Impact of Lighting Conditions on Emotional and Neural Responses of International Students in Cultural Exhibition Halls

Abstract

This study investigates how lighting conditions influence emotional and neural responses in a standardized, simulated museum environment. A multimodal evaluation framework combining subjective and objective measures was used. Thirty-two international students assessed their viewing experiences using 14 semantic differential descriptors, while real-time EEG signals were recorded via the EMOTIV EPOC X device. Spectral energy analyses of the α, β, and θ frequency bands were conducted, and a θα energy ratio combined with γ coefficients was used to model attention and comfort levels. The results indicated that high illuminance (300 lx) and high correlated color temperature (4000 K) significantly enhanced both attention and comfort. Art majors showed higher attention levels than engineering majors during short-term viewing. Among four regression models, the backpropagation (BP) neural network achieved the highest predictive accuracy (R² = 88.65%). These findings provide empirical support for designing culturally inclusive museum lighting and offer neuroscience-informed strategies for promoting the global dissemination of traditional Chinese culture, further supported by retrospective interview insights.

1. Introduction

In the era of globalization, the dissemination of traditional cultures plays a vital role in fostering mutual understanding and preserving intangible heritage. Among various mediums, visual presentation and spatial design—especially lighting—have proven critical in shaping how international audiences emotionally and cognitively engage with cultural artifacts [1]. In exhibitions of traditional Chinese art, lighting not only affects visibility and color perception, but also modulates how viewers experience symbolic meaning and cultural resonance embedded in artworks such as landscape paintings [2]. Meanwhile, advances in neuroaesthetics and affective neuroscience have provided researchers with tools like electroencephalography (EEG) to explore aesthetic and emotional responses at the neural level [3]. EEG has shown particular value in detecting real-time changes in brain activity associated with attention, affect, and visual cognition [4]. Studies have linked α band power to emotional engagement and θ activity to reflective or memory-related processing. Scuello et al. examined how different lighting environments influenced viewers’ emotional responses to museum exhibits [5]. Similarly, Li et al. explored EEG-based emotional responses to traditional Chinese ink paintings, finding that lighting parameters modulated participants’ emotional valence and neural activation patterns [6]. While previous research has explored lighting effects in exhibition contexts and EEG-based responses to art, few studies have specifically examined international students—a population at the intersection of cross-cultural engagement. Even fewer have investigated how lighting conditions affect their emotional and neural responses to traditional Chinese visual culture. This study addresses this gap by investigating how lighting parameters (illuminance and color temperature) influence emotional engagement and neural activity among international students during art viewing. Specifically, it aims to

(1)

Identify lighting conditions that enhance emotional engagement and cultural appreciation;

(2)

Explore EEG-based markers corresponding to positive affective and cognitive responses;

(3)

Construct predictive models that can guide lighting design in intercultural exhibition environments.

We hypothesize that higher illuminance and warmer color temperatures will be associated with increased α activity and heightened emotional valence, particularly among participants with an art background. By bridging lighting design, cognitive neuroscience, and cross-cultural aesthetics, this study contributes to the interdisciplinary understanding of how environmental design facilitates cultural communication. The findings also offer practical implications for exhibition curation, cultural diplomacy, and international education.

2. Literature Review

The intersection of environmental design, emotion, and cultural communication has received growing scholarly attention in recent years. Among environmental variables, lighting has been shown to significantly impact human perception, mood, and cognitive processes. Variations in brightness and color temperature can elicit different emotional responses, such as calmness, alertness, or warmth [1,7]. In the context of art appreciation, ambient lighting alters viewers’ aesthetic judgments and emotional engagement [8]. In Chinese cultural exhibitions, lighting has traditionally been used to enhance symbolic meanings. Peng et al. (2024) [9] and Siniscalco et al. (2021) [10] showed that warm lighting (approximately 2700–3000 K) enhanced feelings of nostalgia and tranquility when international viewers engaged with Chinese calligraphy. However, empirical studies that integrate physiological measurements, such as EEG, with emotional feedback in response to traditional Chinese artwork remain limited—especially in ecologically valid exhibition environments that reflect authentic viewing contexts. Most existing EEG studies focus on general lab-based stimuli, over-looking culturally meaningful environments and underrepresenting international participants. Furthermore, few studies have combined EEG analysis with semantic emotional evaluations, which constrains our understanding of how lighting modulates affective engagement across cultural contexts A (8–13 Hz) and θ (4–7 Hz) frequency bands are of particular interest. A suppression is often interpreted as increased attentional or emotional engagement, whereas θ activity is associated with memory encoding and deep reflection [11,12]. Studies such as Ren et al. (2024) [13] have used EEG to assess emotional valence during art viewing tasks, providing a new layer of insight beyond self-report measures. Despite these advances, relatively few studies have focused on cross-cultural audiences, such as international students, who bring unique cultural schemas and emotional frameworks to their interpretation of visual art. This population offers valuable insight into how lighting might mediate the emotional accessibility of traditional Chinese culture in international contexts.

In recent decades, exhibition design has evolved from static displays to immersive, multisensory environments that actively shape viewer perception and emotional response [14,15]. Among the various environmental factors, lighting design plays a key role, not only in enhancing visibility and preserving artifacts but also in shaping emotional and cultural responses [16]. Lighting conditions have been shown to significantly influence mood, cognitive function, and aesthetic judgments in art museums [17,18]. The psychological effects of lighting, particularly in cultural contexts, have been studied extensively. The foundational study by Kruithof (1941) established a relationship between illuminance and correlated color temperature (CCT), proposing that certain combinations of light parameters are more conducive to emotional comfort [19]. In the context of museum exhibitions, lighting is not merely functional; it also serves to evoke specific emotional atmospheres that reflect cultural and philosophical themes, especially in the presentation of traditional art [20]. Warm lighting (typically around 2700–3000 K) has been shown to elicit emotions such as warmth, nostalgia, and tranquility, which are often associated with traditional cultural values like those found in Chinese calligraphy [21,22]. However, most of the existing research on lighting has relied heavily on subjective assessments of emotional responses to different lighting conditions, often using adjective-based scales or surveys [23,24]. While these approaches provide useful insights, they are limited by biases in self-reporting. As a result, neurophysiological measures, such as EEG, have emerged as an objective tool for exploring emotional engagement and cognitive processing in response to various stimuli, including lighting [25]. EEG has been particularly useful in measuring α (8–13 Hz) and θ (4–7 Hz) brain waves, which are associated with attentional focus and reflective thinking, respectively [26,27]. A suppression is generally seen as an indicator of increased attentional engagement, while θ activity reflects memory encoding and deeper cognitive processing [28]. Studies integrating EEG with emotional feedback in response to visual stimuli, such as art exhibitions, are increasingly common. For instance, research has shown that the θ band is highly sensitive to emotional engagement, while α suppression often occurs in response to stimuli that evoke strong aesthetic or emotional reactions [29,30]. Despite the growing body of work in this area, there is a notable gap in studies focusing on cross-cultural audiences, such as international students, who bring diverse cultural frameworks to their interpretation of visual art and lighting [31]. This population is crucial for understanding how lighting design can facilitate the international dissemination of traditional Chinese culture, as their emotional responses may differ from those of local audiences [32]. Lighting design, particularly in the context of traditional Chinese culture, has been found to play a significant role in influencing how cultural symbols are perceived. The interplay between light intensity and CCT has been studied in the context of art perception, with findings suggesting that specific lighting combinations can enhance cultural accessibility and emotional resonance [33,34]. For instance, when international audiences are exposed to Chinese calligraphy under warm, soft lighting conditions, they report increased feelings of connection and understanding of the cultural symbolism embedded in the artwork [35,36]. Despite these findings, relatively few studies have employed EEG data to assess how lighting affects emotional and cognitive responses to traditional Chinese art. Furthermore, several studies have suggested that combining machine learning techniques with EEG signals can enhance the recognition of how different lighting environments affect individuals. Research shows that the accuracy of predicting lighting comfort using EEG signals reaches 77.0% with the naive Bayes (NB) classifier [37], and up to 98.2% with the support vector machine (SVM) classifier [38]. Deng et al. [39] compared the classification accuracy of random forest (RF) using FAI and artificial neural network (ANN) for job participation, and found that RF achieved an accuracy rate of 83.3% in the three-level classification. To improve the performance of the preferred illuminance prediction model, this study evaluated the performance of four classifiers: LSSVM, XGBoost, RF, and BP. This study aims to fill this gap by integrating EEG data, emotional feedback, and lighting parameters (CCT and illuminance) to evaluate how international students engage with Chinese calligraphy. This approach not only provides objective data on the neurological effects of lighting but also offers a more holistic view of the emotional and cognitive dimensions of cultural experience [40,41]. The empirical focus of this study is on how lighting conditions can optimize the international dissemination of traditional Chinese culture through emotional engagement. By leveraging EEG data and emotional feedback from international students, this research aims to provide actionable insights into lighting design principles that can enhance cultural appreciation and emotional accessibility, thereby supporting the global dissemination of Chinese heritage [42,43].

3. Materials and Methods

3.1. Subjects and Paintings

Thirty-two international students from art and engineering majors (16 males and 16 females) participated in this study. All participants passed the Ishihara test before the experiment. The participants were international students enrolled in Chinese universities, primarily originating from East and Southeast Asia (e.g., Vietnam, Indonesia, and South Korea), with a small subset from Central Asia and Africa. Although this regional distribution implies partial cultural clustering, the group nonetheless encompassed diverse linguistic and perceptual backgrounds, providing preliminary insights into culturally mediated responses to traditional Chinese visual art. The subjects had no color blindness, had normal visual function, and voluntarily participated. The stimulus used is a traditional Chinese landscape painting (1200 mm × 400 mm) created by the Art college. To avoid preference bias, each participant selected their favorite image from the three landscape paintings shown in Figure 1 before the experiment. The landscape paintings with a voting rate of 50% were used in the final test, as shown in Figure 1b.

3.2. Experimental Environment and Equipment

This experiment was conducted in the Optical Environment Evaluation Laboratory of the Institute of Photonics at Dalian Polytechnic University (as shown in Figure 2a), and the walls were all covered with black screens on all sides. In the design of lighting experiments, factors such as the viewpoint height, viewing distance, and the hanging height of the picture frame need to be comprehensively considered to determine parameters such as the appropriate projection direction and the uniformity of the picture illumination. In previous studies, we found that the most ideal visual area for painting is 45° horizontally and 28° vertically, and the distance between the viewpoint and the exhibit is usually between 1.5 h and 2 h (h is the canvas). The average adult eye height of 1600 mm was used as a reference [44] for lighting experiments (as shown in Figure 2a,b). And in order to simulate the suitable physical environment in a real museum, we set the temperature at 22 ± 1 °C and the humidity at 50 ± 5% [45] to improve the simulation environment.

In this experiment, based on the CIE museum lighting standards and referring to the definition of the comfort zone by Kruithof’s rule, four different combinations of illuminance and color temperature light environments were set up, as shown in Figure 1. The color temperature parameters (2700 K, 400 K) refer to the research data and the study of the Kruithof curve by Vienot et al. [24]; the illuminance parameters (150 lx, 300 lx) were set based on the visual load threshold proposed by Kozaki et al. [46], and were divided into four light environments (as shown in Figure 3c to achieve the effect of controlling variables. In order to restore the original colors of the painting to the greatest extent, LED track spotlights with good color rendering performance were selected as the experimental light source. In the dimming section, we adopted an intelligent control system to adjust the color temperature. The Sfim-300 remote spectrometer as shown in Figure 3b was mainly used for light environment testing to obtain the most intuitive lighting data. Before the experiment, the painting was evenly divided into 5 grids using the center distribution method as shown in Figure 3a, and the illuminance values at the center of each grid were tested using an illuminance meter as shown in Figure 3b according to Formulas (1) and (2). The specific formulas are as follows:

$$U_0={\displaystyle \frac{E_{min}}{E_{av}}}$$

(1)

$$E_{av}={\displaystyle \frac{E_1+E_2+E_3+E_4+E_5}{9}}$$

(2)

U₀ is the uniformity of illuminance, the central illuminance value is E_n, the minimum illuminance value is E_min, and E_av is the average illuminance value.

Figure 3 shows the measurement methods of stimuli in four light environments. It can be seen from

Table 1. Illumination of painting point.

Illumination (lx)	2700 k, 150 lx	2700 k, 300 lx	4000 k, 150 lx	4000 k, 300 lx
1	113.5	255.7	132.5	239.7
2	125.0	319.9	132.9	238.1
3	151.6	309.3	167.8	312.4
4	134.8	265.7	146.0	270.3
5	136.4	280.8	151.1	271.5
Illuminance uniformity	0.858158	0.893182	0.907161	0.893769

that the average illuminance of Environment 1 is 132.3 lx, that of Environment 2 is 286.3 lx, that of Environment 3 is 146.1 lx, and that of Environment 4 is 266.5 lx. In the four environments, the uniformity of illuminance is approximately 0.9 and the color rendering index is greater than 95. Therefore, the influence of these two items on this experiment is not discussed.

3.3. Program

Before the experiment, each subject selected one painting from three preselected options, and the painting with the highest number of votes was chosen as the experimental stimulus. As shown in Figure 4, we first introduced the experiment to each subject and conducted an Ishihara color vision test for them. After the test, we measured the illuminance and color temperature of four environments to ensure that other indicators remained unchanged. To minimize the influence of individual differences in light sensitivity and emotional responsiveness, the experiment adopted a within-subject design in which all participants experienced the four lighting conditions in randomized order. This approach enabled direct comparisons of each participant’s responses across conditions and effectively controlled for between-subject variability, thereby enhancing the reliability of the findings. Furthermore, individuals with known visual or neurological disorders were excluded during recruitment, and all participants were instructed to maintain an emotionally neutral state throughout the experiment. We placed the EEG tester on the subject’s head and began the EEG test. To maintain ecological validity and reduce procedural complexity, no dedicated baseline EEG or emotional assessment phase was implemented prior to stimulus onset. However, a 2 min rest interval was provided between each lighting condition to allow participants’ neural and emotional states to return to a relatively stable baseline. This approach helped mitigate potential carryover effects while ensuring consistency across conditions and minimizing participant fatigue. This strategy provided a practical balance between experimental control and participant comfort. The EEG examination lasted for 5 min. During the EEG test, we kept the ambient temperature constant and tried not to move the head. After the electroencephalogram test, subjective questionnaires were distributed and the subjects completed the subjective questionnaires within 2 min. After completing the subjective questionnaire survey, the subjects were transferred to the natural light environment and rested for 2 min before entering the next experimental environment. Each subject experienced all four lighting conditions. After the experiment, subjective questionnaires and objective EEG data were collected. Each lighting condition is random.

In addition to EEG and questionnaire data collection, retrospective interviews were conducted immediately after participants completed all experimental sessions. These semi-structured interviews, lasting approximately 10–15 min per participant, focused on eliciting subjective reflections on emotional and cognitive experiences under different lighting environments. Open-ended questions were used to explore participants’ perceived clarity, comfort, attention, and cultural resonance of the paintings. All interviews were audio-recorded with informed consent and subsequently transcribed for thematic analysis. The experiment received approval from the Research Ethics Committee of the university (Ethics Committee of Biology and Medicine in Dalian University of Technology (protocol code: DUTSICE250411-01, 11 April 2025). Prior to the experiment, all participants provided written informed consent and were informed of their right to withdraw at any time without penalty. All EEG and questionnaire data were anonymized to protect personal privacy and ensure data confidentiality.

3.4. Construction of Subjective and Objective Models

Based on the previous research and the relevant issues of this experiment, we chose 14 pairs of words as shown in

Table 2. Psychological word pair.

No.	Negative Words–Positive Words	Scores	Categories
1	Bright/dark	6–1	Appearance
2	Clear/fuzzy	6–1
3	Natural/unnatural	6–1
4	Color true/unreal	6–1
5	Vivid/not vivid	6–1
6	Warm/cold	6–1
7	Colorful/monotone	6–1
8	Comfortable/uncomfortable	6–1	perception
9	Nervousness/relaxation	6–1
10	inattention/attention	6–1
11	Negative/positive	6–1
12	Dazzling/soft	6–1
13	Like/dislike	6–1
14	Attractive/unattractive	6–1

to specifically describe the degree and psychological feelings of international students towards the appearance and content of paintings. To ensure consistent comprehension among participants with diverse linguistic backgrounds, all emotional descriptors were presented in English. The selected adjectives were characterized by high usage frequency, strong emotional salience, and low semantic ambiguity, which facilitated uniform interpretation across international participants. In addition, brief explanations were provided during the pre-experiment instructions to further clarify the meaning of each term. This multi-faceted approach was designed to uphold the validity of emotional word comprehension within our internationally diverse sample. We assigned scores to different emotional descriptors on a scale from 6 to 1. The more positive the word, the higher the score, and vice versa. From the perspective of emotional theory, an individual’s emotional state is the result of the interplay and combined effects of multiple emotional dimensions. The score of a single emotional word can only reflect the local characteristics of emotions. However, the sum obtained by adding up these scores can more comprehensively and integrally reflect the overall emotional intensity and tendency of the individual, which is in line with the overall characteristics of emotions. The specific calculation formula is as shown in Equation (3):

$$y={\displaystyle \sum _{i=1}^{14}{x}_i}$$

(3)

Here, y represents the target output of the neural network, that is, the total integrated emotional score derived from the participant’s semantic differential ratings. x_i denotes the numerical score assigned to the i-th emotional descriptor (i = 1, 2, …, 14). The sum of all 14 scores reflects the overall emotional intensity and valence experienced under each lighting condition, and serves as the output variable for model training.

In addition, EMOTIV EPOC X (Emotiv, San Francisco, CA, USA) was used to detect the physiological activities of the subjects during the experiment. To ensure high-quality EEG signal acquisition, we followed standard best practices recommended by the manufacturer. Prior to data collection, electrode contact quality was individually verified for each channel using the Emotiv PRO interface, ensuring all sensors achieved optimal or acceptable signal status. Participants were instructed to minimize movement and facial muscle activity during the recording to reduce motion artifacts. The experimental environment was also controlled for electromagnetic interference and ambient noise. EEG signals from different channels were collected and quantified to assess neural correlates of attention and comfort, which were then compared to participants’ subjective evaluations of the lighting environment. We used EEGLAB (Swartz Center for Computational Neuroscience, University of California, San Diego, La Jolla, CA, USA) for signal preprocessing and feature extraction. Visual inspection and manual artifact rejection were conducted, followed by an independent component analysis (ICA) applied to remove common artifacts such as eye blinks, muscle artifacts, and environmental noise. During data processing, the fast Fourier transform (FFT) was applied to analyze both time-domain and frequency-domain characteristics, yielding the spectral energy distribution. By performing FFT using Formula (4), the power values corresponding to the α, β, θ, and δ bands [47] were obtained, denoted as p_δ, p_θ, pα, and p_β. The calculation of the total power (P) is shown in Equation (5), providing a quantitative basis for further statistical analyses, as follows:

$$\mathrm{f}\left(\mathrm{x}\right)={\mathrm{a}}_0+{\sum }_{\mathrm{n}=1}^{\infty }\left({\mathrm{a}}_{\mathrm{n}}\cos {\displaystyle \frac{\mathrm{n}\mathsf{\pi }\mathrm{x}}{\mathrm{L}}}+{\mathrm{b}}_{\mathrm{n}}\sin {\displaystyle \frac{\mathrm{n}\mathsf{\pi }\mathrm{x}}{\mathrm{L}}}\right)$$

(4)

In the formula, a₀ is the DC component, representing the average value of f(x) within one period, and a_n and b_n are the amplitudes of the corresponding frequency components.

$$p=P_{\mathsf{\delta }}+P_{\mathsf{\theta }}+P_{\mathsf{\alpha }}+P_{\mathsf{\beta }}$$

(5)

The total power (P) is obtained by adding up the power of each frequency band.

Since ongoing tonic α activity has been linked to sustained attention processing, this study primarily focuses on analyzing EEG α activity associated with sustained attention performance under varying illumination conditions. To assess this, the ratio of α wave power (8–12.9 Hz) to total power (P, 0–30 Hz) is calculated to reflect the relative power of α waves in different experimental lighting environments, as shown in the following formula:

$${\mathsf{\theta }}_{\mathsf{\alpha }}={\displaystyle \frac{{\mathrm{P}}_{\mathsf{\alpha }}}{\mathrm{P}}}$$

(6)

In this formula, p_α denotes the spectral power of the α frequency band (8–12.9 Hz), and P refers to the total spectral power across the 0–30 Hz range (including δ, θ, α, and β bands). The symbol θ_α represents the relative proportion of α wave activity to total EEG power, and is used here as an indicator of sustained attention. A higher θ_α value indicates a greater degree of attentional engagement under specific lighting conditions. After obtaining the subjective and objective data, we normalize the obtained data through Formula (7).

After obtaining the subjective and objective data, we normalize the obtained data through Formula (7), as follows:

$$z_i={\displaystyle \frac{x_i-\mu }{\sigma }}$$

(7)

In this formula, the entire signal X consists of individual values x_i collected from the subject. The parameters μ and σ correspond to the mean and standard deviation of X, respectively.

Taking the normalized objective electroencephalogram data as the input and the total subjective score as the output, four emotional neural mechanism models were constructed. The performance of this model was evaluated by using the coefficient of determination (R²), mean absolute error (MAE), and mean square error (MSE). The evaluation formulas are shown as in Equations (8)–(10), as follows:

$$R^2=1-{\displaystyle \frac{{\sum }_{i=1}^n{(y_i-{\widehat{\mathrm{y}}}_i)}^2}{{\sum }_{i=1}^n{(y_i-\overline{\mathrm{y}})}^2}}$$

(8)

$$MAE={\displaystyle \frac{1}{n}}{\sum }_{i=1}^n\left|y_i-{\widehat{\mathrm{y}}}_i\right|$$

(9)

$$MSE={\displaystyle \frac{1}{n}}{\sum }_{i=1}^n{(y_i-{\widehat{\mathrm{y}}}_i)}^2$$

(10)

Among them, $\overline{\mathrm{y}}$ is the average value of the true values, ${\widehat{\mathrm{y}}}_i$ is the predicted value, n is the sample size, and the value range of R² is [0, 1]. The closer it is to 1, the better the model and the better the effect.

4. Results

4.1. The Influence of Academic Major on Perceptual Word Pairs

Under four groups of light environments, each participant underwent subjective evaluation and EEG tests. A total of 128 subjective evaluation data and 80 objective evaluation data were obtained and analyzed through SPSS 24.0 (IBM, Armonk, NY, USA) reliability analysis. The Cronbach’s α coefficient of the subjective questionnaire was 0.86 > 0.8, and the experimental data were reliable.

As presented in

Table 3. T-test of major factors.

	Engineering	Art	t	p
	M ± SD	M ± SD
Bright/dark	3.76 ± 1.038	3.83 ± 0.994	−0.381	0.704
Comfortable/uncomfortable	4.04 ± 0.953	4.08 ± 0.671	−0.271	0.787
Dazzling/soft	4.00 ± 0.977	4.17 ± 0.763	−1.484	0.140
attention/inattention	3.88 ± 0.923	3.38 ± 0.904	3.083	0.003
Vivid/not vivid	3.93 ± 0.886	3.80 ± 0.777	0.853	0.395
Clear/fuzzy	3.9 ± 0.872	3.75 ± 0.895	1.504	0.135
Nature/unnatural	4.29 ± 0.774	4.22 ± 0.691	0.594	0.554
Color true/unreal	3.97 ± 0.930	3.83 ± 0.924	0.558	0.405
Nervousness/relaxation	4.16 ± 0.840	3.88 ± 0.885	1.826	0.07
Colorful/monotone	3.82 ± 0.961	3.70 ± 0.869	0.758	0.449
Negative/positive	3.50 ± 0.954	3.73 ± 0.899	−1.418	0.159
Like/dislike	3.69 ± 1.040	3.88 ± 1.010	−1.057	0.292
Attractive/unattractive	3.60 ± 0.979	3.73 ± 0.918	−0.774	0.44
Warm/cold	3.59 ± 1.136	3.57 ± 0.929	0.657	0.512

Note. M—mean; SD—standard deviation.

, the p-values were used to examine the impact of academic major on participants’ emotional and cognitive responses to Chinese landscape paintings under varying lighting conditions, measured through 14 semantic differential word pairs. Among these, only the word pair “attention/inattention” revealed a statistically significant difference between the groups (p < 0.05), indicating a divergence in engagement levels. Specifically, international students majoring in art reported a higher mean score of 3.88, suggesting greater attentional focus when viewing the paintings. This implies that art students, likely due to their training and aesthetic sensitivity, were more visually and emotionally attuned to the cultural nuances of traditional Chinese art when illuminated under dynamic lighting. In contrast, their counterparts from engineering disciplines appeared more distracted, potentially reflecting a less immersive cultural experience. These findings highlight the moderating role of disciplinary background in shaping the emotional resonance and cognitive engagement of international viewers, which is critical for optimizing lighting design strategies in cross-cultural museum settings.

4.2. The Influence of Lighting Environment on Perceptual Responses

As illustrated in Figure 5, among the four tested lighting environments, Environment 4 (4000 K, 300 lx) received the highest overall evaluation scores across most semantic dimensions, with the exception of the “warm/cold” scale. This finding suggests that a lighting setting featuring both high correlated color temperature (CCT) and high illuminance is generally perceived by international participants as more visually engaging and conducive to prolonged observation in a museum context. Such lighting conditions may enhance cognitive alertness and foster deeper interaction with traditional Chinese landscape artworks. However, in the “warm/cold” dimension, higher color temperatures were associated with a cooler visual impression. Consequently, the 4000 K environments were rated lower than the 2700 K settings in this regard. This aligns with findings reported by Luo et al. [48], reinforcing the notion that lower color temperatures—such as 2700 K paired with moderate illuminance (300 lx)—create a warmer, softer atmosphere that is more emotionally resonant for viewers. This emotional warmth may be particularly important in the international dissemination of traditional Chinese culture, as it fosters a sense of intimacy and cultural appreciation among diverse audiences. Additionally, the minimal perceptual difference observed between the 2700 K, 300 lx and 4000 K, 1500 lx environments indicated that both warmth and brightness contribute interactively to affective and aesthetic experiences. These results underscore the importance of carefully calibrated lighting design in enhancing emotional engagement and optimizing the cross-cultural reception of Chinese cultural in curated spaces.

4.3. Factor Analysis of Emotional and Cognitive Dimensions

To better understand the interaction between subjective emotional responses and objective cognitive indicators based on EEG, we conducted a factor analysis, reducing the original 14 pairs of semantic difference words to a smaller set of potential perceptual dimensions. This reduction in dimension enables us to explain more intuitively how lighting design influences the cross-cultural perception of traditional Chinese visual culture. Each extracted factor was marked according to the semantic consistency of the items it contains. A principal component analysis with orthogonal (variance) rotation was adopted. The results are summarized in

Table 4. Rotated component matrix of factor analysis.

0	1 (39.78%)	2 (11.26%)	3 (8.24%)	4 (7.34%)	5 (6.02%)
Meaning	Visibility	Attractiveness	Comfort	Warm	Attention
Bright	0.711	0.240	0.176	0.210	−0.172
Vivid	0.656	0.151	0.327	0.299	−0.022
Clear	0.776	0.215	0.071	0.217	0.008
Nature	0.738	0.044	−0.028	−0.009	0.257
Lively	0.794	0.221	0.051	−0.117	0.142
Colorful	0.689	0.368	0.138	0.121	0.067
Active	0.355	0.673	0.249	0.064	−0.170
Favorite	0.297	0.781	0.338	0.047	−0.006
Attractive	0.147	0.892	−0.008	0.055	0.079
Comfort	0.489	0.246	0.698	−0.084	−0.035
Soft	0.231	0.138	0.799	−0.114	−0.116
Relax	−0.206	0.098	0.658	0.202	0.285
Warm	0.210	0.083	−0.031	0.898	0.089
Attentive	0.210	0.083	−0.031	0.898	0.089

, with the explained variances of each factor listed in parentheses.

Five primary factors were identified: visibility, attractiveness, comfort, warmth, and attention, which together explained 72.6% of the total variance. The visibility factor alone accounted for nearly half of the variance and included descriptors such as bright, vivid, clear, and colorful, reflecting the fundamental visual clarity and salience of the artworks under different lighting conditions. Among the remaining dimensions, attractiveness and warmth correspond to affective responses, whereas comfort and attention align more closely with physiological and cognitive engagement. The analysis of variance revealed that all four lighting environments—combinations of high/low illuminance and high/low correlated color temperature—had significant effects on these perceptual factors. Specifically, the condition featuring high illuminance and high color temperature consistently yielded the highest average scores across all five dimensions, suggesting this lighting setup most effectively enhances both visual and emotional engagement with Chinese landscape paintings among international viewers. As depicted in Figure 6, under low color temperature conditions (e.g., 2700 K), increasing illuminance leads to marked improvements in visibility, attractiveness, comfort, warmth, and attention, with the most pronounced gains observed in visibility. However, gains in attention were more modest. Conversely, under high color temperature conditions (e.g., 4000 K), increasing illuminance also elevated all factors except for warmth, which declined slightly due to the cooler perceived ambiance associated with higher CCTs. Interestingly, regardless of the overall brightness level, the scores for visibility, attractiveness, comfort, and attention generally increased as color temperature rose, again excluding warmth, which consistently favored lower color temperatures. Furthermore, the comparison between Environment 2 (2700 K, 1500 lx) and Environment 3 (4000 K, 300 lx) indicated negligible differences across all five perceptual factors, suggesting a possible interaction effect or perceptual saturation point under these lighting combinations. These findings underscore the importance of tailored lighting strategies in enhancing cross-cultural emotional resonance and perceptual clarity, thereby supporting the more effective international dissemination of traditional Chinese culture through museum and exhibition design.

During the experiment, we observed that the power of the α wave was most obvious in the occipital lobe. Therefore, we chose electrode O2 (Figure 7). The occipital lobe was used for further analysis. Due to the weak EEG caused by mental activities, in addition to calculating α activity, we also analyzed the changes in event-related potentials (ERPs). During the data processing, the continuously recorded EEG data within the experimental period were segmented based on the occurrence time of event-related potentials (ERPs). We superimposed the segmented time windows and calculated the average to obtain more intuitive hotspot map data, as shown in Figure 7c,d. The potential hotspot maps related to α wave events within the 600 ms to 1400 ms time range are presented in Figure 7c,d. Dark blue indicates the lowest level of α wave activity, with amplitude reflecting the degree of brain arousal. It can be observed that in Environment 1, α wave activity is primarily concentrated between 1200 ms and 1450 ms. Only at 250 ms does the area of dark blue reach its maximum. In Environments 2 and 3, α wave activity is mainly concentrated between 550 ms and 950 ms, with dark blue occupying a relatively smaller proportion of the total area. In Environment 4, α wave activity is sustained across almost the entire time range. The proportion of dark blue is the smallest, and fluctuations are relatively minor. Based on these observations, we conclude that α wave activity is highest in Environment 4, lowest in Environment 1, and similar in Environments 2 and 3.

To ensure the reliability of these observations, we performed paired two-tailed t-tests across conditions (n = 32) to evaluate the statistical significance of α power differences. A significance threshold of p < 0.05 was adopted, and only electrode sites exhibiting α power changes exceeding one standard deviation from the baseline (natural light) were retained in the topographic visualizations. These combined statistical and effect-size criteria were used to identify regions of meaningful α modulation, as illustrated in Figure 7.

Table 5. ANOVA of major influence on α power.

	Engineering	Art	t	p
	M ± SD	M ± SD
α power	0.2498 ± 0.08416	0.2797 ± 0.6603	2.068	0.043

further supported this conclusion by highlighting the significant impact of academic major on attention (p < 0.05), which aligned with the subjective findings derived from the questionnaire data. Similarly,

Table 6. ANOVA of environment influence on α power.

	2700 k × 150 lx	2700 k × 300 lx	4000 k × 150 lx	4000 k × 300 lx	F	p
	M ± SD	M ± SD	M ± SD	M ± SD
α power	0.1967 ± 0.0284	0.2701 ± 0.3522	0.2534 ± 0.3013	0.3402 ± 0.484	42.32	0.00

revealed significant differences in α wave power across different lighting environments (p < 0.05). Specifically, attention was highest in the 4000 K, 300 lx environment, as corroborated by the subjective questionnaire responses. The second most attentive environment was 2700 K, 300 lx, followed by 4000 K, 150 lx, and 2700 K, 150 lx. Environments with higher illumination levels consistently exhibited higher α power compared to those with lower illumination levels, underscoring the crucial role that illuminance plays in determining attention. These findings emphasized the substantial influence of lighting design on cognitive processes such as attention, with significant implications for the international dissemination of traditional Chinese culture. By optimizing lighting conditions—such as color temperature and illuminance—cultural spaces can enhance the level of attention and engagement of visitors, thereby promoting more effective cultural communication.

All data were analyzed using a repeated-measures analysis of variance (ANOVA), with the Greenhouse–Geisser correction applied when appropriate. Combining the insights from Figure 7, and Table 5 and Table 6, we found that both lighting and color temperature have a significant impact on attention. Notably, the α power was highest under conditions of high illuminance and high color temperature (4000 K, 300 lx), indicating that attention is most concentrated in such an environment. Furthermore, students majoring in art exhibited a higher level of attention than those majoring in engineering, which further demonstrates the delicate relationship between environmental design and cognitive responses.

4.4. Comfort

The prefrontal lobe plays a crucial role in the regulation of human emotions. In the diverse lighting environments of museums, the primary purpose of lighting is to enhance visitors’ perceptual experience. Based on this, the analysis of participants’ emotional states is considered a key factor in this study. Specifically, the emotional state under standard museum lighting—typically associated with a relaxed condition—is used as a baseline indicator. By comparing emotional responses in other lighting conditions to this standard, we can assess the relative relaxation level induced by each lighting scenario. In other words, by evaluating emotional performance across different lighting environments, we can determine which lighting condition most closely resembles the emotional response elicited by standard lighting. This allows us to infer whether a given environment promotes a relatively relaxed emotional state.

The prefrontal lobe plays a pivotal role in the regulation of human emotions. In the context of museum lighting design, the primary aim of lighting is to accommodate visitors’ perceptual needs and enhance their overall experience. Consequently, analyzing participants’ emotional states becomes a crucial factor in understanding their engagement within cultural spaces. In this study, emotional responses observed under standard museum lighting—typically associated with a relaxed state—are used as a key reference indicator. By comparing the emotional states elicited under various lighting conditions with those in the standard environment, we can evaluate the degree of relaxation each condition promotes. In other words, by examining emotional responses across different lighting environments, we can identify which condition most closely resembles the standard, thereby indicating its effectiveness in fostering a relaxed atmosphere.

Environment 3 (4000 K, 150 lx) was selected as the standard museum lighting condition in accordance with the official guidelines issued by the International Commission on Illumination (CIE) for art museums (CIE 157:2004). These guidelines specify recommended ranges for both illuminance (typically 100–200 lx for general exhibition spaces) and correlated color temperature (within 3000–4000 K) to ensure optimal visual appreciation while protecting artworks. By adopting this officially recognized benchmark, we ensure our experimental conditions align with both industry best practice and conservation standards. Therefore, this lighting environment was adopted in the present study as the benchmark for discrete comfort data. When visitors enter a museum, their primary objective is typically to browse and observe, which is generally associated with a relaxed mental state. Accordingly, we introduced self-perceived relaxation as an indicator for evaluating environmental comfort. The relaxation index of participants under this standard museum lighting condition was calculated as a reference value for assessing comfort across different lighting settings. The calculation formula is provided in Equation (11), as follows:

$$\mathsf{\gamma }={\displaystyle \frac{P_{\alpha }}{P_{\theta }}}$$

(11)

In this formula, γ denotes the quantified comfort level associated with a specific lighting condition, serving as the subjective outcome variable, P_α is the power density of the α wave, and P_θ is the power density value of the θ wave. These neural features are used as indicators of cognitive relaxation (α) and emotional regulation or memory-related engagement (θ) in response to environmental stimuli.

As shown in Figure 8, Environment 3 (150 lx, 4000 K) was used as the standard condition in the experiment. Based on this benchmark, relaxation values under three specific lighting conditions—2700 K at 150 lx, 2700 K at 300 lx, and 4000 K at 300 lx—were calculated using a discrete method, and participants’ subjective perceptions were also compared across these conditions. The emotional state under the standard lighting was normalized to a value of 1, with each participant’s EEG data and α/θ ratios in this environment assigned as the baseline. By comparing the emotional indices under the other lighting conditions to this baseline, we identified which condition most closely approximated the emotional profile of the standard relaxed state. It is important to note that deviations from this baseline do not necessarily indicate an absence of relaxation, but rather reflect relative differences compared to the standard lighting condition. Therefore, this study focuses on relative relaxation in comparison to a commonly accepted industry standard, rather than on absolute emotional states.

According to the analysis, the data from Environment 4 (300 lx, 4000 K) was closest to the standard environment and corresponds to the highest comfort level, which is supported by the results of the subjective questionnaire survey. In contrast, the data from the other two lighting environments showed obvious differences, with significantly stronger emotional responses. This suggested that an individual’s attention performance may vary across different environments, leading to differing comfort states. Some participants reported higher levels of comfort, while others reported lower levels. However, this variation does not undermine the integrity of the experimental results. It is worth noting that in lighting environments with high illuminance and high color temperature, the participants’ emotional states were very similar to those observed under the standard environment.

4.5. Construction of Subjective and Objective Models

We used the four different brain waves recorded from 32 subjects under four different lighting environments as the main inputs, and the model’s output was the estimated total perception score after observation. Four types of emotional neural mechanism models, namely back-propagation neural network (BP), eXtreme gradient boosting (XGBoost), random forest (RF), and least-squares support vector machine (LSSVM), were constructed through MATLAB R2023b (MathWorks, Natick, MA, USA). Before model training, all EEG features and semantic–affective scores were standardized using Z-score normalization to eliminate dimensional differences and ensure comparability among participants. The dataset was randomly divided into training (70%), test (15%), and validation (15%) subsets, with stratification to preserve label distribution. To optimize model performance, hyperparameter tuning was conducted for all four models. Specifically, key hyperparameters such as model complexity, learning rate, and regularization coefficients were systematically adjusted using grid search and cross-validation strategies to enhance generalization and fitting accuracy.

The performance of this model was evaluated using the coefficient of determination (R²), mean absolute error (MAE), and mean square error (MSE). The evaluation results were shown in

Table 7. Comparison of four subjective and objective regression models.

Model	Subset	R2	MAE	MSE
BP	Training Set	0.9193	0.0638	0.080
	Test Set	0.8862	0.0758	0.095
	Validation Set	0.7997	0.1045	0.131
RF	Training Set	0.9059	0.0632	0.0945
	Test Set	0.8055	0.0828	0.1298
	Validation Set	0.6811	0.1496	0.2059
XGBoost	Training Set	0.9120	0.0656	0.0988
	Test Set	0.7475	0.0804	0.1218
	Validation Set	0.6951	0.1000	0.1712
LSSVM	Training Set	0.8994	0.0702	0.1023
	Test Set	0.7042	0.1242	0.1870
	Validation Set	0.6443	0.1338	0.2110

. The ranking of the four models according to the comprehensive score was BP > RF > XGBoost > LSSVM. The comprehensive evaluation score of the BP neural network was superior to the other three models. Among them, the R² of the test set was 0.081 higher than RF, 0.139 higher than XGBoost, 0.1823 higher than LSSVM, with an average increase of 0.1341, and an average increase of 0.1262 in the validation set. Moreover, both the MAE and MSE were smaller than those of the other models. Therefore, we further analyzed the model construction of the BP neural network.

During the training process of the BP neural network model, 70% of the experimental data was used for training (blue), 15% for testing (red), and another 15% for verification (green). The training results of the model were shown in Figure 9, Figure 10 and Figure 11. It can be seen from Figure 8 that after iterative calculation, the curve (training verification test) declined in rounds 0 to 5. In round 5 to 25, the model had the best performance with MES of 0.0089971. In round 31, the training was completed and the model error conformed to an acceptable normal distribution, as shown in Figure 9. The variances and regression values of the three (training verification error) were presented in Table 7. And Figure 10 shows that the regression values (R) of the four parts of training, validation, testing, and the overall data set are all close to 1. This result strongly indicated that there was a high correlation between the model’s predicted output and the corresponding input, and the error was extremely small.

4.6. Qualitative Analysis

To gain a deeper understanding of emotional engagement, we conducted retrospective interviews immediately after all tasks were completed. The interview data were then integrated and analyzed alongside the emotional and EEG results.

The interview results show that some participants exhibited heightened attention and cognitive focus under high illuminance and high color temperature conditions (4000 K, 300 lx). For instance, participant A, an art major from Indonesia, stated the following:

In the brighter and whiter light, I was able to focus more easily on the fine details of the painting. It felt like a real museum, and I didn’t get distracted.

This observation is consistent with EEG findings showing increased α power and higher attention scores under this lighting condition, particularly among participants with an art background.

Some participants exhibited greater emotional comfort and a sense of relaxation under lower illuminance settings, particularly in the 2700 K, 150 lx environment. Participant B, an engineering major from Vietnam, reflected as follows:

The soft, warm light reminded me of home. I wasn’t thinking too much—I just enjoyed looking at the painting in a quiet mood.

This aligns with the elevated comfort ratings and higher γ values under this environment, indicating a more relaxed emotional state.

The interviews also revealed discipline-specific differences in perceptual focus and engagement. Participant C, an art major from South Korea, and Participant D, an engineering major from Pakistan, offered the following contrasting reflections:

I felt emotionally connected to the brushwork and composition, especially under warmer light. It was easier to appreciate the artistic meaning and the rhythm of the ink.

I was more curious about how the light was set up than the painting itself. The environment was interesting, but I didn’t think too much about the artwork.

These reflections support the statistical results showing that art majors scored significantly higher in attention and comfort, and that lighting effects were more emotionally salient to them.

5. Discussion

5.1. The Theoretical and Practical Significance of This Research

This study integrates cross-cultural factors, professional backgrounds, and experimental lighting variables, thereby expanding the theoretical framework of coupled research on the light environment, emotion, and neural mechanisms. Theoretically, this study extends the influence pathway of the visual environment on neural cognition (e.g., high illumination reducing θ power in short-term working memory) [49], while also addressing limitations related to cultural exhibition hall lighting within the context of “cross-cultural participation—neural feedback emotional experience.” It was further observed that professional backgrounds in art and engineering moderate EEG patterns, suggesting an interaction effect between professional training sensitivity to visual cultural stimuli and brain mechanisms. This finding supports previous research on gender and professional differences in eye movement and heart rate variability (HRV) [50]. At a practical level, this study provides a multidimensional basis for lighting strategies in international pavilions. Experimental results confirm that lighting combining high correlated color temperature (CCT) and high illuminance falls within the “Kruithof comfort zone”, enhancing both attention and subjective comfort [51], which holds significant practical implications. Exhibition designers can leverage the “professional–cultural–light environment” model proposed in this study to develop targeted lighting solutions, achieving a dual optimization of visual appeal and emotional resonance. Moreover, the findings offer valuable guidance for the construction of immersive environments in virtual museums. Consistent with psychological index research in virtual reality experiments, different CCTs influence eye movement, heart rate, and preference scores [52]. In conclusion, this study systematically elucidates the combined mechanisms of cultural background, professional factors, and light environment, providing theoretical and methodological support for exhibition design. It contributes to advancing the interdisciplinary field of neurocultural experience design and demonstrates practical value in smart museums and cultural dissemination.

5.2. Subjective Perception and Neural Mechanisms

This study continues and deepens the investigation of the multivariate effects of illuminance and correlated color temperature (CCT) on subjective, objective, and electroencephalogram (EEG) responses. At the subjective level, the findings support the classic illuminance–color temperature comfort range of 150–300 lx and 2700–4000 K [53], and verify the consistent relationships between the semantic dimensions of “comfort,” “clarity,” and “relaxation” and the lighting parameters through subjective ratings. At the objective EEG level, it was found that under high illumination conditions, α and θ power in the frontal and parietal–occipital regions were significantly suppressed, while certain spatial and indirect lighting combinations increased θ power (e.g., at Figure 4 and Figure 8). Changes in the α/θ ratio under the condition of 300 lx at 4000 K showed a high consistency with subjective perceptual scores, further demonstrating that neural signals can serve as mediators of illumination effects. This discovery complements previous studies; for example, high illuminance can reduce θ power in working memory tasks without affecting behavioral performance [54]. Comparisons between indirect/direct mixed lighting indicate that indirect lighting can enhance pleasantness scores and θ power, highlighting its advantage in emotional arousal. Moreover, visual comfort in warm and cool color temperature environments, corresponding to different CCT variations, exhibits an inverted U-shaped correlation with θ wave power, consistent with the CCT–subjective comfort curve observed in virtual reality studies [55]. In conclusion, this study integrates the light environment with neural oscillation mechanisms, providing a technical neural basis for lighting control in museums.

5.3. The Value and Potential of the Prediction Model

This study takes exhibition lighting as the context to investigate the neural and emotional responses to combined illuminance and color temperature patterns on the exhibition viewing experience of international students. It further develops a regression model based on subjective and objective multimodal data to evaluate the predictive power of the light environment for exhibition viewing comfort and attention levels. The established backpropagation (BP) neural network model demonstrates outstanding regression performance, achieving a coefficient of determination (R²) of 88.65%, which significantly surpasses the performance of XG Boost (R² = 74.75%) and random forest (R² = 80.55%) under the same conditions, as well as exceeding the upper limit of linear regression models (R² < 0.75). This result highlights the BP model’s strong capability to accurately fit the nonlinear relationship between EEG signals and subjective perception. Deng et al. (2021) [39] explored the relationship between indoor lighting and work engagement by constructing personalized work state classification models using random forest (RF) and artificial neural networks (ANN), achieving a classification accuracy of 83.3%. Although their task was classification, their predictive accuracy has not reached the level demonstrated by the regression model in the present study. Compared to Peng et al. (2022), who evaluated passenger comfort in high-speed railway environments through EEG signals and constructed a comfort scoring model based on Light GBM with a mean absolute error (MAE) of 0.1261 on the test set, their method appears less sensitive to subtle trends in subjective emotional changes [56]. Similarly, Shi et al. (2025) investigated the influence of preferred illuminance on EEG characteristics during reading tasks, reporting a highest classification accuracy of 77.78% with K-nearest neighbors (KNN) among algorithms such as KNN, SVM, and RF [47]. In contrast, the present study achieves superior accuracy in a regression framework, making it particularly suitable for the continuous prediction of emotional states and attention fluctuations in exhibition scenarios. Overall, the BP model not only outperforms existing methods in statistical metrics but also demonstrates unique advantages in task adaptability and cross-modal data integration. By benchmarking against current EEG-based machine learning models for light environment research, the model proposed here shows strong adaptability and application potential in regulating lighting in cultural exhibition halls and supporting international cultural dissemination. It thus provides a quantifiable and predictive neuroscience foundation for optimizing exhibition and display environments tailored for overseas audiences and offers a data-driven new pathway for lighting strategy in cross-cultural spatial design.

5.4. Limitations and Future Directions

Despite the empirical value of the findings, several methodological constraints may affect the broader applicability of this study. Firstly, this study is relatively small in scale. The limited number of participants may restrict the extent to which the experimental conclusions can be generalized. In the future, it will be necessary to conduct an experiment with a larger sample size to analyze the results further, enhancing the overall generalizability of our findings. Additionally, the current sample is limited to international students and lacks diversity in cultural background, age groups, and gender ratio. Future research should include participants from multicultural backgrounds, across all age groups and genders, to explore the effects of cultural schemas, age-related cognitive differences, and gender hormones on light sensitivity, aiming to develop a more generalizable cross-cultural model. Secondly, to ensure perceptual consistency, only one traditional Chinese landscape painting was used as the visual stimulus. Although this controlled content-related variability, it may have constrained the range of emotional and neural responses elicited from participants. Future studies could incorporate a broader range of artworks—including figurative, abstract, and calligraphic styles—to better capture the spectrum of aesthetic experiences Furthermore, while the laboratory setting enabled precise control over lighting conditions, it lacked the immersive and multisensory qualities of real-world exhibition environments. Employing virtual reality (VR) or conducting on-site experiments in gallery spaces could enhance ecological validity and participant engagement. Finally, the ethical implications of using lighting to influence attention and emotion warrant further consideration. Although such interventions were applied in a minimal-risk, controlled research context in this study, broader applications may raise concerns about psycho-logical autonomy, especially in real-world environments where individuals may not be fully aware of such influences. Future interdisciplinary research is needed to develop ethical design principles for emotionally responsive environments, ensuring transparency, informed consent, and the protection of individual well-being.

6. Conclusions

This study investigates the influence of the lighting environment in cultural exhibition halls on the emotional neural mechanisms of international students. Using a simulated standard museum exhibition environment, 32 international students majoring in art and engineering were recruited as participants. A multimodal evaluation system combining subjective and objective measures was employed. Subjective experience was assessed via a semantic differential method using 14 psychological descriptive terms. Objective neural signals, including the spectral energy of α, β, and θ waves, were recorded using EMOTIV EPOC X equipment to construct indicators of attention and comfort levels. This study explored the effects of four combined illuminance and color temperature conditions (150 lx × 2700 K, 300 lx × 2700 K, 150 lx × 4000 K, 300 lx × 4000 K). Results showed that under high illuminance and high color temperature conditions (300 lx × 4000 K), international students exhibited higher attention and comfort levels. Additionally, α wave power in art majors was significantly higher than in engineering majors. The BP neural network model, constructed using both subjective and objective data, achieved a determination coefficient of 88.65%, effectively simulating the influence of different lighting parameters on the exhibition viewing experience, and was further supported by retrospective interview insights.

This study contributes not only interdisciplinary empirical evidence but also theoretical insight into how environmental lighting conditions shape emotional and neural responses in culturally immersive exhibition contexts. By integrating EEG-based neural indicators with semantic–affective evaluations, this study bridges the fields of neuroaesthetics, lighting design, and intercultural communication. The findings provide a neuroscience-informed foundation for the evidence-based design of lighting environments in international museums, enhancing emotional accessibility and cognitive engagement. These insights also serve as strategic references for future museum planning and cultural policy, promoting emotionally inclusive and cross-culturally resonant approaches to the global dissemination of traditional Chinese culture.