Low-priority items are held in visual working memory: Evidence from flexible allocation in a two-alternative...

Area of Science:

• Cognitive Psychology
• Neuroscience
• Human Perception

Background:

• Visual working memory (VWM) is considered capacity-limited, evidenced by performance declines and flat error distributions in memory tasks.
• Continuous resource models propose that limited VWM resources can be distributed, leading to low-resolution representations with potentially uniform error distributions.

Purpose of the Study:

• To investigate whether low-probability items in VWM are held in memory with low resolution, rather than being absent.
• To test predictions of continuous resource models of VWM using manipulated memory resource allocation.

Main Methods:

• Memory resource allocation was manipulated by varying the probability of an item being probed during recall.
• Responses were collected using a mix of continuous estimation and two-alternative forced-choice (2AFC) trials.
• Discrimination difficulty between target and lure colors was manipulated to compare with continuous response distributions.

Main Results:

• Accuracy on 2AFC trials was influenced by both discrimination difficulty and probe probability.
• Above-chance performance was observed even in the lowest probe probability condition (10%), indicating low-resolution memory representations.
• These findings align with continuous resource models predicting distributed memory resources.

Conclusions:

• VWM may not be strictly capacity-limited but rather utilizes distributed resources, resulting in low-resolution representations for low-priority items.
• The study supports continuous resource and signal detection models of VWM, challenging purely discrete capacity views.