Simulating melodic and harmonic expectations for tonal cadences using probabilistic models

Abstract

This study examines how the mind’s predictive mechanisms contribute to the perception ofcadential closure during music listening. Using the Information Dynamics of Music model(or IDyOM) to simulate the formation of schematic expectations—a finite-context (orn-gram) model that predicts the next event in a musical stimulus by acquiring knowledgethrough unsupervised statistical learning of sequential structure—we predict the terminalmelodic and harmonic events from 245 exemplars of the five most common cadencecategories from the classical style. Our findings demonstrate that (1) terminal events fromcadential contexts are more predictable than those from non-cadential contexts; (2) modelsof cadential strength advanced in contemporary cadence typologies reflect the formation ofschematic expectations; and (3) a significant decrease in predictability follows the terminalnote and chord events of the cadential formula

Statistical learning and probabilistic prediction in music cognition: mechanisms of stylistic enculturation

“…The results reviewed in this paper have been obtained for discrete, symbolic representations of melodic musical styles. To generalize the approach to a wider variety of musical styles, the representational capacity of IDyOM must be expanded to polyphonic music but also to musical cultures that have no written tradition, where the distinction between composition and performance is blurred or nonexistent, or where music is inextricably combined with other modes of communication . Doing so would open up the approach to a much broader range of musical cultures and traditions while also introducing significant computational challenges in modeling statistical learning and probabilistic prediction.…”

“…and ). However, recent research has extended IDyOM to modeling expectations for harmonic movement and has simulated melodic and harmonic expectations separately for tonal cadences in classical string quartets . Current research is also extending IDyOM to polyphonic music represented as parallel sequences, each containing a voice or perceptual stream, for which separate predictions are generated .…”

“…However, recent research has extended IDyOM to modeling expectations for harmonic movement 60 and has simulated melodic and harmonic expectations separately for tonal cadences in classical string quartets. 61 Current research is also extending IDyOM to polyphonic music represented as parallel sequences, each containing a voice or perceptual stream, for which separate predictions are generated. 62 In time, this approach may be capable of modeling complex aspects of polyphonic structure, such as stream segregation, and interactions between harmony and melody (e.g., the ways in which harmonic syntax constrains melodic expectations).…”

Spatiotemporal whole-brain dynamics of auditory patterns recognition

“…Therefore, IDyOM returns an estimation of the predictability of each tone and uncertainty with which it can be predicted, coherently with the human perception [49].…”

“…The data was then high-pass filtered (0.1 Hz threshold) to remove frequencies that were too low for being originated by the brain. We also applied notch filter (48)(49)(50)(51)(52) to correct for possible interference of the electric current. The data was further downsampled to 150 Hz and few parts of the data, altered by large artefacts, were removed after visual inspection.…”