Rhythmic auditory stimulation for motor rehabilitation in Parkinson’s disease

Objectives

This is a protocol for a Cochrane Review (intervention). The objectives are as follows:

To assess the effects of rhythmic auditory stimulation on motor rehabilitation in Parkinson's disease.

Background

Definitions

Parkinson’s disease (PD) is a neurological disorder that results from the degeneration of dopaminergic neurones in the basal ganglia of the brain, with devastating consequences for motor, cognitive, and emotional functioning (WHO 2023). Parkinsonism syndrome refers to a group of neurological disorders that share the clinical features of PD, which are tremors, bradykinesia (slowness of movement), rigidity, and postural instability (Ambar 2020). Parkinsonism can have various presentations, such as drug‐induced parkinsonism, vascular parkinsonism, and atypical parkinsonian syndromes (Aludin 2021; Sung 2016; Udagedara 2019), as well as, most commonly, idiopathic PD. The term "idiopathic" refers to a medical condition or disorder of unknown cause or origin. In the context of PD, it means that the specific cause of the disease is unknown, and it is not attributed to any apparent genetic, environmental, or secondary factors.

Epidemiology

PD affects 1 to 2 per 1000 of the world population at any time. Age of onset is typically between 60 and 65 years, and prevalence rises with age; one per cent of the population over 60 years old is affected worldwide (Tysnes 2017). The disorder occurs more frequently in men than in women (Cerri 2019). According to the World Health Organization (WHO) global estimates, in 2019 there were over 8.5 million individuals with PD. There were 329,000 deaths attributable to PD, a number that has more than doubled since 2000 (WHO 2023).

Aetiology

The causes of PD are largely undetermined. It is established that PD is not a uniform entity but rather a complex set of symptoms with various clinical subtypes, pathogenic genes, and putative environmental agents (Jankovic 2020). Although several mechanisms have been described (Bellou 2016), the most important risk factor for this disease is age (Reeve 2014). Research has identified several genetic factors in families with hereditary PD (Hernandez 2016). Nevertheless, similar incidence between monozygotic and dizygotic twins leaves room for debate regarding the role of genetics in the aetiology of the disease (Balck 2019). Exposure to substances such as magnesium, aluminium, mercury, or pesticides has been related to the physiopathogenesis of PD (Ascherio 2016), and current research exploring the origin of the disease has suggested that it may start in the gut (Borghammer 2019), or the olfactory bulb (Fullard 2017).

Pathophysiology

The underlying pathophysiology of the disease results from the progressive degeneration of dopaminergic neurones in the part of the brain known as the substantia nigra pars compacta (SNc), which impairs dopaminergic input into the striatum (caudate and putamen) and alters the disinhibition mechanisms that are essential for fine‐tuning between the facilitation and suppression of movements (Purves 2012). This leads to excessive inhibition of motor areas in the cortex, subcortex, and brainstem, and a series of functional changes that mediates the main motor features of PD: tremor at rest, bradykinesia, rigidity, and postural instability (Opara 2017). Notably, such dysfunctions involve not only the basal ganglia, but also its dopaminergic corticostriatal circuits and key brain areas such as the primary motor cortex (M1), premotor cortex (PMC), supplementary motor area (SMA), primary somatosensory cortex (SI), prefrontal cortex (PFC), thalamus (THA), and cerebellum (CRBL), as seen in diverse neuroimaging studies (Kim 2017; Tahmasian 2017).

Diagnosis

PD is diagnosed through a comprehensive assessment that involves clinical symptoms, detailed medical history, and neurological examination (Postuma 2015). Blood tests and magnetic resonance imaging (MRI) may be employed to exclude alternative causes and ensure a more accurate diagnosis (Bidesi 2021). Dopamine transporter (DaT) imaging scans allow dopamine levels in the brain to be visualised, which can assist in differential diagnosis. Additionally, a positive response to levodopa, a medication that replaces dopamine, can help confirm the diagnosis (Poewe 2017).

Motor symptoms

The progression of motor symptoms, especially bradykinesia, determines the level of disability and rate of deterioration in quality of life (QoL) in PD (Zhao 2021). Bradykinesia refers to a generalised slowing down of movements, and has further implications including decreased facial expressions, decreased blinking frequency, monotonous and whispering speech, and micrography or small handwriting (Armstrong 2020). The tremor at rest is an oscillatory distal movement of 4 to 6 Hz that predominantly affects the hands, but can also affect lips, tongue, jaw, and lower limbs. The tremor is typically asymmetric in early stages of the disease, and it is the most frequent clinical sign (Bhidayasiri 2005). The characteristic shuffling gait of PD is small, slow steps, with the feet close together, and the body leaning forward (Gong 2023). The normal mechanics of gait are further disrupted by reduced arm swing, festination (short, increasingly rapid steps), and in some cases, freezing of gait (Nonnekes 2019). As the disease progresses, these symptoms can become more severe and can significantly impact a person's ability to perform daily activities, such as walking, dressing, and eating (Leite 2023).

Non‐motor symptoms

Non‐motor symptoms may also have a significant impact on QoL (Schapira 2017). Such symptoms include cognitive impairment, fatigue, depression, anxiety, sleep disorders, constipation, and speech and voice problems. It is important to note that the presentation of PD can vary from person to person, and that the severity of these symptoms can change over time. Individuals with PD work together with healthcare professionals to manage these symptoms and improve QoL. The burden of these symptoms is particularly pronounced in disadvantaged populations, where additional health and QoL challenges may be present (Welsh 2004).

Cognitive decline is a common occurrence amongst individuals living with PD, and can range from mild impairment to more severe dementia, which affects 80% of individuals in later stages of the disease (Jellinger 2018). These cognitive impairments often affect memory, executive function, and attention (Roheger 2018).

Other comorbidities such as depression and anxiety are very common in PD (Sagna 2014). The prevalence of depression in PD is notably higher than in the general population; it varies around 38% depending on the stage of the disease (Cong 2022), and approximately 20% of people with PD develop major depressive disorder (Riedel 2016). Similarly, the prevalence of anxiety varies between 26% and 40%, and has been associated with the worsening of motor symptoms (Upneja 2021).

Sleep disorders are also prevalent in PD and have a significant impact on QoL. Studies suggest that up to 60% to 90% of people with PD may experience sleep‐related problems that can take various forms, including insomnia, fragmented sleep, restless leg syndrome, periodic limb movements, and rapid eye movement sleep behaviour disorder (RBD) (Loddo 2017; Mahmood 2020).

Temporal‐processing dysfunction

Progressive structural and functional changes in the brain strongly suggest that timing and temporal‐processing mechanisms are altered in PD (Grahn 2009). Individuals with PD often exhibit deficits in their ability to accurately perceive and process time intervals, leading to difficulties in performing tasks that require precise timing, such as motor co‐ordination and perception of rhythm (Parker 2013). This temporal impairment can manifest as motor‐timing deficits, which contribute to the characteristic bradykinesia in PD (Wu 2015). Additionally, it may affect cognitive function, such as working memory, attention, and decision‐making processes (Harrington 2011). The underlying neural mechanism behind this temporal processing dysfunction is still unclear, but it is presumed to follow disruptions in the basal ganglia and interconnected brain regions responsible for timing and synchronisation (Jones 2014).

Recent studies comparing people with PD with controls have shown that the brain has specialised networks involved in temporal prediction that seem to work like an internally generated “clock” (Breska 2018; Schwartze 2013). The function of this brain clock seems to depend on effective communication between the SMA and the basal ganglia, which creates a mental model of the temporal structure of successive stimuli (or actions), and produces a “pacing” activating the SMA, PMC, M1, and CRBL (Andersen 2021; Jones 2014), which are all affected to some degree in PD. Interestingly, it has been shown that the degree of timing dysfunction in PD is related to the stage of the disease (Nombela 2013). The emerging understanding of temporal‐processing deficits and their neural mechanisms adds a fascinating dimension to our exploration of this challenging condition, and offers potential avenues for further research and therapeutic interventions.

Treatment

Current evidence‐based treatment for idiopathic PD typically involves a combination of medications and lifestyle adjustments. Commonly prescribed medications for PD include levodopa, dopamine agonists, monoamine oxidase B inhibitors (MAO‐B), and catechol‐O‐methyltransferase inhibitors (COMT). These medications aim to increase levels of dopamine in the brain, which can improve symptoms such as tremors, stiffness, and difficulty with movement (Serva 2022). However, the effect of drug treatments on PD symptoms is modest, and currently there is no therapy that can stop the progression of the disease. Moreover, the combination of multiple medications and the occurrence of side effects contribute to challenges in the effective management of motor symptoms (Bhagavathula 2022; Oonk 2022). Invasive treatment such as deep brain stimulation (DBS), which may be tried when standard treatment is not effective, is not appropriate for all individuals (Rasiah 2022). In light of these limitations, non‐conventional interventions that appear to be safe, are non‐invasive, cheap, and easy to access have gained attention (Ashoori 2015; Bloem 2015).

Description of the intervention

Rhythmic auditory stimulation (RAS) is a technique that employs auditory cues to improve gait and balance issues in PD, or other movement disorders, by promoting temporal processing in the brain through the steady beat of a metronome, or music (Braun 2021; Burrai 2021; Thaut 1996). It is important to note that RAS is not intended as a substitute for conventional drug treatments for PD, but rather as a complement to them.

While there is no standardised way to deliver RAS, there are key elements such as rhythmicity (i.e. rhythmical quality), motor synchronisation, customisation, and the use of gait kinematics as outcome measures. The intervention consists of a healthcare professional delivering auditory cues with beats‐per‐minute (bpm) that are adapted to the individual’s baseline gait measured in steps‐per‐minute (spm). Then, features of RAS, such as the tempo of the beat, or the complexity of the rhythm, can be modified from baseline, often producing a short‐term improvement in gait kinematics, such as speed, cadence, and step length, and others (Wang 2022). The therapy is typically conducted in a supervised setting, such as a physical therapy clinic or hospital, and is usually for a duration of 30 to 60 minutes, two to three times a week for several weeks or months.

The procedure used to deliver and assess the intervention is variously described in research studies (Forte 2021; Wang 2022; Ye 2022), but may include the following.

1. Assessment: before initiating RAS intervention, a comprehensive assessment of the patient's gait, balance, and motor function is conducted to establish baseline measurements and identify specific areas of impairment.

2. Individualised protocol: an individualised RAS protocol tailored to the patient's specific needs is developed, including the type, tempo, intensity, frequency, and duration of the RAS intervention. Consideration is given to the patient's gait characteristics, cognitive abilities, and overall functional status.

3. Selection of auditory cues: appropriate auditory cues for RAS, such as rhythmic music or metronome beats, are selected based on the patient's preferences and ability to synchronise movements with the auditory stimuli.

4. Delivery: the therapist delivers the auditory cues during therapy sessions, through speakers or headphones.

5. Synchronisation: the patient synchronises their movements, such as walking or performing specific exercises, with the auditory cues provided.

6. Monitoring and adjustments: throughout the session, the therapist monitors the patient's response to the auditory cues and makes adjustments as needed to optimise the effectiveness of the intervention.

7. Progress evaluation: progress is evaluated regularly to assess improvements in gait, movement, and overall motor function. Adjustments to the intervention may be made based on the patient's progress and goals.

8. Home rehabilitation consideration: the potential extension of RAS to home rehabilitation is explored under the guidance of healthcare professionals, taking into account the safety and effectiveness of implementing RAS in the patient's daily life activities.

9. Safety precautions: the importance of safety during RAS training is emphasised, particularly when considering the potential risk of falls or balance disturbances. Patients and caregivers are educated on safety measures and potential risks associated with RAS intervention.

10. Long‐term management: depending on the patient's needs and goals, the intervention may be continued over an extended period to maintain and further improve motor function and mobility.

While there are studies that show improvements in gait parameters, the short‐ and long‐term effects of RAS are still under investigation and require further validation (Calabrò 2019). Importantly, while RAS is considered safe, it may not be suitable for all individuals with PD, particularly those in the late stages of the condition, those with hearing impairments, or those with cognitive deficits (Rochester 2009). The effect of RAS can be influenced by various modifying factors such as the type of intervention (e.g. metronome, music, complexity, tempo), which can impact treatment response according to the specific features and mechanisms of each intervention type (Rose 2019). For example, some individuals may benefit from slower and simpler rhythmic patterns, while others may benefit from faster and more complex music (Liu 2018). Additionally, the duration of the intervention provides insights into how the timing and intensity of interventions affect outcomes. By examining outcomes at different intervals, it is possible to study the temporal dynamics of treatment effects, identifying immediate versus sustained benefits and potential fluctuations in response over time (Bella 2018; Erra 2019). Another critical factor is whether the intervention selection is made by the participant or the researcher, as this may influence treatment adherence and efficacy (Garza‐Villarreal 2017; King 2005). Moreover, the intervention provider (e.g. health provider, researcher, self) may play a significant role in treatment delivery and patient engagement (Boone 2021). Furthermore, the time since PD diagnosis may be important, with disease progression and age‐related changes potentially impacting the intervention's effectiveness (Chesnokova 2019).

How the intervention might work

The mechanisms by which RAS may work in PD are still debated (Braun 2019). The idea is that top‐down and bottom‐up integration of information in the brain results from audio‐motor interactions (Pando‐Naude 2021). The concept of audio‐motor coupling, as a process through which the brain uses auditory information to co‐ordinate movement, is a theoretical framework, but this notion is rooted in a substantial body of scientific literature on auditory and motor interactions (Hickok 2003; Novembre 2014; Rodriguez‐Fornells 2012). This process seems to be driven by temporal predictions (Koelsch 2019; Vuust 2022), may be influenced by training (Bermudez 2005; Criscuolo 2022), and has been associated with cognitive and reward processing (Koelsch 2013; Koelsch 2020). The interplay between auditory and motor systems has garnered considerable attention in the neuroscience and psychology fields, with numerous studies providing evidence of the brain's capacity to synchronise motor actions with auditory cues, particularly in the context of timing and rhythm (Chen 2008; Leow 2014; Morillon 2014).

Research has shown the potential of RAS, specifically, to influence motor rehabilitation (Thaut 2010). Regularity of auditory stimuli, such as the metronomic beats employed in RAS, has been shown to affect reaction times and motor synchronisation in both healthy individuals and those with neurological conditions (Thaut 2015). Moreover, the inherent rhythm and musicality of RAS can tap into the unique elements of music that have been known to elicit spontaneous movement responses ‐ an observation deeply rooted in the psychological and neuroscientific study of music and rhythm perception (Large 2000; Nombela 2013; Thaut 2005). Recent studies investigating why and how music makes us want to move have also implicated audio‐motor interactions, driven by predictive processes and embodied affective responses (Vuust 2022). The experience of a pleasurable urge to move to music (PLUMM), colloquially referred to as a "groove" sensation, seems to depend on the interaction between auditory, motor, and emotion‐related areas of the brain, with the basal ganglia and its dopaminergic corticostriatal circuits playing a crucial role (Matthews 2020). Notably, PLUMM is affected by PD; people with PD prefer rhythms with low complexity (i.e. very predictable), compared to controls who prefer medium complexity (i.e. "groove") (Pando‐Naude 2023). Such results suggest that RAS may work in PD by facilitating audio‐motor interactions in the brain and promoting better guidance for the temporal model of movements (Leuk 2020).

Why it is important to do this review

Whilst conventional treatment for motor symptoms in PD has improved considerably, people with PD still experience significant deterioration in movement. Specifically, gait hypokinesia (slow walking) is one of the main movement disorders associated with PD and is a crucial determinant of disability and deterioration in QoL. Studies of RAS for PD have shown mixed results. Some studies have reported positive effects on gait parameters such as velocity, stride length, and motor function (Ghai 2018; McIntosh 1997; Wang 2022), while others have found a negative effect on cadence (Ghai 2018). Despite this, the overall benefit of RAS on gait parameters appears promising, and there is still much research needed to determine the optimal type and dosage of RAS for motor rehabilitation in PD.

We have identified recent systematic reviews investigating the effect of RAS on PD (Forte 2021; Wang 2022; Ye 2022). In our review, the major difference is the definition of outcomes and how these will be analysed. Previous reviews have covered a broad range of studies related to RAS, but have used Part III of the Motor Examination of the Movement Disorder Society's Unified Parkinson's Disease Rating Scale (MDS‐UPDRS‐III) as the sole primary outcome (Goetz 2008). Thus, a comprehensive synthesis of additional gait kinematics and non‐motor outcomes, including cognitive functioning, depression, anxiety, and QoL is of great interest, not only because these are prevalent comorbidities in PD, but also to shed light on the cognitive and emotional mechanisms that underlie the way this intervention might work.

The quality and certainty of evidence in previous reviews has been assessed with a variety of tools, some of which are used only for specific populations and designs. This review will use the original Cochrane risk of bias tool (RoB 1) to assess the risk of bias in the included studies (Higgins 2011), and the GRADEpro GDT tool to assess the certainty of the evidence. The GRADEpro GDT tool is based on the GRADE (Grading of Recommendations Assessment, Development, and Evaluation) method (Guyatt 2008), which provides a structured and transparent approach to evidence assessment and guideline development, ensuring clarity, consistency, and proper documentation. By conducting this review, we are not intending to develop clinical practice guidelines, but rather to provide an assessment of the evidence base that will then allow guideline developers to make evidence‐based recommendations more easily. The tool supports this as it emphasises a comprehensive evaluation of the evidence, including the certainty of evidence and the balance between benefits and risks.

A systematic review and meta‐analysis using gold‐standard methodology will be of wide interest, given that the most common type of parkinsonism syndrome is the idiopathic presentation of PD, which affects one per cent of the population above 60 years. This Cochrane Review will evaluate the current evidence for the effect of RAS on motor rehabilitation in PD, and is expected to:

• provide further evidence for an accessible and potentially effective intervention that could benefit a broad range of individuals, including those who may face barriers to accessing other forms of treatment;

• include a comprehensive assessment of outcomes beyond the normal motor parameters, given that non‐motor aspects of PD can significantly impact an individual's well‐being and may be particularly relevant to disadvantaged populations that might face additional challenges related to their overall health and QoL;

• use strict inclusion criteria, methodology, and analysis according to the Cochrane Handbook for Systematic Reviews of Interventions to enhance the reliability of the conclusions and ensure that the evidence gathered is of high quality and can be more confidently applied in clinical practice (Higgins 2023); and

• address health equity by providing an accurate and up‐to‐date summary of the evidence that guideline developers, health professionals and individuals with PD can use to ensure access to potentially beneficial treatments like RAS is not limited by social, economic, or demographic characteristics.

Objectives

To assess the effects of rhythmic auditory stimulation on motor rehabilitation in Parkinson's disease.

Methods

Criteria for considering studies for this review

Types of studies

We will include randomised controlled trials (RCTs), where participants are randomly allocated to groups using established methods like random number generators. Since participants cannot be blinded to rhythmic auditory stimulation (RAS) treatment, we will include open‐label or single‐blind trials.

Types of participants

Individuals with medically‐diagnosed idiopathic Parkinson's disease (PD). No restrictions will be imposed based on age, sex, ethnicity, prescribed medications, disease severity, duration since diagnosis, or any other treatment being received, including having undergone or currently undergoing deep brain stimulation (DBS) or surgery. We will rely on statements from study authors for confirmation of the medical diagnosis. We will include studies presenting results from PD participants only; we will exclude studies that encompass only a subset of relevant participants (e.g. studies also involving people with other degenerative disorders).

Types of interventions

The intervention to be studied is RAS for motor rehabilitation in PD. The intervention can be self‐administered or administered by research personnel. For an RAS intervention to be eligible for inclusion in the review, it should include the following key elements.

1. Rhythmicity: an essential element of RAS is the presence of rhythmic auditory cues. Traditionally, RAS is administered via music or metronomic beats. Thus, the stimuli should have a clear and consistent rhythmic structure. This rhythm serves as the temporal framework for movement synchronisation.

2. Customisation: RAS may be tailored to an individual's needs and abilities. This customisation includes music genres, tempos, and auditory cues that align with the therapeutic goals and the specific condition being targeted. There will be no restrictions on the customisation of RAS including rhythmic complexity, frequency, duration, or intensity.

3. Synchronisation: the intervention should involve motor synchronisation with auditory stimuli.

4. Outcome measures: objective outcome measures should be used to assess the effectiveness of RAS. These measures may include assessments of motor function (gait and balance), cognitive performance, and emotional well‐being (depression, anxiety, QoL).

5. Duration and dosage: the duration and dosage of RAS can vary widely depending on the specific therapeutic goals, the population being targeted, and the nature of the intervention. We will include interventions of any duration or dosage. Duration refers to the time frame over which the intervention is delivered and outcomes are assessed (e.g. single session, short term (up to one month), medium term (one to three months), or long term (over three months)). Dosage refers to the duration of the stimuli during each session of the intervention (e.g. seconds, minutes).

We will investigate the effect of RAS plus treatment as usual (TAU) versus TAU only or no intervention (e.g. waitlist control). We will not test RAS versus TAU directly. TAU can include physiotherapy or pharmacological, psychological, or alternative complementary treatment, as long as the intervention and control groups receive the same treatment(s).

Types of outcome measures

PD outcomes are assessed by many validated measurement tools. This review will include three primary outcome domains based on a conceptual model of how the intervention might work: PD motor evaluation, gait kinematics, and adverse events. Gait kinematics may include several measured outcomes, which will be analysed independently, but which we will put into a hierarchy to include in a summary of findings table. Secondary outcome domains will include non‐motor symptomatology such as cognitive function, depression, anxiety, and QoL.

We will categorise the time points for outcome assessment as baseline, single session, short term (up to one month), medium term (one to three months), or long term (over three months). These time points will be assessed for all outcomes where available.

The outcomes and specific measurement tools described below are not criteria for study inclusion.

Primary outcomes

1. PD motor evaluation

   1. Movement Disorder Society Unified Parkinson's Disease Rating Scale, Part III Motor Examination (MDS‐UPRDS‐III, score 0 to 132 points) (Goetz 2008)

2. Gait kinematics

   1. Gait velocity (metres per second)

   2. Gait cadence (steps per minute)

   3. Gait distance (metres)

   4. Gait time (seconds)

   5. Step length (centimetres)

   6. Step time (seconds)

   7. Balance (score)

3. Adverse events

   1. Falls

   2. Balance issues

   3. Exacerbation of PD symptoms

   4. Dizziness, vertigo, or nausea

   5. Fatigue

   6. Headache

   7. Interference with hearing or auditory function

Secondary outcomes

1. Cognitive function

   1. Montreal Cognitive Assessment (MoCA), score 0 to 30 points, higher is better (Nasreddine 2005)

   2. Mini‐Mental State Examination (MMSE), score 0 to 30 points, higher is better (Tombaugh 1992)

   3. Frontal Assessment Battery (FAB), score 0 to 18 points, higher is better (Dubois 2000)

2. Depression

   1. Major Depression Inventory (MDI), score 0 to 50 points, lower is better (Olsen 2003)

   2. Beck Depression Inventory ‐ Second Edition (BDI‐II), score 0 to 63 points, lower is better (Rosner 2015)

   3. Hamilton Rating Scale for Depression (HRSD), score 0 to 76 points, lower is better (Sharp 2015)

   4. Patient Health Questionnaire‐9 (PHQ‐9), score 0 to 27 points, lower is better (Kroenke 2001)

3. Anxiety

   1. State‐Trait Anxiety Inventory (STAI), score 20 to 80 points, lower is better (Spielberger 1983)

   2. Hamilton Anxiety Rating Scale (HAM‐A), score 0 to 56 points, lower is better (Thompson 2015)

   3. Beck Anxiety Inventory (BAI), score 0 to 63 points, lower is better (Beck 1988)

4. Quality of Life (QoL)

   1. Parkinson's Disease Questionnaire 39 (PDQ‐39), score 0 to 100, lower is better (Jenkinson 1997)

   2. World Health Organization Quality of Life instrument (WHOQOL), score 4 to 20 points, higher is better (WHO 1982)

   3. Short Form 36 (SF‐36), score 0 to 100 points, higher is better (Ware 1992)

   4. Quality of Life Enjoyment and Satisfaction Questionnaire (Q‐LES‐Q), score 0 to 100, higher is better (Stevanovic 2011)

Search methods for identification of studies

We will search for all applicable studies. There will be no language barriers for either search criteria or trial selection. We will translate non‐English articles as needed. Our comprehensive systematic review and meta‐analysis will follow procedures from the Cochrane Handbook for Systematic Reviews (Higgins 2023), and will be carried out in accordance with the PRISMA statement (Page 2021).

Electronic searches

To identify relevant studies, we will search the following electronic databases, without restricting by date or format of publication.

1. Cochrane Central Register of Controlled Trials (CENTRAL), part of The Cochrane Library

2. PubMed (MEDLINE)

3. Embase (Elsevier)

4. CINAHL (EBSCO)

5. PsycINFO (ProQuest)

6. Web of Science (Science Citation Index Expanded, Social Sciences Citation Index, Arts and Humanities Citation Index, Conference Proceedings Citation Index ‐ Science and Conference Proceedings Citation Index ‐ Social Science and Humanities)

7. Scopus (Elsevier)

8. WHO (International Clinical Trials Registry Platform (ICTRP))

9. ClinicalTrials.gov

10. Current Controlled Trials (www.controlled‐trials.com/)

11. RILM (Répertoire International de Littérature Musicale) Abstracts of Music Literature

Searching other resources

We will search the bibliographies of retrieved articles and relevant reviews to identify trials we may have missed; perform a citation search on ISI Web of Science to discover more recent studies; contact experts in the field; and search conference proceedings databases to identify unpublished trials.

Data collection and analysis

Selection of studies

Two authors (VPN and KVJ) will independently screen all titles and abstracts. Any records that mention a trial in the title or abstract will be retrieved in full, as will those with insufficient information in the title or abstract to make a decision about eligibility. The two authors will then independently assess the full text of the articles using the predetermined inclusion criteria. We will discuss any disagreements and resolve them by involving a third review author (PV), if necessary. We will keep a record of excluded articles and the reasons for their exclusion. We will perform screening of potential studies using Excel spreadsheets and the Covidence tool (Covidence 2023).

Data extraction and management

Two authors (VPN and KVJ) will independently extract data from reports of included studies using a piloted Excel spreadsheet and the Covidence tool (Covidence 2023). We will discuss any disagreements and resolve them by involving a third review author (PV), if necessary. We will extract the following information from each study, and any other relevant details.

1. Information about data extraction from reports: names of data extractors, date of data extraction, publication's digital object identifier (DOI)

2. Eligibility criteria: confirmation of study eligibility for the review, reason for exclusion

3. General information: author, year of publication, title, journal, country

4. Study design: type of study design, inclusion and exclusion criteria, description of control (TAU), method of randomisation and allocation concealment, blinding, loss to follow‐up, analysis method (per protocol or intention‐to‐treat)

5. Participants: setting, region, total sample size, number in experimental group, number in control group, age, sex, ethnicity, diagnosis, comorbidities, duration of disorder, pharmacological treatments, surgical treatments

6. Intervention: type (e.g. metronome, music), characteristics (e.g. tempo, rhythmic complexity, naturalistic), selection (participant‐ or researcher‐selected), provider (health provider, researcher, self), duration (e.g. single session, short‐term intervention (up to one month), medium‐term intervention (one to three months), or long‐term intervention (over three months)), dosage (e.g. seconds, minutes), delivery (e.g. headphones or loudspeakers)

7. Outcomes: for each prespecified primary and secondary outcome domain, we will document:

   1. evidence that the outcome domain was assessed, measurement tool or instrument used, upper and lower limits, whether a high or low score is favourable, definitions of any thresholds, if appropriate

   2. specific metric (e.g. change in score from baseline to a post‐intervention time point)

   3. method of aggregation (e.g. mean and standard deviation of scores)

   4. time frame of outcome measurements

Assessment of risk of bias in included studies

Two authors (VPN and KVJ) will evaluate the risk of bias in the included studies using the risk of bias (ROB) tool and criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We will assess the following domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting bias, other sources of bias, and overall risk of bias in each study. We will categorise the risk of bias as low, high, or unclear based on the method used and the potential for bias. In case of disagreement, we will consult a third author (PV). We will also contact the authors of the studies if information is missing that we require to evaluate trial methodology.

Measures of treatment effect

Two authors (VPN and KVJ) will independently extract data from the included studies. We will use Review Manager 2023 for data entry and analysis.

Dichotomous data

For dichotomous data, we will calculate risk ratios (RR) with 95% confidence intervals (CIs).

Continuous data

For continuous data, we will compare the mean difference (MD) between groups if the same scale (e.g. UPDRS) is used in different trials. If different scales are used for the same outcome, we will use the standardised mean difference (SMD) to combine data from the trials. We will present results with 95% confidence intervals (CI). If a study provides multiple measures of the same construct, we will calculate the average SMD and average variance across these outcomes. If a study reports both continuous and dichotomous data for the same outcome, we will convert the risk ratio (RR) to an SMD to allow for pooling of both continuous and dichotomous data, as described in section 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2023).

Ordinal data

We will analyse ordinal data as continuous data and the intervention effect will be expressed as a difference in means.

Cross‐over trials

We will analyse cross‐over trials using the first‐period data.

Cluster‐randomised trials

If we identify trials using clustered randomisation, we will assess whether the trialists controlled for clustering effects. If not, we will attempt to contact the authors for individual participant data to calculate the intracluster correlation coefficient (ICC). If this is not feasible, we will use external ICC estimates from similar trials or populations. We will use the ICC to re‐analyse the trial data, and conduct sensitivity analyses to evaluate the impact of ICC variations.

Multi‐arm studies

For studies with multiple treatment arms, we will only consider comparisons between RAS + TAU and TAU only. If there are multiple RAS interventions and sufficient information to evaluate their similarity, we will combine similar RAS interventions for a single pair‐wise comparison.

Dealing with missing data

For studies included in the review, we will track levels of attrition and missing data. If any information is missing from the study reports, we will try to retrieve it by contacting the authors. If complete outcome data is not available due to loss of follow‐up or dropout, we will attempt to obtain it from authors. If the authors do not respond or data remain missing, we will fill in missing data by imputing the means as replacement values and treating these as if they were observed. We will assess the impact of missing data on the results of the review by conducting sensitivity analyses and will present the findings in the 'Discussion' section of the review.

Assessment of heterogeneity

We will report the similarity of the interventions (e.g. dose, frequency), participants (e.g. age), trial design (e.g. allocation concealment, blinding), and outcomes. We will assess the heterogeneity of treatment response visually from the forest plot of the RR and using the Chi² test. In addition, we will use the I² statistic to assess heterogeneity statistically for each comparison and outcome. We will assume substantial heterogeneity if I² is greater than 50%, indicating that 50% of the variability in the outcome cannot be explained by sampling variation. For studies with substantial heterogeneity, we will investigate the source of heterogeneity through subgroup analyses based on participant characteristics and intervention characteristics, as described below. We will address the potential impact of heterogeneity on the findings of the review in the 'Discussion' section of the review.

Assessment of reporting biases

We will assess potential reporting bias in studies by generating and inspecting funnel plots. We will conduct visual inspection and statistical tests ‐ such as Bee and Mazumdar (Begg 1994), and Egger (Egger 1997) ‐ to detect asymmetry. If asymmetry is found, we will perform exploratory analysis to determine if it is due to publication bias or a real relationship between trial size and effect size. Our comprehensive search strategy, including evaluation of published and unpublished literature, will minimise the potential for publication bias.

Data synthesis

We will use Review Manager 2023 to enter data from all included studies, and we will check for inputting errors. We will conduct meta‐analyses using the inverse variance method. We will use both fixed‐effect and random‐effects models, and conduct a sensitivity analysis to investigate the impact of the analysis method. If both models produce similar results, we will report the random‐effects model for better representation of variability. If the results are different, we will explore the reasons for heterogeneity amongst trials. Where data is not available for meta‐analysis, we will summarise the results in a narrative form.

Subgroup analysis and investigation of heterogeneity

We will carry out the following subgroup analyses and meta‐regressions (ranked for importance) to investigate heterogeneity, depending on the data reported. Our subgroup analyses and meta‐regressions are exploratory, and we will conduct them as recommended in the Cochrane Handbook for Systematic Reviews of Interventions, sections 10.11.3 and 10.11.4 (Higgins 2023). We require a minimum of 10 studies to conduct meaningful subgroup analyses or meta‐regressions.

Subgroup analyses

1. Type of intervention (metronome/beats, music) (Rose 2019)

2. Tempo of intervention (fast tempo (> 120 bpm), medium tempo (76 to 120 bpm), slow tempo (60 to 76 bpm)) (Liu 2018)

3. Intervention selection (by participant or researcher) (King 2005)

4. Intervention provider (health provider, researcher, self) (Boone 2021)

Meta‐regression analyses

1. Time since PD diagnosis (number of years)

2. Age (number of years)

Sensitivity analysis

1. Risk of bias: we will conduct sensitivity analysis to determine the impact of risk of bias on the results of meta‐analyses by excluding trials rated as being at high risk of bias for random sequence generation, allocation concealment, and blinding of outcome assessment, as recommended in Chapter 13 of the Cochrane Handbook for Systematic Reviews of Interventions (Page 2023). In the context of RAS for PD, selection bias (due to lack of randomisation and concealment), performance bias (due to lack of blinding of participants and personnel), and detection bias (due to lack of blinding in outcome assessment) may have a more substantial impact on the reliability of the results compared to attrition or reporting bias. We will not consider performance bias in sensitivity analysis as this is impossible to avoid in studies of RAS for PD.

2. Choice of meta‐analysis model: we will assess the possible impact of random‐effects models on inflated effect sizes, which may arise when a number of small studies are included, by repeating our analyses using fixed‐effect model as described in Data synthesis above.

Summary of findings and assessment of the certainty of the evidence

We will present a summary of findings (SoF) table to summarise the effects of RAS + TAU compared to TAU for PD. We will use the procedure described by Schünemann 2022. The key findings from the review will include the quantity of data, magnitude of effect size, and overall certainty of evidence. Our SoF table will include the measurement most often used within each of the primary and secondary outcomes: PD motor evaluation (UPDRS‐III), gait kinematics, adverse events, cognitive function, depression, anxiety, and QoL. We will present the end‐of‐treatment results.

Two review authors (VPN and KJV) will assess the certainty of evidence using the GRADEpro GDT approach across domains of study limitations, publication bias, imprecision of results, inconsistency of results, and indirectness of evidence. We will categorise our overall judgement regarding the certainty of the evidence as high, moderate, low, or very low. We will resolve any disagreement by discussion with a third review author (PV).

Appendices

Appendix 1. Search strategies

CENTRAL

((Parkinson's disease OR Parkinson disease) AND (auditory stimulation OR acoustic stimulation OR auditory cue OR acoustic cue OR rhythm OR metronome OR music))

PubMed

(((("Music"[Mesh]) OR ("Acoustic Stimulation"[Mesh])) OR ("auditory stimulation" OR "acoustic stimulation" OR "auditory cue*" OR "acoustic cue*" OR rhythm* OR metronome * OR music*)) AND ("Parkinson Disease"[Mesh] OR parkinson*[Title/Abstract])) AND ((((("Randomized Controlled Trial" [Publication Type]) OR "Controlled Clinical Trial" [Publication Type]) OR (randomized[Title/Abstract] OR placebo[Title/Abstract] OR randomly[Title/Abstract])) OR ("Clinical Trials as Topic"[Mesh])) OR (trial[Title]) ))

Embase

(('randomized controlled trial'/exp OR 'randomized controlled trial' OR 'controlled clinical trial'/exp OR 'controlled clinical trial' OR random*:ti,ab,tt OR 'randomization'/exp OR 'randomization' OR 'intermethod comparison'/exp OR 'intermethod comparison' OR placebo:ti,ab,tt OR compare:ti,tt OR compared:ti,tt OR comparison:ti,tt OR ((evaluated:ab OR evaluate:ab OR evaluating:ab OR assessed:ab OR assess:ab) AND (compare:ab OR compared:ab OR comparing:ab OR comparison:ab)) OR ((open NEXT/1 label):ti,ab,tt) OR (((double OR single OR doubly OR singly) NEXT/1 (blind OR blinded OR blindly)):ti,ab,tt) OR 'double blind procedure'/exp OR 'double blind procedure' OR ((parallel NEXT/1 group*):ti,ab,tt) OR crossover:ti,ab,tt OR 'cross over':ti,ab,tt OR (((assign* OR match OR matched OR allocation) NEAR/6 (alternate OR group OR groups OR intervention OR interventions OR patient OR patients OR subject OR subjects OR participant OR participants)):ti,ab,tt) OR assigned:ti,ab,tt OR allocated:ti,ab,tt OR ((controlled NEAR/8 (study OR design OR trial)):ti,ab,tt) OR volunteer:ti,ab,tt OR volunteers:ti,ab,tt OR 'human experiment'/exp OR 'human experiment' OR trial:ti,tt) NOT (((((random* NEXT/1 sampl* NEAR/8 ('cross section*' OR questionnaire* OR survey OR surveys OR database OR databases)):ti,ab,tt) NOT ('comparative study'/de OR 'controlled study'/de OR 'randomised controlled':ti,ab,tt OR 'randomized controlled':ti,ab,tt OR 'randomly assigned':ti,ab,tt) OR ('cross‐sectional study' NOT ('randomized controlled trial'/exp OR 'controlled clinical trial'/de OR 'controlled study'/de OR 'randomised controlled':ti,ab,tt OR 'randomized controlled':ti,ab,tt OR 'control group':ti,ab,tt OR 'control groups':ti,ab,tt)) OR ('case control*':ti,ab,tt AND random*:ti,ab,tt NOT ('randomised controlled':ti,ab,tt OR 'randomized controlled':ti,ab,tt)) OR ('systematic review':ti,tt NOT (trial:ti,tt OR study:ti,tt)) OR (nonrandom*:ti,ab,tt NOT random*:ti,ab,tt) OR 'random field*':ti,ab,tt OR (('random cluster' NEAR/4 sampl*):ti,ab,tt) OR (review:ab AND review:it)) NOT trial:ti,tt OR ('we searched':ab AND (review:ti,tt OR review:it)) OR 'update review':ab OR ((databases NEAR/5 searched):ab) OR ((rat:ti,tt OR rats:ti,tt OR mouse:ti,tt OR mice:ti,tt OR swine:ti,tt OR porcine:ti,tt OR murine:ti,tt OR sheep:ti,tt OR lambs:ti,tt OR pigs:ti,tt OR piglets:ti,tt OR rabbit:ti,tt OR rabbits:ti,tt OR cat:ti,tt OR cats:ti,tt OR dog:ti,tt OR dogs:ti,tt OR cattle:ti,tt OR bovine:ti,tt OR monkey:ti,tt OR monkeys:ti,tt OR trout:ti,tt OR marmoset*:ti,tt) AND 'animal experiment'/de) OR ('animal experiment'/de NOT ('human experiment'/de OR 'human'/de))) AND o)) AND ('parkinson disease'/exp OR parkinson*:ab,kw,ti) AND ('music'/exp OR 'auditory stimulation'/exp OR ('auditory stimulation':ab,kw,ti OR 'acoustic stimulation':ab,kw,ti OR 'auditory cue*':ab,kw,ti OR 'acoustic cue*':ab,kw,ti OR rhythm*:ab,kw,ti OR metronom*:ab,kw,ti OR music*:ab,kw,ti))

CINAHL (Cumulative Index to Nursing and Allied Health Literature)

(Parkinson's disease OR Parkinson disease) AND (auditory stimulation OR acoustic stimulation OR auditory cue OR acoustic cue OR rhythm OR metronome OR music) AND (randomized controlled trial OR randomised controlled trial)

PsycINFO

("Parkinson's disease" OR "Parkinson disease") AND ("auditory stimulation" OR “auditory stimulation” OR “auditory cue” OR “acoustic cue” OR rhythm OR metronome* OR music*) AND ( "randomized controlled trial" OR "randomised controlled trial")

Web of Science

(Parkinson's disease) AND (auditory stimulation OR acoustic stimulation OR auditory cue OR acoustic cue OR rhythm OR metronome OR music) AND (randomized controlled trial OR randomised controlled trial)

Scopus

TITLE‐ABS‐KEY("Parkinson's disease" OR "Parkinson disease" AND ("acoustic stimulation" OR “auditory stimulation” OR “auditory cue” OR “acoustic cue” OR rhythm OR metronom* OR music*) AND ( "randomized controlled trial" OR "randomised controlled trial"))

ICTRP

(Parkinson's disease) AND (auditory stimulation OR acoustic stimulation OR auditory cue OR acoustic cue OR rhythm OR metronome OR music) AND (randomized controlled trial OR randomised controlled trial)

Clinicaltrials.gov

(Parkinson's disease) AND (auditory stimulation OR acoustic stimulation OR auditory cue OR acoustic cue OR rhythm OR metronome OR music)

Controlled‐trials.com (www.isrctn.com)

(Parkinson's disease OR Parkinson disease) AND (auditory stimulation OR acoustic stimulation OR auditory cue OR acoustic cue OR rhythm OR metronome OR music) AND study_type: Interventional

RIML (Répertoire International de Littérature Musicale)

"Parkinson's disease" OR "Parkinson disease" AND ("auditory stimulation" OR “acoustic stimulation” OR “auditory cue” OR “acoustic cue” rhythm OR metronome OR music) AND "randomized controlled trial"

Contributions of authors

1. Conceptualisation: VPN, KVJ, EJ, and PV

2. Data curation

   1. Search strategy: VPN and KVJ

   2. Selection of trials to include: VPN and KVJ

   3. Risk of bias assessment: VPN and KVJ

   4. Arbitration in the event of dispute regarding study selection and risk of bias assessment: PV

   5. Data extraction: VPN and KVJ

   6. Data entry into RevMan software: VPN and KVJ

3. Formal analysis: VPN and KVJ

4. Funding acquisiton: PV

5. Investigation: VPN and KVJ

6. Methodology: VPN and KVJ

7. Project administration: VPN

8. Resources: none of the contributors provided study materials, reagents, materials, patients, laboratory samples, animals, instrumentation, computing resources, or other analysis tools.

9. Software: none of the contributors were involved in programming, software development, designing computer programmes, implementation of computer code and supporting algorithms, or testing of existing code components.

10. Supervision: EJ and PV

11. Validation: VPN and KVJ

12. Visualisation: VPN and KVJ

13. Writing ‐ original draft: VPN and KVJ

14. Writing ‐ review and editing: VPN and KVJ

Sources of support

Internal sources

• Danish National Research Foundation (grant number DNRF 117), Denmark

  The Center for Music in the Brain (MIB) is supported by The Danish National Research Foundation (grant number DNRF 117).

External sources

• None, Other

  No external sources of support

Declarations of interest

VPN is a health professional and works as a neuroscientist at Aarhus University with a focus on clinical applications of music. VPN reports no known conflicts of interest.

KVJ works as a neuroscientist at Aarhus University with a focus on clinical applications of music. KVJ reports no known conflicts of interest.

EJ works as a health professional (neurologist) at Aarhus University Hospital, Department of Neurology, Aarhus, Denmark. EJ reports no known conflicts of interest.

PV works as a neuroscientist at Aarhus University with a focus on basic music neuroscience. PV reports no known conflicts of interest.

New