Sigmoidal Decoding from Distinct M1 Spiking Populations and LFP Band Power for Locomotion Speed in Mice

Area of Science:

• Neuroscience
• Motor Control
• Computational Neuroscience

Background:

• The primary motor cortex (M1) role in encoding locomotion speed is not fully understood.
• Previous studies suggest linear or gain-modulated tuning for speed representation in M1.
• A sigmoidal framework for speed encoding in M1 has not been previously explored.

Purpose of the Study:

• To investigate the encoding of locomotion speed by neurons in the mouse primary motor cortex (M1).
• To determine if a sigmoidal model can describe the relationship between neural activity and locomotion speed.
• To compare spiking activity and local field potential (LFP) band power for speed encoding.

Main Methods:

• Recorded extracellular neural activity from 8 mice using chronic laminar arrays over 8 weeks.
• Tracked locomotion speed using DeepLabCut and analyzed 5,889 single units and 384 channels.
• Applied unsupervised clustering to identify neural populations and sigmoidal, linear, and quadratic models for tuning analysis.

Main Results:

• Identified two distinct neural populations: 70.8% speed-positively related and 29.2% speed-inversely related units.
• Sigmoidal models significantly outperformed linear and quadratic models for speed-positively related units.
• The speed-inversely related population showed higher decoding accuracy, and LFP band power also exhibited sigmoidal tuning but with lower fidelity.

Conclusions:

• Mouse M1 utilizes a push-pull sigmoidal architecture for spiking-based locomotion speed representation.
• A shared speed threshold suggests a common subcortical locomotor gate influencing both neural populations.
• LFP provides a complementary, stable signal for coarser speed estimation compared to spiking activity.