The Positive Effects of Viewing Gardens for Persons with Dementia

Abstract

Dementia is highly prevalent among the worldwide elderly population. Only a small number of the currently marketed drugs are effective in controlling its symptoms, and none has any effect on its progression. Further, as the condition advances, even these pharmaceuticals lose their efficiency, and new research into interventions that might improve the life quality of patients at the end stage of dementia and their families is increasingly rare. In our previous studies, we explored the benefits of exposure to nature, in the form of a Japanese garden, for persons with advanced dementia. In the current work, we extended our observations to two new locations and a new set of subjects with a different ethnic composition with the goal of identifying interventions that might improve their quality of life. We found that, even in these new settings, garden observation not only relieved physiological stress, it improved qualitative measures such as verbalization and memory retrieval. We present data that viewing the garden is a holistic experience rather a solely visual stimulus. Our new data further support the conclusion that garden observation is worth including in the care planning schedule of advanced dementia patients. Its low cost and easy availability make it an economical adjunct to current pharmacological methods that has the potential to improve the quality of life of people with dementia.

Keywords

Dementia heart rate Japanese garden memory retrieval stress relief

INTRODUCTION

Dementia is the progressive loss of memory and cognitive ability whose risk increases substantially with age [1, 2]. Alzheimer’s disease (AD) is the most common age-related dementia and, due to its prevalence and long disease course, it represents a substantial burden for the affected individuals as well as for their caregivers and the health system on which they are dependent [3]. Current pharmacological approaches to AD memory loss are based primarily on compounds that function as acetylcholinesterase inhibitors (rivastigmine, galantamine, and donepezil) or as a weak NMDA receptor antagonists (memantine). Beyond these agents, behavioral symptoms are managed with drugs including atypical antipsychotics, antidepressants, anxiolytics, and antiepileptics. All of these agents come with associated problems. They are not effective in all people, and often come with unwanted side effects [4, 5]. Recent attempts at disease modifying therapies involving immunotherapy targeting amyloid-β and tau deposition have been intensively explored. Anti-amyloid approaches such as inhibitors of the γ- and β-secretases are also in various stages of development.

From a societal perspective, it will take some time before a new pharmacology-based AD therapy is available for widespread, cost-effective usage. Memantine was the last drug approved by the United States FDA and that approval came almost 15 years ago. Of the more modern approaches involving anti-tau and anti-amyloid treatments, none has yet led to a successful trial in human individuals. Instead, the literature recounts more than two dozen failed Phase II and Phase III trials [6, 7]. In a disease as complex as AD, quick answers were never to be expected; but even if some of the many ongoing clinical trials succeed, the development of widely available therapies based on these strategies will require many years to fully implement. Compounding the problem still further, the complexity of the disease also leads to the prediction that any one therapy may not be effective in addressing all of the various co-morbidities including sleep problems, delusions, circadian rhythm disturbances, aggressiveness, depression, apathy, aberrant motor behavior, etc. [8, 9]. This complexity and the idea that AD may be more syndromal and heterogeneous that previously believed make it likely that there are multiple contributors to the disease and, therefore that multiple targets will need to be engaged simultaneously to provide protection or relief.

We believe that in this emerging disease landscape, with the emphasis on the multifactorial nature of AD etiology, the value of pursuing non-pharmacological approaches is greatly enhanced. They offer potential benefit to persons with dementia, whether used alone or combined with pharmacological agents. While such approaches are unlikely to block the underlying disease progression, they have the advantages of few side effects, potentially significant improvement in patients’ quality of life and convenience of use in a wide range of institutional and residential settings [10, 11]. The most common non-pharmacological approaches for treating dementia include physical exercise [1 , 12–14], music therapy [15 –20], foot massage [21], reminiscence [17, 22] and others [23, 24]. One activity that is consistently cited involves interaction with the natural world [25, 26], in particular horticulture therapy. Gardening activity has repeatedly been shown to be an effective and useful tool for persons with dementia [27 –32]. The ideas are continuous with the observations of Ulrich et al. [26] that a natural environment promotes healing more readily than an urban one. And have been quantitatively probed specifically in an institutional setting [33 –35]. One of the authors (Yutaka Hamano, Congress of the Japan Hospital Association, Tokyo, 2010) has reported on a particularly informative example of this idea. Subjects involved in a horticulture therapy program showed improvement in their MMSE score if they were simultaneously taking cholinesterase inhibitors plus memantine.

Although a “garden” is often thought of as an outdoor space, with ornamental plants, that people can walk through, Japanese gardens are an exception because they are designed to be viewed from indoors. The ultimate purpose of a Japanese garden is to tap into a viewer’s memory of nature by providing views of miniaturized scenery. For example, Ryoan-ji, the world-famous Zen garden in Kyoto, was composed with only fifteen small stones on white sand. The stones abstractly represent natural elements such as islands in the ocean or mountaintops above the clouds. A tea garden is another example as it is designed to evoke a mountain trail within a small landscaped space. Japanese gardens, designed for viewing miniaturized landscapes, have been used to calm the mind for hundreds of years and still attract thousands of visitors who wish to experience this effect. The advantage of this type of garden is that the observer stays indoors instead of walking outside for persons with reduced motor ability or who are wheelchair bound. Our previous studies in institutional settings in New Jersey, both assisted living and dementia care units, have shown that viewing a Japanese garden is effective in improving the quality of life for people with a wide range of cognitive capabilities [36 –38]. Studying the effects of a Japanese garden distinguishes our work from earlier studies, such as those of Gonzalez et al. [30], Detweiler et al. [27], and others [28] because the intervention requires no direct physical contact with, or movement within, the garden for subjects in our study. Our current study complements our earlier work in two additional dementia care facilities with two newly constructed garden installations. Along with a recently published companion study [39], this work has enabled us to extend our investigation to include subjects of different ethnic origin, while suggesting the complex visual stimulus of a Japanese garden may have effect on a person with dementia.

MATERIALS AND METHODS

The gardens

Two Japanese gardens were constructed for the experiment in each of the two main sites between April 15–April 30, 2015. The first site was the 7th floor roof top of Hamano Hospital (Hospital garden); the second was the back yard of Wakaba Terrace Nursing Home (Terrace garden). Both sites were relatively small (12’×20’), such as might be found in any long-term care facility, and both were renovated for a total cost of less than US$30,000 each. Before construction of the Hospital garden, the site was a wooden deck patio. The space was accessed from the main building by two ceiling-height glass doors. The deck was enclosed by a wooden fence, and at its back, between the fence and the exterior windowed wall, was an unstructured planting of shrubs and small trees (Fig. 1A). By contrast, the site of the Terrace garden was just outside of a Japanese-style living room. It was a simple lawn garden and beyond the lawn was a high concrete retaining wall. Behind the wall was a wild area covered with natural hillside vegetation. The area was viewed from the living room through ceiling-height sliding glass doors (Fig. 1D).

Fig. 1.

The appearance of the courtyards. Representative pictures of the courtyards before and after the constructions of the gardens. (A) and (D) illustrate the areas in Hamano Hospital (A) and Wakaba Terrace (D) before the gardens were installed. Note the similarity of the elements in each: water, green plants, and wood all present without deliberate design. (B) and (E) show the same areas in Hamano Hospital (B) and Wakaba Terrace (E) after the construction of the two gardens. (C) and (F) illustrate the two gardens with the glass doors closed. The size of each courtyard is 12’x20’. Each picture is similar to the actual view a participant would have while sitting in the observation room.

After preliminary measurements (see below), we installed gardens in both sites, designed in the classic style of a Japanese tea garden: a viewing garden evoking the sense of a mountain trail (Fig. 1B, E). The foreground of both gardens was composed of traditional Japanese tea garden elements that we used in previous experiments [38]: bamboo, a maple tree, azaleas, moss, gravel, stepping stones, a stone basin, and a stone lantern. Whereas the backdrop of the Hospital garden was bamboo along the wall, two plum trees and three sweet olive trees (Osmanthus fragrans)were planted in the Terrace garden. The plants were kept small and the path narrow to create the sense of miniaturized scenery in such a small space. There were no flowering plants—neither annual nor perennial flowers— whose size would be difficult to control and could potentially destroy the scale of the miniaturized scenery. The gardens required little maintenance beyond routine watering of the plants. These activities were all carried out by the institutional maintenance staff as they required little in the way of additional time.

Three tests

Three tests were conducted to assess the impact of viewing the landscaped space before and after the gardens were constructed. Test 1 (T1) was conducted for two weeks before the construction of the garden, between April 1-April 15 (Fig. 1A, D). We call this the “pre-garden” viewing session. After the construction of the garden in May, Test 2 (T2) was conducted for two weeks, from June 1 to June 15 (Fig. 1B, E). In order to avoid the hot and humid summer, Test 3 (T3) was conducted from October 19 to October 25. The procedure of T3 was identical to that of T2, except that the subjects faced the gardens with the glass doors closed (Fig. 1C, F). For all tests, the subjects were escorted to view the garden for fifteen minutes twice per week at approximately the same time, taking into account their sleeping patterns and planned daily activities. Visits were scheduled only for the daytime hours: 9:30 AM – 12:30 PM and 1:30 PM – 4:30 PM.

For each test, a researcher and a staff caregiver familiar to the subject monitored the subject’s heart rate and behavior. These people were given training sessions prior to the experiment. In order to make the experiment comfortable for the subjects, the staff member most familiar to the subject escorted them to the site. Once the subjects were sitting inside the building at the entrance to the garden, the caregiver placed the heart rate monitor on the subject’s finger; the lead was then attached to a laptop computer situated behind the subject (Fig. 2). Our previous work has shown [37, 38] that within no more than one or two minutes of being seated, a subject’s heart rate tends to be stable. The research assistant then turned on the monitor and asked the subject to observe the garden for fifteen minutes. The caregiver remained seated beside the subject and observed the subject’s expression. The research assistant and the caregiver noted any conversation and behavior of the subject; however, if the action of writing distracted the subject, the caregiver was instructed to stop. The caregiver was also trained not to encourage conversation but only to respond to the subject during the observation. After the session, the caregiver escorted the subject to the living quarters. After this, the caregiver completed the observation check list with the research assistant. The same staff member usually took care of the same subject for consistency, but for approximately 20% of the tests, other staff caregivers were substituted because of scheduling conflicts.

Fig. 2.

The position of the garden, subject, and caregiver. The research assistant (not visible in this scene) sat in front of the computer monitor behind the subject and the caregiver.

Subjects

Sixteen participants each for the Hamano Hospital garden and the Wakaba Terrace garden were recruited from patients and residents with a care-need level of 1–5 according to the standard of the Ministry of Health of Japan [38]. All participants had been at the institution fewer than three years. The average age of the thirty-two participants was 91±4, and their average score on the Mini-Mental State Examination (MMSE) score was 12±6. Six subjects in the Terrace garden and three subjects in the Hospital garden were capable of regular conversation, and five subjects in the Terrace and eight subjects in the Hospital could walk independently. Any person with an MMSE of less than 10 was considered to be suffering from severe dementia; scores between 10 and 18 were assigned to a moderate group; scores ranging from 19 to 23 were assigned as mild dementia. In both Hospital and Terrace, there were 4 severe, 7 moderate, and 4 mild subjects (there were no data on two subjects).

Before the experiment, each participant was interviewed about their willingness to participate in the study. When the subject’s dementia was so advanced as to preclude meaningful consent, we relied on their legal guardian. After obtaining consent, doctors evaluated the condition of the subject and confirmed their ability to be enrolled in the experiment [38]. The exclusion criteria were kept to a minimum and included only severe visual impairment, a pacemaker, or recent history of hospitalization. A medical chart review was used to determine basic demographic characteristics that included age, gender, MMSE, and other clinical notes related to medical conditions and cognitive status. Anonymized case-based records were prepared and used throughout. Summary information concerning the subjects is listed in Table 1. Due to changes in the clinical condition of the residents/patients, and the length of time required to complete the entire series of observations, not all subjects were able to complete all of the tests.

Table 1

Participant’s information. *subject who participated all tests Subjects whose ID number begins with “T” were Terrace residents; those whose number begins with “H” were from the Hospital. AD, Alzheimer’s disease

Terrace subject	Test(s) enrolled	MMSE before	MMSE after	Need care level	Reason of visit/stay	Mobility	Verbal ability	Diagnosis	Medication
T1*	T1-3	12	9	4	illness	wheelchair	seldom	hypoacousie	N/A
T2	T2-3	20	20	5	illness	crutch	normal	hypoacousie /AD	no medication
T3	T2-3	20	20	3	change nursing home	rollator	normal	AD	Donepezil
T4*	T1-3	7	6	5	dementia	wheelchair	seldom	Dementia / Lewy bodies	Rivastigmine 3.5mg/day
T5*	T1-3	7	6	4	stroke	wheelchair	yes/no only	stroke	Trazodone 50 mg/day, Cilostazol 50 mg/day, Rivastigmine 18 mg/day
T6*	T1-3	18	17	1	dementia	rollator	normal	AD	Donepezil 10 mg/day, Trazodone 50 mg/day, Etizolam 0.5 mg/day, Methylcobalamin 0.5 mg/day
T7	T1-3	8	10	2	dementia	crutch	seldom	AD	Donepezil 10 mg/day, Galantamine 20 mg/day
T8*	T1-3	13	14	3	dementia	walk	seldom	AD	Quetiapine 25 mg/day, Chlorpromazine 25 mg/day
T9*	T1-3	13	14	3	dementia	wheelchair	seldom	AD	Galantamine 15 mg/day
T10	T1	0	0	5	dementia	wheelchair	seldom	AD	Galantamine 20 mg/day, Rivastigmine
T11	T1-2	12	11	1	dementia	walk	non-verbal	stroke/AD	N/A
T12	T1-3	12	12	2	dementia	walk	seldom	AD	N/A
T13	T1-3	20	20	4	care	wheelchair	normal	stroke	N/A
T14	T1-3	12	11	1	hypertension	crutch	seldom	no	no medication
T15	T2-3	19	20	1	avoid living alone	walk	normal	AD	Haloperidol 0.75 mg/day, Galantamine 20 mg/day, Sodium valproate 100 mg/day
T16	T1-3	unknown	unknown	1	dementia	walk	normal	AD	N/A
Hospital subject	Test(s) enrolled	MMSE before	MMSE after	Need care level	Reason of visit/stay	Mobility	Verbal ability	Diagnosis	Medication
H1	T1	18	11	4	illness	wheelchair	seldom	AD	Galantamine
H2*	T1-3	0	0	5	illness	wheelchair	non-verbal	AD	Memantine, Donepezil
H3	T1-2	4	2	5	care	wheelchair	seldom	AD	Donepezil
H4*	T1-3	0	0	4	illness	wheelchair	non-verbal	senile dementia	Donepezil
H5	T1-2	13	12	2	illness	wheelchair	seldom	AD	Donepezil 5 mg/day
H6*	T1-3	21	23	5	knee pain	wheelchair	seldom	AD	no medication
H7*	T1-3	8	10	2	illness	walk	seldom	AD	Donepezil 5 mg/day
H8	T2-3	20	21	3	illness	walk	normal	AD	Donepezil 5 mg/day
H9*	T1-2	12	15	1	illness	wheelchair	seldom	AD	Donepezil 5 mg/day
H10	T1-3	11	14	3	illness	walk	non-verbal	AD	no medication
H11	T2-3	20	20	0	illness	walk	normal	AD	Galantamine 12 mg/day
H12	T1-3	13	17	1	care	walk	seldom	AD	N/A
H13	T1-2	20	20	1	care	walk	normal	AD	N/A
H14	T1-2	unknown	unknown	1	care	walk	seldom	AD	N/A
H15	T3	12	12	3	illness	wheelchair	non-verbal	AD	no medication
H16	T3	16	28	1	decreased appetite	walk	seldom	AD	Donepezil 10 mg/day

There was no significant difference in the average age, previous lifestyle, and education between subjects assigned to the Terrace or Hospital gardens; neither was there any significant difference in their MMSE scores. In the Terrace, five participants out of sixteen (two men and fourteen women) were born in rural areas but the current residences of all participants were in urban areas. Five participants had higher education; their past occupations were housewife (8), school teacher (4), employee (2), self-employed (1), and government officer (1). For both Test 1 and Test 2, the average age of all subjects was 90±5, and the average MMSE was 10±7. One man and seven women were enrolled for Test 3 with an average age of 92±5, and an average MMSE of 13±6. Among sixteen participants, ten (two men and eight women) could take all three tests; their average age was 92±6, and the average MMSE was 13±8.

In Hamano Hospital, seven participants out of sixteen (two men and fourteen women) were born and currently live in rural areas. Five participants had higher education, and their past occupations were housewife (8), school teacher (2), employee (2), self-employed (3), and government officer (1). There were eleven persons (two men and nine women) for Test 1. The average age of subjects in Test 1 was 90±5, and the average MMSE was 12±8. Two subjects dropped out because of illness and three new subjects were added for Test 2, and there were twelve persons (two men and ten women) for Test 2. The average age of subjects in Test 2 was 90±4, and the average MMSE was 12±8. Five subjects dropped out because of illness and two new subjects were added. One man and eight women were enrolled for Test 3 with an average age of 91±5, and an average MMSE of 13±6. Among sixteen participants, only six (one man and five women) could take all the tests; their average age was 91±6, and their average MMSE was 16±5 [40 –42].

Heart rate recording

Pulse rate is frequently used, over short periods of times, as an indication of physiological stress as it is responsive to changes in the subject’s emotional state [21 , 44]. While imperfect as a tool, measured by a simple finger-mounted plesthysmograph it is well suited for use with persons suffering from advanced dementia since, as their cognition declines, they become increasingly unable to respond to questions, either written or spoken, and react negatively to more invasive telemetry [38].

The plethysmograph monitor used was an Iworx IWX/404 and PT 100 model as described previously [38]. This equipment detects the slight changes in volume caused by the pulsing of blood through the finger vessels. A signal is sent to a laptop computer situated behind the subject, out of view. The pulse signals were recorded and analyzed by the software provided by the company. For subjects where the contact of the pressure sensitive gauge of the finger monitor was not optimal for the entire session, we manually scanned the traces to find the periods where a reliable pulse rate could be determined. Any test of a subject where reliable pulse data could not be gathered from at least 3 of the 15 one-minute bins was discarded. The protocol for the morning and afternoon sessions was the same. Average pulse rate was calculated in one-minute time periods. Non-parametric tests (Wilcoxon test) (Prism, Graph Pad software, Version 5) was applied to determine the significant of the differences due to small sample size and paired measurements.

Behavioral assessment

The method of behavior assessment was modified slightly from the one described previously [38]. For each session and for each subject, staff caregivers filled out the Behavioral Assessment Check List (Table 2) and noted any specific behavioral changes during the observation with the research assistant. All caregivers were instructed to make note of any verbalizations by the subjects and score the level of subjects’ attention to their environment during the experiment. Entries from these notes were classified into five categories: 1) positive or 2) negative garden-related comments, 3) neutral comments unrelated to the garden, 4) recall of long- or short-term memories, and 5) other – a category that included the utterances with no apparent meaning. Positive comments related to garden were scored 1 while negative comments, scored – 1; neutral comments or comments not directly related to garden scored 0. The final score for any type of comment was determined by summing the scores for each comment they gave. To quantitate the subjects’ responses to the garden and their caregivers, their emotion and their attention to subjects, each activity was given a score ranging from – 2 to +2; +2 represented a strong, continuous focus or positive attitude; 1 represented frequent, but not continuous focus, 0 represented moderate focus, – 1 represented occasional episodic focus, and – 2 represented a complete lack of focus or negative attitude. Informant-based assessment of the subjects was conducted to monitor their response to the garden. Four traits: 1) response to the garden; 2) response to the caregiver; 3) verbal expression; and 4) attention were considered. For the first three measurements, the subject was rated on a five-point scale from positive (2) through neutral (0) to negative (– 2). The numbers of objects they viewed were tallied for the fourth trait, attention. The responses were then reported as the average of the scores from the tests in each section. non-parametric tests (Wilcoxon test) (Prism, Graph Pad software, Version 5) were applied to determine the differences. Samples with p < 0.05 were considered statistically significant.

Table 2

The Behavioral Assessment Checklist (Japanese and English versions)

RESULTS

Pulse rate measurements

The baseline heart rates varied from one session to the next within and between subjects. Therefore, for each session, all values were normalized to the heart rate observed during the first one minute of the test (Fig. 3). Compared to the “pre-garden” space, the viewing of the Japanese garden had a significant effect. In the pre-garden spaces, the average pulse rate remained mostly unchanged during the entire 15-minute session. After the Japanese garden had been installed, however, the subjects’ average pulse rate dropped significantly in both Hospital as shown in Fig. 3A and Terrace as shown in Fig. 3B. In the Terrace, the garden intervention had significant effects in six of the one-minute bins. Grouping the one-minute time bins into three five-minute bins showed that the pulse rates in T2 were significantly lower than those in T1 (Fig. 3C, D). Merely sitting calmly appeared to provide a modest benefit for subjects with dementia as for subjects in both the hospital and the terrace, sitting quietly during T1 led to a slight downward trend in pulse rate over the 15-minute observation period (<10%). This decrease was not statistically significant (Fig. 3A, B, green line). By contrast, in T2 the subjects’ pulse rate was consistently and significantly reduced during each five-minute time bin.

Fig. 3.

Pulse rate during garden viewing. Pulse rate was monitored during T1 (green symbols) and T2 (orange symbols). Observations were made in the Hospital (A) and Terrace (B). The average pulse rate during each one-minute time bin was normalized to the pulse rate in the first minute (the pulse rate ratio). Each point is the mean pulse rate ratio of all subjects in all trials; the error bars represent standard error. The data in panels A and B were replotted using 5-minute instead of 1-minute time bin averages for both the Hospital (C) and the Terrace (D). The data in panels A and B during a single 5-minute time bin (11–15 minutes) for individuals grouped by cognitive status (mild, MMSE 19–23; moderate, MMSE 10–18; severe, MMSE < 10) in the Hospital (E) and the Terrace (F). In all graphs, the significance of different test was determined by non-parametric tests (Wilcoxon test), indicated with *p < 0.05; **p < 0.01; ***p < 0.001. Bonferroni posttest was performed to determine difference between tests for each timepoint/time bin or disease stage, indicated with ^#p < 0.05.

To determine whether the effectiveness of the garden intervention varied with cognitive status, the data were re-plotted based on the subjects’ MMSE scores [45]. Grouped in this fashion, we then compared the ratio of the pulse rates in the two environments during the final 5 minutes of the viewing session by which time the subjects should have fully adapted to the environment (Fig. 3E, F). The first and last five-minute blocks had similar ratios (data not shown). Although separating the subjects into three groups reduces the statistical power of the calculations, the data nonetheless suggest some interesting trends. In the Hospital, the effect of the garden was nearly equivalent at every stage of dementia (Fig. 3E). For the subjects living in the Terrace, however, the effects varied with the cognitive status of the subject (Fig. 3F).

Behavioral symptoms

The attitudes of people with dementia showed improvement with garden intervention, although not a dramatic one. Their responses to both the garden and their caregivers were significantly more positive when observing the garden (Fig. 4A, B, T1 compared with T2).

Fig. 4.

Garden influence on mood and attention. Residents of Wakaba Terrace were assessed by their caregivers, using a 5– point scale ranging from – 2 (negative response) to +2 (positive response). A) Subjects’ response to the garden. B) Subjects’ response to their caregivers. C) Coding of the positive or negative aspects of subjects’ utterances or facial expression during the observation period. D) Numbers of items/elements focused on by the subjects in the garden. Difference by non-parametric tests (Wilcoxon test), indicated with *p < 0.05 (n = 10). Scores were compiled from the responses to the ‘pre-garden’ (T1), the installed garden (T2) and the installed garden viewed through the closed glass doors (T3; see text for details).

Next, we assigned the comments a score and a “valence”— positive, neutral, or negative. The result shows that positive and negative comments both tended to be related to garden, while neutral comments were typically on other topics. For example, a comment about the weather was classified as a neutral comment rather than a positive one. In short, people had more positive comments in T2 than in T1 (Fig. 5A-D). While these effects were significant in both the Hospital and the Terrace via one-way ANOVA analysis, post hoc analysis suggested only a trend for the differences between groups in Hospital (Fig. 5C). Supporting the importance of this trend, we found that in T2 in the Hospital, not only were the number of comments greater, most of them were positive (Fig. 5C) – almost twice as high as the numbers found in the Terrace (Fig. 5D).

Fig. 5.

Analysis of subject comments. A, B) The sum of all subject comments during garden viewing in both the Hospital (A) and the Terrace (B). The contribution of positive (black) neutral (stippled) or negative (white) during each of the three tests is indicated. C, D) These histograms represent the average emotional ‘valance’ of the comments during the three tests. The Hospital scores are shown in (C) and the Terrace in (D). E, F) The total number of recalled memories verbalized by the participants during garden observation in both the Hospital (E) and the Terrace (F). Differences were determined by non-parametric tests (Wilcoxon test), indicated by *p < 0.05.

An additional observation was that the 15-minute duration of T1 was trying for the subjects, many of whom requested to leave the room early or fell asleep. By contrast, in T2 subjects often spontaneously told stories or reflected on their past lives. This phenomenon varied from individual to individual but the 15-minute session was not too long for most subjects in T2. The subject of the related event or experience was usually in the distant past, was related to nature, and notably the emotion attached to the memory was almost universally a positive one. Some subjects had such episodes in more than one session. In these cases, the subjects almost always recalled the same fragment of memory in each session. Quantification of these episodes revealed that the number of recalled events was higher in T2 than in T1 and the effect trended to be stronger and less variable in the Hospital (Fig. 5E) than in the Terrace (Fig. 5F).

Against this background from our T2 observations, the results of T3 stood out as unique. In both the Hospital and the Terrace, the number of comments dropped dramatically (Fig. 5A, B), particularly in the Hospital where the effect of the garden in T2 was large. Recall that the difference between T2 and T3 was primarily that the glass doors to the outdoor area were closed. Although the view of the garden remained largely (though not completely) unobstructed, with the doors closed the emotion behind the comments shifted toward a more negative tone. In the Hospital, the mean comment score actually became negative (purple bars – Fig. 5C), while in the Terrace, the comments remained positive but drifted downward. The effect of closing the doors was even more dramatic on the number of episodes of recall. In both Hospital and Terrace, these episodes essentially ceased.

These behavioral symptoms were reflected in the physiological stress experience as well. The pulse rate of persons viewing the gardens with the door closed was measured (Fig. 6) and compared with the previous values of the subjects who either viewed the garden with the doors open or the space before the Japanese garden was installed. As in Fig. 2, both 1-minute (Fig. 6A) and 5-minute (Fig. 6B) time bins were examined. The results were clear, but not as dramatic as the behavioral changes (Fig. 5). As in T2, the pulse rate dropped during the first few minutes of observation and then remained reduced for the duration of the session. With the doors closed, however, the pulse rate stabilized at a value that was intermediate between those in T1 and T2. This is most clearly seen in the 5-minute time bin plot (Fig. 6B).

Fig. 6.

The effect of a glass door between the garden and the subject. Pulse rate ratios were calculated for all subjects during the viewing of the gardens, but with the glass doors closed (Test 3 – T3). The dark purple points illustrate the data from the Hospital subjects; the light purple points are from the Terrace. The results from T1 (green) and T2 (orange) are replotted here for reference. A) Pulse rate ratios illustrated for one-minute time bins. B) Pulse rate ratios illustrated for 5-minute time bins. All data points represent the average of all subjects in all trials. The data from T1 and T2 are the same as shown in Fig. 2. They are replotted here for reference only. General difference between tests and/or time bins were determined by non-parametric tests (Wilcoxon test); *p < 0.05 when T3 compared to T2; ^#p < 0.05 when T3 compared to T1; ^&p < 0.05 when T2 compared with T1.

DISCUSSION

Our new findings validate and extend our previous studies with respect to the positive effect of viewing a garden designed along the principles of a traditional Japanese garden. The purpose of the design is to create an abstract representation of nature for encouraging meditative thought. For an individual with dementia, it would appear that that these effects result in an improvement of attitude, as suggested by our behavior measures and a calmer physiological state, as reflected in a sustained drop in their pulse rate. In our previous study involving Caucasian subjects in a long term care facility in New Jersey [38], we found that an indoor garden was able to lower this rate by 7–10%. The data presented here showed a comparable reduction among Terrace subjects, but in the Hospital the effect was more pronounced—over 15%. Given our sample size, this relative difference is less meaningful, but the consistent improvements seen in experiments carried out by two different groups of researchers in two different gardens measuring two different ethnic groups suggest that the finding is robust and ethnic origin would seem not to be a factor in the response.

With the caveat of sample size in mind, we also take note of the response differences we observed between the Terrace and the Hospital. We propose that one explanation for these differences is that the Hospital was located in the center of an urban environment with few if any traces of the natural world in its immediate vicinity. The Terrace, like the facility in New Jersey where we conducted our previous study [38], was located in a more rural setting where residents were exposed to nature on a more regular basis. This exposure provided a “background” benefit and hence could lead to the effects of the garden viewing intervention being less dramatic. Yet while Terrace responses were slightly less dramatic than those in the Hospital, there was nonetheless a significant effect on both pulse rate and attitude in the Terrace population. The implication is that even for persons in a supportive care environment with ample exposure to nature, and regardless of their ethnic background, the effects of the garden are still measurable and significant.

We have measured specific features of our subjects’ response to the garden, but there were also individual reactions that point to the limits of quantitative data to fully convey the responses we observed. We noticed that, in keeping with the function of a Japanese garden design to facilitate contemplation, the garden viewing in this study encouraged our subjects to tap into their personal memories. The specific episode that subjects recalled while in the garden depended upon their life history. One of the participants, for example, had been a tea master of the Japanese tea ceremony. Her mood was gloomy before visiting the garden, but once exposed to the view she recalled the process of the tea ceremony, and asked the staff to allow her to perform one, which she did for residents of the Terrace three months after her first garden visit. Another gentleman, who had served as an English translator for the US army before his retirement, had been hospitalized for eight months. After his doctors determined that further pharmacological therapy would not be helpful, they sent him back to the Terrace. As might be expected, when he first entered the facility, his mood was somber. Yet, when he viewed the garden, he talked about his career as an English translator and his pleasant memories of his experiences guiding American soldiers to visit gardens in Kyoto after the war. He told the staff that he wished to conduct an English class at the Terrace. His mood and physical condition both improved. These are only anecdotes, but all of the stories recalled during the garden visits were centered on similarly positive memories. And they are consistent with similar responses reported in our previous study [38]. These qualitative data, while difficult to measure, nonetheless begin to paint a picture of how the simple garden intervention we have described not only relieves stress, but also can also have a noticeable and positive impact on a subject’s quality of life. These findings are consistent earlier studies on the value of well-designed horticultural interventions.

Memory loss is one of the major symptoms of dementia, and treatments that improve memory or stimulate recall are highly desirable. Cholinesterase inhibitors such as galantamine [46] and rivastigmine [47, 48] offer this outcome, but their effects vary from individual to individual and strong side effects in the elderly population are not uncommon. In addition, several non-pharmacological interventions have proven noteworthy for their ability to enhance memory [49, 50]. For example, music intervention designed to improve a subject’s working memory of lyrics has been reported to be effective [50]. This type of intervention, however, led to short-term, highly focused improvements; long-term memories, such as those that define a personality, were unaffected. By contrast, the garden intervention we describe inspired spontaneous autobiographical recall, often of distant memories. These instances, however, also served to highlight the subjects’ continuing limitations. Thus, when a subject recalled an event in two different garden-viewing sessions, the event tended to be the same. Similarly, many subjects repeated the same story when they observed the garden in two different sessions. The narrow focus of memories evoked by the garden might expand with longer-term intervention, but it is also plausible that the complex visual stimulus of the garden is only able to tap a restricted pool of residual memory.

An additional noteworthy feature of the intervention is cost effectiveness. The total construction cost for each garden was less than US$30,000. This is a relatively modest price when compared with the $1000 required to keep a single patient on a cholinesterase inhibitor for one year (http://www.pharmacychecker.com/). This does not even account for the potential savings that might be realized, in staff time and medication cost, if psychotropic medicines could be reduced. Equally important, the intervention is small enough to be easily maintained [38]. The Japanese gardens at Hamano Hospital and the Terrace are watered and cleaned daily by staff members as part of their horticulture therapy program; no added personnel were required. Indeed, staff caregivers have begun to spontaneously bring their patients to the gardens when they find them ‘sundowning’ in the afternoon. They report that the subject’s mood improves in the garden and chats about it, and the positive effects sometimes last until bedtime or into the next day. For the garden to be effective, however, the subject must feel directly connected with it. Imposing even a glass door between the viewer and the garden reduces the positive effects.

Despite this limitation, as an adjunct to current care regimens, we believe that this non-pharmacological approach has great potential to provide a benefit to the millions of people who currently suffer from dementia and for whom there are few truly effective alternatives. This is especially attractive at a time when immunotherapeutic approaches aimed at reducing the amyloid burden in the brain have repeatedly failed to produce any clinically meaningful benefits [51, 52]. These pharmacological strategies will no doubt continue to be explored and some may yet succeed in their mission. Yet it is likely that their effectiveness will vary from person to person, given the heterogeneous nature of AD and a full cure will continue to elude us. In this regard the robust response to the garden viewing intervention we describe— in multiple installations and different ethnic groups—suggests that this relatively simple non-pharmacological approach represents an attractive component of a comprehensive AD treatment regimen.

Conclusions

The findings extend our earlier work on the impact of the complex visual stimulus of viewing a Japanese garden on physiological and behavioral symptoms of AD. The findings are consistent with the idea that the effects of this intervention are not restricted by culture or ethnicity. While cognitive status would appear to alter the response quantitatively, it does not eliminate it. Finally, the blunting of the effect when the viewing was through a glass door hints at the importance of the sense of immersion in the scene. We suggest that there is considerable potential in the further study and further development of this non-pharmacological therapy, in particular because it has proven effective even at very advanced stages of dementia.

Footnotes

ACKNOWLEDGMENTS

The authors wish to express their gratitude to Teppei Fukae for his tireless help with the logistics of this work. Support was provided by National Key Basic Research Program of China (2013CB530900), The Hong Kong University of Science and Technology, and the Research Grants Council, HKSAR (GRF660813, 16124916) to KH.

Authors’ disclosures available online ().

References

Ganguli

, Lee

, Snitz

, Hughes

, McDade

, Chang

(2015) Rates and risk factors for progression to incident dementia vary by age in a population cohort. Neurology 84, 72–80.

Querfurth

, LaFerla

(2010) Alzheimer’s disease. N Engl J Med 362, 329–344.

Schaller

, Mauskopf

, Kriza

, Wahlster

, Kolominsky-Rabas

(2015) The main cost drivers in dementia: A systematic review. Int J Geriatr Psychiatry 30, 111–129.

Schwarz

, Froelich

, Burns

(2012) Pharmacological treatment of dementia. Curr Opin Psychiatry 25, 542–550.

van de Glind

, van Enst

, van Munster

, Olde Rikkert

, Scheltens

, Scholten

, Hooft

(2013) Pharmacological treatment of dementia: A scoping review of systematic reviews. Dement Geriatr Cogn Disord 36, 211–228.

Godyn

, Jonczyk

, Panek

, Malawska

(2016) Therapeutic strategies for Alzheimer’s disease in clinical trials. Pharmacol Rep 68, 127–138.

Venigalla

, Sonego

, Gyengesi

, Sharman

, Munch

(2016) Novel promising therapeutics against chronic neuroinflammation and neurodegeneration in Alzheimer’s disease. Neurochem Int 95, 63–74.

Pinidbunjerdkool

, Saengwanitch

, Sithinamsuwan

(2014) Behavioral and psychological symptoms of dementia. J Med Assoc Thai 97(Suppl 2), S168–174.

Cerejeira

, Lagarto

, Mukaetova-Ladinska

(2012) Behavioral and psychological symptoms of dementia. Front Neurol 3, 73.

10.

Chen

, Liu

, Lin

, Peng

, Chen

, Liu

, Chen

(2014) Non-pharmacological treatment reducing not only behavioral symptoms, but also psychotic symptoms of older adults with dementia: A prospective cohort study in Taiwan. Geriatr Gerontol Int 14, 440–446.

11.

Kaplan

(1994) The role of nature in the context of the workplace. Landsc Urban Plan 26, 193–201.

12.

Brett

, Traynor

, Stapley

(2016) Effects of physical exercise on health and well-being of individuals living with a dementia in nursing homes: A systematic review. J Am Med Dir Assoc 17, 104–116.

13.

Garcia-Mesa

, Pareja-Galeano

, Bonet-Costa

, Revilla

, Gomez-Cabrera

, Gambini

, Gimenez-Llort

, Cristofol

, Vina

, Sanfeliu

(2014) Physical exercise neuroprotects ovariectomized 3xTg-AD mice through BDNF mechanisms. Psychoneuroendocrinology 45, 154–166.

14.

Ohman

, Savikko

, Strandberg

, Pitkala

(2014) Effect of physical exercise on cognitive performance in older adults with mild cognitive impairment or dementia: A systematic review. Dement Geriatr Cogn Disord 38, 347–365.

15.

Arroyo-Anllo

, Diaz

, Gil

(2013) Familiar music as an enhancer of self-consciousness in patients with Alzheimer’s disease. Biomed Res Int 2013, 752–965.

16.

Blackburn

, Bradshaw

(2014) Music therapy for service users with dementia: A critical review of the literature. J Psychiatr Ment Health Nurs 21, 879–888.

17.

Chang

, Chu

, Yang

, Tsai

, Chung

, Liao

, Chi

, Liu

, Chou

(2015) The efficacy of music therapy for people with dementia: A meta-analysis of randomised controlled trials. J Clin Nurs 24, 3425–3440.

18.

Craig

(2014) Music therapy to reduce agitation in dementia. Nurs Times 110, 12–15.

19.

Matthews

(2015) Dementia and the power of music therapy. Bioethics 29, 573–579.

20.

Vink

, Zuidersma

, Boersma

, de Jonge

, Zuidema

, Slaets

(2013) The effect of music therapy compared with general recreational activities in reducing agitation in people with dementia: A randomised controlled trial. Int J Geriatr Psychiatry 28, 1031–1038.

21.

Moyle

, Cooke

, Beattie

, Shum

, O'Dwyer

, Barrett

, Sung

(2014) Foot massage and physiological stress in people with dementia: A randomized controlled trial. J Altern Complement Med 20, 305–311.

22.

Gonzalez

, Mayordomo

, Torres

, Sales

, Melendez

(2015) Reminiscence and dementia: A therapeutic intervention. Int Psychogeriatr 27, 1731–1737.

23.

Abraha

, Rimland

, Trotta

, Dell'Aquila

, Cruz-Jentoft

, Petrovic

, Gudmundsson

, Soiza

, O'Mahony

, Guaita

, Cherubini

(2017) Systematic review of systematic reviews of non-pharmacological interventions to treat behavioural disturbances in older patients with dementia. The SENATOR-OnTop series. BMJ Open 7, e012759.

24.

Goris

, Ansel

, Schutte

(2016) Quantitative systematic review of the effects of non-pharmacological interventions on reducing apathy in persons with dementia. J Adv Nurs 72, 2612–2628.

25.

Ulrich

(1984) View through a window may influence recovery from surgery. Science 224, 420–421.

26.

Ulrich

, Simons

, Losito

, Fiorito

(1991) Stress recovery during exposure to natural and urban environments. J Env Psychol 11, 201–230.

27.

Detweiler

, Warf

(2005) Dementia wander garden aids post cerebrovascular stroke restorative therapy: A case study. Altern Ther Health Med 11, 54–58.

28.

Rivasseau Jonveaux

, Batt

, Fescharek

, Benetos

, Trognon

, Bah Chuzeville

, Pop

, Jacob

, Yzoard

, Demarche

, Soulon

, Malerba

, Bouvel

(2013) Healing gardens and cognitive behavioral units in the management of Alzheimer’s disease patients: The Nancy experience. J Alzheimers Dis 34, 325–338.

29.

Rivasseau Jonveaux

, Fescharek

(2018) When art meets gardens: Does it enhance the benefits? The Nancy Hypothesis of care for persons with Alzheimer’s disease. J Alzheimers Dis 61, 885–898.

30.

Gonzalez

, Kirkevold

(2014) Benefits of sensory garden and horticultural activities in dementia care: A modified scoping review. J Clin Nurs 23, 2698–2715.

31.

Whear

, Coon

, Bethel

, Abbott

, Stein

, Garside

(2014) What is the impact of using outdoor spaces such as gardens on the physical and mental well-being of those with dementia? A systematic review of quantitative and qualitative evidence. J Am Med Dir Assoc 15, 697–705.

32.

Thompson Coon

, Boddy

, Stein

, Whear

, Barton

, Depledge

(2011) Does participating in physical activity in outdoor natural environments have a greater effect on physical and mental wellbeing than physical activity indoors? A systematic review. Environ Sci Technol 45,1761–1772.

33.

Detweiler

, Self

, Lane

, Spencer

, Lutgens

, Kim

, Halling

, Rudder

, Lehmann

(2015) Horticultural therapy: A pilot study on modulating cortisol levels and indices of substance craving, posttraumatic stress disorder, depression, and quality of life in veterans. Altern Ther Health Med 21, 36–41.

34.

Detweiler

, Murphy

, Kim

, Myers

, Ashai

(2009) Scheduled medications and falls in dementia patients utilizing a wander garden. Am J Alzheimers Dis Other Demen 24, 322–332.

35.

Detweiler

, Murphy

, Myers

, Kim

(2008) Does a wander garden influence inappropriate behaviors in dementia residents? Am J Alzheimers Dis Other Demen 23, 31–45.

36.

Goto

, Fritsch

(2011) A pilot study of seniors' aesthetic preferences for garden designs. J Acad Soc Japanse Garden 24, 1–12.

37.

Goto

, Park

, Tsunetsugu

, Herrup

, Miyazaki

(2013) The effect of garden designs on mood and heart output in older adults residing in an assisted living facility. HERD 6, 27–42.

38.

Goto

, Kamal

, Puzio

, Kobylarz

, Herrup

(2014) Differential responses of individuals with late-stage dementia to two novel environments: A multimedia room and an interior garden. J Alzheimers Dis 42, 985–998.

39.

Stephen

, Hongisto

, Solomon

, Lonnroos

(2017) Physical activity and Alzheimer’s disease: A systematic review. J Gerontol A Biol Sci Med Sci 72, 733–739.

40.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

41.

Giordani

, Boivin

, Hall

, Foster

, Lehtinen

, Bluemlein

, Berent

(1990) The utility and generality of Mini-Mental State Examination scores in Alzheimer’s disease. Neurology 40, 1894–1896.

42.

Ideno

, Takayama

, Hayashi

, Takagi

, Sugai

(2012) Evaluation of a Japanese version of the Mini-Mental State Examination in elderly persons. Geriatr Gerontol Int 12, 310–316.

43.

Lupis

, Sabik

, Wolf

(2016) Role of shame and body esteem in cortisol stress responses. J Behav Med 39, 262–275.

44.

Jones

, Chopra

, Grummitt

, Ravindran

, Matthews

, Levitan

(2015) High reactivity of the cortisol awakening response predicts positive treatment outcome in heterogeneous depressed patients completing an alternate milieu inpatient program. Gen Hosp Psychiatry 37, 601–605.

45.

Mungas

(1991) In-office mental status testing: A practical guide. Geriatrics 46, 54–58, 63, 66.

46.

Ohnishi

, Sakiyama

, Okuri

, Kimura

, Sugiyama

, Saito

, Takahashi

, Kobayashi

(2014) The prediction of response to galantamine treatment in patients with mild to moderate Alzheimer’s disease. Curr Alzheimer Res 11, 110–118.

47.

Farlow

, Cummings

, Olin

, Meng

(2010) Effects of oral rivastigmine on cognitive domains in mild-to-moderate Alzheimer’s disease. Am J Alzheimers Dis Other Demen 25, 347–352.

48.

Ismail

, Elmeshad

, Salem

(2013) Potential therapeutic effect of nanobased formulation of rivastigmine on rat model of Alzheimer’s disease. Int J Nanomedicine 8, 393–406.

49.

Hescham

, Lim

, Jahanshahi

, Blokland

, Temel

(2013) Deep brain stimulation in dementia-related disorders. Neurosci Biobehav Rev 37, 2666–2675.

50.

Simmons-Stern

, Deason

, Brandler

, Frustace

, O'Connor

, Ally

, Budson

(2012) Music-based memory enhancement in Alzheimer’s disease: Promise and limitations. Neuropsychologia 50, 3295–3303.

51.

Doody

, Thomas

, Farlow

, Iwatsubo

, Vel-las

, Joffe

, Kieburtz

, Raman

, Sun

, Aisen

, Siemers

, Liu-Seifert

, Mohs

, Alzheimer’s Disease Cooperative Study Steering Committee, Solanezumab Study Group (2014) Phase 3 trials of solanezumab for mild-to-moderate Alzheimer’s disease. N Engl J Med 370, 311–321.

52.

Salloway

, Sperling

, Fox

, Blennow

, Klunk

, Raskind

, Sabbagh

, Honig

, Porsteinsson

, Ferris

, Reichert

, Ketter

, Nejadnik

, Guenzler

, Miloslavsky

, Wang

, Lu

, Lull

, Tudor

, Liu

, Grundman

, Yuen

, Black

, Brashear

, Bapineuzumab 301 and 302 Clinical Trial Investigators(2014) Two phase 3 trials of bapineuzumab in mild-to-moderate Alzheimer’s disease. N Engl J Med 370, 322–333.