Overflow of Working Memory from an Easily Accessible Active State
Ideriha, T.; Ushiyama, J.
Show abstract
Working memory is a system that realizes short-term memory retention and is essential to everyday activities. Recently, it has been suggested that working memory items can be maintained in both an active state with sustained neural activity and a latent state without any explicit neural activity. However, how easily we can retrieve memories in each state and how much information can be retained in the active state remains unknown. Here, we addressed these questions by adopting a novel experimental paradigm for measuring the reaction time required to recall a letter of the alphabet when presented with a corresponding color, based on memorized color/letter pairs. The results demonstrated that when participants retained only two pairs of items, they could recall both pairs similarly quickly. However, when there were two or more memorized pairs, the most recent or internally attended pairs were recalled more quickly than the other pairs. These results suggest that memories in the "active" state are accessed first and memories that have overflowed into the "latent" state are accessed afterward. Additionally, the capacity of the easily accessible active state is extremely limited (to approximately two chunks of information) compared with the traditionally considered working memory capacity (four to seven chunks). These results provide robust behavioral support for the co-existence of multiple states in working memory and underscore the need for a more detailed classification of the working memory capacity. SIGNIFICANCE STATEMENTWe sometimes cannot quickly retrieve what we are sure we remember. Regarding working memory retrieval, we observed a phenomenon that the most recently encoded memory was retrieved quickly whereas others were retrieved afterward, showing an "overflow" of memory items from an easily accessible active state. This observation has two critical implications: 1) when maintaining multiple items in working memory, not all items are functionally the same, and there is a dissociation between easily accessible active items and items that have overflowed to a latent state; 2) the number of items that can be maintained in the active state is two, which is remarkably smaller than the traditionally considered working memory capacity of four, necessitating the update of working memory models.
Matching journals
The top 9 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Interval between two sequential arrays determines their storage state in visual working memory 96%
- When Gaze-Pattern Similarity May Interfere With Future Memory 95%
- Microstructural organization of Superior Longitudinal Fasciculus and Cingulum Bundle support metacognition driven cognitive offloading 94%
Similar papers in this journal
Similar papers in this journal
- Behavioral investigation of allocentric and egocentric cognitive maps in human spatial memory 94%
- Relationships between individual differences in dual process and electrophysiological signatures of familiarity and recollection during retrieval 94%
- Emotion schema effects on associative memory differ across emotion categories at the behavioural, physiological and neural level 93%
Similar papers in this journal
- Encoding-linked pupil response is modulated by expected and unexpected novelty: Implications for memory formation and neurotransmission 96%
- Positive and Neutral Updating Reconsolidate Aversive Episodic Memories via Different Routes 95%
- Transcranial alternating current stimulation at theta frequency to the parietal cortex impairs associative, but not perceptual, memory encoding 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.