Biomarkers of Immersion in Virtual Reality Based on Features Extracted from the EEG Signals: A Machine Learning...

Area of Science:

• Neuroscience
• Computer Science
• Human-Computer Interaction

Background:

• Virtual reality (VR) offers immersive training environments, but objective immersion measurement remains a challenge.
• Current immersion evaluation relies on subjective post-task questionnaires, limiting real-time feedback.
• Accurate immersion metrics are vital for optimizing VR training effectiveness and skill transfer.

Purpose of the Study:

• To develop an objective method for measuring immersion levels during virtual reality (VR) training.
• To investigate the use of electroencephalography (EEG) and machine learning for real-time immersion detection.
• To differentiate between various immersion states based on task difficulty.

Main Methods:

• Collected electroencephalography (EEG) data from 14 participants engaged in VR jigsaw puzzles.
• Utilized machine learning algorithms to analyze EEG data during different task difficulty levels (easy, hard) and baseline conditions.
• Assessed the accuracy of machine learning models in classifying distinct immersion states.

Main Results:

• Machine learning models achieved high accuracy in classifying EEG data.
• Distinguished between easy vs. hard difficulty states with 86% accuracy.
• Differentiated between baseline and VR states with 97% accuracy.

Conclusions:

• EEG and machine learning can provide objective, real-time measures of immersion in VR.
• This approach enables the identification of robust biomarkers for immersion.
• Potential to dynamically adjust VR training difficulty to optimize user engagement and learning outcomes.