A modular theory of multisensory integration for motor control

Area of Science:

• Neuroscience
• Computational Neuroscience
• Motor Control

Background:

• The brain integrates sensory information (vision, proprioception) for targeted movements.
• Conventional models suggest optimal combination of sensory signals for a single estimation.
• Emerging evidence points to a more modular approach in sensorimotor processing.

Purpose of the Study:

• To computationally examine concurrent models of sensory integration.
• To compare concurrent models with conventional converging multi-sensory signal models.
• To investigate the role of signal noise in sensorimotor transformations and sensory modality selection.

Main Methods:

• Computational analysis of two concurrent sensorimotor models.
• Comparison with conventional converging multi-sensory models.
• Review of existing and novel published studies on sensory integration and noise.

Main Results:

• Evidence favors concurrent formulations of sensory integration over conventional models.
• Additive signal noise influences reliance on different sensory modalities.
• Analysis explains shifts in sensory modality preference and recoding of information.

Conclusions:

• The central nervous system (CNS) employs concurrent, modular processing for sensory integration.
• Understanding signal noise is crucial for explaining sensory modality shifts in sensorimotor control.
• Concurrent models provide a better framework for understanding sensorimotor transformations and reference frame selection.