学习是人类最核心的活动之一，通常包括接收、理解和整合外部信息的过程。这一过程与大脑的认知处理能力密切相关，受人类认知结构特征的制约。由于心理资源(包括认知资源和注意资源)的有限性，人们在执行任务时只能处理有限数量的信息。专注度是衡量学生在任务中投入注意力处理资源的程度。教育心理学使用专注度作为评估注意力水平的指标，因其能够反映学生是否真正投入学习，以及学习效率的高低。
本研究通过虚拟现实(VR)视频评估情绪唤醒对学习专注度的影响，并采用改良持续性操作测验任务，评估生理信号模式识别对专注度测量的有效性。本研究共招募了12名受试者，向他们展示了4个不同情感倾向的视频，观看每个视频后，被试需填写情绪量表，完成持续性操作测验任务，本研究同时收集被试脑电信号(EEG)和光电容积脉搏波信号(PPG)，任务结束被试填写心流体验量表。
研究分为两个部分：
研究一分析情绪效价、唤醒度对专注度和心流体验的影响，以及情绪、专注度和心流体验之间的关系。结果显示，VR视频能有效诱发被试的积极情绪和消极情绪，高唤醒和低唤醒情绪。女性在正性情绪下，高唤醒的反应时显著短于低唤醒的反应时，即高唤醒情绪的专注度高于低唤醒专注度。男性在高唤醒情绪下，负性情绪的反应时显著短于正性情绪的反应时，即负性情绪的专注度高于正性情绪。心流体验和专注度高低无显著相关。
研究二提取生理信号特征，探讨使用生理模式客观分析专注度。研究结果表明，在高专注和低专注条件下的脑电信号、心率变异性存在显著差异。专注度较高时，脑电信号α波段在C3通道功率降低，β波段在F7通道、POZ通道的功率升高，γ波段在顶叶区域和前额叶区域功率升高，β/(α+θ)的比值在顶叶区域和前额叶区域升高，心率变异性高频成分降低，非常低频成分升高。
研究结果发现情绪的效价和唤醒度专注度存在影响而且男女存在差异。未来有望通过捕捉学生的多模态数据进行跨媒体计算，可以量化学生的专注状态，为学习分析中的状态评估提供有效手段。

Learning is one of the most fundamental activities for humans, typically encompassing the processes of receiving, comprehending, and integrating external information. This process is closely related to the cognitive processing capacity of the brain and is constrained by the characteristics of human cognitive structures. Due to the limited nature of psychological resources, including cognitive and attentional resources, individuals can only process a limited amount of information when performing tasks. Attentional engagement is a measure of the extent to which learners invest attentional processing resources in tasks, and thus, the attentional engagement of students is particularly crucial. Educational psychology research has found that attentional engagement serves as an indicator for assessing attention levels, and its importance is manifested in its ability to reflect whether students are truly engaged in learning and the efficiency of their learning.
In this study, we examined the impact of emotional arousal on learning attentional engagement using virtual reality (VR) videos and employed a sustained attention to response task to assess the effectiveness of physiological signal pattern recognition for measuring attentional engagement. We recruited 12 participants and exposed them to four VR videos with different emotional valences and arousal. Following the video presentations, participants completed an emotion questionnaire, the sustained attention to response task, and provided electroencephalography (EEG) and photoplethysmography (PPG) data. Upon task completion, a flow experience questionnaire was collected.
The study was divided into two parts:
Analyzing the influence of emotional valence and arousal on attentional engagement and flow experience, as well as the relationships among emotion, attentional engagement, and flow experience. The results showed that the VR videos
ABSTRACT
III
effectively elicited positive and negative emotions, as well as high and low arousal emotions in participants. In females, reaction times were shorter under high arousal positive emotions than under low arousal positive emotions, indicating significantly higher attentional engagement during high arousal states. In males, reaction times were shorter for negative emotions than for positive emotions under high arousal conditions, indicating significantly higher attentional engagement for negative emotions. No significant correlation was found between flow experience and attentional engagement levels.
Extraction of physiological signal features, explores whether physiological patterns can be objectively analyzed for attention levels. The research results indicate that there are significant differences in heart rate variability and brainwave signals under high and low attention conditions. When the attention level is higher, the high-frequency component of heart rate variability decreases, the very low frequency (VLF[abs]) increases, the power of alpha wave in the C3 channel EEG decreases, the power of beta wave in the F7 channel and POZ channel EEG increases, and the power of gamma wave as well as the ratio of β/(α+θ) increases in the parietal and frontal lobe areas.
This study found that emotional valence and arousal have different effects on attentional engagement, and attentional engagement exhibits distinguishable physiological patterns. In the future, it is hoped that by capturing multimodal data of learners for cross-media computation, the attentional state of learners can be quantified, providing an effective means for state assessment in learning analytics.