Machine-Learning Classification of Motor Unit Types in the Adult Mouse

Area of Science:

• Neuroscience
• Skeletal Muscle Physiology
• Motor Control

Background:

• Motor unit diversity stems from muscle fiber and motoneuron properties.
• Quantitative definitions for mouse motor unit types are lacking.
• Current classification methods for mouse motor units are often subjective.

Purpose of the Study:

• To determine if motoneuron electrophysiology can predict the physiological identity of mouse motor units.
• To quantitatively define mouse motor unit properties.
• To establish a predictive framework for motor unit classification.

Main Methods:

• In vivo intracellular recordings were performed in mice.
• Supervised and unsupervised machine learning algorithms were employed.
• Clustering and logistic regression models were used for classification and prediction.

Main Results:

• Unbiased clustering identified four distinct motor unit groups: slow (S), fast fatigue-resistant (FR), intermediate (FI), and fast fatigable (FF).
• A predictive model showed high accuracy, with minor overlap between FI and FF types.
• Four key electrophysiological features (input conductance, rheobase, AHP duration, maximal frequency) were identified as sufficient for prediction.

Conclusions:

• Motoneuron electrophysiology provides a quantitative basis for classifying mouse motor units.
• This study offers a framework for integrating motor unit diversity into neuromuscular research.
• Findings advance the understanding of neuromuscular physiology and disease mechanisms.