Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A comparative study of sensory encoding models for human navigation in virtual reality

Tangyao Li, Qiyuan Zhan, Yitong Zhu, Bojing Hou, Yuyang Wang

TL;DR

This work investigates how sensory encoding underpins human navigation in virtual reality by comparing three computational schemes—Bayesian Efficient Coding (BEC), Fitness Maximizing Code (FMC), and Linear Nonlinear Poisson (LNP)—using a VR navigation dataset with physiological signals. It finds that BEC generally yields the most accurate predictions of navigation behavior when physiological responses are included, while FMC can improve predictions at small error penalties, and LNP is less reliable due to sensitivity to parameter choices. The results support Bayesian Efficient Coding as a robust framework for modeling VR navigation and associated physiological states, offering insight into how sensory encoding shapes experience and cybersickness. These findings have practical implications for designing VR interfaces and interactions that mitigate discomfort and cognitive load during navigation.

Abstract

In virtual reality applications, users often navigate through virtual environments, but the issue of physiological responses, such as cybersickness, fatigue, and cognitive workload, can disrupt or even halt these activities. Despite its impact, the underlying mechanisms of how the sensory system encodes information in VR remain unclear. In this study, we compare three sensory encoding models, Bayesian Efficient Coding, Fitness Maximizing Coding, and the Linear Nonlinear Poisson model, regarding their ability to simulate human navigation behavior in VR. By incorporating the factor of physiological responses into the models, we find that the Bayesian Efficient Coding model generally outperforms the others. Furthermore, the Fitness Maximizing Code framework provides more accurate estimates when the error penalty is small. Our results suggest that the Bayesian Efficient Coding framework offers superior predictions in most scenarios, providing a better understanding of human navigation behavior in VR environments.

A comparative study of sensory encoding models for human navigation in virtual reality

TL;DR

This work investigates how sensory encoding underpins human navigation in virtual reality by comparing three computational schemes—Bayesian Efficient Coding (BEC), Fitness Maximizing Code (FMC), and Linear Nonlinear Poisson (LNP)—using a VR navigation dataset with physiological signals. It finds that BEC generally yields the most accurate predictions of navigation behavior when physiological responses are included, while FMC can improve predictions at small error penalties, and LNP is less reliable due to sensitivity to parameter choices. The results support Bayesian Efficient Coding as a robust framework for modeling VR navigation and associated physiological states, offering insight into how sensory encoding shapes experience and cybersickness. These findings have practical implications for designing VR interfaces and interactions that mitigate discomfort and cognitive load during navigation.

Abstract

In virtual reality applications, users often navigate through virtual environments, but the issue of physiological responses, such as cybersickness, fatigue, and cognitive workload, can disrupt or even halt these activities. Despite its impact, the underlying mechanisms of how the sensory system encodes information in VR remain unclear. In this study, we compare three sensory encoding models, Bayesian Efficient Coding, Fitness Maximizing Coding, and the Linear Nonlinear Poisson model, regarding their ability to simulate human navigation behavior in VR. By incorporating the factor of physiological responses into the models, we find that the Bayesian Efficient Coding model generally outperforms the others. Furthermore, the Fitness Maximizing Code framework provides more accurate estimates when the error penalty is small. Our results suggest that the Bayesian Efficient Coding framework offers superior predictions in most scenarios, providing a better understanding of human navigation behavior in VR environments.
Paper Structure (18 sections, 4 equations, 4 figures, 2 tables)

This paper contains 18 sections, 4 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Human Behavior in VR
  4. Physiological Responses in VR
  5. Sensory Encoding Theories
  6. Methods
  7. Bayesian Efficient Coding
  8. Fitness Maximizing Code
  9. Linear Nonlinear Poisson Model
  10. Data Processing
  11. Results
  12. BEC Model
  13. BEC with FMC loss
  14. LNP Model
  15. Discussion
  16. ...and 3 more sections

Figures (4)

  • Figure 1: The virtual environment in which the participant carries out the navigation task. The participant navigated along a highlighted path in the virtual environment, which guided their movements throughout the task.
  • Figure 2: Cumulative distribution function of the stimulus and the optimized large monopolar cell response (with $p$ = 0.10 to 2.00) for a single participant using the BEC model. The red dashed curve represents the Infomax response derived from the stimulus cumulative distribution function (shown in grey). The blue dots represent the LMC response.
  • Figure 3: Cumulative distribution function of the stimulus and the optimized large monopolar cell response (with $p$ = 0.10 to 2.00) for a single participant using the BEC model with the FMC loss function. The red dashed curve represents the Infomax response derived from the stimulus cumulative distribution function (shown in grey). The blue dots represent the LMC response.
  • Figure 4: Empirical data and nonlinearity for one participant using the LNP model with different $p$ values. The black dashed curve indicates the Poisson-Gaussian process fit of the nonlinear rate function, while the red dashed curve represents the Infomax response.