# Neural decoding of working memory > **NIH NIH R01** · NEW YORK UNIVERSITY · 2020 · $388,592 ## Abstract Despite that many failures of high-level cognition are due to the limited resources that support working memory (WM), we know almost nothing about the neural mechanisms underlying these WM limitations, nor the strategies employed to mitigate the limits of our memory. This gap in our knowledge is a critical problem because a host of psychiatric and neurologic disorders stems from a primary WM dysfunction. Our long-term goal is to understand the mechanisms by which WM representations are limited and how these limitations can be mitigated and remediated. Utilizing Bayesian theory, our overall hypothesis is that noisy population dynamics encode a probability distribution over WM stimulus dimensions, where a greater width in this distribution leads to less certainty about a remembered stimulus. The central aim of the project is to understand the role of uncertainty in the neural encoding of WM representations, including how neural uncertainty limits WM precision, how strategic use of prioritization improves the quality of WM, and how population activity in frontoparietal and visual cortex differentially impact the quality of WM. The rationale for the proposed research is that, as we better understand the neural mechanisms of WM, a strong theoretical framework will emerge within which strategies for understanding and treating cognitive dysfunction will emerge. We test our central hypothesis by pursuing three: With three specific aims, we will test the hypotheses that 1) neural populations encode behaviorally relevant representations of WM uncertainty; 2) sculpting population activity within topographic maps to favor prioritized locations improves the quality of WM representations; and 3) control signals in association cortex, in the form of persistent activity, affect the quality of spatial WM representations in visual cortex. Strong preliminary data demonstrate the feasibility of proposed work as well as initial support for the hypotheses. Under Aim 1, behavioral and modeling data demonstrated that humans use representations of uncertainty and patterns of fMRI activity in retinotopic areas in visual cortex were used to construct generative models of spatial WM that allowed for the estimation of memory uncertainty in neural populations. Under Aim 2, WM resource limitations were overcome by prioritizing some memories, which improved WM by reducing error and uncertainty in the population activities in visual maps where these representations are stored. Under Aim 3, the strength of neural activity during retention intervals in prefrontal and parietal cortex predicted the quality of neural representations of memorized locations decoded in early visual cortex. Overall, the proposed work will generate data needed to test how neural populations encode representations of WM. The approach is innovative because it combines neural and computational modeling to directly test WM theories within a test bed of well-defined topographically organized populations. The p... ## Key facts - **NIH application ID:** 9948656 - **Project number:** 5R01EY027925-04 - **Recipient organization:** NEW YORK UNIVERSITY - **Principal Investigator:** CLAYTON E CURTIS - **Activity code:** R01 (R01, R21, SBIR, etc.) - **Funding institute:** NIH - **Fiscal year:** 2020 - **Award amount:** $388,592 - **Award type:** 5 - **Project period:** 2017-09-01 → 2023-05-31 ## Primary source NIH RePORTER: https://reporter.nih.gov/project-details/9948656 ## Citation > US National Institutes of Health, RePORTER application 9948656, Neural decoding of working memory (5R01EY027925-04). Retrieved via AI Analytics 2026-05-24 from https://api.ai-analytics.org/grant/nih/9948656. Licensed CC0. --- *[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*