Causal inference during closed-loop navigation: parsing of self- and object-motion

Area of Science:

• Cognitive Neuroscience
• Computational Neuroscience
• Perception Psychology

Background:

• Causal inference (CI) is vital for internal world models, but less understood in naturalistic action-perception loops.
• Disambiguating retinal motion from self- and object-motion is crucial for navigation.

Purpose of the Study:

• To investigate causal inference in action-perception loops during closed-loop navigation.
• To model and test human performance in disambiguating self-motion and object-motion cues.

Main Methods:

• Derived a normative model for intercepting moving targets under uncertainty.
• Conducted behavioral experiments during closed-loop navigation tasks.
• Analyzed eye movements to assess gaze control strategies.

Main Results:

• Humans misattribute object motion to self-motion, reporting targets as stationary and steering towards initial positions when self-moving.
• Misattribution increased during passive vs. active self-motion and with eccentric target presentation.
• Eye movement analysis revealed altered gaze pursuit during concurrent self- and object-motion.

Conclusions:

• Demonstrates causal inference within action-perception loops during naturalistic navigation.
• Suggests a protracted temporal unfolding of causal inference computations.
• Highlights the influence of self-motion context on visual perception and action.