Perceptual history biases in serial ensemble representation.

Area of Science:

• Cognitive Psychology
• Neuroscience
• Visual Perception

Background:

• Ensemble perception utilizes summary statistics for efficient object representation.
• Prior research primarily focused on simultaneous ensembles, neglecting temporal dynamics.
• Real-world environments involve temporally presented stimuli, necessitating understanding of dynamic ensemble processing.

Purpose of the Study:

• Investigate the influence of prior experience on implicit perceptual decisions in temporally presented ensembles.
• Differentiate biases arising from the current trial versus previously encountered stimuli.
• Examine how temporal factors shape ensemble statistics and perception.

Main Methods:

• Participants judged stimuli presented serially within a trial, followed by two test items.
• Methods were developed to isolate current trial biases from historical trial biases.
• Trials included equidistant test items from the current mean to probe prior influences.

Main Results:

• Strong bias towards the current trial's ensemble mean was observed.
• When current trial information was ambiguous (equidistant items), judgments shifted towards statistics from recently seen ensembles.
• Eliminated bias towards a presession prototypical diameter, confirming recency effects.

Conclusions:

• Ensemble perception is influenced by both current and recent past ensemble statistics.
• Temporal context and prior experiences play a significant role in shaping perception.
• Individual participant biases show some correlation, suggesting consistent learning patterns.