Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Oct 11;18(10):e0284127.
doi: 10.1371/journal.pone.0284127. eCollection 2023.

Working memory and attention in choice

Aldo Rustichini  1 , Philippe Domenech  2 , Claudia Civai  1   3 , Colin G DeYoung  4
Affiliations

Working memory and attention in choice

Aldo Rustichini et al. PLoS One. .

Abstract

We study the role of attention and working memory in choices where options are presented sequentially rather than simultaneously. We build a model where a costly attention effort is chosen, which can vary over time. Evidence is accumulated proportionally to this effort and the utility of the reward. Crucially, the evidence accumulated decays over time. Optimal attention allocation maximizes expected utility from final choice; the optimal solution takes the decay into account, so attention is preferentially devoted to later times; but convexity of the flow attention cost prevents it from being concentrated near the end. We test this model with a choice experiment where participants observe sequentially two options. In our data the option presented first is, everything else being equal, significantly less likely to be chosen. This recency effect has a natural explanation with appropriate parameter values in our model of leaky evidence accumulation, where the decline is stronger for the option observed first. Analysis of choice, response time and brain imaging data provide support for the model. Working memory plays an essential role. The recency bias is stronger for participants with weaker performance in working memory tasks. Also activity in parietal areas, coding the stored value in working, declines over time as predicted.

PubMed Disclaimer

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Fig 1
Fig 1. Distribution of RT.
Raw response time reported here; the horizontal scale is in ms. For a log-transformation.
Fig 2
Fig 2. Working memory and memory of value.
Horizontal axis: d-prime statistic. Vertical axis: estimated participant-by-participant coefficient in the probit regression of the lag on the probability of choice. Lowess and scatter-plot.
See this image and copyright information in PMC

References

    1. Cover T, Hellman M. The two-armed-bandit problem with time-invariant finite memory. IEEE Transactions on Information Theory. 1970. Mar;16(2):185–95.
    1. Wilson A. Bounded memory and biases in information processing. Econometrica. 2014. Nov;82(6):2257–94.
    1. Glautier S. Recency and primacy in causal judgments: Effects of probe question and context switch on latent inhibition and extinction. Memory & cognition. 2008. Sep;36(6):1087–93. doi: 10.3758/MC.36.6.1087 - DOI - PubMed
    1. Kable JW, Glimcher PW. The neural correlates of subjective value during intertemporal choice. Nature neuroscience. 2007. Dec;10(12):1625–33. doi: 10.1038/nn2007 - DOI - PMC - PubMed
    1. Ballard K, Knutson B. Dissociable neural representations of future reward magnitude and delay during temporal discounting. Neuroimage. 2009. Mar 1;45(1):143–50. doi: 10.1016/j.neuroimage.2008.11.004 - DOI - PMC - PubMed

Publication types