Sustained sensorimotor control as intermittent decisions about prediction errors: computational framework and...

Area of Science:

• Neuroscience
• Cognitive Science
• Robotics

Background:

• Human sensorimotor control models often assume continuous actions.
• Existing models lack a comprehensive framework for intermittent control mechanisms.

Purpose of the Study:

• To propose a conceptual and computational framework for human sensorimotor control, emphasizing intermittency.
• To model the sensorimotor task of steering a car using this novel framework.

Main Methods:

• Developed a framework combining motor primitives, sensory prediction, and evidence accumulation.
• Utilized a driving simulator to collect empirical data for model-fitting.
• Extended existing one-dimensional continuous models to intermittent behavior.

Main Results:

• Human steering control is better described as discrete stepwise adjustments than continuous control.
• Sensory predictions may influence control adjustment amplitudes.
• Evidence accumulation mechanisms show trends related to control onset timing.

Conclusions:

• The proposed framework offers a viable model for intermittent sensorimotor control.
• Findings support discrete control adjustments and evidence accumulation over threshold mechanisms.
• The framework provides a basis for further research into neurobiological control processes.