Supramodal representation of temporal priors calibrates interval timing

Area of Science:

• Cognitive Neuroscience
• Sensory Integration
• Temporal Processing

Background:

• Human timing behavior aligns with Bayesian inference, where prior knowledge and sensory data shape responses.
• The neural representation of temporal priors—whether modality-specific or supramodal—remains unclear when integrating information from different senses.

Purpose of the Study:

• To investigate whether the brain employs modality-specific or supramodal temporal priors.
• To determine how multisensory temporal information is integrated and influences timing judgments.

Main Methods:

• Participants reproduced time intervals in unisensory (visual or auditory) and multisensory contexts.
• Multisensory tasks involved randomly mixed visual and auditory intervals from adjacent distributions.
• Bayesian models with supramodal and modality-specific priors were used to analyze performance.

Main Results:

• Unisensory tasks showed classic central-tendency biases in time reproduction.
• Multisensory tasks revealed biases toward the overall distribution mean, not individual modality means.
• A supramodal prior model significantly outperformed modality-specific models.

Conclusions:

• Temporal priors are encoded via a supramodal mechanism, integrating information across senses.
• The influence of one modality on another in temporal processing varies individually.
• A generalized Bayesian model demonstrates priors are weighted averages of modality-specific temporal distributions.