Optimal observer model of single-fixation oddity search predicts a shallow set-size function

Area of Science:

• Cognitive psychology
• Computational neuroscience
• Visual perception

Background:

• Oddity search tasks, where the target is unique among distractors, often show performance insensitive to set size.
• Previous explanations focused on perceptual and attentional mechanisms for this anomalous set-size effect.

Purpose of the Study:

• To investigate if stimulus information and task demands can explain the shallower set-size function in oddity search.
• To compare computational models with human performance in oddity and target-known search tasks.

Main Methods:

• Developed an ideal-observer model and a difference-coding (standard-deviation) model for single-fixation oddity search.
• Compared model performance to an ideal observer in target-known search and to human accuracy data.

Main Results:

• Both ideal-observer and difference-coding models for oddity search exhibited shallower set-size functions than the target-known ideal observer.
• These models accurately predicted human search accuracy.
• The ideal-observer model was a superior predictor compared to the standard-deviation model for most datasets.

Conclusions:

• Stimulus information and task demands play a crucial role in explaining the oddity search set-size effect.
• Ideal-observer analysis is vital for distinguishing between inherent human perceptual/attentional mechanisms and task-specific stimulus information effects.