Sensory and multisensory reasoning: Is Bayesian updating modality-dependent?

Area of Science:

• Cognitive Science
• Neuroscience
• Psychology

Background:

• Bayesianism posits that probabilistic updating is independent of sensory modality.
• Previous research has largely assumed modality independence in Bayesian updating models.

Purpose of the Study:

• To investigate whether probability judgments align with Bayesian assumptions across different sensory modalities.
• To examine the impact of audio-visual information on Bayesian updating and identify potential limitations in multisensory integration.

Main Methods:

• Five experiments were conducted using visual and auditory stimuli.
• Participants made probability judgments based on single-modality (auditory or visual) and multi-modality (audio-visual) information.
• Experimental conditions included synergy, contrast, and double-contrast scenarios for audio-visual information.

Main Results:

• Probability judgments conformed to Bayesian assumptions with single-modality information.
• Audio-visual information led to a 'synergy-contrast effect,' where accuracy depended on whether individual modalities supported the joint hypothesis.
• Accuracy was significantly reduced in contrast and double-contrast conditions compared to the synergy condition, approaching chance levels in the latter.

Conclusions:

• The modality of information processing influences Bayesian updating, particularly in multisensory contexts.
• A synergy-contrast effect was identified, demonstrating a violation of modality independence in Bayesian updating.
• This finding highlights a significant, previously undocumented limitation of multisensory integration.