TY - JOUR
T1 - Updating, Fast and Slow
T2 - Items, but Not Item-Context Bindings, are Quickly Updated Into Working Memory as Part of Response Selection
AU - Kessler, Yoav
AU - Zilberman, Nitzan
AU - Kvitelashvili, Shalva
N1 - Publisher Copyright:
© 2023 The Author(s).
PY - 2023/1/1
Y1 - 2023/1/1
N2 - It is commonly held that attending to items facilitates their encoding into working memory (WM). This implies that the content of WM is updated with new input as a consequence of directing attention to it. On the other hand, abundant research shows that WM updating is rather slow and effortful, suggesting that shielding WM representation against incoming input, rather than its updating, is the default. To resolve this discrepancy, we suggest that while updating item-to-context associations is costly, updating a single item is fast and is automatically carried out as part of directing attention to items, for example as part of response selection. Participants performed a choice-RT task, in which stimuli appeared within frames, and needed to update their WM with the most recent red item that appeared in each frame. The need for updating was manipulated, so that some trials required updating and others did not. Experiment 1 (N = 25) showed that updating was slower than not updating with a set-size of two items, that required item-context binding, but faster when the set-size only involved one item. Experiment 2 (N = 28) replicated this finding. Experiment 3 (N = 20) showed that the slower no-update RTs are due to the removal of erroneously updated information. In contrast to previous findings, these results suggest that updating can be effortless and obligatory.
AB - It is commonly held that attending to items facilitates their encoding into working memory (WM). This implies that the content of WM is updated with new input as a consequence of directing attention to it. On the other hand, abundant research shows that WM updating is rather slow and effortful, suggesting that shielding WM representation against incoming input, rather than its updating, is the default. To resolve this discrepancy, we suggest that while updating item-to-context associations is costly, updating a single item is fast and is automatically carried out as part of directing attention to items, for example as part of response selection. Participants performed a choice-RT task, in which stimuli appeared within frames, and needed to update their WM with the most recent red item that appeared in each frame. The need for updating was manipulated, so that some trials required updating and others did not. Experiment 1 (N = 25) showed that updating was slower than not updating with a set-size of two items, that required item-context binding, but faster when the set-size only involved one item. Experiment 2 (N = 28) replicated this finding. Experiment 3 (N = 20) showed that the slower no-update RTs are due to the removal of erroneously updated information. In contrast to previous findings, these results suggest that updating can be effortless and obligatory.
KW - Attention
KW - Executive functions
KW - Working memory
UR - http://www.scopus.com/inward/record.url?scp=85170418803&partnerID=8YFLogxK
U2 - 10.5334/joc.257
DO - 10.5334/joc.257
M3 - Article
C2 - 36721798
AN - SCOPUS:85170418803
SN - 2514-4820
VL - 6
JO - Journal of cognition
JF - Journal of cognition
IS - 1
M1 - 11
ER -