Feasibility and acceptability of a group music creativity intervention for adults with varying cognitive function

Abstract

Maintaining healthy cognitive functioning and delaying cognitive decline in cognitively intact and cognitive impaired adults are major research initiatives for addressing dementia disease burden. Music interventions are promising, non-pharmaceutical treatment options for preserving cognitive function and psychological health in older adults with varying levels of cognitive function. While passive, music interventions have attracted considerable attention in the abnormal cognitive aging literature, active, music interventions such as music creativity are less well-studied. Among 58 older adults with different levels of cognitive function (cognitively healthy to mild cognitive impairment), we examined the feasibility and acceptability of Project CHROMA, a Stage 1 clinical trial developed to assess the effects of a novel, music creativity curriculum on various health outcomes. Music intervention participation (93%), overall study retention (78%), and intervention satisfaction (100%) rates were comparable to other similarly designed clinical trials. Exploratory analyses using mixed-level modeling tested the efficacy of the intervention on cognitive and psychological outcomes. Compared to those in the control condition, participants in the music condition showed some improvements in cognitive functioning and socioemotional well-being. Findings suggest that a 6-week music creativity clinical trial with several multi-modal health assessments can be feasibly implemented within a sample of varying cognitive ability.

I. Introduction

As the older adult population increases, preserving cognitive function is becoming a major research initiative for reducing dementia disease burden. Although cognitive decline is a normal part of aging, significant declines in memory and other cognitive functions may reflect pathology. About 20% of older adults aged 70+ develop mild cognitive impairment (MCI) (Ritchie, 2004), a preclinical stage between normal cognition and dementia representing a wide spectrum of abnormal cognitive function ranging from minimal to significant impairment in daily activities. Mild cognitive impairment is a risk factor for Alzheimer’s disease (AD) and related dementias. Ineffective treatment options for dementia have prompted a shift toward dementia prevention in non-demented individuals (Barnes & Yaffe, 2011; Fratiglioni & Qiu, 2011). To this end, researchers and clinicians are increasingly focused on developing interventions that maintain cognitive function or delay cognitive decline and modify risk factors in healthy and cognitively impaired individuals (Fratiglioni & Qiu, 2011; Olanrewaju et al., 2015).

Music interventions may delay abnormal cognitive aging and improve associated biobehavioral symptoms (Jordan et al., 2022). A meta-analysis of 10 randomized controlled music intervention trials demonstrated that music interventions positively affect cognitive function, depressive symptoms, anxiety, and quality of life in healthy, older adults (Xu et al., 2017). Similarly, a systematic review of 9 studies found that music engagement generally improved cognitive functioning in participants with MCI (Jordan et al., 2022).

Music interventions take various forms, from passive music listening to active musical engagement. Depending on the outcome of interest, music listening interventions require participants to listen to either a self-selected playlist based on the participants’ preference (e.g., songs that are emotionally relevant to the participant) or a pre-selected playlist based on specific characteristics (e.g., classical music). Whereas music listening interventions aim to promote relaxation and improve psychological and behavioral disturbances, active music interventions aim to improve mood, motivation, promote socialization, and stimulate motor, sensory, and cognitive domains (Raglio et al., 2015). Active music interventions pair music listening with direct music engagement through body movement (e.g., dance, clapping), instruments (e.g., playing an instrument, singing), or composition (e.g., creating a musical piece). Music training, such as piano training, in music naïve and cognitively healthy older adults improves cognitive performance, enhances quality of life and positive mood state, and decreases depressive symptoms (Seinfeld et al., 2013) . There is some evidence suggesting that active music engagement (i.e., singing, playing songs, composing pieces) compared to passive music engagement (i.e., listening to music) has more substantial positive effects on cognitive functioning, behavioral symptoms, and daily function in AD patients (Gómez-Gallego et al., 2021; Sakamoto et al., 2013).

Within active music interventions, the effect of music composition and improvisation on cognitive functioning is less studied than in other forms of active music engagement (i.e., playing musical instruments, or dancing). Composition is a process of creating novel music, and improvisation is the spontaneous, concurrent creation and performance of music; both are forms of musical creativity that require cognitive flexibility. Moreover, studies show that composition and improvisation engage brain regions involved in motor planning and sequencing (e.g., premotor cortex, supplementary motor area), self-reflection and creativity (e.g., default mode network), and executive functioning (e.g., dorsolateral prefrontal cortex) in musicians (Bashwiner, 2018). Other modes of creativity, such as visual arts therapy, utilize similar cognitive skills as music creativity (e.g., planning, cognitive control, creativity) and promote improvements in cognitive functioning and related psychological symptoms in older adults (Masika et al., 2020). In addition, creativity involves novel work, and there is evidence suggesting that novel activities benefit cognitive aging (Fissler et al., 2013; Fritsche et al., 2005; Oltmanns et al., 2017). Thus, music creativity via composition and improvisation may also improve cognitive performance and psychological well-being. However, few music creativity interventions exist, and limited studies have examined whether music creativity training can improve cognitive functioning in older adults.

Project CHROMA (Cognitive Health Research on Musical Arts) is a Stage 1 clinical trial designed to test the effects of a music creativity intervention on various psychological and biological outcomes in adults with MCI or at higher risk of MCI (i.e., 70 years and older). Stage 1 of the National Institute of Health Stage Model involves developing new/modifying existing interventions (Stage 1a) and feasibility and pilot testing (Stage 1b) (Onken et al., 2014). In contrast to most music interventions that primarily involve passive music listening, the music creativity intervention within Project CHROMA incorporates both music listening and music composition and improvisation; from the first day of class, participants are taught to interact and produce creative pieces with everyday objects and sounds. Novelty is a major theme of the curriculum: it was designed for those with minimal to no musical notation background, and music creativity inherently involves novel work. Moreover, Project CHROMA aims to examine biopsychosocial mechanisms of change, as study outcomes include clinical endpoints (e.g., quality of life domains), secondary cognitive endpoint (e.g., Mini-Mental State Exam), and biological mediators (e.g., functional brain network indices, immune biomarkers) (Biomarkers Definitions Working Group et al., 2001). Project CHROMA’s findings will contribute to the growing interest in quantifying the benefits of music creativity on health.

The current study had two primary objectives. First, Project CHROMA’s feasibility and acceptability were assessed. Compared to most published studies, Project CHROMA requires older adults to participate in more lengthy visits and a diverse array of procedures. Thus, it was important to ensure that despite additional health examination procedures and time commitment, Project CHROMA could recruit and retain participants at comparable rates as other music intervention studies. Second, exploratory analyses were conducted to assess the preliminary effects of the musical creativity curriculum on cognitive functioning and various measures of psychological well-being from pre- to post-intervention.

II. Materials and Methods

Study design and participants

Project CHROMA is a two-armed, semi-randomized controlled study conducted in Houston, Texas since 2019. The study is registered on ClinicalTrials.gov (NCT04137913). Participants were randomized to one of two groups: the music creativity group or the inactive control group. The music creativity group participated in a 6-week group music creativity class, while the control group received no music course. Participants were recruited through flyer distributions at local neighborhoods, medical clinics, community centers, senior residential communities, and community expos/fairs. The research team also shared flyers with adults who participated in previous studies within the lab.

To determine eligibility, interested participants completed a telephone screening with a trained research assistant. Inclusion criteria were as follows: Participants were 1) 70+ years old OR diagnosed with early to moderate MCI (confirmed through physician’s release of medical records using a medical release form), 2) able to read and write in English, 3) cognitively competent to participate (i.e., comprehension was assessed via 3 questions during the consent process), 4) demonstrate an ability to follow instructions. Participants were excluded if they 1) were both below the age of 70 and not diagnosed with MCI; 2) had significant visual or auditory impairment resulting in the inability to read or hear study questionnaires (self-reported by participant); 3) had Class III heart failure, any autoimmune and/or inflammatory disorders, or Parkinson’s disease; 4) had any contraindications for undergoing functional magnetic resonance imaging (MRI) scanning (e.g., any implanted medical device, severe claustrophobia, history of working with metal, bullet wounds as determined by a standard clinical questionnaire) or had dental implants/extensive dental work that would significantly distort functional imaging data; 5) were pregnant or nursing women; 6) weighed 300+ lbs or had a body mass index that exceeded 40 (for inflammatory reasons); or 7) were a current or past professional musician. To ensure that subjects could commit to the 6-week music creativity class, if selected for the course, all participants had to verbally confirm that they could commit to the dates of the music class. ‘Professional musician’ was defined as making money as a musician or regularly volunteering as a teacher (e.g., church choir instructor). All information, except medical diagnosis of MCI (verified through physician’s release of medical records), was dependent on participant self-report.

Eligible participants completed 2 in-person assessments – a baseline visit and a follow-up visit. At both visits, participants underwent a one-hour functional MRI scan, venous blood draw, cognitive testing, creativity task, and psychophysiological heart rate assessments Functional MRI scans were conducted in the Texas Medical Center. At the end of each visit, participants completed self-report questionnaires probing music experience, psychological and physical well-being, and health behaviors. At the end of the baseline visit, participants were informed which group they were randomized to. Participants randomized to the music creativity group had their follow-up visit within 0-3 weeks after the conclusion of the music course. Participants randomized to the control group did not participate in the music class and completed their follow-up visit approximately 7-12 weeks after the baseline visit.

Because it was not possible to conduct a true single-blinded study, measures were taken to minimize bias. First, all participants were told during the phone screening that they would be randomized to either the “creative music” group (treatment group) or the “creative reading” group (control group). Participants in the creative reading group were given a creativity book (i.e., The Runaway Species: How Human Creativity Remakes the World by David Eagleman and Anthony Brandt) to read at their own leisure at the end of their follow-up visit. The book covers the science of creativity through the lens of neuroscience and how creativity can improve daily living. Second, an online survey was sent to participants’ email 2-3 weeks after their follow-up visit. This optional survey probed general questions about the book that did not require participants to have read the entire book (e.g., how much of the book did you read? What was something new or exciting that you learned from the book?). It served to prevent participants from knowing that they were assigned to the control group.

All Project CHROMA study procedures and materials were reviewed and approved by the Institutional Review Board (protocol ID: FY2019-355). Participants provided written informed consent prior to the start of the baseline visit.

Intervention

Development

The 6-week creative music course was designed by co-authors, in association with the award-winning Houston-based contemporary music ensemble Musiqa. The course was designed to be administered within a group setting and intended to be accessible to all participants, regardless of musical background. Thus, no knowledge of music notation was required. Workshops incorporated four main components: listening, theory, performance, and creation (see Table 1a). Class topics became progressively more challenging and sophisticated as the course progressed (see Table 1b). For example, the first week’s listening component focused on short and more familiar works such as folk songs and instrumental etudes. Eventually, the course built towards symphonic movements and more unfamiliar and experimental music. Lessons culminated with creating a final composition. To obviate the need for prior musical experience and enhance the exploratory nature of the activities, the creative projects de-emphasized mimicking the participants’ familiar and/or favorite music, and instead encouraged more personal and experimental approaches to musical expression. Creative projects included a soundscape of a favorite place or event; a theme and variations on a percussion rhythm; a piece using words for their musical value; a musical accompaniment to a narrative; and an improvisation to a silent film. Participants showcased samples of their music compositions in a workshop concert. Invited family, friends, and research collaborators attended the concert.

Table 1a.

Main components and associated objectives of Project CHROMA’s music creativity intervention

Component	Goal and Methods	Example Activities
Listening	We will discuss how to describe musical form, how to follow a work’s primary themes or patterns, how to describe musical textures, and how to articulate one’s emotional and intellectual reactions. Participants will be exposed to a wide variety of repertoire.	We will listen to contemporary music that has expanded the boundaries of the Western tradition, as well as to world music, such as Balinese and Javanese gamelan and African polyphonic singing.
Theory	We will teach the basics of melody, harmony, rhythm, timbre, and how musicians conceptualize music.	The various instrumental families, i.e., strings, woodwinds, brass, percussion, will be introduced.
Performance	We designed these exercises to give the participants confidence in their own music-making. They will practice rhythms using percussion instruments and melodic patterns using their voices.	Sessions will begin with simple repetitive rhythms by the class, and sung scales. Over time, the participants will learn to perform more involved patterns and melodies.
Creation	We designed these exercises to stimulate the participants’ aural imaginations using a variety of prompts. We will use graphic notation to sketch the compositions.	Participants will compose a piece inspired by a work of art, write a piece to accompany a spoken narrative, compose a score for a home movie, and perform a collective improvisation.

Table 1b.

Project CHROMA’s music creativity curriculum schedule by week

Week	Topic
1	Creating using rhythm
2	Creating using pitch and rhythm
3	Creating using harmony, pitch, and rhythm
4	Understanding form and instrument creation
5	Creating an ensemble and writing for the ensemble
6	Rehearsing, editing, and final performances

Procedures

Participants randomized into the music creativity class attended the 6-week class. The class met 3 times a week, 2 hours each day. All class sessions were led by one of the co-authors, an experienced music composer. Participants were enrolled in the class in waves; thus, participants learned and composed pieces with the same cohort throughout the 6-week course.

From 2019-2022, three classes were administered. Class sizes ranged from 6 to 12 participants, and adjustments were made to ensure that small and larger class sizes received similar amounts of 1-on-1 interaction with the instructor.

Design adjustments due to Covid-19

The SARS-CoV-2 outbreak (Covid 19) in March 2020 incited a worldwide shutdown of normal in-person operations for an indefinite amount of time. Project CHROMA took place before and during the COVID-19 pandemic. Several changes were made throughout the study to guarantee safety and preserve research integrity while minimizing risk for participants and researchers. First, some participants during the first wave of the music creativity course completed their follow-up visits via telephone. The first wave of music creativity participants (n=12) completed the creative music course before the COVID-19 pandemic, but only 5 out of 12 subjects were able to complete in-person follow-ups before the shutdown. Second, a covid impact questionnaire was developed and administered to all participants since the start of the pandemic. Third, recruitment was temporarily halted from March 2020 to November 2020 and resumed in Dec 2020. However, only control participants were recruited from Dec 2020 to August 2021: Because our sample population fell within the high-risk group defined by the Centers for Disease Control, it was not possible for us to host a group class with positivity rates constantly changing. From Dec 2020 to August 2021, Project CHROMA was not advertised as having a music creativity portion; instead, participants were told that they would be participating in a creativity study and would be required to participate in two in-person assessments. Fourth, since March 2020, all study personnel were required to wear face masks or shields when interacting with participants. Fifth, randomization with 2 arms resumed Sept 2021 to May 2021 in preparation for two additional music creativity cohorts. Two consecutive music creativity courses took place on Jan 2022 to March 2022 and March 2022 to April 2022; all music classes abided by the university’s current covid-19 safety policies at the time. Sixth, full vaccination was deemed as a requirement for participation. This requirement attempted to minimize spread and exposure to the virus amongst a high-risk group, as the music creativity class had to be conducted in-person in a group setting. Seventh, self-report questionnaires were administered at the participant’s home rather than in the lab beginning in December 2020 to minimize the duration of in-person contact. Eighth, for one music creativity cohort, video conferencing was offered as an option to accommodate 1-2 participants who were out of town or ill during the 6-week course.

Feasibility and acceptability of the intervention

Recruitment, retention, intervention adherence, and participant satisfaction were assessed. Retention rate was estimated to be 80% or higher, based on the research team’s previous success in retaining older participants in year-long longitudinal studies. At screening, participants were required to make 80% or more of the sessions; thus, it was estimated that most participants would be able to attend at least 80% (14/18 classes) of the music course. Satisfaction was assessed at the follow-up visit via a self-report survey. Participants were asked how satisfied they were with the course and whether the course met their expectations. There was also a free-response section for providing specific comments and feedback.

Measures

Cognitive assessments

Previous factor analysis revealed that executive functioning consists of 3 processes: cognitive flexibility, inhibition, and updating/monitoring information in working memory (Miyake et al., 2000). The Color-Word Interference Test (CWIT) from the Delis-Kaplan Executive Function System (D-KEFS), a widely used neuropsychological assessment of executive functioning (Delis et al., 2001; Latzman & Markon, 2010), was used to examine inhibition. Cognitive inhibition is the ability to suppress dominant responses in favor of nondominant responses. The CWIT is comprised of four tasks: (1) Color Naming, (2) Word Reading, (3) Inhibition, and (4) Inhibition/Switching. The Color Naming and Word Reading tasks involve stating and reading color names as quickly as possible. Participants who are color-blind or cannot read English do not continue on to the inhibition and inhibition/switching tasks. In the Inhibition task, word colors are printed in a different colored ink. Participants are asked to inhibit the dominant response, stating the color name rather than reading the word. The Inhibition/Switching task is similar to the Inhibition task (i.e., name the ink color, not read the word) except that when words are printed in a box, participants are told to read the word rather than name the ink color. Error rates and time to completion are calculated for each task. Summary scores and errors are scaled to participant age; age-normalized data comes from a sample of +1,500 individuals aged 8-89 years from the U.S (Delis et al., 2001). A composite score comprising the scaled inhibition, inhibition switching, and error scores for inhibition and inhibition switching subtests was used as an indicator of cognitive inhibition based on evidence that these measures include a factor consistent with inhibition (Latzman & Markon, 2010). Higher composite scores reflect better inhibition.

The computerized version of the Wisconsin Card Sorting Test (WCST) was used to assess cognitive flexibility, which is the ability to shift one’s attention between multiple tasks. The WCST assesses planning, organized searching, and the ability to utilize environmental feedback to shift cognitive sets, direct behavior toward achieving a goal, and modulate impulsive responses (Heaton et al., 1993). Performance on the WCST is associated with cognitive flexibility (Fisk & Sharp, 2004). Four stimulus cards incorporate three stimulus parameters (color, form, and number). Participants match cards to a target card by color, shape, or number but are not told by which characteristic they should be matching. Feedback is provided after each attempt. The rule changes every 10 trials. Subjects were presented with 60 trials total. The WCST was administered through PsyToolKit, a software package for programming psychological experiments (Stoet, 2010, 2017). The number of errors attributed to perseveration or incorrectly matching based on the previous rule, will serve as an index of cognitive flexibility. Lower numbers reflect less error or more cognitive flexibility; higher numbers reflect more error or less cognitive flexibility.

The Digit span subtest of the highly-utilized Wechsler Adult Intelligence Scale-IV Edition (WAIS-IV) was used to assess verbal working memory (Wechsler, 2008). The Digit span subtest is comprised of three conditions: digit span forwards, digit span backwards, and sequencing. During each condition, participants are presented with longer spans of numbers that they repeat back to the administrator. The test administrator discontinues the task condition when the participant incorrectly answers both trials within the same span length. Respondents repeat span lengths in forward, backward, and sequential order (lowest to highest) for the digit span forwards, digit span backwards, and sequencing conditions, respectively. Longer spans correctly recalled indicate better working memory on each task. While the three conditions broadly engage working memory processes, each condition is thought to require different skills. The digit span forwards is conceptualized as a short term memory task whereas the digit span backwards primarily examines working memory (Bowden et al., 2013). The sequencing condition was designed to increase working memory demands (Wechsler, 2008) and is considered to be more cognitively demanding than digit span forwards and backwards (Wisdom et al., 2012). Two summary values, a raw and scaled score, can be derived from the digit span test. Raw total scores are derived from the sum raw scores across the three digit span conditions (forwards, backwards, sequencing). Scaled total scores represent sum age-adjusted scores across the three conditions (Wechsler, 2008). In addition to raw total scores, the longest digit for each task condition was recorded and assessed in post-hoc analyses, when the summary score was found to change significantly over time.

Global cognitive functioning was assessed using The Mini-Mental State Exam (MMSE), a widely used instrument in clinical and research settings. Participants are presented with a series of questions to test orientation, registration, attention, calculation, recall, and language (Folstein et al., 1975). These include asking the participant about the current date and location, repeating three unrelated objects (and recalling them later), counting or spelling backwards, naming objects, reading and following commands, writing a spontaneous sentence, and copying geometric shapes. The number of correct answers is scored out of a total of 30. Higher scores indicate better cognitive function. Scores under 23 are indicative of cognitive impairment. MMSE scores will be modeled as a continuous variable.

Self-report questionnaires

Guidelines for clinical trial research on AD-related research (from preclinical, non-AD to clinical AD) highly recommend including quality of life and other psychological well-being measures as secondary endpoints when considering the efficacy of a treatment or intervention (Committee for Medicinal Products for Human Use (CHMP), 2018). To remain consistent with these guidelines, measures of quality of life, mood, stress, and social support were included. At the end of each research visit, self-report questionnaires were completed via Qualtrics, an online survey platform. Before the COVID-19 pandemic, surveys were completed on-site. Since the pandemic, surveys were completed at home. Research personnel urged participants to complete the survey within 48 hours of receiving the survey link. Beyond the 48-hour window, research personnel reminded participants every 2 days via email and phone until surveys were completed.

Quality of Life.

The RAND 36-Item Health Survey 1.0 provides a non-disease-specific measure of health-related quality of life with excellent normative data (Ware & Sherbourne, 1992). Developed for the Medical Outcomes Study (Ware & Sherbourne, 1992), the RAND 36-Item Health Survey taps eight health concepts: physical functioning, bodily pain, role limitations due to physical health problems, role limitations due to personal or emotional difficulties, emotional well-being, social functioning, energy/fatigue, and general health perceptions. Each item is scored on a 0 to 100 range. Scores represent the percentage of total possible score achieved. Higher values indicate a favorable health state. Each subscale was assessed in this study.

Social support.

The Interpersonal Support Evaluation List (ISEL) includes 40 statements probing the perceived availability of potential social resources (Cohen & Hoberman, 1983). Participants rate how true or false the statements represent their current situation. All answers are given on a 4-point scale ranging from “Definitely true” to “Definitely false.” Items fall within four 10-item subscales. The “tangible” subscale measures perceived availability of material aid. The “appraisal” subscale measures perceived availability of someone to discuss one’s problems with. The “self-esteem” subscale evaluates the perceived availability of a positive comparison when comparing one’s self to others. The “belonging” subscale assesses the perceived availability of people one can do activities with. Overall social support is the sum of all the subscales, with higher values indicating greater social support.

Depressive symptomatology.

The Beck Depression Inventory II (BDI-II) is a revised measure of BDI-I, which assesses depressive symptomatology in the past few weeks. The BDI-II is suitable for clinical and research populations. The 21-item survey evaluates a variety of cognitive and physical symptoms of depression. Each question presents four different states of varying severity (from “I do not feel sad” to “I am so sad or unhappy that I can’t stand it”). Higher values are assigned to responses indicating more acute symptoms of depression; thus higher BDI summary scores indicate more severe depressive symptomatology. The following guidelines have been used to interpret the BDI-II: minimal range = 0-13, mild depression = 14-19, moderate depression = 20-28, severe depression 29-63. The BDI-II is sensitive to change in depression, with 5 points indicating minimal difference, 10-19 points indicating a moderate difference, and ≥20 indicating a large difference (Viljoen et al., 2003).

Perceived stress.

The 10-item Perceived Stress Scale (PSS) was designed to measure the degree to which individuals appraise situations in their life as stressful and includes items that assess perceptions of daily life as unpredictable, uncontrollable, and overloaded (Cohen et al., 1983). The survey was designed to be used within the community with at least a junior high school education. Questions ask about feelings and thoughts during the last month. Higher scores indicate higher stress levels.

Data analysis

The primary aim was to determine feasibility and acceptability of the study protocol. Feasibility was assessed by calculating participation and assessment completion rates at each research visit and music session (e.g., cognitive assessments, self-report questionnaires, functional imaging scan, blood samples, psychophysiological recordings). Acceptability was evaluated through descriptive statistics and qualitative feedback from participants. The secondary aim was to explore the impact of the music creativity intervention on psychological, social, and cognitive outcomes. Mixed level models with random intercept were used to examine group differences in perceived stress, depressive symptomatology, perceived social support, and cognitive performance (inhibition, flexibility, working memory, and overall cognition) from baseline to follow-up. To account for the possible effect of having a medical diagnosis of MCI on study outcomes, MCI status was included as a covariate in post-hoc models. To ensure that adequate randomization occurred and that any effects were not attributed to differences in follow-up duration, independent samples t-tests and Pearson’s chi-squared tests were performed on baseline characteristics and follow-up duration (i.e., time since baseline visit). Pearson correlations were conducted on baseline data for cognitive performance and socio-emotional well-being. Preliminary analyses on biological data were not run, as neuroimaging and blood assay data were not available at the time of writing; notable results from these physiological assessments will be communicated in a separate publication.

III. Results

Feasibility outcomes

Recruitment and retention and group characteristics

Between December 2019 and May 2022, 404 people from the community expressed interest in the study and 55.2% of 404 underwent the full screening procedure. Out of the 223 participants who underwent screening, 103 (46.2%) were eligible. Fifty-six percent (N=58) of those eligible enrolled in the study.

Fifty-eight eligible participants were semi-randomized (28 music creativity intervention, 30 inactive control). Besides the music creativity group having significantly more participants identifying as Hispanic (p = .02), the music creativity group and control group did not differ on demographic characteristics (i.e., age, sex, education, race, employment status, MCI status), baseline psychological state (i.e., perceived stress, depressive symptomatology, quality of life, social support), cognitive performance, and time between baseline and follow-up. Baseline sample characteristics are detailed in Table 2.

Table 2.

Group Descriptives at Baseline

	Control (N=30)	Music (N=28)	Total (N=58)	p value
Age				0.947
Mean (SD)	75.03 (6.31)	74.93 (5.56)	74.98 (5.91)
Range	69.00 - 97.00	59.00 - 84.00	59.00 - 97.00
Sex				0.586
Male	15 (50.0%)	12 (42.9%)	27 (46.6%)
Female	15 (50.0%)	16 (57.1%)	31 (53.4%)
MCI diagnosis				0.257
No	26 (86.7%)	21 (75.0%)	47 (81.0%)
Yes	4 (13.3%)	7 (25.0%)	11 (19.0%)
Education a				0.833
N-Miss	1	0	1
Mean (SD)	4.31 (0.97)	4.25 (1.17)	4.28 (1.06)
Range	2.00 - 5.00	1.00 - 5.00	1.00 - 5.00
Race				0.250
N-Miss	1	0	1
White	27 (93.1%)	23 (82.1%)	50 (87.7%)
Black or African American	1 (3.4%)	4 (14.3%)	5 (8.8%)
Asian	1 (3.4%)	0 (0.0%)	1 (1.8%)
American Indian or Alaskan Native	0 (0.0%)	1 (3.6%)	1 (1.8%)
Ethnicity				0.017
N-Miss	1	0	1
Non-Hispanic	29 (100.0%)	23 (82.1%)	52 (91.2%)
Hispanic	0 (0.0%)	5 (17.9%)	5 (8.8%)
Employment status				0.241
N-Miss	1	2	3
Full-Time	1 (3.4%)	1 (3.8%)	2 (3.6%)
Part-Time	3 (10.3%)	0 (0.0%)	3 (5.5%)
Retired	25 (86.2%)	25 (96.2%)	50 (90.9%)
Days between visits				0.153
N-Miss	8	5	13
Mean (SD)	84.73 (19.83)	76.13 (19.83)	80.33 (20.08)
Range	40.00 - 119.00	47.00 - 121.00	40.00 - 121.00
Days since COVID-19b at Visit 1				0.128
Mean (SD)	441.83 (198.34)	317.29 (391.11)	381.71 (310.51)
Range	−56.00 - 706.00	−113.00 - 737.00	−113.00 - 737.00
Survey completion time, relative to time since Visit 1				0.222
N-Miss	1	0	1
Mean (SD)	3.72 (7.12)	1.89 (3.33)	2.82 (5.62)
Range	0.00 - 29.00	0.00 - 15.00	0.00 - 29.00
MMSE				0.448
Mean (SD)	27.83 (2.80)	27.36 (1.79)	27.60 (2.36)
Range	15.00 - 30.00	24.00 - 30.00	15.00 - 30.00
WCST (perseveration count)				0.919
N-Miss	1	0	1
Mean (SD)	11.93 (5.06)	12.07 (5.27)	12.00 (5.12)
Range	4.00 - 22.00	5.00 - 23.00	4.00 - 23.00
CWIT (composite)				0.947
N-Miss	3	0	3
Mean (SD)	0.01 (0.51)	0.02 (0.48)	0.01 (0.49)
Range	−1.57 - 0.64	−1.19 - 0.89	−1.57 - 0.89
WAIS Digit Span (raw)				0.349
Mean (SD)	24.73 (5.49)	23.32 (5.91)	24.05 (5.69)
Range	15.00 - 39.00	12.00 - 34.00	12.00 - 39.00
WAIS Digit Span (scaled)				0.486
N-Miss	1	0	1
Mean (SD)	10.28 (2.90)	9.71 (3.14)	10.00 (3.01)
Range	4.00 - 18.00	3.00 - 15.00	3.00 - 18.00
Depressive symptomatology (BDI-II)				0.968
N-Miss	2	2	4
Mean (SD)	5.21 (5.97)	5.15 (5.09)	5.19 (5.51)
Range	0.00 - 25.00	0.00 - 20.00	0.00 - 25.00
Perceived Stress (PSS)				0.285
N-Miss	2	1	3
Mean (SD)	10.82 (6.55)	9.07 (5.37)	9.96 (6.01)
Range	1.00 - 25.00	2.00 - 19.00	1.00 - 25.00
QOL (RAND SF-36): General				0.802
N-Miss	2	0	2
Mean (SD)	71.96 (16.41)	70.89 (15.46)	71.43 (15.80)
Range	40.00 - 100.00	40.00 - 95.00	40.00 - 100.00
QOL: Pain				0.275
N-Miss	2	0	2
Mean (SD)	79.55 (21.09)	85.54 (19.45)	82.54 (20.33)
Range	22.50 - 100.00	22.50 - 100.00	22.50 - 100.00
QOL: Emotional functioning				0.242
N-Miss	2	0	2
Mean (SD)	80.29 (14.07)	84.29 (11.04)	82.29 (12.69)
Range	44.00 - 100.00	60.00 - 100.00	44.00 - 100.00
QOL: Role limitations (due to emotional problems)				0.647
N-Miss	2	0	2
Mean (SD)	14.29 (33.25)	10.71 (24.09)	12.50 (28.82)
Range	0.00 - 100.00	0.00 - 66.67	0.00 - 100.00
QOL: Role limitations (due to physical problems)				0.757
N-Miss	2	0	2
Mean (SD)	22.32 (34.25)	19.64 (29.94)	20.98 (31.90)
Range	0.00 - 100.00	0.00 - 100.00	0.00 - 100.00
QOL: Physical functioning				0.762
N-Miss	2	0	2
Mean (SD)	83.57 (22.85)	81.79 (21.09)	82.68 (21.80)
Range	20.00 - 100.00	30.00 - 100.00	20.00 - 100.00
QOL: Fatigue				0.417
N-Miss	2	0	2
Mean (SD)	65.18 (20.57)	69.11 (14.91)	67.14 (17.91)
Range	15.00 - 100.00	25.00 - 95.00	15.00 - 100.00
QOL: Social functioning				0.350
N-Miss	2	0	2
Mean (SD)	87.50 (20.97)	91.96 (13.70)	89.73 (17.69)
Range	37.50 - 100.00	50.00 - 100.00	37.50 - 100.00
Social Support (ISEL)				0.784
N-Miss	2	1	3
Mean (SD)	88.11 (17.25)	86.78 (18.46)	87.45 (17.70)
Range	43.00 - 112.00	34.00 - 109.00	34.00 - 112.00

Note: MMSE = Mini Mental State Exam. WCST = Wisconsin Card Sorting Task. CWIT = Color Word Inference Test. WAIS = Wechsler Adult Intelligence Scale. BDI-II = Beck Depression Inventory II. PSS = Perceived Stress Scale. RAND SF-36 = RAND 36-item health survey. QOL: quality of life. ISEL = Interpersonal Support Evaluation List.

^aEducation: 0 = Less than 7 years, 1= 7 to 12 years (non-graduate), 2 = High school graduate, 3 = Up to 3 years of college, 4 = 3 or more years of college, 5 = Graduate/professional training

^bCOVID-19 reference point: March 11, 2020

The study maintained a 78% retention rate across both groups, with 45 out of 58 participants completing the entire study (23 music creativity intervention, 22 inactive control). Out of the 13 that discontinued participation midway, three subjects’ follow-up visits were intentionally canceled by the research team due to unprecedented circumstances of COVID-19, and one subject completed the music creativity class but did not complete the follow-up visit. Compared to those who remained in the study, participants who failed to complete the study did not differ significantly on baseline demographic characteristics, psychological well-being (i.e., perceived stress, depressive symptomatology, quality of life), and most cognitive tests (MMSE, CWIT, WCST). However, they had moderately lower scores on the Digit span test and less perceived social support. See Table 3 for means, standard deviations, and significance testing.

Table 3.

Dropout Descriptives

	Completed Visit 1 & Visit 2 (N=45)	Dropped out after Visit 1 (N=13)	Total (N=58)	p value
Age				0.704
Mean (SD)	74.82 (5.16)	75.54 (8.23)	74.98 (5.91)
Range	59.00 - 88.00	62.00 - 97.00	59.00 - 97.00
Sex				0.974
Male	21 (46.7%)	6 (46.2%)	27 (46.6%)
Female	24 (53.3%)	7 (53.8%)	31 (53.4%)
MCI				0.218
No	38 (84.4%)	9 (69.2%)	47 (81.0%)
Yes	7 (15.6%)	4 (30.8%)	11 (19.0%)
Education a				0.912
N-Miss	0	1	1
Mean (SD)	4.29 (1.10)	4.25 (0.97)	4.28 (1.06)
Range	1.00 - 5.00	2.00 - 5.00	1.00 - 5.00
Race				0.546
N-Miss	0	1	1
White	38 (84.4%)	12 (100.0%)	50 (87.7%)
Black	5 (11.1%)	0 (0.0%)	5 (8.8%)
Asian	1 (2.2%)	0 (0.0%)	1 (1.8%)
American Indian/Alaskan Native	1 (2.2%)	0 (0.0%)	1 (1.8%)
Ethnicity				0.952
N-Miss	0	1	1
Non-Hispanic	41 (91.1%)	11 (91.7%)	52 (91.2%)
Hispanic	4 (8.9%)	1 (8.3%)	5 (8.8%)
Employment Status				0.532
N-Miss	2	1	3
Full-Time	1 (2.3%)	1 (8.3%)	2 (3.6%)
Part-Time	2 (4.7%)	1 (8.3%)	3 (5.5%)
Retired	40 (93.0%)	10 (83.3%)	50 (90.9%)
Time between Visit 1 and Visit 2
N-Miss	0	13	13
Mean (SD)	80.33 (20.08)	NA	80.33 (20.08)
Range	40.00 - 121.00	NA	40.00 - 121.00
Time since COVID-19 at Visit 1 b				0.719
Mean (SD)	389.69 (303.99)	354.08 (343.61)	381.71 (310.51)
Range	−113.00 - 737.00	−61.00 - 734.00	−113.00 - 737.00
Survey completion time, relative to time since Visit 1				0.002
N-Miss	0	1	1
Mean (SD)	1.64 (2.14)	7.25 (10.73)	2.82 (5.62)
Range	0.00 - 9.00	0.00 - 29.00	0.00 - 29.00
MMSE				0.984
Mean (SD)	27.60 (2.47)	27.62 (2.02)	27.60 (2.36)
Range	15.00 - 30.00	24.00 - 30.00	15.00 - 30.00
WCST (perseveration count)				0.464
N-Miss	1	0	1
Mean (SD)	11.73 (5.20)	12.92 (4.91)	12.00 (5.12)
Range	4.00 - 23.00	5.00 - 22.00	4.00 - 23.00
CWIT Composite				0.873
N-Miss	2	1	3
Mean (SD)	0.01 (0.50)	0.03 (0.49)	0.01 (0.49)
Range	−1.57 - 0.89	−1.03 - 0.64	−1.57 - 0.89
WAIS digit span (raw)				0.234
Mean (SD)	24.53 (5.94)	22.38 (4.54)	24.05 (5.69)
Range	12.00 - 39.00	15.00 - 33.00	12.00 - 39.00
WAIS digit span (scaled)				0.197
N-Miss	0	1	1
Mean (SD)	10.27 (3.04)	9.00 (2.76)	10.00 (3.01)
Range	3.00 - 18.00	4.00 - 15.00	3.00 - 18.00
Depressive symptoms (BDI-II)				0.566
N-Miss	1	3	4
Mean (SD)	4.98 (4.91)	6.10 (7.89)	5.19 (5.51)
Range	0.00 - 20.00	0.00 - 25.00	0.00 - 25.00
Perceived Stress (PSS)				0.264
N-Miss	0	3	3
Mean (SD)	9.53 (5.62)	11.90 (7.58)	9.96 (6.01)
Range	1.00 - 20.00	3.00 - 25.00	1.00 - 25.00
QOL (RAND SF-36): General				0.741
N-Miss	0	2	2
Mean (SD)	71.78 (15.64)	70.00 (17.18)	71.43 (15.80)
Range	40.00 - 100.00	40.00 - 90.00	40.00 - 100.00
QOL: Pain				0.601
N-Miss	0	2	2
Mean (SD)	81.83 (20.33)	85.45 (21.03)	82.54 (20.33)
Range	22.50 - 100.00	45.00 - 100.00	22.50 - 100.00
QOL: Emotional functioning				0.131
N-Miss	0	2	2
Mean (SD)	83.56 (11.35)	77.09 (16.79)	82.29 (12.69)
Range	48.00 - 100.00	44.00 - 100.00	44.00 - 100.00
QOL: Role limitations (due to emotional problems)				0.666
N-Miss	0	2	2
Mean (SD)	13.33 (28.78)	9.09 (30.15)	12.50 (28.82)
Range	0.00 - 100.00	0.00 - 100.00	0.00 - 100.00
QOL: Role limitations (due to physical problems)				0.952
N-Miss	0	2	2
Mean (SD)	21.11 (32.40)	20.45 (31.26)	20.98 (31.90)
Range	0.00 - 100.00	0.00 - 100.00	0.00 - 100.00
QOL: Physical functioning				0.217
N-Miss	0	2	2
Mean (SD)	80.89 (23.65)	90.00 (8.94)	82.68 (21.80)
Range	20.00 - 100.00	70.00 - 100.00	20.00 - 100.00
QOL: Fatigue				0.731
N-Miss	0	2	2
Mean (SD)	67.56 (18.51)	65.45 (15.88)	67.14 (17.91)
Range	15.00 - 100.00	40.00 - 85.00	15.00 - 100.00
QOL: Social functioning				0.821
N-Miss	0	2	2
Mean (SD)	90.00 (16.56)	88.64 (22.68)	89.73 (17.69)
Range	37.50 - 100.00	37.50 - 100.00	37.50 - 100.00
Social Support (ISEL)				0.011
N-Miss	0	3	3
Mean (SD)	90.27 (16.80)	74.80 (16.82)	87.45 (17.70)
Range	34.00 - 112.00	43.00 - 96.00	34.00 - 112.00

Note: MMSE = Mini Mental State Exam. WCST = Wisconsin Card Sorting Task. CWIT = Color Word Inference Test. WAIS = Wechsler Adult Intelligence Scale. BDI-II = Beck Depression Inventory II. PSS = Perceived Stress Scale. RAND SF-36 = RAND 36-item health survey. QOL: quality of life. ISEL = Interpersonal Support Evaluation List.

^aEducation: 0 = Less than 7 years, 1= 7 to 12 years (non-graduate), 2 = High school graduate, 3 = Up to 3 years of college, 4 = 3 or more years of college, 5 = Graduate/professional training

^bCOVID-19 reference point: March 11, 2020

Assessment completion

Cognitive assessments.

Baseline CWIT data was acquired from 56 out of 58 subjects. Incomplete baseline data was attributed to one participant being color-blind and one participant being outside the age range for computing a scaled CWIT inhibition and inhibition switching score. Besides data lost to dropout (n=13), incomplete CWIT data at follow-up was attributed to the cancellation of in-person visits due to the unprecedented COVID-19 shutdown in 2020 (n=7), color-blindness (n=1), and marked difficulty with task completion (n=1). For the Digit Span subtests, there was no missing baseline data, and missing data on follow-up was only attributed to dropout (n=13) and canceled visits due to the unprecedented COVID-19 shutdown in 2020 (n=7). For the WCST, 57 out of 58 subjects completed the baseline assessment; one subject could not complete the task due to color blindness. Incomplete follow-up WCST data was attributed to dropout (n=13), no access to a working computer during COVID-19 shutdown (n=3), misplaced data (n=1), and color-blindness (n=1).

Questionnaires.

Fifty-six out of 58 participants completed the baseline questionnaire packet. One participant did not complete the entirety of the questionnaire prior to the start of the music course, and one participant dropped out before completing their baseline questionnaire. Incomplete questionnaire data at follow-up was attributed to drop-out (n=13) and failure to complete questionnaires within 3 weeks of the follow-up visit (n=1). Self-report questionnaires across baseline and follow-up visits were completed on average 2.75 days after the in-person assessment (SD=5.68, Median = 1, Range = 0-32 days). Across both visits, there was a significant negative correlation between questionnaire completion time and cognitive inhibition (r = −.25, p = .02), which was mainly due to the strong correlation observed at follow-up (r = −.47, p = .005).

Functional MRI.

Baseline functional MRI scans were successfully acquired for 55 out of 58 subjects. Incomplete or missing scans at baseline were attributed to participant discomfort (i.e., claustrophobia, prolonged scan duration) (n=2) and technical issues (i.e., scanner dysfunction and no opportunity to reschedule the scan) (n=1). Twenty-two out of 58 subjects did not have a follow-up scan. Incomplete or missing scans at follow-up were attributed to dropout (n=13), cancellation of in-person visits due to the unprecedented COVID-19 shutdown in 2020 (n=7), claustrophobia (n=1), and technical issues at baseline (n=1).

Blood draws.

Blood was successfully acquired from 57 out of 58 participants: Needle shortage due to pandemic-related supply chain issues prevented us from acquiring blood from one participant with small veins. For follow-up visits, blood was successfully obtained for 37 subjects. Missing blood data was attributed to dropout (n=13), cancellation of in-person visits due to the unprecedented COVID-19 shutdown in 2020 (n=7), and unsuccessful blood draw attempts (n=1).

Psychophysiological data.

Baseline heart rate recordings were acquired from 53 out of 58 participants. Missing data (n=5) is wholly attributed to the pandemic, as psychophysiological data collection was temporarily halted as part of a restricted, COVID-19 research protocol. For follow-up visits, heart rate recordings were acquired from 33 participants. Missing data at follow-up was attributed to COVID-19 restrictions (n=12) and dropout (n=13).

Music creativity intervention adherence

Of the 24 music creativity participants who completed the music creativity intervention, 20 individuals participated in the concert that occurred during the final class session. Class attendance across the 6-week course was high (mean percentage of classes attended= 92.36%, mode = 100, range = 55.6-100%). Participants who dropped from the music creativity course completed < 50% of the classes (Mean: 26.5%, range: 11.7%-44.4%). Health issues (self or family) (n=3) and disinterest in the music classes (n=2) were the reasons for discontinuation. One class took place before the pandemic (Jan-March 2020) and two classes took place during the COVID-19 pandemic (Jan-March 2022, March-April 2022). Follow-up visits took place, on average, 17.13 days after the music concert (range: 3-60 days, median: 11 days).

Participant satisfaction, expectations, and feedback

Participants in the music creativity class, on average, rated their impression of the class (1 = poor, 2 = fair, 3 = good, 4 = excellent) as 3.83 out of 4 (100% rated the class as good or excellent). The course met or exceeded expectations for 94% of participants. Participants described the experience as “enlightening: a new way of thinking about music”, “challenging yet enjoyable”, and “very unique and creative.” When asked about their favorite part of the class, participants highlighted the following aspects of the course: the collaborative group effort when composing and performing a piece, the musical excerpts they listened to before composing their own pieces, the opportunity to compose their own pieces, and performing their pieces at the concert. Among (the few) stated concerns, two participants commented that they expected the class to be based on musical scales and theory, not just percussion.

Exploratory findings: Intervention effects on cognitive performance and socio-emotional well-being

Cognitive performance

Digit Span.

A significant time x group effect on raw summary digit span scores was found (b = 1.90, CI[.24, 3.55], p = .03). Simple slopes tests revealed that music creativity participants showed a moderate increase in digit span performance over time (b = 1.32, CI[−.14, 2.88], p = .09) whereas control participants showed no changes in digit span performance over time (b = −.67, CI[−1.63, .34], p = .18) (see Figure 1a). When the effect of time x group on scaled summary scores was tested, the association was attenuated (b = . 76, CI[−.12, 1.64], p = .10). Upon examination of individual subtests, significant time x group effects were found only for sequencing (b = 1.15, CI[.08, 2.25], p = .04), not digit span forwards (b = .31, CI[−.42, 1.02], p < .41) or backwards (b = .54, CI[−.25, 1.31], p = .18). Simple slopes tests revealed that music creativity participants showed an increase in total raw sequencing scores over time (b = .85, CI[.19, 1.64], p = .03) while control participants showed no change in sequencing performance scores over time (b = −.06, CI[−.79, .70], p = .87) (see Figure 1b). Similar, but moderate, associations were found when evaluating longest sequencing span length (b = .60, CI[−.03, 1.23], p = .07). Controlling for MCI status in post-hoc analyses did not change overall conclusions (see supplemental material).

Figure 1a.

Performance on digit span over time by treatment (raw summary scores)

Figure 1b.

Performance on digit span sequency across visits by treatment (raw total score)

Mini Mental State Exam (MMSE).

There was no time x group effect on MMSE scores (b = −.22, CI[−1.1, .67], p = .62). However, 7 music creativity participants’ MMSE scores at follow-up were based on their performance on the Telephone Interview for Cognitive Status (TICS) instead of the MMSE. Further examination revealed that being administered the TICS had a significant effect on change in MMSE scores (b = −2.62, CI[−3.68, −1.56], p < .001), such that those who received the TICS showed a moderately significant decrease in MMSE scores from baseline to follow-up (b = −2.2, CI[−4.15, −.33], p = .06) whereas those who received the MMSE showed an increase in MMSE scores over time (b = .38, CI[.05, .72], p = .03). When 7 participants were excluded from the sample (n=51), there was a trending effect of time x group on MMSE performance (b = .57, CI[−.07, 1.22], p = .09). In models controlling for MCI status, the time x group effect was moderately predictive of MMSE performance (b = .63, CI[.00, 1.28], p = .06); simple effects tests revealed that music creativity participants showed a significant increase in MMSE scores from baseline to follow-up (b = 1.05, CI[.33, 1.78], p = .01) while control participants showed no change over time (b = .16, CI[−.44, .76], p = .60) (see Figure 2).

Figure 2:

Performance on the Mini Mental State Exam across visits by treatment

Color Word Interference Test.

There was no time x group effect on inhibition composite (b = .03, CI[−.12, .19], p = .65; see supplemental material)

Wisconsin Card Sorting Test.

There was no time x group effect on perseveration errors (b = −1.04, CI[−3.24, 1.13], p = .36; see supplemental material).

Socioemotional well-being

There were nonsignificant relationships between treatment group and depressive symptomatology over time (b = −.84, p = .18), perceived stress (b = .40, p = .66), and perceived social support (b = −2.35, CI[ −6.04, 1.36], p = .22). There were also nonsignificant time x group relationships for all quality of life subscales (p value range = .48-.96) except the subscale related to role limitations related to emotional/personal problems (b = 18.67, CI[5.29, 32.20], p = .01). Simple effects revealed that music participants showed a significant improvement in role limitations related to emotional/personal problems over time (b = 16.36, CI[5.66, 27.31], p = .01) while control participants showed no change in this domain over time (b = −2.36, p = .59) (see Figure 3).

Figure 3.

Role functioning (emotional) across visits by treatment

IV. Discussion

Project CHROMA demonstrated good feasibility and acceptability among an older adult sample with and without MCI. Despite circumstances surrounding the COVID-19 pandemic, 400+ community members expressed interest and 56% of those who were eligible were enrolled. The study maintained a 78% retention rate (or 81% if we exclude participants whose visits were intentionally canceled due to the 2020 COVID-19 shutdown). Functional MRI scans, blood specimens, cognitive scores, and self-report socio-emotional questionnaires were successfully acquired for 95% of the participants who completed the entire study. An 82% retention rate was maintained for the music creativity group. All participants in the music creativity group rated the class positively; over 94% thought it met or exceeded their expectations. Music creativity and control groups did not differ on demographic characteristics (except ethnicity), MCI prevalence, baseline psychosocial health, and baseline cognitive function. Importantly, preliminary data analyses suggest that the music creativity intervention may have promising effects on cognitive function: participants in the music creativity group showed some improvements in working memory ability and global cognitive functioning, while participants in the control group showed no differences from baseline to follow-up. Music creativity training was associated with reduced role limitations attributed to emotional problems, but it did not affect other aspects of socioemotional well-being.

Feasibility and acceptability studies are crucial for developing interventions that can have lasting health effects. Because delaying cognitive decline and maintaining healthy cognitive function likely requires sustained interventions, ensuring that music interventions with high time commitment can be successfully completed and can garner high participant satisfaction is vital. The present findings demonstrate that a 6-week novel, music creativity intervention is attractive to and well-received by older adults with varying cognitive abilities and can be successfully implemented within this community.

Project CHROMA’s retention and participation rates are comparable to other Stage 1 music interventions with a similar study design. A double-armed, group song-writing intervention study attracted 47% of eligible dementia family caregivers, maintained a 100% retention rate among both music and control groups, and achieved a 91% participation rate across 6 songwriting sessions (Baker et al., 2018). Domínguez-Chávez et al. (2019) reported an 84% retention rate among 16 older adults with MCI who were enrolled into a single-armed group music therapy intervention focused on music expression through body movements. Participants in Domínguez-Chávez et al. (2019) study received the same amount of training as participants in Project CHROMA (i.e., 36 hours total) but at a different frequency: The intervention was administered for 1 hour, 3 days a week for 12 weeks. A group-singing therapy program designed by Fu et al., (2018) was administered 75 min/week for 12 weeks among community older adults aged 60+. On average, Fu and colleagues observed an 86% retention rate, with participants reporting a 9 out of 10 satisfaction rate. In sum, group music interventions such as Project CHROMA attract interest from the older community, and retention rates typically fall between 80-90%, especially for programs that require longer time commitment. Considering the additional procedures (i.e., functional MRI scans, blood draws) embedded within Project CHROMA that are absent from prior studies, Project CHROMA’s feasibility and acceptability are excellent.

The music creativity curriculum improved performance on some cognitive tests (i.e., global cognitive functioning, working memory) but not others (i.e., set-shifting, inhibition) – a finding that aligns with the current state of the literature. In general, music training benefits cognition but there is less consensus about which cognitive domain is affected and how strong the effect is. For example, some studies report significant improvements in overall cognition (Biasutti & Mangiacotti, 2018; Doi et al., 2017; Han et al., 2020), executive function (Bugos et al., 2007), working memory (Bugos et al., 2007), and memory (Doi et al., 2017) after the music intervention, relative to the control condition, while other studies failed to find improvements in these domains (Hars et al., 2014; Shimizu et al., 2017). Moreover, even within the same study, some measures of working memory/executive functioning showed improvement while others did not (Bugos et al., 2007). Heterogeneity in the target population, music curriculum, type of control group, and neuropsychological assessments administered likely contribute to observed differences across studies.

Separate meta-analyses of existing music intervention studies reveal different standard mean differences in overall cognitive improvement between music and control groups. A meta-analysis on music therapy and MMSE performance in older adults (mean age range: 71.4-82.0 years) revealed that while studies individually showed significant improvements in the MMSE, pooled short-term effects across 5 studies diminished evidence of improvement (std mean difference: .73, CI[.−.07, 1.54], p = .08) (Li et al., 2015). In contrast, a recent meta-analysis of 9 studies examining people with MCI or mild dementia randomized to the control group or active music-making intervention (i.e., improvisation or recreating music through singing/playing instruments), people receiving the music-making intervention showed significant positive changes in their cognitive function relative to those who did not receive the intervention (std mean difference: .30, CI [.10, .51], p = .004)(Dorris et al., 2021). Differing meta-analytic results underscore the heterogeneity across studies and the need for more replication with high-quality clinical trials. In addition, the degree of improvement may depend on the sample’s characteristics, such as the level of cognitive functioning or impairment at baseline.

Contrary to previous work, the group music creativity course did not produce noticeable changes in socioemotional well-being, except in one quality of life domain. While music participants reported their daily functioning being less impacted by emotional problems (such as feeling depressed or anxious), there was no evidence that the music course altered depressive symptomatology or stress. However, given that most participants in the sample were relatively healthy and minimally distressed at baseline, null findings are, perhaps, not surprising. Indeed, previous groups reporting significant intervention effects on depressive symptoms and quality of life (Chan et al., 2011, 2012; Dorris et al., 2021; Yang et al., 2022) had recruited patients or participants who had poorer mental health than the current participant sample. The potential effectiveness of the present music creativity course on socioemotional well-being may be better captured with a more distressed sample – a direction for future work.

This study poses some considerations for future research. First, an important next step is to ascertain the appropriate exposure time needed to acquire a potent change in outcomes of interest and whether participants can commit to longer training periods. Based on participant feedback and staff observations, most participants were heavily invested in the class; these observations suggest that elongating intervention exposure (e.g., increasing it from 6 weeks to 8 weeks or 12 weeks) may be feasible. Second, identifying the essential components of the intervention will require different control conditions. Based on the control condition used, it is unclear whether the benefits of the class were attributed primarily to the musical, social, or the creative aspect of the class or whether benefits required a combination of all three. It is possible that the observed intervention benefits were attributed primarily to enhanced weekly social interaction rather than enhanced music creativity. However, within the music group, perceptions of social support did not change over time (p = .35), suggesting that intervention effects may not come directly from enhanced social interaction. Future studies incorporating a creativity task before and after the intervention and/or an active control condition (e.g., a group-based relaxation or health information class, music-listening class) could help isolate the mechanism of action (i.e., social engagement, creativity, music). Third, a larger MCI sample is needed to determine 1) whether the music intervention can be completed at reasonable completion rates and 2) whether it selectively or more strongly benefits those with or without MCI. Fourth, modest music creativity effects on cognitive and socioemotional well-being may suggest that while the program has some benefits, it may need to be combined with other interventions to alter one’s cognitive health trajectory. For example, it would be worthwhile to examine whether music creativity interventions enhance the effects of concurrent behavioral or pharmaceutical treatment for mild cognitive impairment (Petersen et al., 2018).

Some limitations should be noted. Fist, although preliminary efficacy testing regardless of power considerations are frequently performed in feasibility and acceptability studies (Orsmond & Cohn, 2015), findings should be interpreted with caution. Rather than extrapolating preliminary findings as conclusive evidence for a treatment effect, we suggest using the present findings as evidence to support further efficacy testing in fully-powered clinical trials on music creativity. Second, the present sample consisted primarily of participants without an MCI diagnosis. Although including participants with an MCI diagnosis in the sample did not significantly alter overall conclusions, music creativity effects may look different in a cognitively homogeneous sample. Third, this study lacked an active control group; it is possible that the effects we observed were due to supplemental features of the music creativity course (i.e., behavioral activation, social engagement), rather than the curriculum itself. Fourth, questionnaires were completed at varying times (i.e., range from 0-29 days) and in different contexts (i.e., in the laboratory or in the participants’ home), which may have increased within-group variability. Fifth, recruitment strategies and extensive time commitment may have biased the sample to be higher in socioeconomic status and less cognitively impaired. Higher SES is a protective factor against MCI and Alzheimer’s disease (Sattler et al., 2012). Music intervention effects may be especially beneficial for a lower SES sample, as risk for MCI doubles for the lowest SES quartiles when compared to the highest SES quartile (Fernández-Blázquez et al., 2021). Sixth, whether participants considered music to be an effective intervention for health was not assessed. Because participant expectations are well-known to influence study outcomes (Linde et al., 2007), future iterations of this study should assess participant expectations prior to the start of the study. Lastly, some participant data were collected before the COVID-19 pandemic, with a portion of the music participants completing their baseline visits and follow-up visits during the pandemic. While no pandemic-related group differences were observed, the long-term effects of the COVID-19 pandemic on cognitive and socioemotional factors are still being studied.

In conclusion, Project CHROMA demonstrated good feasibility and acceptability among a sample of older adults with varying degrees of cognitive functioning. Despite incorporating a wide range of health procedures and requiring participants to commit to 6-week training period, Project CHROMA successfully attracted and retained participants at comparable rates as other music interventions. High satisfaction rates and positive participant feedback further emphasize Project CHROMA’s success, posing the possibility that elongating the training duration may be feasible in future studies. Moreover, the music creativity curriculum, relative to the inactive control condition, improved some aspects of cognitive functioning and socioemotional well-being. Preliminary findings substantiate the need to probe music creativity’s effects on these outcomes with a larger sample size, more MCI participants, and a sample with wider range of psychological well-being.