Multisensory causal inference is feature-specific, not object-based.

Area of Science:

• Cognitive Neuroscience
• Psychology
• Sensory Perception

Background:

• Multisensory integration is crucial for perception.
• Causal inference mechanisms guide how sensory signals are combined.
• Previous research has not clarified whether causal inference operates at the object or feature level.

Purpose of the Study:

• To investigate whether implicit causal inference in multisensory perception applies to whole objects or specific features.
• To determine if feature-specific processing influences the integration of visual and haptic information.

Main Methods:

• Participants performed a two-alternative forced-choice task judging virtual surfaces (visual, haptic, visual-haptic) on slant and roughness.
• Response data were modeled to quantify the reliance on integrated visual-haptic information for each feature.
• The study examined how mismatches between visual and haptic information for one feature affected judgments of another feature.

Main Results:

• The degree of visual-haptic information integration varied unsystematically across different features (e.g., slant and roughness).
• A mismatch in perceived roughness did not prevent integration of visual and haptic slant information.
• Judgments were based on a feature-specific selection of sensory information.

Conclusions:

• Multisensory causal inference operates at the level of individual object features, not on the object as a whole.
• Perceptual judgments are guided by a selective integration of sensory information relevant to specific features.
• This feature-level processing suggests a flexible and adaptive mechanism for multisensory perception.