Balancing Flexibility and Interference in Working Memory

By Timothy J. Buschman. 2021

43 citations recorded by the source corpus.

Working memory is central to cognition, flexibly holding the variety of thoughts needed for complex behavior. Yet, despite its importance, working memory has a severely limited capacity, holding only 3–4 items at once. Here, I review experimental and computational evidence that the flexibility and limited capacity of working memory reflect the same underlying neural mechanism. I argue working memory relies on interactions between high-dimensional, integrative, representations in prefrontal cortex and structured representations in sensory cortex. Together, these interactions allow working memory to flexibly maintain arbitrary representations. However, the distributed nature of working memory comes at the cost of causing interference between items in memory, resulting in a limited capacity. Finally, I discuss several mechanisms used by the brain to reduce interference and maximize the effective capacity of working memory.

Subfields: Cognitive Neuroscience

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cognitive Neuroscience

1. Qian Li
2. Yi Zhong
3. Liyuan Wang
4. Xuchen Zhang

Citations in the corpus

• Genetic dissection of mutual interference between two consecutive learning tasks in Drosophila (2022, background): “suggesting that there is a trade-off between flexibility and interference resistance 189 (Buschman, 2021).”
• Human intracranial correlates of dynamic coding in auditory working memory (2026, unknown): “A recent interpretation combines these two models for WM maintenance and argues that the underlying mechanism of WM maintenance consists of interactions between high-dimensional, integrative, representations in prefrontal cortex and structured representations in sensory cortex (Buschman, 2021).”
• Radical flexibility of neural representation in frontoparietal cortex and the challenge of linking it to behaviour (2024, unknown): “For example, in computational models where PFC neurons are randomly connected to sensory neurons, PFC responses represent arbitrary combinations of stimuli [11,12].”
• Back in time: Reversed speech as a window into cognitive language processing—A systematic review (2026, unknown): “A proposed neural mechanism for these network-level interactions are cortical oscillations and the cross-coupling between them (Aktürk et al., 2022; Riddle et al., 2020, 2021).”
• Non-Target Suppression Supports the Formation of Representational Prioritization Under High Working Memory Load (2025, unknown): “Given the flexible nature of working memory representations [30], such coupling may also support cognitive flexibility via dynamic reconfiguration of attentional priorities [31].”