almost all of the world’s music, chord-based harmony is not, despite its recent expansion into popular music around the globe3,4,5. Vuust et al. correctly note that “Whereas tonality is known in music from all cultures studied, neuroscientific studies have concentrated mainly on Western harmony,” but then they continue to focus predictions from the PCM hypothesis on Western chord-based harmony, limiting the generality of this hypothesis (see Vuust et al. Figs. 1 and 3–5). We propose instead that a more general PCM framework could focus on tonality, rather than harmony. Western chord-based harmony could thus be seen as a special case of generalized tonal relationships between notes, which can take the form of chords, scales or non-chord-based simultaneous tones (for example, South Asian drone or Central African hocket)4,6.

Rhythm, too, could benefit from a more cross-culturally general framework. Currently, predictions from the PCM hypothesis regarding syncopation and groove are focused on the 4/4 (‘common time’) and 3/4 (‘waltz’) metres common in Western music (see Vuust et al. Figs. 1–5). While these metres are also found in much non-Western music, so are non-isochronous and unmetered musics (for example, Middle Eastern aksak, Hindustani alap and Japanese shakuhachi)3,7,8, for which the predictions of the PCM hypothesis are less clear.

Expanding the PCM hypothesis beyond chord-based harmonies and isochronous metres could allow its predictions to be tested cross-culturally using paradigms such as transmission chains, corpus studies and probe-tone perception7,9,10. We hope that broadening predictions from the PCM hypothesis will allow them to be more easily tested against alternative or complementary explanatory frameworks such as statistical learning9,10 or social bonding11. We look forward to seeing the results.

There is a reply to this letter by Vuust, P., Heggli, O. A., Friston, K. J. & Kringelbach, M. L. Nat. Rev. Neurosci. https://doi.org/10.1038/s41583-022-00621-5 (2022).