On-line processing of uncertain information in visuomotor control

Area of Science:

• Neuroscience
• Motor Control
• Computational Neuroscience

Background:

• Sensory observations are inherently delayed and noisy, leading to instability and uncertainty.
• The brain must overcome these limitations for effective environmental interaction and motor control.

Purpose of the Study:

• To test the theory that the brain uses probabilistic state estimation to process sensory information.
• To investigate how the brain integrates predictions with delayed sensory feedback during motor tasks.

Main Methods:

• Participants performed reaching movements towards targets with varying blurriness (uncertainty).
• Target position or blurriness was occasionally altered mid-movement.
• Motor response trajectories were analyzed to assess the integration of sensory information and predictions.

Main Results:

• Increased target uncertainty led to longer reach reaction times.
• Motor responses to subsequent targets were influenced by prior target uncertainty.
• Movement trajectories aligned with predictions of a minimum variance state estimator.

Conclusions:

• The brain employs a minimum variance state estimator to generate motor commands.
• Motor output is a continuous combination of predicted sensory states and delayed sensory measurements.
• This strategy enables robust movement control in the presence of sensory noise and delays.