Attractor dynamics of working memory explain a concurrent evolution of stimulus-specific and decision-consistent biases in visual estimation.

Journal: Neuron
Published Date:

Abstract

Sensory evidence tends to be fleeting, often unavailable when we categorize or estimate world features. To overcome this, our brains sustain sensory information in working memory (WM). Although keeping that information accurate while acting on it is vital, humans display two canonical biases: estimates are biased toward a few stimuli ("stimulus-specific bias") and prior decisions ("decision-consistent bias"). Integrative-especially neural mechanistic-accounts of these biases remain scarce. Here, we identify drift dynamics toward discrete attractors as a common source of both biases in orientation estimation, with decisions further steering memory states. Behavior and neuroimaging data reveal how these biases co-evolve through the decision-steered attractor dynamics. Task-optimized recurrent neural networks suggest neural mechanisms that enable categorical decisions to emerge from WM for continuous stimuli while updating their trajectory, warping decision-consistent biases under stimulus-specific drift.

Authors

  • Hyunwoo Gu
    Department of Brain and Cognitive Sciences, Seoul National University, 1 Gwanak-ro, Seoul 08826, Republic of Korea; Department of Psychology, Stanford University, Stanford, CA 94305, USA; Wu Tsai Neurosciences Institute, Stanford University, Stanford, CA 94305, USA.
  • Joonwon Lee
    Department of Neurology, Inje University College of Medicine, Inje University Haeundae Paik Hospital, Busan, Republic of Korea.
  • Sungje Kim
    Department of Brain and Cognitive Sciences, Seoul National University, 1 Gwanak-ro, Seoul 08826, Republic of Korea.
  • Jaeseob Lim
    Department of Brain and Cognitive Sciences, Seoul National University, 1 Gwanak-ro, Seoul 08826, Republic of Korea.
  • Hyang-Jung Lee
    Department of Brain and Cognitive Sciences, Seoul National University, 1 Gwanak-ro, Seoul 08826, Republic of Korea.
  • Heeseung Lee
    Department of Brain and Cognitive Sciences, Seoul National University, 1 Gwanak-ro, Seoul 08826, Republic of Korea; Department of Psychological and Brain Sciences, Dartmouth College, Hanover, NH 03755, USA.
  • Min Jin Choe
    Department of Brain and Cognitive Sciences, Seoul National University, 1 Gwanak-ro, Seoul 08826, Republic of Korea.
  • Dong-Gyu Yoo
    Department of Brain and Cognitive Sciences, Seoul National University, 1 Gwanak-ro, Seoul 08826, Republic of Korea.
  • Jun Hwan Joshua Ryu
    Department of Psychology, Stanford University, Stanford, CA 94305, USA; Wu Tsai Neurosciences Institute, Stanford University, Stanford, CA 94305, USA.
  • Sukbin Lim
    Shanghai Frontiers Science Center of Artificial Intelligence and Deep Learning, NYU Shanghai, 567 West Yangsi Road, Shanghai 200126, P.R. China; Neural Science, NYU Shanghai, 567 West Yangsi Road, Shanghai 200126, P.R. China; NYU-ECNU Institute of Brain and Cognitive Science, NYU Shanghai, 3663 Zhongshan Road North, Shanghai 200062, P.R. China. Electronic address: sukbin.lim@nyu.edu.
  • Sang-Hun Lee
    Division of Gynaecologic Oncology, Department of Obstetrics and Gynaecology, University of Ulsan College of Medicine, Ulsan University Hospital, Ulsan, Republic of Korea.

Keywords

No keywords available for this article.