A Dynamic Bayesian Observer Model Reveals Origins of Bias in Visual Path Integration

Kaushik J. Lakshminarasimhan, Marina Petsalis, Hyeshin Park, Gregory C. DeAngelis, Xaq Pitkow, Dora Angelaki

Research output: Contribution to journalArticle

Abstract

Path integration is a strategy by which animals track their position by integrating their self-motion velocity. To identify the computational origins of bias in visual path integration, we asked human subjects to navigate in a virtual environment using optic flow and found that they generally traveled beyond the goal location. Such a behavior could stem from leaky integration of unbiased self-motion velocity estimates or from a prior expectation favoring slower speeds that causes velocity underestimation. Testing both alternatives using a probabilistic framework that maximizes expected reward, we found that subjects’ biases were better explained by a slow-speed prior than imperfect integration. When subjects integrate paths over long periods, this framework intriguingly predicts a distance-dependent bias reversal due to buildup of uncertainty, which we also confirmed experimentally. These results suggest that visual path integration in noisy environments is limited largely by biases in processing optic flow rather than by leaky integration. Humans are typically biased while navigating by integrating their self-motion. Using virtual reality and probabilistic modeling, Lakshminarasimhan et al. demonstrate that systematic errors in visual path integration originate from the combined influence of a slow-velocity prior and growing position uncertainty.

Original languageEnglish (US)
Pages (from-to)194-206.e5
JournalNeuron
Volume99
Issue number1
DOIs
StatePublished - Jul 11 2018

Fingerprint

Optic Flow
Uncertainty
Reward

Keywords

  • Bayesian model
  • leaky integration
  • optic flow-based navigation
  • path integration bias
  • virtual reality

ASJC Scopus subject areas

  • Neuroscience(all)

Cite this

Lakshminarasimhan, K. J., Petsalis, M., Park, H., DeAngelis, G. C., Pitkow, X., & Angelaki, D. (2018). A Dynamic Bayesian Observer Model Reveals Origins of Bias in Visual Path Integration. Neuron, 99(1), 194-206.e5. https://doi.org/10.1016/j.neuron.2018.05.040

A Dynamic Bayesian Observer Model Reveals Origins of Bias in Visual Path Integration. / Lakshminarasimhan, Kaushik J.; Petsalis, Marina; Park, Hyeshin; DeAngelis, Gregory C.; Pitkow, Xaq; Angelaki, Dora.

In: Neuron, Vol. 99, No. 1, 11.07.2018, p. 194-206.e5.

Research output: Contribution to journalArticle

Lakshminarasimhan, KJ, Petsalis, M, Park, H, DeAngelis, GC, Pitkow, X & Angelaki, D 2018, 'A Dynamic Bayesian Observer Model Reveals Origins of Bias in Visual Path Integration', Neuron, vol. 99, no. 1, pp. 194-206.e5. https://doi.org/10.1016/j.neuron.2018.05.040
Lakshminarasimhan KJ, Petsalis M, Park H, DeAngelis GC, Pitkow X, Angelaki D. A Dynamic Bayesian Observer Model Reveals Origins of Bias in Visual Path Integration. Neuron. 2018 Jul 11;99(1):194-206.e5. https://doi.org/10.1016/j.neuron.2018.05.040
Lakshminarasimhan, Kaushik J. ; Petsalis, Marina ; Park, Hyeshin ; DeAngelis, Gregory C. ; Pitkow, Xaq ; Angelaki, Dora. / A Dynamic Bayesian Observer Model Reveals Origins of Bias in Visual Path Integration. In: Neuron. 2018 ; Vol. 99, No. 1. pp. 194-206.e5.
@article{c48c2ff4ca84432ebd45385c829102c1,
title = "A Dynamic Bayesian Observer Model Reveals Origins of Bias in Visual Path Integration",
abstract = "Path integration is a strategy by which animals track their position by integrating their self-motion velocity. To identify the computational origins of bias in visual path integration, we asked human subjects to navigate in a virtual environment using optic flow and found that they generally traveled beyond the goal location. Such a behavior could stem from leaky integration of unbiased self-motion velocity estimates or from a prior expectation favoring slower speeds that causes velocity underestimation. Testing both alternatives using a probabilistic framework that maximizes expected reward, we found that subjects’ biases were better explained by a slow-speed prior than imperfect integration. When subjects integrate paths over long periods, this framework intriguingly predicts a distance-dependent bias reversal due to buildup of uncertainty, which we also confirmed experimentally. These results suggest that visual path integration in noisy environments is limited largely by biases in processing optic flow rather than by leaky integration. Humans are typically biased while navigating by integrating their self-motion. Using virtual reality and probabilistic modeling, Lakshminarasimhan et al. demonstrate that systematic errors in visual path integration originate from the combined influence of a slow-velocity prior and growing position uncertainty.",
keywords = "Bayesian model, leaky integration, optic flow-based navigation, path integration bias, virtual reality",
author = "Lakshminarasimhan, {Kaushik J.} and Marina Petsalis and Hyeshin Park and DeAngelis, {Gregory C.} and Xaq Pitkow and Dora Angelaki",
year = "2018",
month = "7",
day = "11",
doi = "10.1016/j.neuron.2018.05.040",
language = "English (US)",
volume = "99",
pages = "194--206.e5",
journal = "Neuron",
issn = "0896-6273",
publisher = "Cell Press",
number = "1",

}

TY - JOUR

T1 - A Dynamic Bayesian Observer Model Reveals Origins of Bias in Visual Path Integration

AU - Lakshminarasimhan, Kaushik J.

AU - Petsalis, Marina

AU - Park, Hyeshin

AU - DeAngelis, Gregory C.

AU - Pitkow, Xaq

AU - Angelaki, Dora

PY - 2018/7/11

Y1 - 2018/7/11

N2 - Path integration is a strategy by which animals track their position by integrating their self-motion velocity. To identify the computational origins of bias in visual path integration, we asked human subjects to navigate in a virtual environment using optic flow and found that they generally traveled beyond the goal location. Such a behavior could stem from leaky integration of unbiased self-motion velocity estimates or from a prior expectation favoring slower speeds that causes velocity underestimation. Testing both alternatives using a probabilistic framework that maximizes expected reward, we found that subjects’ biases were better explained by a slow-speed prior than imperfect integration. When subjects integrate paths over long periods, this framework intriguingly predicts a distance-dependent bias reversal due to buildup of uncertainty, which we also confirmed experimentally. These results suggest that visual path integration in noisy environments is limited largely by biases in processing optic flow rather than by leaky integration. Humans are typically biased while navigating by integrating their self-motion. Using virtual reality and probabilistic modeling, Lakshminarasimhan et al. demonstrate that systematic errors in visual path integration originate from the combined influence of a slow-velocity prior and growing position uncertainty.

AB - Path integration is a strategy by which animals track their position by integrating their self-motion velocity. To identify the computational origins of bias in visual path integration, we asked human subjects to navigate in a virtual environment using optic flow and found that they generally traveled beyond the goal location. Such a behavior could stem from leaky integration of unbiased self-motion velocity estimates or from a prior expectation favoring slower speeds that causes velocity underestimation. Testing both alternatives using a probabilistic framework that maximizes expected reward, we found that subjects’ biases were better explained by a slow-speed prior than imperfect integration. When subjects integrate paths over long periods, this framework intriguingly predicts a distance-dependent bias reversal due to buildup of uncertainty, which we also confirmed experimentally. These results suggest that visual path integration in noisy environments is limited largely by biases in processing optic flow rather than by leaky integration. Humans are typically biased while navigating by integrating their self-motion. Using virtual reality and probabilistic modeling, Lakshminarasimhan et al. demonstrate that systematic errors in visual path integration originate from the combined influence of a slow-velocity prior and growing position uncertainty.

KW - Bayesian model

KW - leaky integration

KW - optic flow-based navigation

KW - path integration bias

KW - virtual reality

UR - http://www.scopus.com/inward/record.url?scp=85048798138&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048798138&partnerID=8YFLogxK

U2 - 10.1016/j.neuron.2018.05.040

DO - 10.1016/j.neuron.2018.05.040

M3 - Article

VL - 99

SP - 194-206.e5

JO - Neuron

JF - Neuron

SN - 0896-6273

IS - 1

ER -