Chang Eunhee, Billinghurst Mark, Yoo Byounghyun
Empathic Computing Laboratory, University of South Australia, Mawson Lakes, SA Australia.
Center for Artificial Intelligence, Korea Institute of Science and Technology, 5 Hwarangro14-gil Seongbuk-gu, Seoul, 02792 South Korea.
Virtual Real. 2023 Apr 12:1-25. doi: 10.1007/s10055-023-00795-y.
Virtual reality (VR) experiences can cause a range of negative symptoms such as nausea, disorientation, and oculomotor discomfort, which is collectively called cybersickness. Previous studies have attempted to develop a reliable measure for detecting cybersickness instead of using questionnaires, and electroencephalogram (EEG) has been regarded as one of the possible alternatives. However, despite the increasing interest, little is known about which brain activities are consistently associated with cybersickness and what types of methods should be adopted for measuring discomfort through brain activity. We conducted a scoping review of 33 experimental studies in cybersickness and EEG found through database searches and screening. To understand these studies, we organized the pipeline of EEG analysis into four steps (preprocessing, feature extraction, feature selection, classification) and surveyed the characteristics of each step. The results showed that most studies performed frequency or time-frequency analysis for EEG feature extraction. A part of the studies applied a classification model to predict cybersickness indicating an accuracy between 79 and 100%. These studies tended to use HMD-based VR with a portable EEG headset for measuring brain activity. Most VR content shown was scenic views such as driving or navigating a road, and the age of participants was limited to people in their 20 s. This scoping review contributes to presenting an overview of cybersickness-related EEG research and establishing directions for future work.
The online version contains supplementary material available at 10.1007/s10055-023-00795-y.
虚拟现实（VR）体验会引发一系列负面症状，如恶心、迷失方向和动眼不适，这些症状统称为网络晕动症。以往的研究试图开发一种可靠的检测网络晕动症的方法，而非使用问卷调查，脑电图（EEG）被视为可能的替代方法之一。然而，尽管关注度不断提高，但对于哪些大脑活动与网络晕动症始终相关，以及应采用何种类型的方法通过大脑活动来测量不适感，人们知之甚少。我们通过数据库检索和筛选，对33项关于网络晕动症和脑电图的实验研究进行了范围综述。为了理解这些研究，我们将脑电图分析流程组织为四个步骤（预处理、特征提取、特征选择、分类），并考察了每个步骤的特点。结果表明，大多数研究对脑电图进行频率或时频分析以提取特征。部分研究应用分类模型预测网络晕动症，准确率在79%至100%之间。这些研究倾向于使用基于头戴式显示器（HMD）的虚拟现实技术，并配备便携式脑电图耳机来测量大脑活动。所展示的大多数VR内容是驾车或道路导航等场景视图，参与者年龄限于20多岁的人群。本范围综述有助于概述与网络晕动症相关的脑电图研究，并为未来工作确立方向。
在线版本包含可在10.1007/s10055-023-00795-y获取的补充材料。
