A model of reaching dynamics in primary motor cortex

Area of Science:

• Neuroscience
• Computational Neuroscience
• Motor Control

Background:

• The primary motor cortex (M1) is crucial for planning and executing voluntary movements.
• Understanding the dynamic representation of movement in M1 is key to deciphering motor control.
• Previous studies focused on static representations, but movement dynamics are complex.

Purpose of the Study:

• To investigate the temporal dynamics of movement representation in the primary motor cortex.
• To explore how the motor cortex represents changes in movement direction and magnitude.
• To model neural activity during reaching movements with changing targets.

Main Methods:

• Trained a fully recurrent neural network (RNN) to simulate motor cortical activity.
• The RNN outputted movement direction and magnitude towards randomly changing targets.
• Analyzed model neuron properties and ensemble representations for similarities to biological M1.

Main Results:

• Model neurons developed preferred directions and response properties analogous to real motor cortical neurons.
• Ensemble representations of movement changed nearly monotonically when targets shifted.
• Individual neurons exhibited strong, nonmonotonic transients during target changes.

Conclusions:

• Nonmonotonic transients in individual neurons act as internal recurrent signals.
• These transients accelerate the ensemble representation's adaptation to new movement goals.
• Experimental verification of these transients is possible with minor modifications to existing paradigms.