Measuring vocal motor skill with a virtual voice-controlled slingshot

Area of Science:

• Motor Control
• Speech Science
• Voice Disorders

Background:

• Voice training and rehabilitation require learning complex vocal behaviors.
• Current metrics for vocal motor learning and skill persistence are lacking.
• Understanding vocal motor learning is crucial for effective voice therapy.

Purpose of the Study:

• To develop metrics for describing and predicting vocal motor learning.
• To investigate similarities between vocal and limb motor learning.
• To adapt a limb motor control paradigm for vocal skill acquisition.

Main Methods:

• Developed a virtual task using vocal pitch and loudness to control a projectile.
• Ten healthy subjects practiced the vocal task for five days.
• Analyzed variability using Tolerance, Noise, Covariation (TNC) costs, Lag-1 autocorrelation (AC1), and detrended fluctuation analysis scaling index (SCI).

Main Results:

• Vocal data replicated limb-based motor learning findings.
• TNC costs positively correlated with target error.
• AC1 and SCI were modulated by the task's solution manifold, indicating similar learning processes.

Conclusions:

• Vocal and limb motor learning share fundamental principles in navigating solution spaces.
• The virtual task shows potential for quantifying vocal motor learning.
• Future research should explore its application in voice disorder diagnosis and treatment.