Statistical Properties of Musical Creativity: Roles of Hierarchy and Uncertainty in Statistical Learning

Area of Science:

• Neuroscience
• Cognitive Science
• Psychology

Background:

• Creativity, the generation of original and valuable output, drives societal innovation and artistic expression.
• The neural mechanisms underlying creativity are not fully understood, though statistical learning is increasingly implicated.
• Statistical learning is an innate, implicit brain function crucial for development and processing structured information like music.

Purpose of the Study:

• To review neural and computational evidence linking statistical learning to creativity, termed 'statistical creativity'.
• To propose a hierarchical model of statistical learning (shallow and deep) and its role in creativity.
• To explore the influence of perceptual uncertainty and brain development on statistical creativity.

Main Methods:

• Review of existing neural and computational studies on statistical learning and creativity.
• Proposal of a hierarchical statistical learning model (shallow and deep learning).
• Discussion of factors such as perceptual uncertainty and neurodevelopmental variations.

Main Results:

• Deep statistical learning is hypothesized to be a dominant factor in musical creativity.
• Temporal dynamics of perceptual uncertainty may also causally influence statistical creativity.
• Brain development, both typical and atypical, modulates hierarchical statistical learning and creativity.

Conclusions:

• Implicit statistical knowledge, particularly through deep statistical learning, plays a key role in enabling creativity, especially in music.
• Understanding statistical learning offers insights into the implicit and sometimes subconscious nature of creative 'eureka' moments.
• Further research into statistical learning mechanisms can illuminate the fundamental processes of creativity in the brain.