本研究旨在探討室內閱讀環境下，照明光源之黑視素有效輻照度對使用者專注度、覺醒狀態及視覺舒適度之影響。研究中導入國際照明委員會推行的視錐響應基礎函數與CIE S 026光響應指標，以此改善傳統光參數量測之生理學偏差。利用基因演算法作為光譜調控與優化工具，維持照明之亮度、色彩等視覺刺激，設計出三組生理色度相似，但黑視素有效輻照度具備顯著差異之實驗情境，分別為0.67、0.86和1.01 W∙m^(-2)。 為了驗證照明情境之人因效應，本研究沿用實驗室既有之人因照明評估架構，結合心理學進行實驗設計。共招募30名研究參與者，在三種照明情境下進行靜態活動，包含智力測驗、d2 test注意力測試以及閉眼休息，同時使用NeuroSky腦波儀即時記錄生理數據。腦電波訊號經預處理剔除市電干擾與眨眼偽跡，並使用希爾伯特-黃轉換萃取訊號特徵。由α、β與γ頻段之瞬時功率，推導出歸一化之專注度特徵指標〖DSA〗_(3-100)、〖DSB〗_(3-100)和〖DSG〗_(3-100)。利用接收者操作特徵曲線(ROC)之曲線下面積(AUC)驗證指標之分類可靠性，並以重複量數變異數分析(Repeated Measures ANOVA)進行情境間之數據對比分析。 實驗結果顯示三種特徵指標在區分工作與休息狀態上皆具備良好的分類能力(AUC>0.7)，其中以〖DSG〗_(3-100)分類效果最佳。然而，在變異數分析中，不論工作、d2 test或休息狀態，三種專注度指標在不同照明情境間均未發現統計學上的顯著差異。中文智力測驗與d2 test之工作成效(完成率、正確率等)，在各情境間亦無顯著差異。此外，數據分析揭露練習效應與題目複雜度不均對分析準確度之干擾，成為未來實驗設計之重要參考。在主觀問卷回饋中，參與者評價情境一(低黑視素刺激)為偏冷白，情境三(高黑視素刺激)則偏暖白，證實本研究成功在色溫相近的情況下分離黑視素響應與光源顏色之強關聯性。 綜上所述，本研究結果指出在短期暴露的閱讀情境下，照明光源之黑視素有效輻照度差異不足以對使用者的大腦即時生理活動、行為績效及疲勞程度造成顯著干擾。研究成果可為未來人因照明與智慧動態照明設計提供實務設計參考，並為未來研究在延長暴露時間、擴大輻照度差異或優化腦波指標方面奠定科學基礎。;This study aims to investigate the effects of the melanopic irradiance of indoor reading illumination on users′ concentration， alertness， and visual comfort. Incorporating the cone fundamentals and CIE S 026 photoreceptor response metrics promulgated by the International Commission on Illumination， this research addresses the physiological deviations inherent in traditional optical parameter measurements. Using a genetic algorithm as a spectral regulation and optimization tool， while maintaining constant visual stimuli such as luminance and color， three experimental lighting environments with similar psychological chromaticity but distinct melanopic irradiances (0.67， 0.86， and 1.01 W∙m^(-2)) were designed. To validate the human-centric effects of these illumination scenarios， this study utilizes the existing human-centric lighting evaluation framework established by our laboratory， combined with psychological experimental design. A total of 30 research participants were recruited to perform static activities—including an intelligence test， a d2 test of attention， and eyes-closed rest—under the three lighting conditions， while their physiological data were recorded in real-time using a NeuroSky EEG device. The EEG signals were preprocessed to eliminate power-line interference and eyeblink artifacts， and the Hilbert-Huang Transform (HHT) was employed to extract signal features. Normalized concentration feature indicators， namely 〖DSA〗_(3-100)， 〖DSB〗_(3-100)， and 〖DSG〗_(3-100)， were derived from the instantaneous power of the α， β， and γ bands. The classification reliability of these indicators was validated using the Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC)， and Repeated Measures ANOVA was conducted for cross-conditional data comparative analysis. The experimental results indicate that all three feature indicators demonstrate robust binary classification performance (AUC > 0.7) between the "work" and "rest" states， 〖DSG〗_(3-100) yielding the highest accuracy. However， within the Analysis of Variance， no statistically significant differences were observed among the different lighting conditions for any of the three concentration indicators during work， d2 testing， or rest states. Similarly， task performance metrics (such as completion rate and accuracy rate) for both the Chinese intelligence test and the d2 test showed no significant variations across the conditions. Furthermore， the data analysis revealed that practice effects and imbalances in task complexity interfered with analytical precision， serving as critical methodological insights for future experimental designs. Regarding subjective questionnaire feedback， participants rated Condition 1 (low melanopic stimulus) as cooler white and Condition 3 (high melanopic stimulus) as warmer white， confirming that the spectral design successfully decoupled the melanopic response from visual color appearance under similar correlated color temperatures. In conclusion， the findings suggest that under short-term exposure within a reading context， variations in the melanopic irradiance of the illumination source do not induce significant immediate alterations in users′ cortical electrophysiological activities， behavioral performance， or fatigue levels. This study provides empirical design references for future human-centric lighting and dynamic intelligent illumination design， and establishes a scientific foundation for future research exploring prolonged exposure， expanded melanopic variations， or refined EEG feature indicators.
