本研究聚焦於人因照明中光譜設計於室內閱讀環境之應用，旨在探討於 6500K 高色溫條件下，照明光譜中「藍光占比」之變化對使用者專注表現與視覺感受的潛在影響。本研究模擬室內閱讀與辦公情境，透過精確控制光譜能量分布，評估不同藍光強度對使用者在作業過程中所產生的生、心理負荷差異。 在實驗光源建構階段，本研究導入基因演算法（Genetic Algorithm）進行參數最佳化。在嚴格控制色偏差（Duv）、相關色溫及桌面照度（1000 lx）之前提下，計算 14 通道 LED 燈箱之最佳光譜權重，並精確擬合出藍光（480±10 nm）占比分別為 4.14%（低）、5.14%（中）與 6.03%（高）之三種實驗情境。 本研究招募研究參與者進行重複量測之心理物理學實驗。參與者需於各照明情境下依序完成中文智力測驗、d2 注意力測驗及閉眼休息任務，並全程透過 NeuroSky 腦波儀同步擷取其腦電波訊號。在訊號處理方面，原始腦波經自動化眨眼偽影剔除後，接續運用經驗模態分解與希爾伯特-黃轉換提取時頻特徵。本研究結合機率密度函數與接收者操作特徵曲線下面積，將各頻帶功率（DSA3-45、DSB3-45、DSG3-45）量化為客觀專注度指標；並以情境體驗問卷作為視覺舒適度之主觀評估基準。 經重複量數變異數分析（RM-ANOVA）檢定，整體研究結果顯示，三種藍光占比情境在參與者群體平均之腦波專注度指標與客觀行為績效（智力測驗與 d2-test 之答題數、正確率）上皆未呈現統計顯著落差。在此實驗設定之藍光微調變化範圍內，並未對研究參與者之專注力造成顯著波動。 在主觀感受方面，問卷統計顯示各情境之視覺安全性極佳，藍光比例之微調並未增加參與者的視覺負擔，且獲得了「適中」至「滿意」之整體正向評價。綜合主、客觀數據，於本研究6500K 照明條件與短時間暴露下，適度之藍光比例微調能在不干擾整體持續性專注力的同時，穩定維持使用者之視覺舒適度，為未來高色溫智慧室內照明設計提供了深具價值的科學實證與光譜配置參考。;This study focuses on the application of spectral design in human-centric lighting within indoor reading environments， aiming to explore the potential impact of variations in the "blue light ratio" within the lighting spectrum on users′ con-centration performance and visual perception under a high color temperature con-dition of 6500K. By simulating indoor reading and office scenarios and precisely controlling the spectral power distribution， this study evaluates the differences in physiological and psychological workloads induced by varying blue light intensi-ties during task execution. In the light source construction phase， a Genetic Algorithm (GA) was intro-duced for parameter optimization. Under the strict constraints of controlling chromaticity difference (Duv)， correlated color temperature， and desktop illumi-nance (1000 lx)， the optimal spectral weights for a 14-channel LED light box were calculated. This allowed for the precise fitting of three experimental scenarios with blue light (480±10 nm) ratios of 4.14% (low)， 5.14% (medium)， and 6.03% (high)， respectively. Participants were recruited to undergo a repeated-measures psychophysical experiment. Under each lighting scenario， participants sequentially completed a Chinese intelligence test， the d2 Test of Attention， and a closed-eye resting task， while their electroencephalogram (EEG) signals were continuously recorded using a NeuroSky device. Regarding signal processing， after the automated removal of blink artifacts from the raw EEG data， Empirical Mode Decomposition (EMD) and the Hilbert-Huang Transform (HHT) were subsequently applied to extract time-frequency features. By combining Probability Density Functions (PDF) with the Area Under the Receiver Operating Characteristic Curve， this study quantified various band powers (DSA3-45， DSB3-45， DSG3-45) as objective indicators of con-centration; additionally， a situational experience questionnaire was utilized as the subjective assessment baseline for visual comfort. Examined via Repeated Measures Analysis of Variance (RM-ANOVA)， the overall results indicate that the three blue light ratio scenarios showed no statisti-cally significant differences in the group′s average EEG concentration indicators or objective behavioral performance (number of answered questions and accuracy rate in both the intelligence test and d2-test). Within the fine-tuned variation range of blue light set in this experiment， no significant fluctuations were caused in the participants′ concentration levels. In terms of subjective perception， questionnaire statistics indicated excellent visual safety across all scenarios. The fine-tuning of the blue light ratio did not in-crease the participants′ visual burden， and it received overall positive evaluations ranging from ′moderate′ to ′satisfied.′ Synthesizing both objective and subjective data， under the 6500K lighting conditions and short-term exposure of this study， moderate fine-tuning of the blue light ratio can stably maintain users′ visual com-fort without interfering with overall sustained attention. This provides valuable scientific evidence and spectral configuration references for future high-color-temperature smart indoor lighting design.
