Training-dependent transfer within a set of nested tasks.

Area of Science:

• Cognitive Psychology
• Neuroscience
• Human Performance

Background:

• Cognitive task practice can improve performance on untrained tasks.
• This performance benefit, known as transfer, is typically specific to the trained task.
• The precise nature of task similarity that drives transfer remains an open question.

Purpose of the Study:

• To investigate the type of task similarity that best predicts cognitive transfer.
• To determine if low-level versus high-level training influences the specificity of transfer.
• To explore how training effects cascade through a hierarchy of nested tasks.

Main Methods:

• A randomized controlled trial with 175 participants and an adaptive control group.
• Development of a hierarchical set of nested assessment tasks.
• Training on either a low-level (simultaneous judgments) or high-level (delayed judgments, switching paradigm) cognitive task.

Main Results:

• Shared specific features between trained and assessment tasks were the strongest predictors of transfer for both training groups.
• For low-level training, overall feature overlap also predicted transfer.
• High-level training transfer was not predicted by overall feature overlap; pre-training correlations did not predict transfer patterns.

Conclusions:

• Specific feature overlap is crucial for cognitive transfer, regardless of training level.
• Low-level training benefits from broader feature similarity, suggesting different mechanisms for transfer.
• Findings elucidate the specificity of transfer and the critical role of task overlap in performance improvements.