A Study on the Restorative Effects of Hydrangea Flower Color and Structure on Human Psychology and Physiology

Abstract

Amid growing “nature deficit” associated with urbanization and indoor living, flowering plants are increasingly used to support psychological restoration. Yet evidence on how floral color and structural morphology jointly shape restorative outcomes remains limited. This study employed a within-subjects, repeated-measures design, utilizing physiological instruments and psychological questionnaires to investigate the physiological and psychological restorative benefits of Hydrangea macrophylla and to quantify the differences in restorative effects across five colors (blue, pink, white, mauve, red), two inflorescence types (mophead, lacecap), and two petal structures (single, double). Twenty-eight healthy young adults viewed 15 live hydrangea stimuli under controlled laboratory conditions. Multimodal outcomes combined objective measures—eye-tracking and single-channel EEG—with subjective measures (SD; POMS). Hydrangea exposure significantly reduced negative mood, and color and structure exerted distinct and interactive effects on visual attention and arousal. Red and mauve elicited larger pupil diameters than white and pink, while lacecap inflorescences were associated with lower cognitive load and improved attentional recovery relative to mophead. Double-petaled forms showed greater attentional dispersion than single-petaled forms. Interactions indicated that morphology modulated color effects. The mauve lacecap double-flowered cultivar (M02) showed the strongest observed restorative potential within this sample. These findings highlight the importance of integrating color and structural cues when selecting flowering plants for restorative environments and horticultural therapy, and they motivate field-based replications with broader samples and higher-density physiology.

1. Introduction

Urban populations currently face multiple substantial health challenges [1]. The prevalence of mental disorders, including anxiety and depression, exacerbated by societal pressures, shows a marked increase [2]. According to the Global Burden of Disease 2010 (GBD 2010), a notable rise occurs globally in the number of individuals affected by mental, neurological, and substance use disorders (MNSD), with the MNSD burden—predominantly driven by depression—exerting a substantial impact on public health [3,4,5]. Projections suggest that by 2030, stress will emerge as the leading contributor to the global disease burden [6].

Restorative benefits specifically refer to the stress relief and cognitive recovery benefits provided by natural environments through specific psychological mechanisms [7,8]. In urban environments, green vegetation demonstrates significant regulatory potential in this regard [9,10,11]. Roadside trees such as cherry blossoms, London plane trees, and dawn redwoods [12]; flowering trees and autumn street trees [13]; ornamental bamboo [14]; broadleaf forests [15,16] and coniferous forests [17,18]; as well as plant landscapes dominated by trees and herbaceous plants [19] have all been shown to have positive effects on human health.

Among plant varieties, ornamental flowering species with affective significance function as pivotal intermediaries for urban residents to reconnect with nature and improve their living environments [20,21]. Studies reveal that exposure to flowering plants effectively mitigates human stress responses and facilitates relaxation [22,23,24]. Two studies find that in care settings, white chrysanthemums enhance patients’ nutritional intake and verbal communication [25], and blue hydrangeas prove superior to mauve variants in inducing relaxation among financial professionals [26]. These restorative outcomes may originate from the “soft fascination” elicited by flowering plants—manifested through their chromatic richness, delicate floral structures, and dynamic movement. When such mild, naturally occurring stimuli guide human attention, attentional mechanisms recover, negative cognitions are attenuated, and affective states transition toward more positive valence [27].

Research systematically establishes the role of plant color variation in stress recovery processes [28,29]. Orange, red, yellow, magenta, and pink blossoms are associated with positive emotional responses, while mauve, white, blue, and analogous shades connect with relaxation effects [30,31,32,33,34]. Evidence also suggests that spatial structure and form shape restorative responses, as reflected in preferences for skeletal elements in Zen gardens [35,36], trees with high crown-to-trunk ratios [37,38,39], and bamboo structures and understory spatial organization that differentially affect demographic groups [40,41]. In contrast, the influence of flowering plant structure on perceptual preference and psychophysiological restoration remains underexplored. A notable case comes from studies on camellias, where Electroencephalogram (EEG) data reveal that stimuli from specific floral types—semi-double, double, anemone-form, and rose-form double—induce significantly stronger relaxation responses compared to others.

Quantifiable metrics are crucial for deciphering restorative mechanisms. Eye-tracking studies based on Attention Restoration Theory (ART) [7] and Stress Recovery Theory (SRT) [8] use fixation duration on natural elements to indicate effortless attention and recovery from mental fatigue, while pupil dynamics and fixation frequency reflect restorative potential [42,43,44,45,46,47,48,49,50]. Electroencephalography (EEG) captures neural responses associated with emotion and arousal, for instance, increased alpha activity when viewing flowers is often interpreted as relaxation [51,52,53,54]. Self-report tools complement these measures: The Semantic Differential (SD) scale quantifies perceptual dimensions for cross-comparison [55,56,57,58], and the Profile of Mood States (POMS) effectively detects transient mood changes before and after exposure [59].

Despite this progress, most investigations focus on specific groups (e.g., physicians, office workers, students, patients) [23,24,25], broad plant categories (blooming versus non-blooming and foliage species) [60,61,62,63,64], and chromatic variation in floral and foliar tissues [65,66,67]. Critically, plant attributes are often treated uni-dimensionally (color only), while the independent and interactive roles of floral structure remain largely overlooked. To date, relatively few studies have explicitly addressed the distinct effects of structural variation on recovery [68]. Accordingly, a systematic investigation into how color and structure in flowering plants jointly shape restorative efficacy is both timely and necessary (

Table 1. Key characteristics of studies included in the literature review.

Author	Plant Material	Flowering	Physiological Metrics	Subjective Measures	Color	Structure	Live Plants	Main Conclusions
Igarashi M. et al. [22]	Pansy	√	√	√	-	-	√	Induction of psychophysiological relaxation
Song C. et al. [23]	Rose	√	√	√	-	-	√	Improved mood state
Talbott J.A. et al. [25]	Chrysanthemum	√	√	-	-	-	√	Enhanced food intake and vocalization
Nakamura R. et al. [54]	Begonia, Geranium	√	√	-	-	-	√	Increased relaxation
Ai L. et al. [65]	Camellia	√	√	√	√	√	-	Restoration influenced by planting Structure and flower color
Yamane K. et al. [60]	Pansy	√	√	√	-	-	√	Flowering plants exert stronger positive effects
Park S.H. et al. [61]	Dendrobium, Cestrum aurantiacum	√	√	√	-	-	√	Accelerated post-operative recovery
Adachi M. et al. [62]	Lathyrus odoratus, Larkspur	√	-	√	√	-	-	Elevated pleasure and positive affect
Kim E. et al. [63]	Geranium	√	√	√	-	-	√	Promoted stress recovery in high-stress women
Renlin Z. et al. [64]	Peony	√	√	√	-	-	√	Mitigation of tension
Elsadek M. et al. [65]	Hydrangea	√	√	√	√	-	√	Blue hydrangeas enhance relaxation

).

Building on this foundation, the present study addresses two key questions: (1) How do structural traits of flowering plants affect their restorative benefits? (2) How do these structural variations influence human color preference? We examine the independent and interactive effects of floral color, inflorescence structure, and petal structure on restoration, by integrating subjective scales and objective measures (e.g., eye-tracking, EEG). The study is guided by the following hypotheses:

(1)

Floral color significantly influences restorative outcomes.

(2)

Inflorescence type and petal structure affect visual attention and relaxation.

(3)

Color and structure interactively affect restorative outcomes.

2. Materials and Methods

2.1. Plant Materials

Hydrangea macrophylla (Thunb.) Ser. is a widely cultivated ornamental species valued for its pronounced floral abundance and is extensively utilized in landscape gardening, indoor potted displays, and cut flower production. Hydrangea’s long flowering period enhances the ecological and ornamental value of urban green spaces [69,70,71,72]. It also serves as a seasonal marker—recognized in Japan as a symbol of the rainy season—strengthening cultural identity and well-being through regional associations [73]. Hydrangea flower color results from the interaction between genetic traits and soil chemistry. Research indicates that over 60% of H. macrophylla varieties exhibit color variation: in acidic, aluminum-rich soils, anthocyanins bind with aluminum ions to form stable blue/purple hues, representing the most distinct and adjustable color feature. Under alkaline or aluminum-deficient conditions, the same pigments appear red/pink. White hydrangeas belong to a separate biochemical category, lacking anthocyanins entirely, which makes their color stable and unaffected by soil environment [74,75].

Therefore, this investigation employs 15 distinct H. macrophylla specimens in a comparative design. The selected specimens represent five distinct floral colors (blue, pink, white, mauve, and red), two inflorescence types (mophead and lacecap), and two petal structures (single and double).

(1)

Color Classification: Field color sampling is performed on the day of experimentation under clear skies and sufficient sunlight. RGB values are recorded and calibrated against the Natural Colour System (NCS) standard. To minimize environmentally induced chromatic variation, two intact florets from the primary bloom of each specimen are identified as representative color samples.

(2)

Inflorescence Types: Based on the ratio and arrangement of sterile and fertile flowers, the species is classified into mophead type and lacecap type. Mophead hydrangeas produce inflorescences predominantly composed of large, showy sterile flowers, with few or no visible fertile florets. These sterile florets are densely arranged into solid, spherical clusters. In contrast, lacecap hydrangeas form inflorescences characterized by a central dome of numerous small, fertile florets, encircled by one or more whorls of conspicuously larger sterile flowers, resulting in a flattened or gently concave, saucer-shaped structure [26,76]. This classification system is defined by the characteristics of individual florets within the inflorescence, specifically focusing on the number and structure of sepals in the decorative sterile flowers.

(3)

Petal Structure: Based on the sepal architecture and phyllotaxis of ornamental florets (sterile flowers), specimen types are categorized as single or double. The single type presents a single whorl of 4–5 sepals that spread radially without overlapping, constituting a single-layered floret structure. In contrast, the double-type develops more than two whorls (8–20 sepals) arranged in imbricate layers, where sterile sepals exhibit petaloid transformation and hyperplastic growth.

Each floral color category is represented by both single- and double-petaled spherical inflorescences, as well as double-petaled lacecap-type inflorescences. The mean pot height measures 30 cm, with mean plant height and crown width reaching 42.5 cm and 50 cm, respectively. Mophead-type specimens develop ≥ 3 flowering globes, whereas lacecap-type specimens develop ≥ 5 inflorescences. To minimize variability in visual stimuli, plants within the same color category were selected to have a comparable number of flower heads (i.e., spherical inflorescences) and similar overall crown dimensions. All plants display normal anthesis, well-defined petal characteristics, and readily distinguishable floral colors, without exhibiting pathologies or damage such as wilting, lodging, or scorching, detailed information is presented in

Table 2. Summary of hydrangea samples characteristics.

No.	Sample ID	Color & Structure	Crown Width (cm)	Height (cm)	Flower Head	Bloom Size (cm)
1	B01	Blue Mophead Single	46.5	35	4	3.6
2	B02	Blue Lacecap Double	50	41	13	4
3	B03	Blue Mophead Double	45.5	34	4	3.2
4	P01	Pink Mophead Double	51	44	5	4
5	P02	Pink Mophead Single	55	40.5	5	6
6	P03	Pink Lacecap Double	51.5	42	5	5
7	W01	White Mophead Double	49	41	4	3.2
8	W02	White Mophead Single	51	37	5	3.4
9	W03	White Lacecap Double	56	40.5	5	4.5
10	M01	Mauve Mophead Single	53	48	9	3.8
11	M02	Mauve Lacecap Double	56.5	50	12	5
12	M03	Mauve Mophead Double	58	46.5	11	3.5
13	R01	Red Mophead Single	55.5	49	10	4
14	R02	Red Mophead Double	51.5	38.5	4	6
15	R03	Red Lacecap Double	56	45	14	5.2

(see Appendix A

Table A1. Summary of sample characteristics.

No.	Sample ID	Image	Color Data (RGB/NCS)	Color & Structure	No.	Sample ID	Image	Color Data (RGB/NCS)	Color & Structure
1	B01		58, 65. 137 NCS S4050-R70B	Blue Mophead Single	9	W03		228, 209, 200 NCS S1002-R	White Lacecap Double
2	B02		131, 143, 170 NCS S3020-R90B	Blue Lacecap Double	10	M01		77, 51, 114 NCS S4050-R50B	Mauve Mophead Single
3	B03		104, 114, 147 NCS S4030-R70B	Blue Mophead Double	11	M02		128, 94, 155 NCS S2040-R50B	Mauve Lacecap Double
4	P01		187, 79, 97 NCS S1060-R10B	Pink Mophead Double	12	M03		167, 104, 168 NCS S3030-R50B	Mauve Mophead Double
5	P02		206, 112, 122 NCS S2040-R10B	Pink Mophead Single	13	R01		217, 0, 61 NCS S1575-R10B	Red Mophead Single
6	P03		191, 104, 113 NCS S0550-R10B	Pink Lacecap Double	14	R02		153, 29, 70 NCS S2060-R20B	Red Mophead Double
7	W01		240, 228, 216 NCS S0500-N	White Mophead Double	15	R03		184, 39, 70 NCS S1070-R10B	Red Lacecap Double
8	W02		255, 224, 219 NCS S0502-R	White Mophead Single

for the extended dataset). Furthermore, during the experiment, all participants viewed the same selected specimen for each plant type, and were instructed to observe from a fixed, consistent frontal angle to control for potential differences in visual presentation.

2.2. Study Participants

To minimize interference from light and noise, the experiment is conducted in a laboratory located in Building 4 at Sichuan Agricultural University (Chengdu Campus). Indoor temperature was 24 ± 1 °C; illuminance at the specimen plane was 800–900 lx (5000 K), verified by a lux meter before each session; ambient noise was ≤35 dB(A). Specimens were positioned on an observation platform 55 cm above the floor at a fixed viewing distance of 50 cm; background was matte white (3 m × 2 m). The study employs a within-subjects repeated-measures design, in which all 28 participants view each of the 15 hydrangea specimens. Participants have a mean age of 22 years (range: 18–30). All are in good general health, free from color vision deficiencies (including color blindness or weakness), and have normal or corrected-to-normal visual acuity (≥1.0). None report a history of mental or neurological disorders. All participants provide voluntary informed consent after receiving comprehensive details regarding the experimental objectives, procedures, and potential risks. Throughout the study, two laboratory technicians deliver standardized procedural training and guidance. The study protocol adheres to the principles of the Declaration of Helsinki and is approved by the Ethics Committee of Sichuan Agricultural University.

2.3. Data Acquisition

2.3.1. Eye-Tracking

Research on the perceptual restoration effects of plants typically examines both psychological and physiological dimensions. At the physiological level, this study employs the Tobii Glasses 2 wearable eye-tracker (Figure 1, left; manufacturer: Tobii AB, Danderyd, Sweden; sampling rate: 100 Hz; head movement range: 30–50 cm; operating distance: 60–100 cm) to record eye movements. Data are acquired via the Tobii Pro Glasses Controller (v1.14) software and subsequently processed and analyzed using Tobii Pro Lab (v1.232).

During stress response, the human body activates the sympathetic nervous system, stimulating the adrenal glands to release stress hormones, including cortisol and catecholamines, thereby inducing a cascade of physiological changes [77]. Researchers assess stress levels through specific physiological metrics such as total fixation duration (TFD), changes in average pupil diameter (APD), and saccade frequency (SF) [45]. Building on this evidence, the present study selects five eye-tracking parameters for analysis: TFD, Number of Blinks (NB), APD, SF, and Average Saccade Amplitude (ASA). The operational definitions and physiological significance of each metric are summarized in

Table 3. Metrics and their interpretations.

Metric	Abbr.	Definition	Significance	Unit
Total Fixation Duration	TFD	Cumulative duration of all fixations	Overall cognitive processing time	ms
Number of Blink	NB	Total blink count within a specified period	Attentional dispersion & information processing load	counts
Average Pupil Diameter	APD	Mean diameter of pupil constriction and dilation	Level of mental workload	mm
Saccade Frequency	SF	Number of saccades per unit time	Visual search behavior; stimulus salience	s−1
Average Saccade Amplitude	ASA	Distance between two consecutive fixation points	Information acquisition range; feature distinctiveness	mm
Alpha Wave	α	Neural oscillations at 8–13 Hz	Relaxed wakefulness; mental calmness and internalized attention;	μV/Hz
Beta Wave	β	Neural oscillations at 14–30 Hz	Active cognition, concentration, and alertness	μV/Hz
Theta Wave	θ	Neural oscillations at 4–7 Hz	Drowsiness, introspective states, memory retrieval	μV/Hz

.

2.3.2. EEG Data

The NeuroSky MindWave Mobile-EEG headset (Figure 1, right; manufacturer: Wuxi Lansu Intelligent Technology Co., Ltd., Wuxi, China; sampling rate: 512 Hz) is employed to monitor physiological responses during stress recovery testing, capturing direct emotional variations and stress-reduction effects. The device acquires EEG signals from the Fp1 position above the eye (prefrontal lobe) using a forehead-mounted sensor [78]. Signal detection and filtering are performed by two dry electrodes, while integrated ThinkGear and ASIC modules suppress electrical interference and filter artifacts by preserving the principal frequency components of EEG and removing power-line noise [79]. Data are streamed via Bluetooth to a computing device, with raw EEG waveforms collected at 1-s intervals for subsequent power spectral analysis across frequency bands and attention metrics. Based on the absolute power values of each frequency band output by the device, a moving time-window averaging method is applied: the continuous EEG data stream is segmented into multiple 10-s epochs, and spectral analysis is performed on each epoch to calculate the average power of frequency bands such as α, β, and θ.

Among derived parameters, enhanced alpha-band activity typically reflects a relaxed and pleasant state [80]. High-frequency beta oscillations indicate cortical activation linked to logical reasoning and analytical processing—commonly observed during focused problem-solving and often associated with elevated tension and anxiety [81]. The α/β ratio serves as an arousal index for stress assessment, demonstrating a decrease under stressful conditions [82]. Theta rhythm enhancement correlates with improved creative performance and imaginative capacity, whereas its attenuation signifies heightened focused attention [83].

2.3.3. Subjective Psychological Assessment

The POMS questionnaire, developed by McNair et al. [59], is administered before and after exposure to assess participants’ affective states following hydrangea viewing. The instrument consists of six subscales: Tension-Anxiety (T), Depression-Dejection (D), Anger-Hostility (A), Vigor-Activity (V), Fatigue-Inertia (F), and Confusion-Bewilderment (C). Each subscale contains six items represented by multiple adjectives describing distinct emotional states—such as tense, unpleasant, angry, and energetic—totaling 65 adjectives across the questionnaire. Items from all subscales are randomly interspersed. Respondents rate emotional intensity on a 5-point scale: 0 (Not at all), 1 (A little), 2 (Moderately), 3 (Quite a bit), and 4 (Extremely). Additionally, seven positive emotion items (e.g., friendly, thoughtful, helpful) serve as distractor items to verify response authenticity and participant cooperation. Higher scores are used as indicators of changes [84].

For perceptual evaluation, the SD method developed by Osgood is implemented. Twelve pairs of bipolar adjectives (e.g., beautiful-ugly, vibrant-decaying, relaxed-anxious) are rated on a 7-point scale ranging from −3 to +3 to capture subtle perceptual variations across specimens.

2.4. Experimental Methods

The experiment was conducted from 1 June to 8 June 2025, daily between 9:00–12:00 and 13:00–19:00. Participants were told to take part in an “aesthetic viewing experience” but were not told the order of the hidden hydrangea samples. This method helped mitigate expectation effects regarding the study’s purpose or restorative outcomes, ensuring that participant responses reflected only their immediate perceptual experience. Given that early neural activity for visual recognition and emotional processing completes within hundreds of milliseconds [85], minutes-long viewing of a single plant specimen is sufficient to ensure stable initial perception and emotional experience in participants, while effectively capturing positive changes in autonomic nervous system and brain activity [67,86,87]. During monotonous sustained-performance tasks (e.g., Sustained Attention to Response Task SART), attention lapses and performance decline occur rapidly [88]. Therefore, to examine differences in overall and color-specific attentional restoration effects of hydrangeas under induced “directed attention fatigue” (primarily stemming from the sustained directed attentional effort required by the experimental task), participants were instructed to view 15 flowerpots consecutively. Single hydrangea plant as the visual stimulus, the restorative effect of the plant itself can be effectively isolated, thereby avoiding potential stress that complex plant landscapes may induce due to their excessive scale, high density, or tendency to trigger biophobia in individuals. Each pot was viewed for 2 min (6 min per color category, totaling approximately 30 min).

The specific procedure included the following stages:

(1)

Preparation: On arrival at the laboratory, participants rested for 3 min, provided demographics and consent, and completed the pre-test POMS while experimenters calibrated equipment.

(2)

Stress induction: Participants completed a challenging mental arithmetic task within one minute, which included 15 problems involving three-digit addition, subtraction, multiplication, and division. They were informed that their performance would be recorded and monitored. A minimum of six correct answers was required, without the use of formulas or calculators. Previous studies have shown that performing mental arithmetic under time-limited or complex conditions significantly increases cognitive load and stress levels [89].

(3)

Instrument Setup: Participants wore the Tobii Pro Glasses 2 eye-tracker and performed a 5-s calibration by fixating on a 5-mm black dot, verified via Tobii Pro Glasses Controller (v1.14). The NeuroSky MindWave Mobile-EEG headset was then fitted to ensure firm contact of the forehead sensor and stable signal acquisition.

(4)

Experimental Trial: Seated comfortably in a reclining chair, participants viewed 15 hydrangea specimens presented in random order (2 min each). Eye movement and EEG data were continuously recorded during viewing. After observing each specimen, participants filled out the SD questionnaire and rested briefly. After viewing all specimens, they completed the post-test POMS questionnaire.

(5)

Conclusion: Participants received a gift and were debriefed. The total duration of the session was approximately 53 min (Figure 2).

2.5. Data Analysis

Participants were excluded based on Tobii Pro Lab (v1.232) criteria, which included those with poor data reliability, pupil detection rates below 80%, more than 20% incomplete or missing eye-tracking data, or EEG recordings containing excessive artifacts. Data from six participants were removed, resulting in a final sample of 22 subjects.

Twenty-two valid eye-tracking datasets were imported into Excel and SPSS 29.0 for statistical analysis. Five colors (blue, pink, white, mauve, and red), two inflorescence types (lacecap and mophead), and two petal structures (single and double) were treated as independent variables. Five oculometric parameters served as dependent variables: TFD, NB, APD, SF, and ASA. A one-way analysis of variance was used to examine the influence of hydrangea structural characteristics on eye-tracking metrics.

To control for type I error inflation associated with univariate testing, repeated-measures MANOVA was conducted to assess the main and interaction effects of hydrangea structural traits on eye-tracking parameters, with post hoc comparisons adjusted using Bonferroni correction. Paired t-tests were used to evaluate the overall differences between inflorescence types and petal types across oculomotor measures. Statistical significance was defined as p < 0.05. Effect sizes were reported using partial eta squared (η²) for MANOVA results and Cohen’s d for paired comparisons. All statistical tests met assumptions of normality and homogeneity of variance.

Electroencephalographic (EEG) metrics included alpha wave activity, beta wave activity, the alpha-to-beta ratio, and theta wave activity. The amplitude values of elevated alpha, elevated beta, and theta waves were averaged over 10-s epochs. Non-parametric Kruskal–Wallis H tests were employed to analyze differences in EEG signals across color conditions, with mean rank used to compare inter-color mean differences.

Twenty-two POMS questionnaires were preprocessed in Excel. Total Mood Disturbance (TMD) scores were calculated for both pre- and post-test measurements using the formula TMD = (T + D + A + F + C) − V, where the Vigor-Activity (V) subscale score was subtracted from the sum of the negative subscales. Emotional improvement (ΔTMD) was calculated as pre-test TMD minus post-test TMD. ΔTMD results were combined with eye movement parameters to understand the recovery effect of each sample through Pearson correlation analysis.

The SD method was used to assess participants’ subjective evaluations of the hydrangea specimens. Multiple comparison procedures were applied to analyze the mean scores and identify significant differences. The complete analytical workflow is detailed in Figure 3.

3. Results

3.1. Self-Reported Emotional Data

The internal consistency of the SD and POMS was assessed using Cronbach’s alpha. Both exhibited acceptable reliability (α = 0.723 > 0.70, SD; α = 0.849 > 0.80, POMS).

SD analysis showed that white and pink hydrangeas were associated with positive emotions (bright, soft, relaxing), whereas red and mauve hues clustered with negative terms (dark, harsh, anxious). Blue hydrangeas were linked to decay and instability. These results demonstrated that color acted as a potent stimulus, with white/pink and red/mauve spectra showing marked differences in emotional dimensions. Pink differed significantly from blue, white, and red on the beautiful–ugly dimension, while blue differed significantly from white and pink on the bright–dark dimension. Pink was perceived as more distinctive compared to mauve and red, whereas white was the most ordinary. Pink and red were described as warm colors, while blue was considered a cool color (Figure 4;

Table 4. Mean Scores and Significant Differences in the Hydrangea SD Questionnaire.

	Blue& White	Blue& Pink	Blue&Mauve	Blue& Red	White& Pink	White& Mauve	White& Red	Pink& Mauve	Pink& Red	Mauve& Red
Beautiful	1.697	2.121 **	1.924	1.682	2.121 **	1.924	1.697	4.045	2.121 **	1.924 *
Ugly	0.110	0.110 *	0.110	0.110	0.080	0.080	0.080	0.050	0.060	0.060
Bright	1.894 ***	1.818 ***	1.121	1.121	1.894	1.894 ***	1.894 ***	1.818 ***	1.818 ***	1.091
Dark	0.350 ***	0.350 ***	0.350	0.350	0.000	0.270 **	0.200 *	0.270 **	0.200 *	0.270
Special	1.242	1.773 **	1.485	1.303	1.773 ***	1.485 *	1.303	1.773	1.773 *	1.485
Ordinary	0.520	0.330	0.330	0.330	0.520 ***	0.520 **	0.520 **	0.210	0.210	0.210
Warmth	0.730	0.620 ***	0.800	1.380 ***	1.5 ***	0.800	1.380 ***	1.500 ***	1.500	1.380 ***
Cool	0.485	0.485 ***	0.530	0.485 ***	0.333 **	0.530	0.333	0.530 ***	0.152	0.530 ***
Pleasant	1.455	1.803 *	1.485	1.576	1.804 *	1.485	1.576	1.803	1.803	1.576
Unpleasant	0.200	0.200 *	0.200	0.200	0.080	0.110	0.080	0.110	0.080	0.110
Enrich	1.591 **	1.591	1.924	1.591	1.530 *	1.924 ***	1.470 *	1.924 *	1.530	1.924 *
Monotonous	0.380 *	0.180	0.180	0.180	0.380 *	0.380 **	0.380 *	0.150	0.180	0.180
Vitality	1.621	1.955 **	1.682	1.803 *	1.955	1.682	1.803	1.955	1.955	1.803
Decay	0.300 ***	0.300 ***	0.300 *	0.300 ***	0.060	0.170	0.060	0.170 *	0.050	0.170
Stable	1.636 **	1.546 *	1.182	1.515	1.636	1.636 *	1.636	1.546 *	1.546	1.515
Unstable	0.360 **	0.360 *	0.360 *	00.360 **	0.150	0.200	0.110	0.200	0.150	0.200
Soft	1.710 **	1.700 **	1.260	1.210	1.710	1.710 **	1.710 ***	1.700 **	1.700 ***	1.260
Rough	0.182	0.182	0.182	0.303	0.091	0.152	0.303 ***	0.152	0.303 **	0.303 *
Harmony	1.909 **	1.758 *	1.455	1.561	1.909	1.909 *	1.909 *	1.758	1.758	1.561
Disharmony	0.260 *	0.260	0.260 *	0.260 *	0.140	0.110	0.090	0.140	0.140	0.110
Relax	1.742 **	1.758 **	1.546	1.379	1.758	1.742	1.742 *	1.758	1.758 *	1.546
Anxious	0.150	0.150 *	0.150	0.150	0.060	0.090	0.060	0.090	0.060	0.090
Natural	1.773	1.727	1.697	1.697	1.773	1.773	1.773	1.727	1.727	1.591
Artificial	0.170	0.150	0.270	0.230	0.170	0.270	0.230	0.270	0.230	0.230

***. Significant at the 0.001 level (two-tailed); **. Significant at the 0.01 level (two-tailed); *. Significant at the 0.05 level (two-tailed).

).

Viewing hydrangeas significantly alleviated negative emotions, with particularly pronounced effects on reducing anxiety, hostility, and depression (

Table 5. Comparison of POMS Questionnaire Scores Before and After the Experiment.

	Mean ± SD	df	t	p	F	Cohen’s d
Pre-TMD	52.77± 32.318	26.239	5.606	<0.001	18.636	24.257
Post-TMD	11.77 ± 11.505

). ΔTMD was calculated as TMD (pre-test) minus TMD (post-test). Higher values indicated a greater cumulative restorative effect. The total pre-test score decreased from 1161 (mean = 52.77) to 259 (mean = 11.77) after the intervention, resulting in a mean recovery (ΔTMD) of 41 points. TMD was significantly reduced (t (26.239) = 5.60, p < 0.05). Significant reductions were observed in Tension-Anxiety (T, t (32.649) = 7.038, p < 0.001), Anger-Hostility (A, t (27.898) = 4.872, p < 0.001), Depression-Dejection (D, t (31.552) = 4.586, p < 0.001), Fatigue-Inertia (F, t (36.830) = 3.997, p = 0.006), and Confusion-Bewilderment (C, t (31.552) = 4.586, p < 0.001). In contrast, Vigor-Activity (V, t (42) = 0.667, p = 0.502) showed no significant change (Figure 5).

3.2. Restorative Benefits of Hydrangea Based on Eye Movement Analysis

Pearson correlation analysis revealed that eye-tracking parameters were significantly associated with the restorative benefits reported by participants. ΔTMD showed significant positive correlations with blink rate (NB) and TFD (p < 0.01), SF (p < 0.05), and ASA (p < 0.001). A significant negative correlation was found between ΔTMD and APD (p < 0.001) (all tests two-tailed) (

Table 6. Pearson Correlation Between Eye Movement Indicators and ΔTMD.

	Mean ± SD	df	t	p	F	Cohen’s d
ΔTMD	SF	330	41.000 ± 29.949	0.112 *	0.042	SF
	ASA	330	7.052 ± 1.322	0.294 **	<0.001	ASA
	TFD	330	110.994 ± 12.889	0.149 **	0.007	TFD
	NB	330	113.712 ± 45.098	0.158 **	0.004	NB
	APD	330	3.129 ± 0.301	−0.168 **	0.002	APD

**. Significant at the 0.01 level (two-tailed); *. Significant at the 0.05 level (two-tailed).

).

These results indicate that greater perceived attentional restoration (higher ΔTMD) corresponded to oculomotor patterns reflecting more relaxed, free, and engaged visual exploration. Specifically: (1) Increased blink rate (NB) and TFD are generally associated with lower cognitive load, reduced stress, higher relaxation, and sustained visual engagement, suggesting more comfortable and prolonged viewing behavior in individuals with better restoration. (2) Increased SF and amplitude (ASA) reflect more active and extensive visual exploration, indicating enhanced interest in exploration and greater stimulus engagement, which may facilitate attentional release. (3) The negative correlation with APD suggests that greater restoration is linked to pupil constriction, typically associated with reduced cognitive load, relaxed states, and responses to pleasant or low-arousal stimuli. In summary, individuals reporting higher subjective restoration demonstrated objective oculomotor patterns characteristic of relaxation, active engagement, and reduced tension or fatigue.

3.3. Physiological Data

Color systematically influenced visual attention patterns, with significant differences in APD detected among colors (F (4) = 4.437, p = 0.002). Specifically, red (p = 0.005) and mauve (p = 0.013) induced significantly larger pupil diameters than white and pink specimens (

Table 7. Analysis of Variance—Effects of Color on Eye Movement Metrics.

	N	Mean ± SD	df	F	p	η2
SF	330	1.062 ± 0.438	4	1.723	0.144	0.021
ASA	330	7.052 ± 1.322	4	1.594	0.176	0.019
TFD	330	110.994 ± 12.889	4	0.088	0.986	0.001
NB	330	113.712 ± 45.098	4	1.972	0.098	0.024
APD	330	3.129 ± 0.301	4	4.437	0.002	0.052

; Figure 6). These findings indicate that color, as a visual stimulus, systematically modulates observers’ physiological arousal and cognitive engagement. The significant differences in APD directly reflect varying intensities of neural activation and autonomic nervous system responses elicited by different colors. Specifically, red and mauve induced larger pupil diameters, suggesting higher physiological arousal and emotional activation, which aligns with their previously associated negative emotional descriptors such as “tense” and “anxious.” In contrast, white and pink corresponded to smaller pupil diameters, indicating lower arousal states and more relaxed emotional experiences, consistent with their positive emotional attributes such as “soft” and “relaxing.”

The study found significant interaction effects between color and structural characteristics on oculometric parameters. (1) Both blink count (NB) and SF were influenced by the color × inflorescence type interaction (NB: F (4) = 3.668, p = 0.006, ηp² = 0.045; SF: F (4) = 3.393, p = 0.010, ηp² = 0.018), with inflorescence type serving as the dominant factor (NB: F (1) = 7.838, p = 0.005, ηp² = 0.024; SF: F (1) = 5.755, p = 0.017, ηp² = 0.018). (2) Saccade amplitude (ASA) was modulated by both color × inflorescence type (F (4) = 5.365, p < 0.001, ηp² = 0.064) and color × structure (F (4) = 5.091, p < 0.001, ηp² = 0.061) interactions, with color (F (4) = 2.242, p = 0.065) emerging as the predominant factor. (3) APD remained unaffected by any color × structure interactions (

Table 8. Interactive Effects of Floral Structure, Color, and Petal Shape on Eye-Tracking Metrics.

Dependent Variable		Type III Sum of Squares	df	Mean Square	F	p	ηp2
NB	Clour	17,764.752	4	4441.188	2.332	0.056	0.029
	Inflorescence	14,924.291	1	14,924.291	7.838	0.005	0.024
	Clour × Inflorescence	27,939.845	4	6984.961	3.668	0.006	0.045
SF	Clour	1.504	4	0.376	2.039	0.089	0.025
	Inflorescence	1.061	1	1.061	5.755	0.017	0.018
	Clour × Inflorescence	2.502	4	0.626	3.393	0.010	0.041
ASA	Clour	14.570	4	3.643	2.241	0.065	0.028
	Inflorescence	5.323	1	5.323	3.274	0.071	0.010
	Structure	1.105	1	1.015	0.680	0.410	0.002
	Clour × Inflorescence	34.887	4	8.722	5.365	<0.001	0.064
	Clour × Structure	33.107	4	8.277	5.091	<0.001	0.061

) (See Appendix A

Table A2. Interactive Effects of Floral Structure, Color, and Petal Shape on Eye-Tracking Metrics.

Dependent Variable		Type III Sum of Squares	df	Mean Square	F	p	ηp2
NB	Modified Model	69,344.424 a	14	4953.173	2.601	0.001	0.104
	Clour	17,764.752	4	4441.188	2.332	0.056	0.029
	Inflorescence	14,924.291	1	14,924.291	7.838	0.005	0.024
	Structure	9.618	1	9.618	0.005	0.943	0.000
	Clour × Inflorescence	27,939.845	4	6984.961	3.668	0.006	0.045
	Clour × Structure	6705.973	4	1676.493	0.880	0.476	0.011
SF	Modified Model	5.317 b	14	0.380	2.059	0.014	0.084
	Clour	1.504	4	0.376	2.039	0.089	0.025
	Inflorescence	1.061	1	1.061	5.755	0.017	0.018
	Structure	0.013	1	0.013	0.069	0.793	0000
	Clour × Inflorescence	2.502	4	0.626	3.393	0.010	0.041
	Clour × Structure	0.827	4	0.207	1.121	0.347	0.014
ASA	Modified Model	63.499 c	14	4.536	2.790	<0.001	0.110
	Clour	14.570	4	3.643	2.241	0.065	0.028
	Inflorescence	5.323	1	5.323	3.274	0.071	0.010
	Structure	1.105	1	1.015	0.680	0.410	0.002
	Clour × Inflorescence	34.887	4	8.722	5.365	<0.001	0.064
	Clour × Structure	33.107	4	8.277	5.091	<0.001	0.061
APD	Modified Model	2.421 d	14	0.173	1.978	0.019	0.081
	Clour	1.617	4	0.404	4.619	0.001	0.055
	Inflorescence	3.682 × 10−5	1	3.682 × 10−5	0.000	0.984	0.000
	Structure	0.001	1	0.001	0.014	0.908	0.000
	Clour × Inflorescence	0.416	4	0.104	1.189	0.316	0.015
	Clour × Structure	0.508	4	0.127	1.451	0.217	0.018

a: R² = 0.104, b: R² = 0.084, c: R² = 0.071, d: R² = 0.081.

for the extended dataset).

A significant difference in blink count (NB) was found between inflorescence types (mophead vs. lacecap) (t (328) = −3.129, p = 0.002) (

Table 9. Differences in Flower and Petal Types in Eye-Tracking Data.

		N	Mean ± SD	df	F	t	p	Cohen’s d
NB	Mophead-type	220	108.290 ± 44.992	328	0.283	−3.129	0.002	44.507
	Lacecap-type	110	124.554 ± 43.517
	Single-type	110	108.500 ± 46.460	328	0.236	−1.487	0.138	45.015
	Double-type	220	116.318 ± 44.278

, Figure 7) (See Appendix A

Table A3. Differences in Flower and Petal Types in Eye-Tracking Data.

		N	Mean ± SD	df	F	t	p	Cohen’s d
NB	Mophead-type	220	108.290 ± 44.992	328	0.283	−3.129	0.002	44.507
	Lacecap-type	110	124.554 ± 43.517
	Single-type	110	108.500 ± 46.460	328	0.236	−1.487	0.138	45.015
	Double-type	220	116.318 ± 44.278
SF	Mophead-type	220	1.019 ± 0.438	328	0.154	−2.584	0.10	0.435
	Lacecap-type	110	1.15 ± 0.428
	Single-type	110	1.026 ± 0.456	328	0.237	−1.059	0.290	0.438
	Double-type	220	1.119 ± 0.440
ASA	Mophead-type	220	6.972 ± 1.327	328	0.006	−1.568	0.118	1.319
	Lacecap-type	110	7.212 ± 1.304
	Single-type	110	7.043 ± 1.250	328	1.310	−0.089	0.929	1.324
	Double-type	220	7.056 ± 1.359

for the extended dataset). The main findings were: (1) Among mophead-type inflorescences, red specimens elicited consistently higher values (red > pink > white > mauve > blue); (2) For lacecap-type inflorescences, mauve specimens evoked the highest blink counts (mauve > red > blue > pink > white); (3) Lacecap-type inflorescences produced higher blink counts than mophead-type, and double-petaled varieties exceeded single-petaled ones. No significant differences between structural types were detected for SF or saccade amplitude (ASA), though lacecap types consistently surpassed mophead types, and double-petaled varieties exceeded single-petaled ones. Across all specimens, the mauve lacecap hydrangea (M02) elicited stronger oculomotor responses than the red mophead hydrangea (R01) (Figure 8).

The raw data were converted into ranks, and statistical analysis was conducted based on the mean rank differences across color groups. Theta wave activity during the 70–80 s epoch showed significant variation (H = 9.980, p = 0.041, ε² = 0.035). Comparison of mean rank scores revealed that mauve (182.21) and blue (181.36) elicited higher values than white (151.29) and red (140.91). The α/β ratio varied significantly during the 90–100 s interval (H = 15.216, p = 0.004, ε² = 0.018), with white (197.17) showing significantly higher values than other color conditions (

Table 10. Differences in EEG α/β Waves and θ Waves Across Color Groups.

Wave	Time (s)	Mean ± SD	Median (p25, p75)	H	p	ε2	Clour	Mean Rank
α/β	90	1.851 ± 1.072	1.6500 (1.1250, 2.3400)	15.216	0.004	0.035	Blue	133.67
							Pink	162.14
							White	197.17
							Mauve	172.98
							Red	161.55
θ	70	141,715.840 ± 101,854.303	112,111.45 (77,792.47, 175,295.27)	9.980	0.041	0.018	Blue	181.36
							Pink	171.73
							White	151.29
							Mauve	182.21
							Red	140.91

).

4. Discussion

4.1. The Emotional Impact of Viewing Hydrangea

This study investigated the interactive effects of hydrangea structural traits—including inflorescence type (mophead/lacecap), petal type (single/double), and color (blue, pink, white, mauve, and red)—using live plant specimens. It aimed to explore how variations in color and structural characteristics might influence participants’ oculomotor behavior, EEG signals, and restorative outcomes.

Short-term exposure to hydrangeas significantly improved participants’ emotional states, evidenced by a 77.71% reduction in the total POMS score. This finding is consistent with prior research on the mood-enhancing effects of blue and purple flowers, particularly regarding the reported significant decreases in negative POMS subscales (T, A, F, D, C) and the increase in the Vigor (V) subscale [65]. However, the emotional benefits of hydrangeas may extend across the multiple colors tested, suggesting their restorative effect likely stems from the synergistic effect of integrated aesthetic features like color and form, rather than from the sole contribution of individual color attributes.

4.2. Color Effects on Emotional and Physiological Responses

Flower color was an important factor associated with restorative benefits. White hydrangeas induced the lowest stress levels and the highest α/β EEG ratio. Neurophysiologically, a higher α/β ratio indicates a marked relaxation state characterized by enhanced α-wave (8–13 Hz) activity and relative suppression of β-waves (13–30 Hz). This finding aligns with previous studies reporting that white flowers (e.g., white roses, white Kalanchoe) elicit positive emotions and increased relaxing brainwave activity [31,67], with white chrysanthemums noted for their calming properties [90]. Although some literature suggests its positive effects can be surpassed by other colors, white is generally considered a stable, low-burden color with positive emotional valence.

Plant coloration has been suggested to modulate human emotional states [91], though the same color may evoke varied experiences. For instance, red is commonly categorized as both a “high-arousal positive color” and a “warm color” [28]. This study observed that red induced high arousal and pupil dilation, interpreted as an alert state requiring cognitive resources. Previous studies indicate that red plants may be associated with excitement but also with negative affect. In their investigation of red gerberas, the flowers were linked to cheerfulness and excitement in office workers; they also correlated with elevated tension [92], and saturated red may elevate cortical arousal levels [93]. Pink hydrangeas are described as bright, soft, and relaxing. Such hues are typically associated with joyful and tranquil emotions, evoking positive feelings [92].

Blue and mauve hydrangeas elicit heightened physiological arousal (pupil dilation) and enhance theta wave (4–8 Hz) activity, which is typically associated with deep relaxation, subconscious activity, creativity, and memory processing. A dedicated study on hydrangeas similarly found that bluish-purple flowers increase alpha waves, linked to wakeful relaxation, mental calmness, and clarity of consciousness [65]. The enhancement of either alpha or theta waves shares a common physiological basis—the activation of the parasympathetic nervous system, which promotes bodily relaxation and recovery. Thus, blue and purple (mauve) flowers demonstrate beneficial effects on the restoration of attention and cognitive performance [90].

However, a difference noted is that while the cited research indicates blue elicits more pronounced changes in alpha waves compared to purple—with blue hydrangeas reducing skin conductance levels and evoking feelings of “cuteness” and “relaxation” [65]—the present study found blue to be weaker than purple, described as gloomy, decaying, and cold. This discrepancy may be related to multidimensional attributes of color, such as hue, saturation, and brightness, and their interactive effects on emotion [92]. For instance, while blue is often connected with natural elements such as the ocean and sky and is generally associated with relaxation and calmness, it may also be linked to associations with nighttime or overcast conditions, potentially leading to feelings of oppression [94]. Darker flowers with low brightness are generally less favored, whereas colors with higher brightness tend to elicit more positive responses [91].

4.3. Structural Effects and Interactive Influences

Plant morphology is a key variable driving differentiated responses in human perceptual preferences [95]. Naturally formed umbrella-type hydrangeas exhibit stronger restorative efficacy compared to regular globular varieties. This aligns with research proposing that irregular plant forms possess enhanced aesthetic appeal [96]. Studies exploring the perception of various trees found that pyramidal and fan-shaped trees are rated higher in visual quality and intensity, whereas isolated round trees are perceived as having lower visual quality [97].

Perceptual differences are shaped by multiple attributes including plant size, texture, and color [64,98], along with a plant’s unique structural diversity and dynamic color complexity. This study found that within globular hydrangeas, red most easily triggers attentional dispersion, followed by pink, while blue and mauve help maintain sustained attention. Conversely, within umbrella-type hydrangeas, mauve is most likely to cause distraction, followed by red, whereas pink and white are associated with greater attentional persistence. Overall, globular hydrangeas generally promote sustained attention, while umbrella-type hydrangeas more effectively facilitate attentional restoration and mental relaxation by reducing cognitive load. However, a study on peonies found that flower color had a greater influence on attentional concentration (beta waves) than petal type [64]. This discrepancy may stem from differences in measurement metrics and psychological constructs: eye-tracking fixation patterns primarily reflect early, automatic visual allocation, which is more influenced by overall contour (structure), whereas beta-wave-associated attentional concentration involves deeper cognitive appraisal, which is more readily modulated by the emotional-semantic attributes of color.

Different structural characteristics of monoecious flowering plants exerted different effects on restorative benefits. Among lacecap-type hydrangeas, mauve (especially the mauve lacecap hydrangea M02) showed the strongest restorative effect, while pink and white exhibited weaker effects. Overall, lacecap-type and double-flowered hydrangeas showed superior restorative effects compared to the other two structures. Hydrangea varieties with distinctive features and lower cognitive load yielded greater perceived restorative benefits.

4.4. Theoretical Interpretation (SRT, ART, Evolutionary Perspective)

These findings could be interpreted through three theoretical lenses. 1. Evolutionary perspectives: Human sensory mechanisms evolved through natural selection in natural environments [99], inherently possessing tendencies and desires to connect with nature and other life forms [100]. Human plant preferences were neither random nor entirely culturally determined but were substantially influenced by psychological mechanisms shaped during evolution. These mechanisms instinctively favored environmental features that signaled survival opportunities and well-being. 2. The SRT framework: The natural irregular distribution of branches, leaves, and flower clusters in lacecap-type hydrangeas mirrored natural woodland complexity. In contrast, the highly regular mophead-type spherical structure, while conveying “health” and “controllability” signals, also suggested artificial pruning and intervention (as indicated by SD questionnaire responses), making their stress-reduction effects less potent than more natural structures. 3. The ART perspective: Regarding fascination, the loose, organic, asymmetrical inflorescence structure of lacecap-type varieties more closely resembled natural vegetation, possessing greater organic “soft fascination.” Conversely, mophead-type fascination initially stemmed from regular geometric structures and rich coloration, creating strong but often transient appeal. Once the brain processed their regular patterns, exploration tended to conclude quickly, lacking the continuous detail revelation of natural structures. Concerning extent, the natural lacecap structure resembled a “miniature world,” drawing the gaze into detailed exploration spaces and offering rich, coherent visual scenes. The regular sphere, however, presented a closed, self-contained structure where the gaze glided across the surface but could not “enter” the interior, resulting in relatively flat, limited visual worlds that lacked the depth and detail of lacecap types. Mophead types were readily “seen through,” hindering sustained soft fascination. Regarding being away, natural irregular structures readily evoked distance from everyday environments defined by straight lines, right angles, and artificial order. For compatibility, the lacecap hydrangea structure did not force the gaze along a single path—attention flowed freely and compatibly with personal preferences—effortlessly capturing attention without inducing cognitive overload or psychological rejection, thereby promoting directed attention restoration and cognitive fatigue alleviation. Although no obvious difference in petal shape existed between double-petaled hydrangeas and their simple-structured single-petaled counterparts, the interaction between their visual features and human cognitive mechanisms made double-petaled varieties more likely to cause distraction. According to SRT, the complex layered structure and blurred boundaries of double hydrangeas prompted the brain to perform continuous, high-load analysis, making it difficult to quickly match a clear prototype. This cognitively demanding process captured involuntary attention, but this capture did not facilitate easy and pleasant recovery; instead, it tended to lead to expendable distraction. The issue lay not in the floral details themselves, but in the relatively chaotic and disordered presentation of details in double hydrangeas (compared to the harmonious and orderly presentation in single-petaled varieties), which contributed to distraction. Furthermore, the more complex structure of double hydrangea was not prevalent in nature; it was the result of human horticultural breeding, with its design inherently aimed at attracting and sustaining human attention.

4.5. Study Limitations

This study had limitations, including a small sample size and limited participant diversity, which restricted analysis of moderating variables like age and gender that may affect vegetation-restoration relationships [101]. Ecological validity was limited by brief laboratory sessions, and EEG accuracy was constrained by the single-channel device. The use of a single potted plant in a controlled indoor environment and a brief, fixed viewing duration may not fully capture the restorative experiences encountered in more complex, real-world settings. Furthermore, although plant selection aimed for visual consistency, inherent variations between individual specimens—such as subtle differences in flower count, color saturation, and growth form—could potentially influence participants’ perceptions and responses, representing an uncontrolled source of variability. Consequently, the results should not be directly generalized to outdoor environments, groupings of plants, different seasonal or lighting conditions, or longer-term exposures. Despite these limitations, this study provides scientific support for utilizing flowering plants as a low-cost, easily implementable health intervention, confirming their potential for stress reduction and mental health enhancement. Future research should explore the effects of different cultivars and colors of Hydrangea macrophylla, examine the impact of group plantings and combinations with other species, and validate findings in outdoor or more naturalistic settings with longer exposure times. Additionally, incorporating multisensory experimental designs and conducting long-term tracking in real landscape contexts will be valuable to advance evidence-based restorative environment design and implementation.

5. Conclusions

This study employed hydrangeas as a model system to investigate how structural characteristics—including color, inflorescence type, and petal structure—influence human well-being. Using a multimodal assessment framework that integrated subjective measures (POMS/SD questionnaires) with objective physiological monitoring (eye-tracking and EEG), we examined the interplay between floral pigmentation and structural attributes beyond conventional plant classifications. Results showed that hydrangea color and structure jointly affected restoration, with white and pink flowers promoting calmness, red and mauve inducing arousal, and natural, irregular structures such as lacecap enhancing attention restoration. These findings provided empirical evidence to guide plant selection in restorative landscape and horticultural therapy design.

In practical application, hydrangea selection could be tailored to different environmental needs based on these findings. In interior restorative environments such as healthcare facilities or retirement homes, varieties that reduced cognitive load and promoted relaxation—for example, mauve double-flowered lacecap hydrangeas—were particularly suitable. White or pink lacecap varieties also served as excellent choices for establishing tranquil atmospheres. For spaces requiring sustained concentration, such as offices or study areas, blue or mauve mophead hydrangeas that supported focused attention were preferable, while strongly arousing red mophead types could be used more selectively. In exterior restorative gardens and healing landscapes, naturally formed, soft-colored lacecap hydrangeas could be prioritized to maximize therapeutic potential. For vibrant commercial plazas where energy stimulation was desired, intensely colored red, blue, or mauve mophead hydrangeas were more appropriate. Optimal outcomes required consideration of both environmental functions and psychological requirements rather than assuming brighter coloration or greater flower density automatically improved effectiveness.