Visual recency bias is explained by a mixture model of internal representations.

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Visual Perception

Background:

• Human behavior often shows a bias towards recent events, a phenomenon observed even in visual perception tasks.
• The underlying internal models and reasons for this suboptimal recency bias remain poorly understood.

Purpose of the Study:

• To investigate the internal model underlying human recency bias in a visual orientation estimation task.
• To frame recency bias within the context of incremental Bayesian inference.

Main Methods:

• Utilized a standard orientation estimation task to probe human visual perception.
• Applied Bayesian inference modeling to analyze the observed recency bias.

Main Results:

• The study identified a weighted mixture of past states as the only Bayesian model that can explain recency bias.
• Results suggest the bias stems from participants' inability to infer a data model within visual short-term memory.

Conclusions:

• Recency bias in visual perception may be a consequence of how information is represented and updated internally.
• The findings offer insights into the internal representations and inference processes governing human decision-making under uncertainty.