The effectiveness of dance movement interventions for older adults with mild cognitive impairment, Alzheimer’s disease, and dementia: A systematic scoping review and meta-analysis

Abstract

Objectives:

To synthesize evidence and summarize research findings related to the effectiveness and feasibility of dance movement intervention (DMI) in older adults with mild cognitive impairment (MCI), Alzheimer’s disease (AD), and dementia; to systemically map existing research gaps and research directions for future practice.

Methods:

A systematic search was conducted using six electronic databases: Web of Science, PubMed, PsycINFO, MEDLINE, ScienceDirect, and Cochrane Central Register of Controlled Trials. The methodological quality of included studies was assessed using the Cochrane Risk of Bias Tool for Randomized Trials (RoB 2) and The Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I).

Results:

29 dance intervention studies (13 RCT studies) were included in the scoping review: 62% of MCI, 10% of AD, and 28% of dementia; a total of 1708 participants (Female=1247; Male=461) aged from 63.8 (± 5.24) to 85.8 (± 5.27) years old. Eight RCT studies were included in the meta-analysis; results indicated that dance interventions had a significant effect on global cognition, memory, balance, and significantly decreased depression. No significant effects were found for executive function.

Conclusions:

Dance is a non-pharmacological, effective, affordable, and engaging intervention that can be used as a complementary treatment for older adults with MCI, AD, and dementia.

1. Introduction

1.1. Background

Mild cognitive impairment (MCI) is an intermediate stage between healthy older adults and dementia that affects 10%−15% of people aged over 65 years (Anderson, 2019). Hypoperfusion and hypometabolism in temporoparietal cortices are neurobiological characteristics of MCI. In addition, there is medial temporal lobe atrophy, notably in rhinal cortices, decreased Aβ42 and high and phosphorylated tau in cerebrospinal fluid and brain deposition (Anderson, 2019). Diverse etiologies of MCI, presenting as systemic diseases, neurological diseases, and psychiatric problems, result in a wide range of results. Moreover, uncontrolled cardiovascular factors, depression, and polypharmacy are risks for cognitive impairment (Kenneth and Langa, 2014). Mild cognitive impairment phenomenology includes decreased quality of life, increased symptoms of depression, and withdrawal from social involvement. The prevalence of MCI in community dwellers across the world is over 15% (Bai et al., 2022). Around 30–40% of people with MCI, especially those experiencing difficulty with memory, may develop Alzheimer’s diseases (AD) and other dementias than people without MCI within five years of diagnosis (Kantarci et al., 2009; Mitchell and Shiri-Feshki, 2009; Ward et al., 2013). According to the recommendations and criterion to diagnose AD, MCI may be classified as an early stage of AD or other types of dementia under certain circumstances (Gaugler et al., 2016). Mild cognitive impairment might be the best time to implement pharmacological and non-pharmacological interventions to delay progression of AD and dementia (Lazarou et al., 2017). The increasing frequency and tendency toward AD and dementia suggest that effective treatments are required to prevent future decrease in cognitive function with older adults (Chang et al., 2021).

Alzheimer’s disease is an illness that causes cell degeneration in the brain and is responsible for around 60–80% of dementia cases (2020 Alzheimer’s disease facts and figures, 2020); it is characterized by a deterioration in thinking and independence in daily tasks (Breijyeh and Karaman, 2020). Alzheimer’s disease is a multifaceted disease with two primaries hypothesized causes: the cholinergic and amyloid(Breijyeh and Karaman, 2020). Briefly, the cholinergic hypothesis describes a malfunction of acetylcholine-containing neurons in the brain contributing significantly to the cognitive decline seen in people with advanced age and AD. The amyloid hypothesis suggests that cell-to-cell communication is disrupted and immune cells are activated during the stages of beta-amyloid accumulation (Breijyeh and Karaman, 2020). The activated immune cells cause inflammation, which leads to the destruction of brain cells. Furthermore, the condition is influenced by various risk factors, including increased age, genetic variables, head injuries, vascular diseases, infections, and environmental factors (Breijyeh and Karaman, 2020). Alzheimer’s disease is a degenerative syndrome, which means that it deteriorates over time (2020 Alzheimer’s disease facts and figures, 2020; Gaugler et al., 2016). and is likely to begin at least twenty years before symptoms occur, with undetectable alterations in the brain of the individual affected (2020 Alzheimer’s disease facts and figures, 2020). Diagnosing AD requires a careful and comprehensive medical evaluation. The evolution of AD progresses from imperceptible brain alterations to memory impairments and, eventually, physical incapacity (2020 Alzheimer’s disease facts and figures, 2020). However, dementia is an umbrella term for a specific set of symptoms. Dementia symptoms include deficits of memory impairment, language issues, problem-solving, and other conceptual skills, which can impair the ability of a person to conduct daily activities (2020 Alzheimer’s disease facts and figures, 2020). Although doctors can tell if a person has dementia, pinpointing the exact reason can be challenging (Gaugler et al., 2016). Except for AD, the common causes of dementia are also related to Frontotemporal lobe degeneration, Parkinson’s disease, Lewy body disease, Hippocampal sclerosis, Cerebrovascular disease and Mixed pathologies (2020 Alzheimer’s disease facts and figures, 2020; Gaugler et al., 2016). The World Health Organization (WHO) has designated dementia as a worldwide public health concern and has urged national governments to address the growing issues at global, national, and local levels (Chiesi et al., 2021). The global population living with dementia in 2015 was 46.8 million. This number is estimated to rise to 74.7 million by 2030 and 131.5 million by 2050 (prince, martin, 2015). Additionally, the majority of informal caregivers are family members; the yearly global amount of informal care hours provided to people with dementia living at home was around eighty-two billion hours, or six hours each day in 2015 (Wimo et al., 2018). Dementia is a major health concern in aging communities, with considerable costs to individuals and society for medical and informal care (Ho et al., 2020).

There are currently no pharmacological treatments (medications) that can cure MCI, AD, or dementia (Chan et al., 2020). Non-pharmacological therapies do not include the use of medication; they are normally used for patients with the purpose of sustaining or increasing cognitive function, general quality of life, or the ability to conduct daily living tasks (2020 Alzheimer’s disease facts and figures, 2020) with low-cost, low-risk, easy-to-use characteristics (Alwan et al., 2023; Baker et al., 2021; Tao et al., 2023). For example, aerobic activity and anaerobic exercise have been found in previous studies to improve cognitive function, and they may reduce the speed of cognitive deterioration in individuals with AD or dementia (Groot et al., 2016; Ho et al., 2020). Non-pharmacological therapies for violence and agitation in people with dementia also appear to be more successful than pharmaceutical interventions for symptom reduction (Watt et al., 2019).

Dance activity has been recommended previously to postpone cognitive decline, reduce agitation, enhance mood, improve joy, and boost social contacts (Tao, Gao et al., 2021; Tao, Gao, Li et al., 2022; Yamada and Kawano, 2021). Dance movement intervention (DMI) can stimulate the brain areas involved in memory, executive function, and motor skills. Dance improvisation, for example, makes use of frontal lobe resources, while movement relies on visual tracking and path integration (Tao, Supriya et al., 2021). The mechanisms may engage the entorhinal and hippocampal networks involved in spatial memory, which is severely impaired in cognitive decline and dementia (Ho et al., 2020). Dance has also been shown to be beneficial for Parkinson’s disease, as a result of music providing an external reference point for movements and the repetitive exercise strategy improving balance, flexibility, and endurance abilities (Barnish and Barran, 2020; Carapellotti et al., 2020; Ismail et al., 2021; Simpkins and Yang, 2023).

1.2. Objectives

It is unclear what kind of information is available in the literature about the effectiveness and feasibility of DMI used in conjunction with MCI, AD, and dementia treatments and rehabilitations. Therefore, this study’s objectives were:

1. To comprehensively map the research that has been completed in this field; and to identify any current knowledge omissions by drawing conclusions from existing literature.

2. To synthesize evidence about the efficiency and feasibility of DMI for older adults with MCI, AD, and dementia; and to perform a meta-analysis using eligible RCT studies. As a result, a mechanism for summarizing and communicating research findings to policymakers and practitioners who may lack time or resources to examine the information themselves will be provided.

3. To consider the research directions of DMI for MCI, AD and dementia treatment and rehabilitation for future practice.

The specific questions to be addressed in the review were:

• What are the involved studies’ characteristics (design and location of study, dance intervention type, main outcomes etc.)?

• Which types of dance intervention are more popular and effective for older adults with MCI, AD, and dementia in included studies?

• How to identify the effectiveness of DMI (outcomes, measurements, methods) for controlling or reversing MCI, AD, and dementia conditions?

• Has DMI any positive effects on physiological, psychological, and psychiatric aspects for participants?

• How satisfied are participants and caregivers with dance classes?

• Are there any risks and adverse events during the dance interventions process?

2. Methodology

2.1. Protocol and registration

This scoping review adhered to the guidelines provided from the Joanna Briggs Methods Manual for Scoping Review (Peters et al., 2015; Tricco, Cardoso et al., 2016) and followed the PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation (see Supplementary Table 1 https://osf.io/cve7u) (Tricco et al., 2018). The study protocol was developed and registered on the Open Science Framework at https://osf.io/fe8my/ on 25th February 2023 prior to commencing this review. The meta-analysis section followed The PRISMA Statement for Reporting Systematic Reviews and Meta-analyses of Studies That Evaluate Health Care Interventions: Explanation and Elaboration (see Supplementary Table 1 https://osf.io/cve7u) (Liberati et al., 2009). The protocol was registered on PROSPERO (CRD 42023432363) on 4th June 2023 before data extraction procedures.

2.2. Eligibility criteria

2.2.1. Inclusion criteria

A search of the literature was carried out using both PICOS and SPIDER tools for quantitative and qualitative studies to comply with our research aims (Methley et al., 2014). Studies were included if they matched the following criterion: 1) Population and Sample: participants diagnosed with MCI, AD, or dementia; no restrictions were made for stage of disease, age and sex, cultural background, geographic location, or race of the participants. 2) Intervention and Phenomenon of Interest: studies mainly evaluating dance intervention, or other types of intervention to compare with dance effectiveness as the primary outcome target; the dance interventions were required to have a clear structure; no restrictions were made for dance type chosen, intervention duration and frequency, and intervention setting (e.g., hospital, care home, rehabilitation center, or patient’s home). 3) Comparisons and Design: no restriction was made for comparison groups. 4) Outcome and Evaluation: studies were required to measure and explain outcomes linked to our research aims, whether assessed by self-report or objective measurement by clinicians, researchers, or caregivers. 5) Study design and Research type: quantitative or qualitative primary/original research; only RCT studies were eligible for inclusion in the meta-analysis.

2.2.2. Exclusion criteria

1) Studies that included participants who were at risk of MCI, AD or dementia. 2) Video-based dance intervention or other virtual intervention. 3) Studies not published in English, or not in a peer-reviewed journal. 4) Reviews, secondary analysis, conference articles, opinion articles, commentaries, and policy documents. 5) Studies for which the authors could not be reached to obtain full texts. 6) More than one article described the same experiment and presenting duplicate data (the latest or quantitative study was included).

2.3. Information sources

An initial PubMed search was conducted to discover relevant papers. This ensured validity of the presented hypothesis and confirmed search items. The purpose of this stage was to avoid duplication of addressed questions and to ensure that there were adequate articles available to undertake the study. In the initial search, no scoping review and meta-analysis focusing on DMI in older adults with AD, MCI, and dementia was found. Following the initial exploration, a comprehensive systematic search was carried out independently by two authors using six electronic databases on the 26th February 2023. PubMed, Web of Science, PsycINFO, MEDLINE, ScienceDirect, and the Cochrane Central Register of Controlled Trials were explored to retrieve all relevant literature published in peer-reviewed journals between 26th February 2003 and 26th February 2023. The final search outcomes were exported to EndNote 20 where duplicate items were removed.

2.4. Search

MeSH terms (‘dance’ OR ‘dancing’) AND (‘mild cognitive impairment’ OR ‘Alzheimer’s disease’ OR ‘dementia’) were used to search the published studies. Candidate articles were selected by screening words in titles and abstracts. Furthermore, the reference lists of the eligible articles and previous similar review studies were examined to find additional potentially appropriate publications that were not discovered during the electronic databases exploration.

2.5. Selection of sources of evidence

The titles and abstracts were initially screened to classify records. If the eligibility conditions were met, the full texts were obtained for further screening. To increase consistency among reviews, two independent reviewers were involved in each phase of the screening process. Disagreements on research selection were solved through agreement and conversation with another senior author if necessary.

2.6. Data charting process

Based on earlier research recommendations, the review team created and pilot-tested a data charting form for recording pertinent information from included articles (Arksey and O’Malley, 2005). Three reviewers classified the eligible studies, addressed differences, and then finalized the data items throughout the pilot preparation. After the final version of the data charting form was completed, two authors individually plotted the information and then iteratively adjusted the data form.

2.7. Data items

The following items were abstracted: summary of the included studies characteristics (e.g., year of publication, author, study location, study aim and population, study design and setting, dance intervention details, outcome measurements, and important results); numerical summary of the studies’ characteristics (i.e., diseases categories, year of publication, study location, gender characteristics, age, and study design); summary of dance intervention protocol details (e.g., dance intervention type and duration, training load and monitoring, sample size, intervention setting and instructor, attrition rate, and assessment time point); summary of main outcomes and instruments across studies; and existing research limitations and future research directions. The meta-analysis further analyzed global cognition, executive function, memory, balance, and depression outcomes.

2.8. Critical appraisal of individual sources of evidence

The methodological quality of the 13 RCT studies was assessed using the Cochrane Risk of Bias Tool for Randomized trials (RoB 2) by Cochrane Review Manager 5.4.1 software (Julian PT Higgins et al., 2022). Risk of bias were divided into three categories: Low risk of bias, High risk of bias, and Unclear risk of bias. The rest of the sixteen non-RCT studies’ quality were assessed using the Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool (Jonathan et al., 2022) by robvis (online visualization tool). Five judgements including Low risk of bias, Moderate risk of bias, Serious risk of bias, Critical risk of bias and No information were evaluated for bias domains. Two authors independently evaluated the included studies; any disagreements were resolved in consultation with a third author.

2.9. Synthesis of results

The findings were interpreted using quantitative and qualitative analysis methods to represent the existing literature. Firstly, fundamental numerical analysis was prioritized for the studies included in this review, which produced tables and graphs. Secondly, a narrative data synthesis from the key results was completed. Thirdly, a meta-analysis was undertaken with eight RCT studies following Cochrane guidance (Deeks et al., 2022) to improve precision and provide more evidence to answer the research questions. Review Manager (version 5.4.1) was used to manage and analyze the extracted data from the various studies. The meta-analysis includes outcomes measured in at least two studies. Mean difference (MD) and standard deviation (SD) were used to calculate the effectiveness of the dance intervention. Because of possible clinical heterogeneity, a random-effect model was used. Chi-squared test and the I² statistic were used to examine the heterogeneity of the studies. Heterogeneity was interpreted using an I² index: 0%–40% might not be important; 30%–60% may represent moderate heterogeneity; 50%–90% may represent substantial heterogeneity; and 75%–100% represents considerable heterogeneity (Cochrane Handbook, n.d.). Forest plots were generated to observe effects, confidence intervals, and the weight in each study. A funnel plot was not employed to assess publication biases, because the quantity of studies considered for meta-analysis were fewer than ten. In addition, to determine the causes of heterogeneity, subgroup analyses and sensitivity analyses were performed.

3. Results

3.1. Selection of sources of evidence

From the literature search, a total of 672 articles were retrieved. Following the removal of duplicates (n = 381), title and abstract screening (n = 297), full-text screening (n = 45) and manual search (n = 5), twenty-nine articles were included in this study. The reasons for excluding articles are reported in the PRISMA flow diagram below (see Fig. 1).

Fig. 1.

PRISMA Flow Diagram of Screening and Selection Process.

3.2. Characteristics of sources evidence

Supplementary Table 2 (https://osf.io/awqmz) summarizes the characteristics of the included studies. In total, twenty-nine studies focused on three disease types: mild cognitive impairment (MCI) (Adam et al., 2016; Ammar et al., 2021; Bisbe et al., 2020; Blumen et al., 2022; Bracco et al., 2023; Chang et al., 2021; Doi et al., 2017; Dominguez et al., 2018; Douka et al., 2019; Kropacova et al., 2019; Lazarou et al., 2017; Parial et al., 2022; Qi et al., 2018; Thumuluri et al., 2022; Wang et al., 2020; Zhao et al., 2021; Zhu et al., 2018, 2022), Alzheimer’s diseases (AD) (Abreu and Hartley, 2013; Chiesi et al., 2021; Yamada and Kawano, 2021), and dementia (Choo et al., 2019; Debbie Duignan, Lynne Hedley, 2009; Ghadiri et al., 2022; Guzmán et al., 2016; Guzmán-García et al., 2013; Hamill et al., 2012; Ho et al., 2020; Koh et al., 2020) (see Fig. 2); 1708 participants were included (Female = 1247; Male = 461) aged from 63.8 (± 5.24) to 85.8 (± 5.27) years old.

Fig. 2.

Diseases Categories of Selected Studies.

Included studies were published from 2009 to 2023. There were three years where more studies were published than other years for this research topic: these were 2019 (Choo et al., 2019; Douka et al., 2019; Koh et al., 2020; Kropacova et al., 2019; Qi et al., 2018; Wang et al., 2020), 2021 (Ammar et al., 2021; Chang et al., 2021; Chiesi et al., 2021; Yamada and Kawano, 2021; Zhao et al., 2021), and 2022 (Blumen et al., 2022; Ghadiri et al., 2022; Parial et al., 2022; Thumuluri et al., 2022; Zhu et al., 2022) (see Fig. 3). One article from 2023 was excluded from the summary of publication trends, because the literature search ended on 26th February 2023; therefore, we did not include any studies published after this date. The research designs of the included studies consisted of five methodologies: randomized control trial (RCT) (Ammar et al., 2021; Bisbe et al., 2020; Blumen et al., 2022; Chang et al., 2021; Doi et al., 2017; Ho et al., 2020; Kropacova et al., 2019; Lazarou et al., 2017; Parial et al., 2022; Qi et al., 2018; Wang et al., 2020; Zhu et al., 2018, 2022), quasi-experimental study (Adam et al., 2016; Dominguez et al., 2018; Douka et al., 2019; Ghadiri et al., 2022; Koh et al., 2020; Zhao et al., 2021), pilot study (Choo et al., 2019; Debbie Duignan, Lynne Hedley, 2009; Guzmán-García et al., 2013; Hamill et al., 2012; Thumuluri et al., 2022), case study (Abreu and Hartley, 2013; Guzmán et al., 2016; Yamada and Kawano, 2021), and multicenter intervention study (Bracco et al., 2023; Chiesi et al., 2021) (see Fig. 3).

Fig. 3.

Publication Trend by Year and Research Design.

The dance intervention topic was popular among Chinese researchers. There are seven studies that were conducted in China (Chang et al., 2021; Ho et al., 2020; Qi et al., 2018; Wang et al., 2020; Zhao et al., 2021; Zhu et al., 2018, 2022); Fig. 4 displays the studies’ geographical distribution. A numerical summary of the involved studies’ characteristics is outlined in Supplementary Table 3 (https://osf.io/vhbrf).

Fig. 4.

Geographical Location of Publications.

3.3. Critical appraisal within sources of evidence

3.3.1. RCT studies

The methodological quality of the thirteen RCT studies was evaluated using the RoB 2 tool. Two studies were graded with “Overall high risk of bias”; the rest of the RCT studies showed a high quality of methodology study design, operation, and reporting issues (see Fig. 5). Regarding the “Other bias (Self-reported outcomes)” item, if the study combined objective and subjective measurements it was considered as “Low risk of bias”; if the study only used subjective (self-reported) measurements it was considered as “High risk of bias”. As a dance intervention cannot be blinded to the participants, the “High risk of bias” with “Blinding of participants and personnel” can be accepted. Reasons for absence of adherence across studies were reported, such as medical problems (Ammar et al., 2021; Blumen et al., 2022; Chang et al., 2021; Doi et al., 2017; Kropacova et al., 2019; Wang et al., 2020; Zhu et al., 2022), changed residence (Bisbe et al., 2020; Parial et al., 2022; Wang et al., 2020; Zhu et al., 2022), loss of interest or refusal to participate (Bisbe et al., 2020; Blumen et al., 2022; Chang et al., 2021; Doi et al., 2017; Lazarou et al., 2017; Parial et al., 2022; Zhu et al., 2022), time constraints (Blumen et al., 2022; Parial et al., 2022), and missed testing sessions (Ammar et al., 2021). Around 38% of studies had “High risk of bias” results with “Incomplete outcome data”. If the attrition rate did not affect data analysis this was considered as “Low risk of bias”.

Fig. 5.

Risk of Bias Summary - RCT Studies.

3.3.2. Non-RCT studies

The sixteen non-RCT studies were assessed using the ROBINS-I tool. Two studies were graded with an “Overall serious risk of bias”; the rest of the non-RCT studies demonstrated good quality (see Fig. 6). Some studies were graded with “Serious risk of bias” in “Bias due to confounding” item, as they compared healthy people and MCI participants (Douka et al., 2019), or compared mild, moderate and advanced dementia participants (Choo et al., 2019). These confounders are more likely to be associated with a beneficial outcome for some groups. Some other “Serious and Moderate risk of bias” were found in “Bias in measurement of outcomes” because of the study design: for example, case studies and pilot studies with no control group that cannot achieve a blinding measurement (Abreu and Hartley, 2013; Bracco et al., 2023; Chiesi et al., 2021; Debbie Duignan, Lynne Hedley, 2009; Guzmán et al., 2016; Guzmán-García et al., 2013; Hamill et al., 2012; Yamada and Kawano, 2021). Our study aimed to represent evidence related to our research aims, therefore we included all relevant studies in the scoping review. This strategy adhered to normal scoping review methodologies (Peters et al., 2015; Tricco, Lillie et al., 2016).

Fig. 6.

Risk of Bias Summary - Non-RCT Studies.

3.4. Results of individual sources of evidence

3.4.1. Dance intervention type

A total of twenty types of dance interventions were used in the included studies. For eighteen MCI studies, the most used interventions were Aerobic dance (Bisbe et al., 2020; Qi et al., 2018; Zhu et al., 2018, 2022), Ballroom dance (Blumen et al., 2022; Doi et al., 2017; Dominguez et al., 2018; Lazarou et al., 2017), and Chinese square dance (Chang et al., 2021; Wang et al., 2020; Zhao et al., 2021). These three types of dance are also popular with healthy people across society, and can be conducted under many different environmental conditions with people from different cultural backgrounds. Some traditional or local dance forms such as Greek traditional dance (Douka et al., 2019), poco-poco dance (Adam et al., 2016), Tango with choreography (Kropacova et al., 2019), Zumba gold dance (Parial et al., 2022), and Fitness and dance training programmes (Thumuluri et al., 2022) were used once only. One study used improvisational dance (Thumuluri et al., 2022); this may be a challenge for participants who have no dancing experience. Regarding the three AD studies, three dance types were included: Salsa dance (Abreu and Hartley, 2013), Biodanza dance (Chiesi et al., 2021), and Traditional bon dance (Yamada and Kawano, 2021). Among the eight dementia studies, Danzón Latin ballroom dance (Guzmán et al., 2016; Guzmán-García et al., 2013) was used twice; local feature dance such as Iranian dance (Ghadiri et al., 2022), Circle dance (Hamill et al., 2012), Wu Tao dance (Debbie Duignan, Lynne Hedley, 2009) were used once respectively; Creative dance (Koh et al., 2020), Modified dance programme (Ho et al., 2020), and Intuitive movement re-embodiment programmme (Choo et al., 2019) also were used once (see Supplementary Table 4 https://osf.io/pdftg).

3.4.2. Training load

Intervention durations ranging from 6-week (Adam et al., 2016) to 100-week (Yamada and Kawano, 2021) were used. Most of the researchers conducted 12-week (Bisbe et al., 2020; Bracco et al., 2023; Parial et al., 2022; Qi et al., 2018; Wang et al., 2020; Zhu et al., 2018, 2022) intervention durations and found positive results. Dance classes lasted from 30-minute sessions (Chang et al., 2021) to 120-minute sessions (Chiesi et al., 2021); most of the classes lasted 60-minute (Abreu and Hartley, 2013; Adam et al., 2016; Bisbe et al., 2020; Blumen et al., 2022; Bracco et al., 2023; Choo et al., 2019; Doi et al., 2017; Dominguez et al., 2018; Douka et al., 2019; Ho et al., 2020; Kropacova et al., 2019; Lazarou et al., 2017; Parial et al., 2022; Thumuluri et al., 2022; Zhao et al., 2021) and were managed twice per week (Ammar et al., 2021; Bisbe et al., 2020; Blumen et al., 2022; Dominguez et al., 2018; Douka et al., 2019; Guzmán et al., 2016; Guzmán-García et al., 2013; Ho et al., 2020; Thumuluri et al., 2022) or three times per week (Chang et al., 2021; Ghadiri et al., 2022; Kropacova et al., 2019; Parial et al., 2022; Qi et al., 2018; Wang et al., 2020; Zhu et al., 2018, 2022). Only nine studies monitored training load during the dance classes; the methods included: the target heart rate established at 60%–80% of maximal heart rate (Qi et al., 2018; Zhao et al., 2021; Zhu et al., 2018, 2022), controlled exercise heart rate of 100–140 beats per minute (Chang et al., 2021), based on levels of perceived exertion not surpassing 2–3 points on the Borg Scale (1–10) (Bisbe et al., 2020), systolic blood pressure staying at ≤ 150 mmHg, while the perceived exertion not greater than ≤ 14 on the Borg scale (6–20) (Blumen et al., 2022), or using portable heart rate monitors to keep a consistent level of activity throughout exercise (40%−60% of VO_2max value) (Ghadiri et al., 2022; Ho et al., 2020), the rest of the studies did not monitor participants’ training load. Since the review included all study designs, the sample size showed large differences. From one participant (Abreu and Hartley, 2013; Yamada and Kawano, 2021) in the case studies to 207 participants (Dominguez et al., 2018) in a quasi-experimental study; three studies used the same sample size (sixty participants) (Douka et al., 2019; Parial et al., 2022; Zhu et al., 2018). Numbers in dance classes ranged from one (in two case studies) (Abreu and Hartley, 2013; Yamada and Kawano, 2021) to sixteen (Doi et al., 2017; Zhu et al., 2022). Study settings were different, such as in hospital (Bisbe et al., 2020; Choo et al., 2019; Ghadiri et al., 2022; Zhu et al., 2018), nursing home (Bracco et al., 2023; Chang et al., 2021; Ghadiri et al., 2022; Wang et al., 2020), community center (Doi et al., 2017; Dominguez et al., 2018; Parial et al., 2022), outpatient clinic (Abreu and Hartley, 2013), day care center (Chiesi et al., 2021; Debbie Duignan, Lynne Hedley, 2009; Douka et al., 2019; Koh et al., 2020), care home (Guzmán-García et al., 2013), group home (Yamada and Kawano, 2021), university (Kropacova et al., 2019), government residential homes (Adam et al., 2016), dance studio (Thumuluri et al., 2022), or outdoor square (adjacent community) (Zhao et al., 2021). The attrition rate ranged from 0% (Abreu and Hartley, 2013; Adam et al., 2016; Chiesi et al., 2021; Debbie Duignan, Lynne Hedley, 2009; Douka et al., 2019; Guzmán-García et al., 2013; Hamill et al., 2012; Thumuluri et al., 2022; Yamada and Kawano, 2021) to 57.1% (Ammar et al., 2021); only five studies (Ammar et al., 2021; Blumen et al., 2022; Bracco et al., 2023; Ghadiri et al., 2022; Guzmán et al., 2016) observed a more than 20% attrition rate, thus showing a generally high adherence for the dance intervention classes. There were four types of assessment schedule: the basic model was pre- and post-test (Adam et al., 2016; Bisbe et al., 2020; Blumen et al., 2022; Bracco et al., 2023; Chiesi et al., 2021; Choo et al., 2019; Debbie Duignan, Lynne Hedley, 2009; Doi et al., 2017; Dominguez et al., 2018; Douka et al., 2019; Ghadiri et al., 2022; Hamill et al., 2012; Kropacova et al., 2019; Lazarou et al., 2017; Qi et al., 2018; Zhu et al., 2022); three studies did pre-test, middle point test, and post-test (Abreu and Hartley, 2013; Chang et al., 2021; Wang et al., 2020); six studies conducted pre- and post-test, and a follow-up assessment after the post-test (Abreu and Hartley, 2013; Ammar et al., 2021; Guzmán et al., 2016; Parial et al., 2022; Zhao et al., 2021; Zhu et al., 2018); one study carried out pre- and post-test, and two follow-up assessments after the post-test (Ho et al., 2020). Dance intervention protocol details are summarized in Supplementary Table 4 (https://osf.io/pdftg).

3.4.3. Measurement instruments

In MCI studies, ten measurement tools were most commonly used; these included the Geriatric Depression Scale (GDS) (Blumen et al., 2022; Chang et al., 2021; Dominguez et al., 2018; Douka et al., 2019; Lazarou et al., 2017; Parial et al., 2022; Thumuluri et al., 2022; Wang et al., 2020; Zhao et al., 2021; Zhu et al., 2018, 2022), the Mini-Mental State Examination (MMSE) (Adam et al., 2016; Bisbe et al., 2020; Bracco et al., 2023; Doi et al., 2017; Douka et al., 2019; Lazarou et al., 2017; Qi et al., 2018; Wang et al., 2020; Zhu et al., 2022), the Montreal Cognitive Assessment (MoCA) (Chang et al., 2021; Dominguez et al., 2018; Kropacova et al., 2019; Parial et al., 2022; Qi et al., 2018; Wang et al., 2020; Zhao et al., 2021; Zhu et al., 2018, 2022), the Trail Making Test (TMT) (Bisbe et al., 2020; Doi et al., 2017; Douka et al., 2019; Lazarou et al., 2017; Parial et al., 2022; Qi et al., 2018; Zhu et al., 2018, 2022), the Wechsler Memory Scale-Revised Logical Memory test (WMS-RLM) (Bisbe et al., 2020; Kropacova et al., 2019; Qi et al., 2018; Zhu et al., 2018, 2022), Magnetic Resonance Imaging (MRI) scanning (Blumen et al., 2022; Kropacova et al., 2019; Qi et al., 2018; Thumuluri et al., 2022; Zhu et al., 2022), the Forward and backward Digit Span Task (DST) (Parial et al., 2022; Qi et al., 2018; Zhu et al., 2018, 2022), the Neuropsychiatric Inventory (NPI) (Dominguez et al., 2018; Douka et al., 2019; Lazarou et al., 2017; Thumuluri et al., 2022), the Berg Balance Scale (BBS) (Bisbe et al., 2020; Chang et al., 2021; Qi et al., 2018; Wang et al., 2020), and the Quality of life in Alzheimer’s Disease (QOL-AD) (Adam et al., 2016; Bracco et al., 2023; Douka et al., 2019; Thumuluri et al., 2022). In Dementia studies, four studies were conducted via interviews (Choo et al., 2019; Guzmán et al., 2016; Guzmán-García et al., 2013; Koh et al., 2020); two of the most popular measurement tools were the Quality of Life Scale in Alzheimer’s Disease (QoL-AD) (Hamill et al., 2012; Koh et al., 2020), and the Mini-Mental State Examination (MMSE) (Guzmán et al., 2016; Hamill et al., 2012). In the AD studies, every measurement was only used once (see Supplementary Table 5 https://osf.io/uf9wt).

Mild cognitive impairment studies commonly explored cognitive function (Adam et al., 2016; Ammar et al., 2021; Bracco et al., 2023; Chang et al., 2021; Doi et al., 2017; Dominguez et al., 2018; Douka et al., 2019; Kropacova et al., 2019; Qi et al., 2018; Zhao et al., 2021; Zhu et al., 2018), depression (Adam et al., 2016; Chang et al., 2021; Dominguez et al., 2018; Douka et al., 2019; Lazarou et al., 2017; Parial et al., 2022; Wang et al., 2020; Zhao et al., 2021; Zhu et al., 2018, 2022), executive function (Doi et al., 2017; Douka et al., 2019; Kropacova et al., 2019; Lazarou et al., 2017; Parial et al., 2022; Qi et al., 2018; Zhu et al., 2018, 2022), memory performance (Doi et al., 2017; Douka et al., 2019; Kropacova et al., 2019; Lazarou et al., 2017; Parial et al., 2022; Qi et al., 2018; Zhu et al., 2018, 2022), quality of life (Adam et al., 2016; Bisbe et al., 2020; Chang et al., 2021; Douka et al., 2019; Parial et al., 2022; Thumuluri et al., 2022; Zhu et al., 2018, 2022), attention (Doi et al., 2017; Douka et al., 2019; Kropacova et al., 2019; Lazarou et al., 2017; Qi et al., 2018; Thumuluri et al., 2022; Zhu et al., 2018), balance status and falls risk (Bisbe et al., 2020; Blumen et al., 2022; Chang et al., 2021; Qi et al., 2018; Thumuluri et al., 2022; Wang et al., 2020), functional capacity (Bracco et al., 2023; Dominguez et al., 2018; Douka et al., 2019; Lazarou et al., 2017), and global cognition (Bisbe et al., 2020; Parial et al., 2022; Wang et al., 2020; Zhu et al., 2022). In the dementia studies, cognitive function (Debbie Duignan, Lynne Hedley, 2009; Guzmán et al., 2016; Hamill et al., 2012; Ho et al., 2020), quality of life (Choo et al., 2019; Hamill et al., 2012; Koh et al., 2020), physical fitness performance (Ghadiri et al., 2022; Koh et al., 2020), behavioral and psychological symptoms of dementia (Guzmán et al., 2016; Ho et al., 2020), memory (Choo et al., 2019; Ho et al., 2020), and mood (Guzmán et al., 2016; Ho et al., 2020) were the common outcomes. Regarding the AD studies, they also paid attention to physical fitness performance, mental status (Abreu and Hartley, 2013), neuropsychiatric symptoms, cognitive function, agitation level, depression (Chiesi et al., 2021), and wisdom (Yamada and Kawano, 2021) (see Supplementary Table 5 https://osf.io/uf9wt).

3.4.4. Key outcomes

Dance interventions were found to have positive effects on physiological, neuropsychological, and psychiatric parameters in older adults with MCI, AD, and dementia. Regarding physiological aspects, participants in dance intervention groups showed an increase in quality of life (Adam et al., 2016; Bracco et al., 2023; Chang et al., 2021; Choo et al., 2019; Douka et al., 2019; Koh et al., 2020; Parial et al., 2022; Thumuluri et al., 2022; Wang et al., 2020), balance ability (Abreu and Hartley, 2013; Chang et al., 2021; Koh et al., 2020; Thumuluri et al., 2022), gait, distance walked and speed (Abreu and Hartley, 2013; Ghadiri et al., 2022; Koh et al., 2020), right hippocampal volume and total hippocampal volume (Zhu et al., 2022), amplitude of low-frequency fluctuation (ALFF) in several brain areas (Qi et al., 2018), variety of brain network properties including global efficiency and modularity (Thumuluri et al., 2022), diurnal cortisol slope (Ho et al., 2020), mobility (Parial et al., 2022), cardiac adaptations (Ammar et al., 2021), figural fluency (Kropacova et al., 2019), physical condition (Douka et al., 2019), motion (Abreu and Hartley, 2013), strength (Abreu and Hartley, 2013), and daily functioning (Ho et al., 2020). As for neuropsychological outcomes, participants in dance intervention groups demonstrated enhancement in memory (Bisbe et al., 2020; Doi et al., 2017; Lazarou et al., 2017; Qi et al., 2018; Zhu et al., 2018, 2022), cognition (Adam et al., 2016; Doi et al., 2017; Dominguez et al., 2018; Lazarou et al., 2017; Parial et al., 2022; Zhao et al., 2021), immediate and delayed recall (Bisbe et al., 2020; Lazarou et al., 2017; Parial et al., 2022), executive function (Blumen et al., 2022; Douka et al., 2019; Kropacova et al., 2019), attention (Douka et al., 2019; Lazarou et al., 2017), emotional states (Debbie Duignan, Lynne Hedley, 2009; Guzmán-García et al., 2013), processing speed (Qi et al., 2018; Zhu et al., 2018), verbal fluency (Douka et al., 2019; Lazarou et al., 2017), leaning ability (Lazarou et al., 2017), concentration and communication (Hamill et al., 2012), and wisdom (Yamada and Kawano, 2021). For psychiatric aspects, participants in dance intervention groups decreased in depression (Adam et al., 2016; Chang et al., 2021; Dominguez et al., 2018; Ho et al., 2020; Lazarou et al., 2017; Wang et al., 2020; Zhao et al., 2021), negative mood (Guzmán et al., 2016; Hamill et al., 2012; Ho et al., 2020), anxiety (Adam et al., 2016; Douka et al., 2019), agitation (Chiesi et al., 2021), loneliness (Ho et al., 2020), and positive effects on neuropsychiatric behaviors (Chiesi et al., 2021) (see Supplementary Table 2 https://osf.io/awqmz).

However, there were several outcomes that changed in a favorable direction but were not statistically significant in some studies. These included depressive symptoms, working memory, blood pressure, waist circumference, body mass index (Parial et al., 2022), physical fitness performances, functional capacity (Bracco et al., 2023), and nonmemory cognitive tests (Doi et al., 2017). The baseline hippocampus-to-cortex volume (HV: CTV) ratio was not correlated with cognitive changes (Kropacova et al., 2019); less hippocampal atrophy was observed following dance intervention, but there was no statistical difference (Blumen et al., 2022). There was no adverse event reported across included studies, except one participant in a case study who had a fall during week-five of dance intervention; however, this resulted in no severe injuries, and at the time of discharge, the participant had not reported further falls (Abreu and Hartley, 2013).

Follow-up assessment showed outcome parameters in dance intervention groups increased gradually with time in three studies (Ho et al., 2020; Parial et al., 2022; Zhao et al., 2021). However, some follow-up tests indicated that the benefits achieved after dance intervention were returning to baseline level (Ammar et al., 2021; Zhu et al., 2018). This may be due to decreased training intensity or lack of participation in dance practice. It is essential to maintain dance practice and to achieve moderate intensity following the dance intervention classes (see Supplementary Table 2 https://osf.io/awqmz).

Another meaningful result was observed from caregivers during the dance interventions. The finding of Guzman’s study discovered that a dance intervention improved both participants’ and caregivers’ positive emotional status and overall levels of satisfaction (Guzmán-García et al., 2013). Hamill’s research also supported using therapeutic circle dance as a self-oriented psychological intervention for individuals with dementia and their caregivers (Hamill et al., 2012). The engagement of the caregivers or families have a crucial role in the attendance rate and safe support during the dance intervention process.

3.4.5. Existing Studies’ Limitations

The existing studies contained several limitations. First, four studies (two case studies) only focused on female participants with MCI (Abreu and Hartley, 2013; Chang et al., 2021; Ghadiri et al., 2022; Yamada and Kawano, 2021), while 69% of all the studies considered were female-dominated in terms of their participants (Bracco et al., 2023; Chiesi et al., 2021; Choo et al., 2019; Debbie Duignan, Lynne Hedley, 2009; Doi et al., 2017; Dominguez et al., 2018; Douka et al., 2019; Guzmán et al., 2016; Guzmán-García et al., 2013; Hamill et al., 2012; Ho et al., 2020; Koh et al., 2020; Kropacova et al., 2019; Lazarou et al., 2017; Parial et al., 2022; Qi et al., 2018; Wang et al., 2020; Zhao et al., 2021; Zhu et al., 2018, 2022). Second, 62% of the studies reported small sample sizes (under 100 participants) (Abreu and Hartley, 2013; Adam et al., 2016; Ammar et al., 2021; Bisbe et al., 2020; Choo et al., 2019; Douka et al., 2019; Ghadiri et al., 2022; Guzmán et al., 2016; Guzmán-García et al., 2013; Hamill et al., 2012; Parial et al., 2022; Qi et al., 2018; Thumuluri et al., 2022; Wang et al., 2020; Yamada and Kawano, 2021; Zhao et al., 2021; Zhu et al., 2018, 2022). Third, 79% of the studies did not conduct a follow-up assessment (Abreu and Hartley, 2013; Adam et al., 2016; Bisbe et al., 2020; Blumen et al., 2022; Bracco et al., 2023; Chang et al., 2021; Chiesi et al., 2021; Choo et al., 2019; Debbie Duignan and Lynne Hedley, 2009; Doi et al., 2017; Dominguez et al., 2018; Douka et al., 2019; Ghadiri et al., 2022; Guzmán-García et al., 2013; Hamill et al., 2012; Koh et al., 2020; Kropacova et al., 2019; Lazarou et al., 2017; Qi et al., 2018; Thumuluri et al., 2022; Wang et al., 2020; Yamada and Kawano, 2021; Zhu et al., 2022). Another common limitation was that there was no professional dance instructor among 55% of the studies (Adam et al., 2016; Ammar et al., 2021; Bisbe et al., 2020; Blumen et al., 2022; Choo et al., 2019; Debbie Duignan, Lynne Hedley, 2009; Douka et al., 2019; Guzmán et al., 2016; Guzmán-García et al., 2013; Hamill et al., 2012; Koh et al., 2020; Kropacova et al., 2019; Qi et al., 2018; Wang et al., 2020; Zhu et al., 2018, 2022). Further limitations and future research directions are outlined in Supplementary Table 6 https://osf.io/5fxvt.

3.5. Synthesis of results

Meta-analysis results (see Fig. 7) across the eight RCT studies, based on the assessment results and outcome measurements, including 600 participants with MCI examined global cognition, executive function, memory, balance, and depression (see Table 1). These eight studies’ outcome data were reported in mean difference (MD) and standard deviation (SD) format and could therefore be included in the meta-analysis (one study (Ho et al., 2020) reported data in an unstandardized coefficient and standard error format, but we contacted the first author and received the raw data). Studies compared dance interventions to control groups (Chang et al., 2021; Lazarou et al., 2017; Parial et al., 2022; Qi et al., 2018; Wang et al., 2020; Zhu et al., 2018). One of the studies compared a dance intervention group to a physical therapy group(Bisbe et al., 2020); the physical therapy group was then used as a control group in the meta-analysis.

Fig. 7.

Forest Plot Outcome Results of Dance Intervention Group vs Control Group.

Table 1.

Summary of Outcome Measurements.

Outcome	Studies	Measurements
Global cognition	Chang et al. (2021) (Chang et al., 2021) Parial et al. (2022) (Parial et al., 2022) Lazarou et al. (2017) (Lazarou et al., 2017) (Qi et al., 2018) (Wang et al., 2020) Zhu et al. (2018) (Zhu et al., 2018) Bisbe et al. (2020) (Bisbe et al., 2020)	Montreal Cognitive Assessment (MoCA) Mental State Examination (MMSE)
Executive function	Bisbe et al. (2020) (Bisbe et al., 2020) (Qi et al., 2018) Zhu et al. (2018) (Zhu et al., 2018) Parial et al. (2022) (Parial et al., 2022) (Ho et al., 2020)	Trail Making Test (TMT-A) Trail Making Test (TMT-B)
Memory	(Qi et al., 2018) Zhu et al. (2018) (Zhu et al., 2018) (Ho et al., 2020)	Wechsler Memory Scale-Revised Logical Memory test (WMS-RLM) Digit Span Task (DST)
Balance	Chang et al. (2021) (Chang et al., 2021) Bisbe et al. (2020) (Bisbe et al., 2020) (Qi et al., 2018) (Wang et al., 2020)	Berg Balance Scale (BBS)
Depression	(Chang et al., 2021) (Parial et al., 2022) (Lazarou et al., 2017) (Wang et al., 2020) (Zhu et al., 2018)	Geriatric Depression Scale (GDS)

3.5.1. Global cognition

There are six studies (a total of 440 participants) which measured MoCA as an outcome in the meta-analysis. The results indicated that dance intervention had a significant effect on global cognition (MD = 1.83; 95% CI: [0.35, 3.32]; P = 0.02), with substantial heterogeneity (Chi² = 45.53, df = 5 (P < 0.00001); I² = 89%). Subgroup analysis was conducted to evaluate the effect of dance intervention durations on the outcome. Four studies conducted dance interventions for 12-weeks (Parial et al., 2022; Qi et al., 2018; Wang et al., 2020; Zhu et al., 2018), and two studies intervention durations longer than 12-weeks (18 and 40-week respectively) (Chang et al., 2021; Lazarou et al., 2017). The subgroup of 12-week interventions (MD = 0.87; 95% CI: [0.22, 1.51]; P = 0.008) showed significant improvements in global cognition. The subgroup of intervention durations longer than 12-weeks (MD = 3.62; 95% CI: [−1.33, 8.56]; P = 0.15) produced no significant difference between the dance intervention and control groups; however, two studies in this subgroup demonstrated significant differences between the dance intervention group and control group separately. The effect size of intervention durations in the longer than 12-week subgroup was larger than with the 12-week interventions subgroup. Heterogeneity was considerable in the subgroup with intervention durations longer than 12-weeks (Chi² = 35.93, df = 1 (P < 0.00001); I² = 97%); while there was no heterogeneity with the 12-week interventions subgroup (Chi² = 1.56, df = 3 (P = 0.67); I² = 0%) and there was no subgroup difference (Chi² = 1.17, df = 1 (P = 0.28); I² = 14.2%).

There are four studies (a total of 253 participants) which measured MMSE as an outcome in the meta-analysis. The results indicated that dance interventions also had a significant effect on global cognition (MD = 1.36; 95% CI: [0.28, 2.43]; P = 0.01) with substantial heterogeneity (Chi² = 9.42, df = 3 (P = 0.02); I² = 68%).

3.5.2. Executive function

The effect of dance interventions on executive function, particularly cognitive flexibility, and processing speed were assessed by TMT-A and TMT-B tests. There are four studies (a total of 250 participants) which measured TMT-A as an outcome in the meta-analysis. The results indicated that dance intervention had no effect in enhancing executive function score (MD = −0.72; 95% CI: [−9.56, 8.12]; P = 0.87) with no heterogeneity (Chi² = 0.83, df = 3 (P = 0.84); I² = 0%). There are five studies (a total of 301 participants) which measured TMT-B as an outcome in the meta-analysis. The results indicated that dance interventions also had no effect in enhancing executive function score (MD = −2.64; 95% CI: [−21.67, 16.39]; P = 0.79) with no important heterogeneity (Chi² = 5.75, df = 4 (P = 0.22); I² = 30%).

3.5.3. Memory

The effect of dance interventions on memory was assessed using two studies (a total 90 participants) by WMS-RLM (MD=3.01; 95% CI: [1.36, 4.66]; P = 0.0003) with no heterogeneity (Chi² = 0.08, df = 1 (P = 0.77); I² = 0%). The results indicated that dance interventions had a significant effect on memory improvement. Results for memory (immediate recall and working memory, verbal memory), executive function, and attention were assessed using three studies (a total 219 participants) by DST. The results indicated that dance interventions had no effect in enhancing the outcomes of DST (MD = −0.26; 95% CI: [−0.92, 0.40]; P = 0.44) with no heterogeneity (Chi² = 0.26, df = 2 (P = 0.88); I² = 0%).

3.5.4. Balance

There are four studies (a total of 233 participants) which measured BBS as an outcome in the meta-analysis. The results indicated that dance interventions had a significant effect on balance (MD = 1.33; 95% CI: [0.20, 2.46]; P = 0.02) with moderate heterogeneity (Chi² = 5.62, df = 3 (P = 0.13); I² = 47%).

3.5.5. Depression

There are five studies (a total of 408 participants) which measured GDS as an outcome in the meta-analysis. The results indicated that dance interventions had a significant effect on reducing depression (MD = −0.95; 95% CI: [−1.68, −0.22]; P = 0.01) with substantial heterogeneity (Chi² = 11.42, df = 4 (P = 0.02); I² = 65%). Subgroup analysis was conducted to evaluate the effect of intervention durations on the outcome. Three studies conducted dance interventions for 12-week (Parial et al., 2022; Wang et al., 2020; Zhu et al., 2018), and two studies for longer than 12-week (18 and 40-week respectively) (Chang et al., 2021; Lazarou et al., 2017). The 12-week interventions subgroup (MD = −1.10; 95% CI: [−3.12, 0.93]; P = 0.29) showed no significance in reducing depression, but the subgroup of intervention durations longer than 12-weeks (MD = −0.82; 95% CI: [−1.24, −0.39]; P = 0.0002) had a significant effect between dance intervention and control groups. The effect size of intervention durations of longer than 12-week subgroup was larger than the 12-week interventions subgroup. Heterogeneity was substantial in the subgroup of 12-week interventions (Chi² = 10.76, df = 2 (P = 0.005); I² = 81%), but there was no heterogeneity within the subgroup of intervention durations longer than 12-weeks (Chi² = 0.33, df = 1 (P = 0.57); I² = 0%), and there was no subgroup difference (Chi² = 0.07, df = 1 (P = 0.79); I² = 0%).

3.6. Sensitivity analysis

Sensitivity analysis was conducted by omitting each study one by one, revealing that the findings were not significantly altered by any individual study within the analysis of TMT-A, TMT-B, and DST. However, when the study of Chang et al. (2021) was omitted from the MoCA outcome analysis, the effect changed to non-significant (P = 0.05); when the study of Lazarou et al. (2017) was omitted in the same outcome analysis, heterogeneity changed to I² = 0. When the Lazarou et al., 2012 and Wang et al., 2019 studies were omitted from the MMSE outcome analysis, the effect changed to non-significant (P = 0.09 and P = 0.10 respectively); there was no change in heterogeneity by omitting studies one by one in the MMSE analysis. When the Chang et al. (2021) and Bisbe et al. (2020) studies were omitted in the BBS outcome analysis, the effect changed to non-significant (P = 0.08 respectively); when the Qi et al., 2019 and Wang et al., 2019 studies were omitted, heterogeneity changed to I² = 19% and I² = 0% respectively. In the GDS outcome analysis, when the Chang et al. (2021) and Lazarou et al. (2017) studies were omitted the effect changed to non-significant (P = 0.07 and P = 0.09 respectively); when the study of Wang et al., 2019 was omitted, heterogeneity changed to I² = 0%.

4. Discussion

4.1. Principal results

The effectiveness of dance intervention for older adults with MCI, AD and dementia was examined by a systematic scoping review and a meta-analysis. According to quantitative and qualitative evidence, dance intervention is a promising, efficient, successful, and entertaining physical therapy intervention with intrinsic motivation sources for long-term neurological patients to carry out in hospital, community, and home settings (Abreu and Hartley, 2013). Dance intervention significantly enhanced participants’ quality of life (Adam et al., 2016; Bracco et al., 2023; Chang et al., 2021; Choo et al., 2019; Douka et al., 2019; Koh et al., 2020; Parial et al., 2022; Thumuluri et al., 2022; Wang et al., 2020), memory (Bisbe et al., 2020; Doi et al., 2017; Lazarou et al., 2017; Qi et al., 2018; Zhu et al., 2018, 2022), cognition (Adam et al., 2016; Doi et al., 2017; Dominguez et al., 2018; Lazarou et al., 2017; Parial et al., 2022; Zhao et al., 2021), balance ability (Abreu and Hartley, 2013; Chang et al., 2021; Koh et al., 2020; Thumuluri et al., 2022), gait distance and speed (Abreu and Hartley, 2013; Ghadiri et al., 2022; Koh et al., 2020), immediate and delayed recall (Bisbe et al., 2020; Lazarou et al., 2017; Parial et al., 2022), executive function (Blumen et al., 2022; Douka et al., 2019; Kropacova et al., 2019), processing speed (Qi et al., 2018; Zhu et al., 2018), and verbal fluency (Douka et al., 2019; Lazarou et al., 2017); it also reduced depression (Adam et al., 2016; Chang et al., 2021; Dominguez et al., 2018; Ho et al., 2020; Lazarou et al., 2017; Wang et al., 2020; Zhao et al., 2021), negative mood (Guzmán et al., 2016; Hamill et al., 2012; Ho et al., 2020), and anxiety (Adam et al., 2016; Douka et al., 2019). The findings were consistent with earlier systematic reviews and meta-analysis studies (Patricia et al., 2021; Yuan et al., 2022). Our meta-analysis results indicated significant effects on some outcomes including global cognition, memory, balance, and depression, but no statistically significant findings were obtained for executive function and the DST test.

The results from the sensitivity analysis indicated that some changes in statistical significance may have occurred as a result of the omitted study having a large weight in the meta-analysis as outlined in the forest plots (Chang et al., 2021; Lazarou et al., 2017) and the omitted study thus causing the weight of the study with no statistical difference to increase (Bisbe et al., 2020); some studies’ effect sizes may have been too large to affect the results in the outcome analysis group (MoCA, MMSE and GDS) (Lazarou et al., 2017; Wang et al., 2020). As for heterogeneity changes in BBS and GDS analyses, the reason is that the omitted study occupied a large proportion of weight in the meta-analysis as outlined in the forest plots (Qi et al., 2018); and the omitted study’s effect size was too large to affect the results (Wang et al., 2020). Potential causes may include sample sizes, participants inclusion criterion, baseline parameters, or dance intervention intensities.

Three individual studies in the scoping review observed that dance intervention enhanced executive function (Blumen et al., 2022; Douka et al., 2019; Kropacova et al., 2019). However, our meta-analysis failed to find dance benefit effects for executive function; these results are consistent with a previous meta-analysis (Liu et al., 2021). The inconsistency in the outcomes may be due to differences in study design and training intensities. Nevertheless, a previous systematic review suggested that dance can improve executive function, working memory capacity, and cognitive flexibility in children (Tao, Gao, Cole et al., 2022). Older adults with neurological diseases may have lower sensitivity than young people, or may be unable to reach the required training load during the intervention because of physical ability limitations. Despite some measurements providing no statistically significant results, qualitative studies using interview and questionnaire methods pointed to actual positive feedbacks, indicating that participants received satisfying experiences.

Dancers must pay close attention to observe and organize their motions to perform a dance routine correctly, exerting neurological stimulation on the memory system and inducing neuroplasticity. Ballroom dancers, for example, have more resting activity in the middle and inferior frontal gyri (which are neural substrates implicated in attention control) than non-dancers (Chan et al., 2020). Using structural MRI, researchers have discovered higher hippocampus volumes following dance intervention in MCI participants (Qi et al., 2018; Zhu et al., 2022), and increased amplitude of low-frequency fluctuation in par hippocampal area (Qi et al., 2018) – memory processes are strongly linked to the hippocampus. Dance movement interventions frequently challenge attentional systems, which may result in adaptive alterations in attentional skills.

Dance with music may be seen as a pleasurable activity for participants, increasing involvement and adherence to intervention (Abreu and Hartley, 2013). Listening to familiar rhythms and lyrics, as well as engaging in traditional dance actions that are familiar from childhood or young adulthood, may stimulate memories and bring participants’ awareness into the present time (Yamada and Kawano, 2021). Additionally, the rhythmic concept’s time-energy-space principles are used to increase the participants’ spontaneous bodily responses by improving their concentration while listening to music and thinking (Chéour et al., 2023; Choo et al., 2019). The same benefits may be observed with cultural dance, such as traditional Greek dance (Sofianidis et al., 2009), or Turkish folklore dance-based exercise (Eyigor et al., 2009), and these are also practical types of exercise activity that enhance physical and mental health with MCI, AD, or dementia participants that often have poorer levels of exercise activity involvement. The advantage of cultural dance is that the participants were able to follow one- to two-step instructions and complete a familiar task due to their cultural background despite participants’ memory problem (Abreu and Hartley, 2013). Participants’ cultural backgrounds may also influence the style of proposed dance intervention, as well as intervention attendance and involvement (Chiesi et al., 2021). Dance movement interventions can be tailored to participants in various regions, and a variety of dance styles can be used to meet participants’ health status with the emphasis on participation rather than performance. Participants were either partnered to join the dance sessions, or dancing in groups to allow for social contacts that they are lacking in daily life. Group dance, in particular, fosters a sense of integration and connection, which helps to alleviate the difficulties of orienting oneself in time and space (Ho et al., 2015).

Employing professional dance instructors instead of physiologists to manage dance intervention classes is another key factor for the programmes’ success. Additionally, although keeping some regular routine is essential for memory function, in order to keep the participants’ interest fresh creative aspects must be added regularly when designing the dance intervention protocol. It may be advantageous to establish programmes that encourage the use of dance as a form of exercise activity in elderly people with cognitive neurological diseases, so as to raise activity level, exercise compliance, function and quality of life (Abreu and Hartley, 2013). Dance movement intervention could also be implemented into a structural rehabilitation program to provide early psychosocial assistance to vulnerable older adults.

Based on the evidence from this study, future research needs high methodological quality RCT trials with large sample sizes to assess DMI effectiveness for people with MCI, AD and dementia. Also needed is investigation relating to the protective effect of DMI on the deterioration in everyday functioning and neuroendocrine function, by investigating putative pathways such as insulin metabolism, inflammatory processes, cholinergic transmission, and synaptic plasticity. Consideration of diagnostic tests such as neuropsychological recordings by electroen-cephalography, which captures the brain’s functioning, particularly its electrical activity, for outcome measures requires further investigation. Additionally, MRI scanning and related professional software could be used to detect any changes in brain activity. Besides, it is important to include physical measures of balance, fall risk, walking speed, and muscle strength tests in studies, as these are important measures to assess frailty in the elderly. Moreover, adopting validated alternate formats in future studies would be critical to minimizing the potential practice effects from repeated testing. Furthermore, consideration needs to be given to follow-up tests after the intervention to monitor the long-term dance intervention effects. Studies also need to be designed to determine whether there are thresholds in terms of intervention duration, or the number of courses required to achieve clinically significant treatment effects. There are still research gaps related to gender and education level differences in dance interventions; future studies should combine these research gaps with other measurements to investigate DMI effectiveness.

4.2. Limitations

This study combined scoping review and meta-analysis methodologies to examine the effectiveness and feasibility of DMI in older adults with MCI, AD and dementia. The scoping review mapped research gaps and identified future research directions for health practitioners and policymakers; the meta-analysis results provided solid support for conclusions regarding the possibility of implementation of DMI in the clinical and rehabilitation practice for participants. Nonetheless, there are three limitations in this study. First, the scoping review included three case studies with only one participant, or very small sample sizes, for inclusion of all related evidence. All selected studies were included in the scoping review without the risk of bias assessment results. Second, 62% of the included studies were MCI studies; this may be because it is difficult to conduct an intervention with AD and dementia participants who have severe diseases status and are physically inactive. Third, only eight studies were included in the meta-analysis, and since the number was small, publication bias could not be tested.

5. Conclusions

The implications of DMI have long-term positive influences for future research in clinical and rehabilitation practice for MCI, AD, and dementia. Dance interventions pose few hazards because adverse outcomes were rare and demonstrated good adherence. Dance is an ecologically and culturally appropriate activity that is applicable to the physiological and psychological constraints of older people, while also providing multimodal stimulations including medical, mental, and social experiences. Dance movement interventions do not require large financial expenditure, or training, so it may be simply adopted and integrated into current medical services for senior people in communities, care homes, and clinics. Future studies are needed to explore further the evidence for DMI as a complimentary treatment for older adults with MCI, and especially in AD and dementia.

Abbreviations:

DMI

Dance movement intervention

MCI

Mild cognitive impairment

AD

Alzheimer’s disease

ROBINS-I

Risk of Bias in Non-randomized Studies of Interventions

RoB 2

Cochrane Risk of Bias Tool for Randomized trials

GDS

Geriatric Depression Scale

MMSE

Mini-Mental State Examination

MoCA

Montreal Cognitive Assessment

TMT

Trail Making Test

WMS-RLM

Wechsler Memory Scale-Revised Logical Memory test

MRI

Magnetic Resonance Imaging

DST

Digit Span Task

NPI

Neuropsychiatric Inventory

BBS

Berg Balance Scale

QOL-AD

Quality of life in Alzheimer’s Disease

Data Availability

I have shared the link to my data.