Collaborative Research: The computational and neural basis of statistical learning during musical enculturation

音乐无处不在，在每一种已知的人类文化中演奏和享受，无论是过去还是现在。大多数人自发地接触和享受音乐。尽管不同文化之间存在重大的结构性差异(在旋律、和声、音色、节奏等方面)，但听众能够欣赏并迅速学习他们从未接触过的音乐。熟悉并享受一种陌生的音乐文化意味着要获得管理它的规则的隐性知识，例如旋律和节奏。人们在听音乐时也会自动建立自己的知识：他们对任何音乐的反应--无论熟悉与否--都反映了他们自己的音乐文化。这个项目结合了计算和大脑研究技术，以阐明形成这一引人注目的人类现象的基础的机制：文化融合。音乐作为一种文化对象，也作为一种可以量化的信号，构成了研究这一重要类型学习的理想领域。理解形成音乐文化基础的机制将有助于我们更广泛地解释我们如何获得一个文化领域的规则和规范。音乐信号可以被描述为一系列符号：特定音调和持续时间的音符，最终形成了我们听到的旋律。这些音乐序列可以被用作计算模型(例如，名为IDyOM的公知模型)的输入，该计算模型在许多大型乐谱语料库上计算它们的统计。当对来自不同文化的乐谱进行训练时，这位模特会“学习”不同文化的特点。这个项目利用这个模型的力量和分析透明度来研究音乐熏陶背后的神经和计算机制：人们对音乐的反应反映了他们长期接触自己的文化，但他们也可以迅速学习和享受他们从未接触过的音乐。研究人员在动物模型中进行了一系列基于行为、神经成像和神经生理学的实验，以检验长期接触熟悉的音乐文化和短期接触不熟悉的音乐文化的神经基础。这些数据不仅告诉我们音乐是如何在文化内部和跨文化学习的，而且还提供了对反映在文化领域的复杂统计信息是如何获得和使用的机械性洞察。研究人员还将与公众接触，并向高中生进行外联，让他们参与音乐和文化的研究。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Music is everywhere, played and enjoyed in every known human culture, past and present. Most people spontaneously engage with and enjoy music. Despite major structural differences across cultures (in melodies, harmonies, timbres, rhythms, etc.), listeners are able to appreciate and rapidly learn music to which they were never exposed. Becoming familiar with – and enjoying – an unfamiliar musical culture implies acquiring implicit knowledge of the rules that govern it, for example melodies and rhythms. People also automatically build on their own knowledge when listening to music: their reaction to any music – familiar or not – reflects their own musical culture. This project combines computational and brain research techniques to elucidate the mechanisms that form the basis for this compelling human phenomenon: enculturation. Music as a cultural object, but also as a signal that can be quantified, constitutes an ideal domain to investigate this important type of learning. Understanding the mechanisms that form the basis for music enculturation will help us explain how we acquire the rules and norms of a cultural domain more broadly. A musical signal can be described as a sequence of symbols: notes of specific pitches and durations that ultimately form the melody we hear. These musical sequences can be used as input to a computational model (e.g., the well-known model named IDyOM) that computes their statistics, over many large corpora of musical scores. When trained on musical scores from different cultures, the model ‘learns’ the cultural specificities. This project harnesses the power and analytic transparency of this model to investigate the neural and computational mechanisms that underlie musical enculturation: people’s responses to music reflect their long-term exposure to their own culture, but they can also rapidly learn and enjoy a music they were never exposed to. The researchers implement a series of experiments, based on behavior, neuroimaging, and neurophysiology in an animal model, to examine the neural underpinnings of long-term exposure to a familiar musical culture and short-term exposure to a non-familiar musical culture. The data tell us not just about how music is learned within and across cultures but also provide mechanistic insight into how complex statistical information reflected in cultural domains is acquired and put to use. The researchers will also engage with the public and conduct outreach to high school students to engage them in research on music and culture.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.