Biased guessing in a complete-identification visual-working-memory task: Further evidence for mixed-state models.

Area of Science:

• Cognitive Psychology
• Neuroscience

Background:

• Visual working memory (VWM) is crucial for cognitive tasks.
• Understanding the underlying mechanisms of VWM, including potential biases and errors, is essential.

Purpose of the Study:

• To investigate the role of mixed states and guessing processes in visual working memory.
• To compare the explanatory power of mixed-state models versus variable-resources (VR) models in a complete-identification VWM task.

Main Methods:

• Participants completed a complete-identification VWM task with 16 colors.
• Payoffs for correct responses were manipulated across trials.
• Individual-subject identification-confusion data were analyzed and modeled.

Main Results:

• Evidence for a systematic guessing process was found, where participants used high-payoff responses when memory was absent.
• Mixed-state models, incorporating memory and guessing, accurately accounted for the data.
• Standard VR models without a guessing state struggled to explain the observed performance patterns.

Conclusions:

• Discrete, mixed states play a significant role in visual working memory.
• Future VR models should incorporate sophisticated decision rules for tasks like complete identification.