本研究探討了基於虛擬實境（VR）的認知訓練系統在改善注意力不足過動症（ADHD）兒童的注意力和工作記憶方面的有效性。與傳統的治療方法相比，如藥物治療和認知行為治療（CBT），VR 系統提供了一個非侵入性且沉浸式的訓練環境，能夠更準確地捕捉認知和行為變化。該系統整合了多模態生理感測器，包括腦電圖（EEG）、眼動追蹤、頭部/身體運動和面部表情分析，實時提供與任務相關和自發性行為的數據。結果顯示，在任務表現和量表分數上均有顯著改善，並且多項生理信號顯示出與訓練相關的變化，表明注意力和抑制控制得到了更廣泛的改善。此外，使用多模態特徵的機器學習模型達到了93%的準確率，這表明這些生理信號可能作為治療反應的潛在生物標記。這些發現突顯了 VR 輔助、數據驅動的介入方法在 ADHD 治療中的巨大潛力，並為未來更具個性化和客觀性的評估框架奠定基礎。;This study investigates the effectiveness of a virtual reality (VR)-based cognitive training system in improving attention and working memory in children with Attention-Deficit/Hyperactivity Disorder (ADHD). Compared to traditional treatment methods, such as behavioral assessments and medication, the VR system provides a non-invasive and immersive training environment that more accurately captures cognitive and behavioral changes. The system integrates multimodal physiological sensors, including EEG, eye tracking, head/body movement, and facial expression analysis, to provide real-time data on both task-related and spontaneous behaviors. Results showed significant improvements in both task performance and scale scores, with several physiological signals revealing training-related changes, indicating broader improvements in attention and inhibitory control. Furthermore, machine learning models using multimodal features achieved an accuracy rate of 93%, suggesting that these physiological signals may serve as potential biomarkers for treatment response. These findings highlight the promising potential of VR-assisted, data-driven interventions for ADHD rehabilitation and pave the way for more personalized and objective evaluation frameworks in the future.
