Working memory training revisited: A multi-level meta-analysis of n-back training studies

Area of Science:

• Cognitive Psychology
• Neuroscience
• Human Behavior

Background:

• The efficacy of working memory (WM) training remains controversial despite numerous studies and meta-analyses.
• Previous meta-analyses have not fully resolved the debate on cognitive transfer effects from WM training.

Purpose of the Study:

• To conduct a multi-level meta-analysis on cognitive transfer effects in healthy adults after WM updating training using n-back tasks.
• To differentiate between task-specific transfer and transfer to other cognitive domains like fluid intelligence and cognitive control.

Main Methods:

• A multi-level meta-analysis of 203 effect sizes from 33 randomized, controlled trials.
• Separation of transfer effects into task-specific (untrained n-back tasks), other WM tasks, fluid intelligence (Gf), and cognitive control tasks.
• Analysis of potential moderators including age, training dose, training type, and task content.

Main Results:

• A medium-sized transfer effect was observed for untrained n-back tasks.
• Very small effect sizes were found for transfer to other WM tasks, fluid intelligence, and cognitive control.
• Moderator analyses revealed no significant effects of age, training dose, training type, or task content.

Conclusions:

• A substantial portion of cognitive transfer following n-back WM training is task-specific.
• The findings have significant implications for understanding the limitations and potential of WM training research and application.