This is the authors' abstract. We don't add key points for this paper.
Virtual Reality (VR) provides immersive and interactive experiences in healthcare, education, entertainment, and defense. However, cybersickness remains a major barrier to its widespread adoption, reducing user comfort and engagement. Early and accurate detection of cybersickness is critical to developing adaptive VR systems that ensure safety and improve usability. In this study, we propose a novel real-time cybersickness detection approach using Bidirectional Long Short-Term Memory (Bi-LSTM) networks trained on electroencephalography (EEG) signals. Power Spectral Density (PSD) and Signal Magnitude Area (SMA) features were extracted to capture frequency- and amplitude-related characteristics of cybersickness. EEG data were collected from six electrodes across frontal (F3–F4), prefrontal (FP1–FP2), and central parietal (P3–P4) regions during VR exposure. The proposed Bi-LSTM model achieved 95% classification accuracy, significantly outperforming baseline methods. Results indicate that cybersickness can be reliably detected with a compact EEG setup, supporting resource-efficient, real-time monitoring for adaptive VR environments.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Neelakandan et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: