Technical Program

Paper Detail

Paper: PS-2A.3
Session: Poster Session 2A
Location: Symphony/Overture
Session Time: Friday, September 7, 17:15 - 19:15
Presentation Time:Friday, September 7, 17:15 - 19:15
Presentation: Poster
Publication: 2018 Conference on Cognitive Computational Neuroscience, 5-8 September 2018, Philadelphia, Pennsylvania
Paper Title: Signal power as the limited resource of working memory
Manuscript:  Click here to view manuscript
Authors: Thomas Christie, Paul Schrater, University of Minnesota, United States
Abstract: Working memory is famously capacity-limited, but the nature of the capacity is still a matter of debate. While recent research supports a continuous-resource model of working memory, these models do not account for the phenomenology of cognitive effort or the capacity’s apparent sensitivity to task demands and incentives. We suggest that describing working memory as a noisy information channel accounts for these phenomena. In this view, capacity is a function of continuous signal power that can be allocated across multiple signals. We used a computational simulation to infer signal power for subjects performing the N-back task, for N = 1,2,3. We found that as task difficulty increases, signal power per response is lower, but overall signal power increases, suggesting that subjective feelings of cognitive effort may be related to signal power. Our approach provides a parsimonious explanation for working memory effects that grounds the continuous resource view in an established theory while providing a plausible account of subjective effort and cognitive control.