Scholarlink

<p>This study examines the relationship between subjective cognitive load assessments and objective neurophysiological indicators in virtual reality (VR) educational environments. We integrated three validated psychometric scales—NASA Task Load Index (NASA-TLX), Leppink Cognitive Load Scale, and Multidimensional Cognitive Load Scale for Virtual Environments (MCLSVE)—with continuous 32-channel EEG monitoring to identify objective markers of cognitive–affective states during VR art education tasks. 32 participants completed three cognitive load conditions (low, medium, high) in an immersive virtual art studio, yielding N = 134 observation units for Bayesian structural equation modeling (SEM) analysis. The Bayesian structural model demonstrated good fit with posterior R<sup>2</sup> = 0.713 (95% CI: 0.698–0.720). NASA-TLX emerged as the strongest predictor (β = 0.781), followed by Leppink (β = 1.014) and MCLSVE (β = 0.933). Neurophysiological comparisons revealed significant differences between high and low load conditions: theta/alpha ratio (d = 2.00), frontal theta power (d = 1.88), and posterior alpha suppression (d = 1.75). The multimodal approach outperformed single-modal methods by 18.2% (ΔR<sup>2</sup> = 0.111), with NASA-TLX contributing 45.2% unique variance. This framework provides methodological foundations for developing adaptive VR educational systems with cognitive state monitoring capabilities.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multimodal assessment of cognitive–affective states in virtual reality education: integrating psychometric scales with EEG biomarkers for cognitive load monitoring

  • 期刊论文

  • 发表日期：

    2026年3月13日

查看全文
  • Jian Teng,
  • Sukyoung Cho

摘要

This study examines the relationship between subjective cognitive load assessments and objective neurophysiological indicators in virtual reality (VR) educational environments. We integrated three validated psychometric scales—NASA Task Load Index (NASA-TLX), Leppink Cognitive Load Scale, and Multidimensional Cognitive Load Scale for Virtual Environments (MCLSVE)—with continuous 32-channel EEG monitoring to identify objective markers of cognitive–affective states during VR art education tasks. 32 participants completed three cognitive load conditions (low, medium, high) in an immersive virtual art studio, yielding N = 134 observation units for Bayesian structural equation modeling (SEM) analysis. The Bayesian structural model demonstrated good fit with posterior R2 = 0.713 (95% CI: 0.698–0.720). NASA-TLX emerged as the strongest predictor (β = 0.781), followed by Leppink (β = 1.014) and MCLSVE (β = 0.933). Neurophysiological comparisons revealed significant differences between high and low load conditions: theta/alpha ratio (d = 2.00), frontal theta power (d = 1.88), and posterior alpha suppression (d = 1.75). The multimodal approach outperformed single-modal methods by 18.2% (ΔR2 = 0.111), with NASA-TLX contributing 45.2% unique variance. This framework provides methodological foundations for developing adaptive VR educational systems with cognitive state monitoring capabilities.