1 Abstract
Knowledge of speech and music depends upon the ability to perceive relationships between sounds in order to form a stable mental representation of statistical structure. Although evidence exists for the learning of musical scale structure from the statistical properties of sound events, little research has been able to observe how specific acoustic features contribute to statistical learning independent of the effects of long-term exposure. Here, using a new musical system, we show that spectral content is an important cue for acquiring musical scale structure. In two experiments, participants completed probe-tone ratings before and after a half-hour period of exposure to melodies in a novel musical scale with a predefined statistical structure. In Experiment 1, participants were randomly assigned to either a no-exposure control group, or to exposure groups who heard pure tone or complex tone sequences. In Experiment 2, participants were randomly assigned to exposure groups who heard complex tones constructed with odd harmonics or even harmonics. Learning outcome was assessed by correlating pre/post-exposure ratings and the statistical structure of tones within the exposure period. Spectral information significantly affected sensitivity to statistical structure: participants were able to learn after exposure to all tested timbres, but did best at learning with timbres with odd harmonics, which were congruent with scale structure. Results show that spectral amplitude distribution is a useful cue for statistical learning, and suggest that musical scale structure might be acquired through exposure to spectral distribution in sounds.
Competing Interest Statement
The authors have declared no competing interest.
Footnotes
I have now revised the manuscript with a follow-up experiment that more specifically tests spectral contributions to statistical learning by comparing learning with odd and even harmonics. This new Experiment 2 complements the results of the original Experiment 1 well in that both show a contribution of the spectrum to learning the statistical structure of musical scale. I have also addressed statistical / effect size concerns raised by the reviewers, and expanded the Introduction and Discussion as they suggested. I acknowledge the helpful comments of the anonymous reviewers in the revised Acknowledgements section.
