The Effect of Music interventions on Cognitive Function in Older Adults with Mild Cognitive Impairment: A Systematic Review

Abstract

The purpose of the current systematic review was to examine the effects of music interventions on cognitive function in older adults with mild cognitive impairment (MCI). A systematic search of CINAHL, PubMed, PsycINFO, and Web of Science databases was performed. Studies examining the effects of music interventions on cognitive function in older adults with MCI were included. Narrative synthesis for cognitive outcomes postintervention was performed. A total of 11 articles met inclusion criteria. Music interventions significantly improved global cognitive function, verbal fluency, executive function, and spatial function in older adults with MCI. Included studies were heterogeneous in terms of the type of intervention, cognitive assessment tool, and intervention duration. Six studies were at risk of bias due to missing data and confounding factors. Our findings suggest that music interventions can be an effective strategy to improve cognitive function for older adults with MCI. However, findings should be interpreted with caution. More rigorous studies with various types of music interventions investigating cognitive domain–specific effects are needed.

In an aging society with increased life expectancy, a significant number of people have cognitive impairment, with a 19% average prevalence worldwide (^{Pais et al., 2020}). A recent study identified that two of three American individuals experience cognitive problems by age 70 years (^{Hale et al., 2020}). Cognitive impairment is a serious public health concern, given that medical costs are 44% higher for patients with cognitive impairment than for those without cognitive impairment (^{Zhu et al., 2013}), and the average cost burdens related to patients with cognitive impairment were estimated to be $280,000 per person in 2021 (^{Favreault & Johnson, 2021}).

Cognitive impairment ranges from mild to severe. Mild cognitive impairment (MCI) is the stage of transition between a normal cognitive state for age and a state of dementia (^{Domínguez-Chávez et al., 2019}). Two subtypes of MCI are amnestic (aMCI) and nonamnestic MCI (naMCI). Although memory impairment is prominent in aMCI, in naMCI, memory remains intact, but one (single-domain) or more other cognitive domains (multiple-domain) are impaired (^{Csukly et al., 2016}; ^{Dunn et al., 2014}). For example, the cognitive domains impacted by naMCI include attention, language ability, executive function, and visuo-spatial function (^{Taler & Phillips, 2008}). People with aMCI are more likely to progress toward Alzheimer’s disease (AD), with a 17% to 18% conversion rate, and those with naMCI are more likely to develop other forms of dementia (^{Ferman et al., 2013}; ^{Michaud et al., 2017}). At the same time, with either type of MCI, patients may recover cognitive function to the normal level for age, with reversion rates of 58% (^{Overton et al., 2019}). Therefore, early interventions for cognitive improvement are imperative to prevent progression of MCI to dementia.

However, there are no pharmacological solutions available to prevent or significantly delay the progression of dementia, and some medications can increase the risks for serious outcomes, such as stroke and embolism (^{Fink et al., 2018}). On the other hand, using music as a therapeutic intervention has received attention as an easily accessible and cost-effective intervention (^{Amor Gaviola et al., 2022}) to improve cognitive function and has shown positive effects in various populations, including older adults with and without dementia (^{Cheung et al., 2018}; ^{Diaz Abrahan et al., 2019}). Th ese beneficial effects of the interventions may stem from the impact of music on neural reorganization (^{Altenmüller et al., 2009}) and structural brain changes (^{Habibi et al., 2018}; ^{Hyde et al., 2009}). Facilitated social interaction through music interventions (^{Waters et al., 2022}) may also contribute to improving cognitive function as daily social interactions are related to better cognitive function in older adults (^{Zhaoyang et al., 2021}).

Music interventions are categorized as music medicine and music therapy. Music medicine is passive listening to music offered by clinicians, whereas music therapy involves a therapeutic relationship between participants and trained music therapists during the music intervention (^{Bradt & Dileo, 2014}). Music interventions include a variety of methods, such as listening to music, singing, music with movement, dancing, and playing musical instruments. Evidence for which types of music interventions, active or passive, are more effective is limited, as results of previous systematic reviews on dementia are inconsistent: Fusar-Poli et al. (²⁰¹⁸) found only active music intervention to be effective, whereas passive music intervention had the greatest effect in Moreno-Morales et al.’s (²⁰²⁰) study.

Although music interventions have been widely used for patients with cognitive impairment, the beneficial effect of the interventions on global and subdomains of cognitive function for older adults with MCI is not well established. Prior reviews have focused on people with dementia and excluded those with MCI (^{Fang et al., 2017}; ^{Fusar-Poli et al., 2018}; ^{Moreno-Morales et al., 2020}), and some reviews examined only global cognitive function or memory as targets of the music interventions (^{Moreno-Morales et al., 2020}; ^{Xu et al., 2017}). However, music interventions are more likely to be effective for protecting cognitive function in the early stages of dementia because it is hard for people with severe dementia to cooperate with intervention providers and sustain the intervention due to physical or cognitive problems (^{Fang et al., 2017}). Th us, it is essential to explore their effects on people with MCI so that they can recover cognitive function to the normal level for age. In addition, the most common MCI subtype is impairment in nonmemory single domains (^{Palmer et al., 2008}), so a more comprehensive review should be done to examine the beneficial effects of music interventions on subdomains of cognitive function affected in naMCI. Therefore, the purpose of the current review was to examine the effects of music interventions on global cognitive function as well as on each subdomain of cognitive function in older adults with MCI.

Method

Search Strategies

A systematic review was performed using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (^{Moher et al., 2009}). Broad search terms (e.g., music, music therapy, listening to music, cognition, cognitive dysfunction, memory) were used in consultation with a health-research librarian to allow for a comprehensive review without missing available data; and we used medical subject headings (MeSH) for music and music therapy (Appendix A, available in the online version of this article). Keywords were searched through CINAHL, PubMed, PsycINFO, and Web of Science databases through May 2021.

Eligibility Criteria

Inclusion criteria were as follows: studies that (a) investigated effects of music interventions (music medicine and/or music therapy) on global cognitive function or at least one cognitive subdomain (e.g., verbal fluency, memory, attention), (b) included participants aged >50 years with MCI, and (c) were peer-reviewed and written in English. Publication date was not limited so we could fully understand the state of the science. We excluded studies that (a) did not focus on music interventions (e.g., focused on art therapy), (b) included participants who had AD, dementia, or other neurological problems (e.g., stroke, Parkinson’s disease, multiple sclerosis), or (c) included only cognitively intact participants.

Data Extraction and Risk-of-Bias Assessment

Covidence (^{Veritas Health Innovation, n.d.}) was used to manage the search, remove duplicate studies, and screen studies. With the identified publications, data extraction and risk-of-bias assessment were performed by two reviewers (J.Y.K., S.T.W.), and any disagreement was resolved through discussion with another reviewer (H.C.). Th e main variables of interest were sample characteristics, intervention characteristics, cognitive assessment tools used, and cognitive outcomes. Data from included studies were organized using a structured template (Table A, available in the online version of this article).

Table A.

Data Extraction of Included Studies

	Authors, year, country	Groups (analyzed sample size)	Age (mean)	Diagnostic tool for mild cognitive impairment	Type of music intervention	Music selection/Method of delivery	Duration of music intervention	Cognitive outcome assessment	Result
1	Adam et al. 2016 Peninsular Malaysia	1) Experimental group (n=44) 2) Control group (n=40)	70.9	Mild-moderate cognitive impairment identified via MMSE	1) Poco-poco dance + relaxation 2) Relaxation only	Not detailed	60 minutes twice a week for 6 weeks	Cognitive function (MMSE)	The intervention group significantly improved compared to the control at 6 weeks on the MMSE (p < .001).
2	Biasutti & Mangiacotti 2018 Italy	1) Experimental group (n=18) 2) Control group (n=17)	83.6	Mild-moderate cognitive impairment identified via MMSE and healthy aging	1) Music training program: imitation, creation, and executive rhythm – body & instrument percussion/scat singing 2) Gymnastic activities	Modern music (top-ten hits from the period of training), famous Italian music (belonging to the participants’ youth)	70 minutes twice a week for 12 weeks	• General cognitive function (MMSE) • Access skills and lexical retrieval (VFT) • Spatial planning, selective attention, psychomotor speed (TMT A) • Visual selective attention (AMT) • Praxis skills, mental representation, planning, logical skills (CDT)	The intervention group significantly improved compared to the baseline at 12 weeks on the MMSE (p < .001); VFT (p < .013); and CDT (p < .001). There was a significant reduction in AMT in the control group (p < .05).
3	Cacciafesta et al. 2010 Italy	1) N=12 2) N=10 3) N=11 (another group)	Age range 66–77	MCI identified via MMSE or subjective memory disorder	Listening to music	1) K448 by Mozart and Für Elise by Beethoven 2) K448 by Mozart 3) Other compositions with a different index of periodicity than K448	1) once 2) 6 months daily 3) once	• Spatial-temporal abilities (PFC) • Episodic learning (3 objects, 3 places) • Ideational-praxis abilities (Clock test) • Immediate recall (Rey’s 15 words test) • Attention (TMT A&B) • Memory for numbers (Digit span)	Significant improvement in 1) Mozart’s K448: PFC test – all 12 patients Rey’s 15 words test — 10 out of 12 patients Other tests did not show significant improvement. Beethoven’s Für Elise: All tests did not show significant improvement. 2) Mozart’s K448: After 6 months, 10 of all remaining patients showed improvement in the PFC test. 3) Other compositions with a different index of periodicity with K448: No significant variation.
4	Doi et al. 2017 Japan	1) Dance (n=67) 2) Playing musical instruments (n=67) 3) Health education control group (n=67)	76	MCI identified via subjective cognitive complaints, objective cognitive impairment assessed by tests of memory and nonmemory domains	1) Dancing with an instructor (ballroom dance, including salsa, rumba, waltz, cha-cha, blues, jitterbug, and tango) 2) Playing percussion instruments (e.g., conga), improvised on the musical scores	Not detailed	60 minutes once a week for 40 weeks	• Memory (Story memory and word list memory tests) • General cognitive function (MMSE) • Attention & executive function (TMT A &TMT B)	The dance intervention group significantly improved compared to the control at 10 months on the story memory test (p = .011) and the MMSE (p = .026) The playing instruments intervention group significantly improved compared to the control at 10 months on the MMSE (p = .008) The dance intervention aMCI subgroup significantly improved compared to the control aMCI (amnestic) group at 10 months on the story memory test (p = .001) but not in the naMCI (nonamnestic) subgroup.
5	Domínguez-Chravez et al. 2019 Mexico	N=16	71.3	MCI identified via MoCA	Music with movement intervention (walking with rhythmic schemes/reading of rhythmic patterns/listening to music while doing activities and movement/playing instruments/singing)	Songs from Mexican folklore (patients’ musical preferences were taken into account)	60 minutes three times a week for 12 weeks	• Global cognition state (MoCA) • Attention (digit span subset) • Immediate memory (Texts I from the Wechsler Memory Scale) • Delayed memory (Texts II from the Wechsler Memory Scale) • Executive function (FAB)	The intervention group significantly improved compared to the baseline at 12 weeks on the MoCA (p = .001); digit span subset (p = .007); Wechsler Memory Scale Texts I (p < .001); Texts II (p = .001); FAB (p = .002).
6	Douka et al. 2019 Greece	1) Cognitively intact group (n=30) 2) MCI group (n=30)	1) median: 65.5 2) median: 67.5	MCI identified via MMSE	Traditional dances selected from all over Greece	Not detailed	60 minutes two times a week for 24 weeks	• Six domains of cognitive and functional performance (CRD) • Executive functioning (FUCAS, TMT) • Attention (TEA) • Memory (ROCF, RAVLT, RBMT) • Verbal fluency (VFT)	The intervention group of MCI significantly improved compared to baseline at 24 weeks on the TEA (S4viac) (p < .001); TEA (S4viti) (p < .001); FUCAS (p = .03); VFT (Verflx: p = .023; Verfls: p = .001; and Verfmo: p = .003). However, a significant decrease at 24 weeks was found on the RBMT 1&2 (p’s < .001).
7	Lake & Goldstein 2011 USA	1) MCI group (n=12) 2) Cognitively intact group (n=12)	70.2	Amnestic MCI diagnosed by experienced neurologists and neuropsychologists (subjective memory complaint/objective memory impairment)	Listening to music via headphones	“Spring” movement of Four Seasons by Vivaldi	10 minutes for one time	• Attention (RBANS: Digit Span & Coding; DRS)	There was no significant difference between the music listening and the silence condition.
8	Lazarou et al. 2017 Greece	1) Ballroom dancing group (n=66) 2) Control group (n=63)	66.9	Amnestic MCI based on Petersen criteria: (1) memory complaint, (2) normal activities of daily living, (3) normal general cognitive function, (4) abnormal memory for age, and (5) not demented; stage 3 of the disease according to Global Deterioration Scale	1) International Ballroom Dancing with special music (tango, waltz, Viennese waltz, fox trot, rumba, cha-cha, swing, salsa, merengue, disco–hustle, Greek traditional ballroom dancing) 2) No intervention	Special music depending on the type of dance	60-minute sessions twice a week for 10 months	• Global cognition (MMSE, MoCA) • Daily function (FUCAS) • Memory (RBMT) • Verbal fluency (FAS) • Executive function (ROCFT-copy and delayed recall; TMT B) • Learning (RAVLT) • Attention (TEA)	The intervention group significantly improved compared to the control at 10 months on the MMSE (p < .01); MoCA (p = .03); TEA (p = .002); RAVLT (p = .003); FAS (p = .005); RBMT (p’s < .005); and ROCFT (p’s < 0.01). The intervention group significantly improved compared to the baseline at 10 months on the FAS (p < .01); ROCFT (p = .004); RAVLT (p = .001); TEA (p = .03); and RAVLTI (p = .002). The control group significantly worsened at 10 months on the MMSE, FUCAS, ROCFT delay recall, RBMT delay recall, FRSSD, RAVLT, and TEA map.
9	Shimizu et al. 2018 Japan	1) Movement music therapy (MMT; n= 30) 2) Exercise group (n=9)	74.6	MCI defined by Pertersen criteria; memory impairment checklist ≥ 1 item	1) Repetitive rhythmic movements with music and the use of the Naruko clapper 2) Same exercises performed in the MMT group without background music	Various songs (e.g., Astro Boy, Samba of ladybug, Yosakoi Icchorai Ondo, Kiyoshi’s Zundoko)	60 minutes once a week for 12 weeks	Frontal lobe function (FAB)	The intervention group significantly improved compared to the baseline at 12 weeks on the FAB (p = .008).
10	Wang et al. 2020 China	1) Square dancing group (n=33) 2) Control group (n=33)	81.1	MCI identified based on four criteria: (1) subjective cognition decline; (2) MMSE>24; (3) MoCA<26; (4) activities of daily living < 26	1) Stepping exercise and following the rhythm of the music 2) Usual lifestyle	Not detailed	40 minutes three times a week for 12 weeks	Global cognition (MMSE, MoCA)	There was no significant difference after the intervention at 6 weeks on the MoCA or MMSE. The intervention group significantly improved compared to the baseline at 12 weeks on the MoCA (p = .001) and MMSE (p < .001). The intervention group significantly improved compared to the control at 12 weeks on MMSE (p = .006).
11	Zhu et al. 2018 China	1) Aerobic dance + usual care group (n=29) 2) Control group (usual care; n=31)	69.6	(1) Diagnosed with amnestic MCI according to NIA-AA; (2) MMSE≥25; (3) MoCA≤26	Dance in synchronization with a musical phrase of eight rhythmical meters	Not detailed	35 minutes three times a week for 12 weeks	• Logical memory (WMS-RLM) • Overall cognitive function (MoCA) • Attention, processing speed, working memory (SDMT) • Visual perception, motor speed, speed processing, mental flexibility (TMT A&B) • Short-term verbal memory (DST-F&B) • Processing speed (P300 latency) • Attention (P300 amplitude)	The intervention group significantly improved compared to the baseline at 3 months on the MoCA (p < .001); WMSR-LR (p < .05); SDMT (p < .05); and P300 latency (p < .05); and at 6 months on the MoCA (p < .001); WMSR-LR (p < .05); SDMT (p < .05); and TMT-A (p < .05). The intervention group significantly improved compared to the control at 3 months on the WMSR-LR (p < .001); TMT B (p < .05); and at 6 months on the P300 latency (p < .05).

Note. MMSE: Mini-Mental State Examination; VFT: verbal fluency test; TMT A & B: Trail Making Test A and B; AMT: attentional matrices test; CDT: clock-drawing test; MCI: mild cognitive impairment; PFC: paper-folding and cutting test; MoCA: Montreal Cognitive Assessment; FAB: Frontal Assessment Battery; CRD: Clinical Dementia Rating; FUCAS: Functional and Cognitive Assessment Test; FRSSD: Functional Rating Scale for Symptoms of Dementia; TEA: Test of Everyday Attention; ROCF: Rey–Osterreith Complex Figure Test; RAVLT: Rey Auditory Verbal Learning Test; RBMT: Rivermead Behavioral Memory Test; RBANS: Repeatable Battery for the Assessment of Neuropsychological Status; DRS: Mattis Dementia Rating Scale; FAS: the Verbal Fluency F-A-S test; ROCFT-copy and delayed recall: Rey Osterrieth Complex Figure Test copy and delay recall; NIA-AA: National Institute on Aging and Alzheimer’s Association; WMS-RLM: Wechsler Memory Scale-Revised Logical Memory; SDMT: symbol digit modalities test; DST-F & B: Digit Span Test-forward and backward.

Risk-of-bias assessment was done using the revised Cochrane risk-of-bias tool for randomized trials (RoB 2) for the randomized controlled trial (RCT) studies (^{Sterne et al., 2019}), evaluating “low risk of bias,” “some concerns,” and “high risk of bias” according to judgment on five domains of RoB 2. Risk of Bias in Non-randomized Studies–of Interventions (ROBINS-I) criteria were used for quasi-experimental studies (^{Sterne et al., 2016}), appraising the overall risk-of-bias judgments based on the seven domains of the ROBINS-I.

Considering the risk of bias in the included studies and variability of cognitive outcome assessment tools and intervention duration among studies, integrating the data with meta-analysis was unnecessary and could generate misleading results (^{Trkulja & Hrabač, 2020}). Th erefore, narrative synthesis for cognitive outcomes after interventions was performed.

Results

Study Selection

We identified 4,636 articles through database searches and six additional records from the reference lists of reviewed articles. After removing duplicates, 3,824 articles were screened by title and abstract, resulting in 48 articles for full-text review. Thirty-seven articles were excluded as they did not meet inclusion/exclusion criteria; therefore, a total of 11 articles were included in the final review. Figure 1 shows the study selection process.

Figure 1.

PRISMA Diagram

Of the 11 articles, five studies were RCTs with two parallel arms except for one study with three parallel arms (^{Doi et al., 2017}); and six were quasi-experimental studies. Studies were conducted in various countries, including China (^{Wang et al., 2020}; ^{Zhu et al., 2018}), Japan (^{Doi et al., 2017}; ^{Shimizu et al., 2018}), Greece (^{Douka et al., 2019}; ^{Lazarou et al., 2017}), Italy (^{Biasutti & Mangiacotti, 2018}; ^{Cacciafesta et al., 2010}), Malaysia (^{Adam et al., 2016}), Mexico (^{Domínguez-Chávez et al., 2019}), and the United States (^{Lake & Goldstein, 2011}).

Study Participants

A total of 695 older adults with MCI and 42 cognitively intact older adults (as a part of control groups) were included in the current review. Three studies (^{Lake & Goldstein, 2011}; ^{Lazarou et al., 2017}; ^{Zhu et al., 2018}) included patients with aMCI only. Sample sizes ranged from 16 (^{Domínguez-Chávez et al., 2019}) to 201 (^{Doi et al., 2017}), and mean ages ranged from 66.9 years to 83.6 years. Participants’ baseline cognitive function was assessed and identified as MCI before the initiation of interventions. Participants’ chronic conditions (e.g., hypertension, diabetes, arthritis) were reported in only two studies (^{Doi et al., 2017}; ^{Domínguez-Chávez et al., 2019}). No studies reported adverse events during the music interventions. Table A provides details of participants’ characteristics.

Intervention Characteristics

Music Medicine.

Two included studies (^{Cacciafesta et al., 2010}; ^{Lake & Goldstein, 2011}) used music listening as an intervention. In Cacciafesta et al. (²⁰¹⁰), different types of music (Mozart’s K448 and Beethoven’s Für Elise) were provided, and cognitive function was examined before and after onetime listening to the specific music. In Lake and Goldstein (²⁰¹¹), participants listened to the “Spring” movement of Vivaldi’s Four Seasons via headphones for 10 minutes, and the same participants sat in silence for 10 minutes.

Music Therapy.

Various types of music therapy were implemented, including dancing (^{Adam et al., 2016}; ^{Doi et al., 2017}; ^{Douka et al., 2019}; ^{Lazarou et al., 2017}; ^{Wang et al., 2020}; ^{Zhu et al., 2018}), musical improvisation (^{Biasutti & Mangiacotti, 2018}), playing musical instruments (^{Doi et al., 2017}), and movement with a musical instrument (^{Shimizu et al., 2018}). In one study (^{Domínguez-Chávez et al., 2019}), participants took part in five music interventions, such as movement with music, reading of rhythmic patterns, music listening, playing instruments, and singing, one after another. Seven studies had a control group, and four (^{Adam et al., 2016}; ^{Biasutti & Mangiacotti, 2018}; ^{Doi et al., 2017}; ^{Shimizu et al., 2018}) used active control groups, which included relaxation therapy, gymnastic activities, health education, and the same exercises as those used in the intervention group but performed without music, whereas three control groups participated in usual care or no intervention (^{Lazarou et al., 2017}; ^{Wang et al., 2020}; ^{Zhu et al., 2018}).

The duration of interventions varied from 12 to 48 sessions, but Lazarou et al. (²⁰¹⁷) conducted dancing interventions for 80 sessions over 10 months. Th e most common therapy duration was 3 months with an average of 28 sessions (^{Biasutti & Mangiacotti, 2018}; ^{Domínguez-Chávez et al., 2019}; ^{Shimizu et al., 2018}; ^{Wang et al., 2020}; ^{Zhu et al., 2018}). All interventions were carried out one to three times per week, whereas the duration of each intervention session ranged from 35 minutes (^{Zhu et al., 2018}) to 70 minutes (^{Biasutti & Mangiacotti, 2018}). Six studies (^{Biasutti & Mangiacotti, 2018}; ^{Cacciafesta et al., 2010}; ^{Domínguez-Chávez et al., 2019}; ^{Lake & Goldstein, 2011}; ^{Lazarou et al., 2017}; ^{Shimizu et al., 2018}]) identified the specific music used for the music interventions (e.g., classical music, songs from Mexican folklore, Italian music), whereas the other studies did not provide information about the type of music used. Moreover, Domínguez-Chávez et al. (²⁰¹⁹) took patients’ music preferences into account when designing the intervention program. More information about the intervention characteristics is provided in Table A.

Cognitive Outcomes After Interventions

Changes in Global Cognitive Function.

Seven studies (^{Adam et al., 2016}; ^{Biasutti & Mangiacotti, 2018}; ^{Doi et al., 2017}; ^{Domínguez-Chávez et al., 2019}; ^{Lazarou et al., 2017}; ^{Wang et al., 2020}; ^{Zhu et al., 2018}) assessed global cognitive function with either the Mini-Mental State Examination or Montreal Cognitive Assessment as an outcome. These studies implemented various music interventions, including dancing (^{Adam et al., 2016, Doi et al., 2017}; ^{Lazarou et al., 2017}; ^{Wang et al., 2020}; ^{Zhu et al., 2018}), playing percussion instruments (^{Doi et al., 2017}), musical improvisation (^{Biasutti & Mangiacotti, 2018}), and music with movement (^{Domínguez-Chávez et al., 2019}). Global cognitive function was improved in all seven studies after the music interventions. Global cognitive function of the control groups did not change or significantly worsened, which led to significant differences between the music intervention and control groups (^{Adam et al., 2016}; ^{Biasutti & Mangiacotti, 2018}; ^{Doi et al., 2017}; ^{Lazarou et al., 2017}; ^{Wang et al., 2020}), but Zhu et al. (²⁰¹⁸) did not demonstrate that the aerobic dance intervention was significantly effective compared to usual care, as global cognitive function was improved in both groups. However, the improvement was substantially greater for participants in the intervention group.

Changes in Specific Cognitive Domains.

Specific cognitive domains were tested in nine studies (^{Biasutti & Mangiacotti, 2018}; ^{Cacciafesta et al., 2010}; ^{Doi et al., 2017}; ^{Domínguez-Chávez et al., 2019}; ^{Douka et al., 2019}; ^{Lake & Goldstein, 2011}; ^{Lazarou et al., 2017}; ^{Shimizu et al., 2018}; ^{Zhu et al., 2018}). Eight studies showed significant improvement in more than one specific cognitive domain after the music intervention, except for one study (^{Lake & Goldstein, 2011}) evaluating the effect of music listening on attention.

Memory.

Four of five studies that included memory as an outcome showed memory improvement after the music interventions (^{Doi et al., 2017}; ^{Domínguez-Chávez et al., 2019}; ^{Lazarou et al., 2017}; ^{Zhu et al., 2018}). In Domínguez-Chávez et al. (²⁰¹⁹), participants engaged in music with movement. After 3 months, their immediate and delayed memory was improved compared to their baseline assessment. In addition, Doi et al. (²⁰¹⁷) demonstrated that memory was significantly improved after 10 months of a dancing intervention (p = 0.011). Interestingly, this improvement in memory was seen only in people with aMCI but not in those with naMCI. In the same vein, memory was significantly improved after dancing interventions in the studies that included only people with aMCI (^{Lazarou et al., 2017}; ^{Zhu et al., 2018}).

However, one study (^{Douka et al., 2019}), which did not differentiate the type of MCI, differed from the previous studies in that memory was significantly decreased (p < 0.001) after the dancing intervention compared to baseline.

Attention.

Four of eight studies demonstrated a better outcome in attention after the music intervention. Significant improvement in attention was shown after 3 months of music with movement (p = 0.007; ^{Domínguez-Chávez et al., 2019}), a 6-month dance program (p < 0.001; ^{Douka et al., 2019}), 10 months of ballroom dancing with music (p = 0.002; ^{Lazarou et al., 2017}), and 3 months of aerobic dance (p < 0.05; ^{Zhu et al., 2018}). These studies used various tools, such as the Digit Span subtest (^{Domínguez-Chávez et al., 2019}), Test of Everyday Attention (^{Douka et al., 2019}; ^{Lazarou et al., 2017}), and Symbol Digit Modalities Test (^{Zhu et al., 2018}), to assess attention.

However, there were inconsistent results in that no significant improvement in attention was found in four other studies (^{Biasutti & Mangiacotti, 2018}; ^{Cacciafesta et al., 2010}; ^{Doi et al., 2017}; ^{Lake & Goldstein, 2011}). Two studies (^{Cacciafesta et al., 2010}; ^{Lake & Goldstein, 2011}) did not show the beneficial effects of music listening on attention when assessed with Trail Making Test A and B (TMT-A and B) and Repeatable Battery for the Assessment of Neuropsychological Status, respectively. Doi et al. (²⁰¹⁷) demonstrated that attention score (TMT-A and B) improved only in 10 months of the dancing group, whereas the score worsened in playing percussion instruments and the health education control group, but the difference in scores failed to reach statistical significance. Similarly, in Biasutti and Mangiacotti (²⁰¹⁸), although the TMT-A score improved after 3 months of musical improvisation (e.g., singing, percussion) compared to baseline, this change did not reach statistical significance.

Verbal Fluency.

Verbal fluency improved after the music interventions in all studies that assessed verbal fluency (^{Biasutti & Mangiacotti, 2018}; ^{Douka et al., 2019}; ^{Lazarou et al., 2017}; ^{Shimizu et al., 2018}). Biasutti and Mangiacotti (²⁰¹⁸) and Douka et al. (²⁰¹⁹) assessed lexical retrieval with a verbal fluency test (VFT), and participants had better VFT scores compared to their baseline scores after a 3-month musical improvisation and 6-month dancing intervention. In Lazarou et al. (²⁰¹⁷), dancing significantly improved verbal fluency assessed with the F-A-S Test compared to baseline, which led to significant differences between the dancing group and no intervention group (p < 0.01). Shimizu et al. (²⁰¹⁸), which implemented rhythmic movements with music, demonstrated significant improvement in the Frontal Assessment Battery (FAB) score, which assesses frontal lobe functions, including lexical fluency, compared to baseline score; however, no significant difference was shown between the intervention group and control group (same exercises without music).

Executive Functioning.

All studies that assessed executive functioning demonstrated that music interventions had positive effects (^{Domínguez-Chávez et al., 2019}; ^{Douka et al., 2019}; ^{Lazarou et al., 2017}; ^{Shimizu et al., 2018}). Two studies (^{Domínguez-Chávez et al., 2019}; ^{Shimizu et al., 2018}) found that music with movement intervention significantly improved MCI participants’ executive functioning measured by FAB after 3 months of intervention (p = 0.002 and p = 0.008, respectively). In addition, dancing interventions significantly enhanced executive function (^{Douka et al., 2019}; ^{Lazarou et al., 2017}). Douka et al. (²⁰¹⁹) provided traditional dancing interventions for 6 months, whereas Lazarou et al. (²⁰¹⁷) implemented ballroom dancing for 10 months, both of which improved executive functioning (p = 0.03 and p = 0.004, respectively).

Spatial Function.

Spatial abilities were assessed in two studies (^{Biasutti & Mangiacotti, 2018}; ^{Cacciafesta et al., 2010}), both of which showed significant improvement in spatial functioning after music interventions compared to baseline. In Biasutti and Mangiacotti (²⁰¹⁸), participants who received musical improvisation intervention scored better on the Clock-Drawing Test, which is used to assess visual-spatial function, than their baseline scores (p < 0.001), whereas the gymnastic activities group showed no difference in test scores.

Cacciafesta et al. (²⁰¹⁰) was the only study that investigated the effectiveness of specific musical pieces and tested the effects of two different kinds of music on cognitive function. In their study, although participants listening to Beethoven’s Für Elise did not improve in any tested cognitive domains, those listening to Mozart’s K448 significantly improved their spatial-temporal abilities. Th e improvement in spatial-temporal abilities was seen constantly over 6 months when stimulation was maintained. Th e study authors assumed that certain pieces of music may activate areas of the brain used for spatial reasoning (^{Cacciafesta et al., 2010}).

Risk-of-Bias Assessment

Using RoB 2 for RCT studies, three studies (^{Doi et al., 2017}; ^{Shimizu et al., 2018}; ^{Zhu et al., 2018}) were found to have a low risk of bias, whereas two studies (^{Biasutti & Mangiacotti, 2018}; ^{Lazarou et al., 2017}) were categorized as “having some concerns” due to missing outcome data as participants were lost to follow up or withdrew from the study. As evaluated by ROBINS-I, two quasi-experimental studies (^{Domínguez-Chávez et al., 2019}; ^{Wang et al., 2020}) were assessed as having a low risk of bias, and the other four studies (^{Adam et al., 2016}; ^{Cacciafesta et al., 2010}; ^{Douka et al., 2019}; ^{Lake & Goldstein, 2011}) had a moderate risk. Th e lower-bias scores in those four studies were due to confounding factors and outcome measurement.

Discussion

Main Findings

The current review examined the effects of music interventions on global cognitive function and each subdomain of cognitive function in older adults with MCI. Th e result of our review provides the potential beneficial effects of music intervention on global cognitive function. Furthermore, the use of music interventions is promising to improve verbal fluency, executive function, and spatial function, but evidence of its effect on memory and attention is limited. However, this finding should be interpreted cautiously due to the heterogeneity of intervention types and the risk of bias of the included studies.

Global Cognitive Function

The current review’s finding is consistent with previous reviews regarding the cognitive benefits of music interventions on global cognitive function (^{Bian et al., 2021}; ^{Moreno-Morales et al., 2020}). However, those reviews are on people with dementia, focusing on global cognitive function, whereas the current review further examined subdomains of cognitive function in addition to global cognitive function and focused on older adults with MCI. In this review, global cognitive function improvement was shown after dancing, musical improvisation, and music with movement, which are all active music therapy. Th ese findings are supported by a previous meta-analysis demonstrating that active music therapy improved global cognitive function in people with dementia (^{Fusar-Poli et al., 2018}). However, the studies included in our review, which used passive music interventions (i.e., listening to music), did not assess global cognitive function. Thus, more investigation is needed to evaluate the effectiveness of passive music interventions on global cognitive function among people with MCI.

Specific Cognitive Domains

Subdomains of cognitive function, such as verbal fluency, executive function, and spatial function, were also improved in all included studies, if assessed, after each music intervention. Th is finding is similar to prior systematic reviews in which music interventions had significant effects on verbal fluency (^{Lam et al., 2020}) and executive function (^{Ito et al., 2022}) in people with dementia/MCI. However, Ito et al. (²⁰²²) also demonstrated a significant improvement in memory after music interventions, whereas our review showed inconsistent findings. Th is discrepancy may be due to differences in participants’ severity of memory deficits in the included studies, as the previous review (^{Ito et al., 2022}), which had only two studies in the analysis for the effects on memory, included people with AD. The current review found that older adults with aMCI, characterized by memory impairment, are more likely to benefit from music interventions than those with naMCI. Further research is necessary to support the idea that music interventions have different effects depending on the subtypes of MCI and the severity of impairment, which will contribute to designing music interventions to prevent the progression of aMCI to AD.

Verbal fluency improved after musical improvisation (e.g., singing, percussion), dancing, and music with movement (^{Biasutti & Mangiacotti, 2018}; ^{Douka et al., 2019}; ^{Lazarou et al., 2017}; ^{Shimizu et al., 2018}). However, in Shimizu et al.’s (²⁰¹⁸) study, there were no significant differences in FAB scores between the music with movement group and control group, who were involved in the same exercises without music, which implies that this positive effect may come from physical movement, as higher physical activity is associated with better verbal fluency (^{Cheval et al., 2021}).

Similarly, studies that demonstrated an enhancing effect on executive function used music interventions that involve physical movement, such as music with movement (^{Domínguez-Chávez et al., 2019}; ^{Shimizu et al., 2018}) and dancing interventions (^{Douka et al., 2019}; ^{Lazarou et al., 2017}). Physical activity, which was shown to be beneficial in improving executive function in patients with AD (^{Zhu et al., 2020}), might be behind this improvement. Thus, more studies with different types of music interventions without physical activity, such as singing, listening to music, and playing instruments, or dancing interventions with active control groups (e.g., same physical movement without music), with a focus on verbal fluency and executive function in older adults with MCI, should be done.

Spatial function was improved after musical improvisation (^{Biasutti & Mangiacotti, 2018}) and listening to specific pieces of music (^{Cacciafesta et al., 2010}), whereas it was not changed after gymnastic activity in the study by Biasutti and Mangiacotti (²⁰¹⁸). The positive emotion evoked by listening to music may contribute to the improvement, as pleasant music listening arouses a positive mood and has a therapeutic effect in improving spatial attention in patients with unilateral spatial neglect (^{Tsai et al., 2013}). However, evidence for the effects of music interventions on spatial function is still limited. Previous research findings are inconsistent regarding the relationship between music interventions and spatial function (^{Miendlarzewska & Trost, 2014}). In the current review, only two included studies examined spatial function as an outcome with a moderate risk of bias, which represents limited evidence. Further music intervention studies with a rigorous research design should include spatial function as a cognitive outcome.

The inconsistent outcomes in attention in the current review may stem from the differences in the intensity of music interventions. The music interventions that were implemented one time or once per week did not have positive effects on attention, whereas those that were conducted twice or three times per week showed favorable effects on attention. Another possible explanation is that the differences in the types of music used in the included studies affected the outcomes of attention. Indeed, the type of music influenced the outcome of cognitive function in one study (^{Cacciafesta et al., 2010}) that showed improvement in spatial-temporal abilities after listening to Mozart’s K448 but not after different pieces of music. More research on the effects of different types of music on each subdomain of cognitive function is needed, as the influence of the type of music differs by cognitive subdomains (^{Bottiroli et al., 2014}). In addition, the effect of physical movement cannot be excluded, as evidence has shown that physical activity is beneficial for cognitive function (^{Mandolesi et al., 2018}). Music interventions that contain physical movement were more likely to improve attention scores than passive music listening and playing a musical instrument. Th us, music interventions without physical activity should be further explored to clarify the unique effect of music itself.

Limitations

Although music interventions showed promising effects on cognitive function in older adults with MCI, findings from the current review should be interpreted with caution due to several limitations. First, included studies were heterogeneous in terms of the types of implemented music interventions, frequency and duration of the interventions, domains of cognitive function assessed, and cognitive assessment tools. In addition, this review included studies with all types of music interventions, but dancing accounted for more than one half of the included studies. As a result, it was difficult to compare the included studies and to describe which intervention affects each cognitive domain the most. Therefore, future studies of diverse music interventions should use standardized cognitive assessments, such as the NIH Toolbox for Assessment of Neurological and Behavioral Function (^{Weintraub et al., 2014}) to improve the analysis of music interventions’ cognitive domain–specific effects. Second, many of the included studies had small sample sizes and were limited by missing outcome data, which can affect results of the review. Th ird, we believe that the systematic search in four databases following PRISMA reporting guidelines found all available studies related to this field, but there is a possibility that studies from other databases were not included.

Implications for Practice and Research

Findings of the current review have important implications for future research and practice to improve cognitive function in older adults with MCI. Albeit with limited evidence, our review provides the promising effects of music interventions on cognitive function, especially on global cognitive function, verbal fluency, executive function, and spatial function. Considering the potential benefits of music interventions with minimal risk of an adverse event, music interventions may be one nonpharmacological strategy to manage cognitive function and prevent the progression of MCI to dementia, further contributing to recovering normal cognitive function. The type of music interventions, the subtype of MCI, and participants’ music preferences could be considered when implementing music interventions. However, more research is required to strengthen the evidence of music interventions’ effects before they can be implemented in the primary care or outpatient setting.

The current review highlights the need for rigorous studies using various types of music interventions without physical activity, such as music listening, singing, and playing musical instruments, to clarify the effects of musical experiences themselves without physical movement. In addition, future research investigating cognitive domain–specific effects with respect to the subtype of MCI should be done to identify optimal treatment strategies. Moreover, individualized music selection should be further investigated to establish robust evidence of its effectiveness on cognitive function and use in practice. Using preferred music for each participant may be more beneficial to their cognitive function given that preferred music listening is effective in managing emotional and cognitive symptoms in older adults with dementia (^{Eggert et al., 2015}; ^{Pérez-Ros et al., 2019}). In the current review, only one study (^{Domínguez-Chávez et al., 2019}) considered participants’ music preferences, demonstrating significant improvement in all five cognitive subdomains assessed in the study.

None of the included studies examined the effects of music interventions on people with MCI with consideration given to their specific chronic conditions. Only two studies (^{Doi et al., 2017}; ^{Domínguez-Chávez et al., 2019}) reported participants’ histories of chronic disease (e.g., hypertension, diabetes, arthritis), but these studies did not examine the effects of music related to certain disease-induced cognitive impairment. Music interventions may have different effects on different chronic conditions as each chronic disease increases the risk of cognitive impairment in a particular mechanism (^{Hill et al., 2021}). More research examining the effects of music interventions on disease-related cognitive decline is recommended to explore and introduce optimal music interventions for patients with cognitive impairment in combination with chronic diseases.

Of note, only one study (^{Lake & Goldstein, 2011}) investigated or controlled participants’ exposure to music in their daily lives (e.g., listening to music while driving) apart from the music interventions provided, which may impact study results. Listening to music while doing another task or activity influences cognitive performance (Angel et al., 2010; ^{Bottiroli et al., 2014}) and brain activation (^{Leggieri et al., 2018}). Therefore, researchers may also consider background music exposure when designing music intervention studies and further investigate how it influences cognitive function among people with MCI, which may advance the knowledge in this field and provide useful implications for practice.

Conclusion

The current systematic review suggests that music interventions significantly improve global cognitive function, verbal fluency, executive function, and spatial function in older adults with MCI. However, findings should be interpreted cautiously because of the heterogeneity and risk of bias of the included studies. More rigorous studies with various types of music interventions investigating cognitive domain–specific effects are needed to better understand the effects of music interventions on cognitive function among older adults with MCI.