Task-uninformative visual stimuli improve auditory spatial discrimination in humans but not the ideal observer

Area of Science:

• Cognitive Neuroscience
• Auditory Perception
• Multisensory Integration

Background:

• Human survival depends on integrating environmental information from multiple senses.
• Previous research suggested Bayesian causal inference models human multisensory integration.
• Existing models often use simplified tasks, not reflecting naturalistic complex sensory processing.

Purpose of the Study:

• To investigate how spatially aligned, task-irrelevant visual stimuli affect auditory spatial discrimination.
• To test whether Bayesian ideal observer models accurately predict human performance in complex multisensory tasks.

Main Methods:

• An auditory spatial discrimination task was used.
• Participants identified the side of concurrently presented sounds.
• Two visual conditions were compared: centrally presented shapes and shapes spatially aligned with auditory stimuli.

Main Results:

• Auditory spatial discrimination performance improved when visual stimuli were spatially aligned with auditory stimuli.
• This improvement occurred even when visual stimuli were task-uninformative.
• A Bayesian ideal observer model performing causal inference could not explain the observed performance enhancement.

Conclusions:

• Human multisensory integration deviates from predictions of Bayesian ideal observer models.
• Spatially congruent, even irrelevant, visual cues can enhance auditory spatial perception.
• This suggests a more complex, less 'ideal' mechanism for how the brain integrates sensory information.