Capacity estimates in working memory: Reliability and interrelationships among tasks.

Area of Science:

• Cognitive Psychology
• Neuroscience

Background:

• Working memory (WM) is often conceptualized as a limited-capacity system.
• Estimates of WM capacity from complex span tasks predict real-world performance.
• The term 'capacity' is also applied to change detection tasks, potentially referring to a different cognitive ability.

Purpose of the Study:

• To investigate whether capacity estimates from different working memory tasks reflect the same underlying cognitive capability.
• To differentiate the cognitive constructs measured by change detection, complex span, and self-ordered WM tasks.

Main Methods:

• Developed maximum-likelihood models to estimate capacity from change detection, complex span, and self-ordered WM tasks.
• Analyzed correlations between capacity estimates from these different tasks.
• Utilized exploratory factor analysis to examine the structure of performance across tasks.

Main Results:

• Capacity estimates from change detection and complex span tasks were not correlated.
• Capacity estimates from change detection tasks were correlated with those from the self-ordered WM task.
• Factor analysis indicated that self-ordered WM and change detection tasks load on one factor, while complex span tasks load on a separate factor.

Conclusions:

• The findings suggest that 'working memory capacity' is not a unitary construct.
• At least two distinct cognitive capabilities contribute to performance across different WM tasks.
• Differentiating these capabilities is crucial for understanding working memory.