Internal models in active self-motion estimation: role of inertial sensory cues

Area of Science:

• Neuroscience
• Human Motor Control
• Perception

Background:

• Self-motion perception relies on sensory input and motor output predictions.
• In driving, steering commands can predict inertial motion cues if an internal model of steering dynamics exists.

Purpose of the Study:

• To investigate if humans can build an internal model of steering dynamics.
• To determine if steering behavior adapts to the predictability of steering-to-motion gain.

Main Methods:

• A closed-loop steering experiment was conducted on a motion platform in darkness.
• Steering-to-motion gain varied unpredictably (white noise), moderately (random walk), or predictably (constant gain) across trials.
• Participants steered to align with a memorized target, and their reliance on previous gain versus inertial feedback was assessed.

Main Results:

• Participants utilized previous trial gain more when it followed a random walk compared to unpredictable variations.
• Fast corrective responses to gain jumps occurred regardless of predictability, indicating reliance on inertial feedback.
• Findings support the internal model hypothesis over simple path integration.

Conclusions:

• The brain constructs internal models of steering dynamics to predict sensory consequences.
• These internal models are crucial for accurate self-motion estimation, particularly in dynamic environments like driving.
• The study highlights the interplay between predictive processing and sensory feedback in motor control.