Did you see the sound? A Bayesian Perspective on Crossmodal Perception in Low Vision

Area of Science:

• Neuroscience
• Vision Science
• Auditory Perception

Background:

• Multisensory integration is crucial for perception, with assumptions that degraded visual input amplifies cross-modal effects.
• Previous research has not clarified if low vision increases susceptibility to cross-modal illusions or if local visual reliability is key.

Purpose of the Study:

• To investigate whether low vision enhances audiovisual interactions, specifically the Double Flash Illusion.
• To determine if cross-modal illusion susceptibility is influenced by local variations in visual reliability.

Main Methods:

• Tested low vision and sighted participants on the Double Flash Illusion across 24 visual-field locations.
• Measured flash-detection accuracy independently for each location.
• Employed Bayesian Causal Inference modeling to analyze sensory precision and visual noise.

Main Results:

• Both groups experienced auditory-driven increases in perceived flash numerosity, but only sighted controls consistently perceived the double-flash.
• Illusion strength correlated with local flash detection accuracy, not visual field eccentricity.
• Bayesian Causal Inference models showed weaker, more variable fits for low vision observers, linked to reduced visual sensitivity.

Conclusions:

• Low vision does not globally amplify audiovisual interactions or cross-modal illusions.
• Auditory enhancement of visual perception relies on the local reliability of the visual signal.
• Degraded vision leads to reduced adherence to Bayesian-optimal multisensory integration predictions.