Using cortico-cerebellar structural patterns to classify early- and late-trained musicians.

Area of Science:

• Neuroscience
• Cognitive Science
• Music Psychology

Background:

• Evidence suggests a sensitive period for musical training, with early starters (before age 7) showing superior musical skills and distinct brain structures.
• Differences are noted in motor cortical and cerebellar regions between early-trained (ET) and late-trained (LT) musicians.

Purpose of the Study:

• To investigate structural brain differences between ET and LT musicians using machine learning.
• To identify the age boundaries of the sensitive period for early musical training.

Main Methods:

• Support vector machine (SVM) models, a supervised machine learning technique, were employed.
• Recursive feature elimination with cross-validation was used to select key brain regions.
• Regions of interest were focused on the cerebellum and cortical sensorimotor areas.

Main Results:

• An optimal SVM model accurately classified ET and LT musicians, identifying 17 critical regions (9 cerebellar, 8 sensorimotor).
• The model demonstrated high accuracy, sensitivity, and specificity in distinguishing between early and late trainees.
• The classification was most accurate when defining ET musicians as those starting training before age 7, outperforming models with different age cutoffs (5-10 years).

Conclusions:

• Musical training before age 7 significantly influences cortico-cerebellar structure in adulthood.
• Findings support the hypothesis of reciprocal interactions between connected brain regions during development, affecting brain and behavioral maturation.
• The study provides robust evidence for a sensitive period in musical development impacting neuroplasticity.