A social dancing pilot intervention for older adults at high risk for Alzheimer's disease and related dementias

Abstract

Close to 6 million older US adults have Alzheimer's disease or related dementias, yet there is currently no cure or effective treatment. This single-blind randomized controlled trial (clinicaltrials.gov: NCT03475316) aims to establish feasibility, and explore the relative efficacy, of a 6-month social ballroom dancing intervention versus a 6-month active control intervention (treadmill walking) for improving executive function in 32 older adults at increased risk for Alzheimer's disease or related dementias. Dementia-at-risk status is determined with cut-scores on the memory impairment screen (≥3 to ≤6) and/or the AD8 Dementia Screening Interview (≥1). The primary outcome is a composite executive function score from digit-symbol substitution, flanker interference and walking-while-talking tasks. The secondary outcome is functional neuroplasticity during fMRI-adapted versions of digit-symbol substitution, flanker interference and walking-while-talking.

Practice points.

• Alzheimer's disease and related dementias (ADRD) is an increasingly important public health problem, with no cure or effective treatment.

• Identifying novel, noninvasive and cost-effective interventions for ADRD is therefore imperative.

• This pilot randomized clinical trial protocol aims to establish the feasibility of a social dancing intervention in older adults at increased risk for ADRD.

• This social dancing intervention will be contrasted with an active control intervention (treadmill walking).

• It is hypothesized that the social, cognitive and physical demands of social dancing will lead to greater benefits on cognition (executive functions) and greater neuroplasticity than treadmill walking.

• The results of this pilot will be used to develop a more definitive large-scale randomized clinical trial.

Executive functions (EF) monitor and coordinate complex behaviors – including planning, reasoning, attention allocation and selection/inhibition of actions [1] – and are negatively affected by aging, mild cognitive impairment (MCI) and Alzheimer's disease and related dementias (ADRD [2]). Aerobic exercise interventions result in modest improvements in EF among older adults [3–5], likely because they induce neuroplasticity in hippocampal and prefrontal cortex (PFC) regions [6–8]. Neuroplasticity is the brain's ability to alter its structure and function during development, and in response to learning, environmental challenges or pathology [9]. Long-term adherence to exercise programs, however, is low, particularly among older adults [10–13] – and only 17.6% of older US adults met aerobic and muscle strengthening guidelines in 2017 [13]. Cognitive training, particularly computerized cognitive training, also results in modest improvements in EF, and neuroplasticity in PFC [7,14,15]. Cognitive training interventions that incorporate social interactions may also improve EF and neuroplasticity in PFC [16], but evidence is limited. Identifying enjoyable activities that are physically, cognitively and socially engaging to older adults may be key to bring about sustained behavioral and brain changes, and maximize improvements in EF.

Social dancing is a potentially enjoyable – yet complex and challenging – social, physical and cognitive activity that could be used to induce considerable and long-lasting improvements in EF and neuroplasticity in older adults. We have shown that older adults who report dancing socially on a regular basis have a reduced risk for dementia, better balance, and more stable walking patterns than those who do not dance [17,18]. Limited, but emerging, evidence suggests that dance in general (individual or social) improves mobility, memory and EF in older adults – particularly dual-task walking performance [19–28]. Some evidence further suggests that social dancing may increase structural brain connectivity in the fornix of older adults, yet such neuroplasticity was not accompanied with improved cognition [29].

The current pilot, single-blind, randomized controlled clinical trial (RCT; clinicaltrials.gov: NCT03475316) contrasts a 6-month social (ballroom) dancing intervention with a 6-month active control/treadmill walking-to-music intervention in 32 older adults at increased risk for ADRD. Dementia-at-risk status is defined as meeting cut-scores on any one of two established cognitive screening instruments: the 4-item memory-impairment screen (MIS; ≥3≤6 [30]) or the AD8 Dementia Screening Interview (AD8; ≥1 [31]). The first aim of this pilot is to obtain preliminary data regarding the feasibility and relative efficacy of social dancing and treadmill walking on EF. A composite EF score derived from three validated and reliable standardized (digit symbol substitution test [DSST]), computerized (Flanker interference) and dual-task walking (walking-while-talking [WWT]) tasks is the primary outcome. We predict that social dancing will result in greater intervention-related improvements in EF than treadmill walking. The second aim is to explore intervention-related functional neuroplasticity – or relative changes in the function of the brain following dancing and walking. Functional activation/deactivation patterns during our validated imagined WWT task [32] as well as flanker interference and DSST tasks are the secondary outcomes. We predict that dancing will result in greater intervention-related changes than walking in functional activation/deactivation patterns, particularly in supplementary motor, anterior cingulate and PFC regions.

Methods

Study design

We will recruit 32 community-dwelling older adults to participate in a pilot single-blind RCT (NCT03475316). Participants will be randomized to a 6-month (90 min, twice a week) social dancing or an active control/treadmill walking to music intervention. The study flow and procedures are summarized in Figure 1.

Figure 1. . Study flow diagram.

AD8: Ascertain dementia eight-item informant questionnaire; MIS: Memory-impairment screen; MRI: Magnetic resonance imaging.

Recruitment, screening & randomization

A random sample of older adults (65 years and older; stratified by age [65–80 years; >80 years] and sex) will be identified from local population lists (voter registration lists), the clinical population in the Montefiore–Einstein health system and previous studies at Albert Einstein College of Medicine (Division of Cognitive and Motor Aging). Those expressing interest will be initially screened over the phone using our ‘dementia-at-risk’ cut-scores on either the MIS (>3 to ≤6 [30]) and the AD8 (≥1 [31,33]). The MIS is a brief four-item dementia screen, and the AD8 is an eight-item dementia screen that assesses perceived and recent changes or problems with judgement, loss of interest in hobbies or activities, repeating themselves, learning, forgetting, handling finances, remembering appointments and daily problems with thinking and memory. The MIS was developed in our Bronx population, accounts for various education levels, and has a sensitivity of 85% and a specificity of 86% for detecting dementia [30,34–37]. The AD8 has a sensitivity of 74% and a specificity of 84% for detecting dementia [31,33]. Performance on both the MIS and AD8 are associated with dementia pathology [38,39]. Those expressing interest in the trial will also be screened about their willingness and safety to undergo magnetic resonance imaging (MRI). This additional criterion was included in this pilot trial to avoid recruiting a larger sample, but will be loosened in future trials. Older adults with dementia will be excluded based on a previous physician diagnosis of dementia, or dementia diagnosed during phone screening (MIS <3) or at the initial visit after review of all cognitive and function assessments. Additional exclusion criteria include serious chronic or acute illness, and mobility or health limitation that prevent participation in intervention. Table 1 includes a complete list of inclusion and exclusion criteria. To reduce attrition following enrolment as might occur if participants wait to start classes if done by blocks, we will have rolling admission into the dance and walking classes. Randomization will take place after the participant has completed all baseline procedures. Group assignment is generated by our statistician, Dr Wang, who is not involved in participant testing or interventions.

Table 1. . Eligibility criteria.

Inclusion criteria

Exclusion criteria

• Adults aged 65 and older • A score of ≤6 on the memory impairment screen or ≥1 on the AD8 • Plan to be in area for next year or more • English speaking • Willing to complete MRI

• Presence of dementia based on previous physician diagnosis of dementia or dementia diagnosed by the study clinician during phone screening or initial visit • Serious chronic or acute illness such as cancer (late stage, metastatic or on active treatment), chronic pulmonary disease on ventilator or continuous oxygen therapy or active liver disease • Mobility limitations solely due to musculoskeletal or cardiovascular conditions that prevent participation in the intervention programs • Any medical condition or chronic medication use (e.g., neuroleptics) in the judgment of the screening clinician that will compromise safety or affect cognitive functioning • Terminal illness with life expectancy less than 12 months • Presence of progressive, degenerative neurologic disease (e.g., Parkinson's disease or amyotrophic lateral sclerosis) • Severe auditory or visual loss • Active psychoses or psychiatric symptoms (such as agitation) noted during the clinic visit that will prevent completion of study protocols • Either participation in competitive dancing or recreational dancing at a frequency >1/month in the past 6 months • Participation in other interventional study that overlaps with intervention period of this study

Study measures

A complete list of outcomes and additional study measures are summarized in Table 2. Primary and secondary outcomes are discussed in more detail below. Additional study measures aim to comprehensively characterize participants at baseline, identify potential confounders and outcomes for consideration in future studies, and to determine the feasibility of this RCT in older adults at risk for dementia. Potential bias will be reduced by: concealing treatment allocation until participants have enrolled in the trial [40], administering social dancing and treadmill walking interventions at different times, instructing study staff not to disclose assignment or details about the interventions and using different study staff to administer pre- and post-intervention assessments and social dancing and treadmill walking interventions (to ensure single blinding). Pre- and post-intervention assessments take approximately 8 h, but are completed over the course of 2 days (a week apart) to avoid fatigue. Primary outcomes (but not additional study measures) are also administered 2 and 4 months into the intervention, and 3 months after completing the intervention. Primary outcome assessments take approximately 1 h to complete. Each intervention session lasts for 90 min, including warm-up, dance or treadmill walking and cool down, and are completed twice a week for 6 months (48 sessions).

Table 2. . Outcomes and study measures.

Measures	Pre-intervention	Randomization after successful completion of baseline measures	During intervention		Post-intervention
			Month 2	Month 4	Month 6	Month 9
Phone screen
Demographic history	✓
Medical history/dementia screen	✓
Memory impairment screen (MIS)	✓
Ascertain Dementia Eight-Item Informant Questionnaire (AD8)	✓
MRI safety (verbal form)	✓
In-person screen
MRI safety (paper form)	✓
Montreal cognitive assessment (MoCA)	✓
Primary outcomes
Walking-while-talking (WWT) task: gait velocity (cm/s) during WWT	✓		✓	✓	✓	✓
Flanker interference task: time, incongruent trials minus congruent trials	✓		✓	✓	✓	✓
Digit symbol substitution task (DSST): correct trials (paper form)	✓		✓	✓	✓	✓
Secondary outcomes
Imagined WWT task: blood-oxygen-level-dependent (BOLD) signal during imagined WWT	✓				✓
Flanker interference task: BOLD signal during incongruent trials minus congruent trials	✓				✓
DSST: BOLD signal during correct trials (computer form)	✓				✓
Tertiary outcomes
Community Healthy Activities Model Program for Seniors (CHAMPS)	✓				✓	✓
Modified Katz Disability Scale	✓				✓	✓
Quantitative gait performance: GaitRite and PKMAS instrumented walkways	✓		✓	✓	✓	✓
Balance: unipedal stance	✓				✓	✓
Geriatric depression scale (GDS-30)	✓				✓	✓
Other study measures
Wide range achievement test (WRAT)	✓				✓	✓
Free cued selective reminding test (FCSRT)	✓				✓	✓
Trail making test A & B	✓				✓	✓
Phonemic fluency (FAS)	✓				✓	✓
Semantic fluency (CAT)	✓				✓	✓
Hopkins verbal learning test (HVLT)	✓				✓	✓
Repeatable battery for the assessment of neuropsychological status (RBANS): figure copy & delay	✓				✓	✓
WAIS digit symbol substitution test (paper)	✓				✓	✓
Beck anxiety inventory	✓				✓	✓
Instrumental activities of daily living (ADL) questionnaire	✓				✓	✓
Medical history	✓				✓	✓
Sensory screen	✓				✓	✓
Neurological gait exam	✓				✓	✓
Medications	✓				✓	✓
Height/weight/body mass index	✓				✓	✓
Blood pressure	✓			✓	✓	✓
Short physical performance battery	✓				✓	✓
Floor maze test	✓				✓	✓
Multi-sensory integration	✓				✓	✓
Berg balance scale	✓				✓	✓
Stair climbing	✓				✓	✓
Grip strength	✓				✓	✓
Falls efficacy scale & falls	✓				✓	✓
Duke activity status index	✓				✓	✓
General mobility questionnaire	✓				✓	✓
Activity balance confidence scale (ABC)	✓				✓	✓
Apathy evaluation scale	✓				✓	✓
Leisure activity questionnaire	✓				✓	✓
Social network index	✓				✓	✓
MOS social support survey	✓				✓	✓
Stroop interference	✓		✓	✓	✓	✓
Letter number sequencing	✓		✓	✓	✓	✓
White matter integrity (from diffusion-weighted images)	✓				✓
White matter hyperintensities (from FLAIR images)	✓				✓
Adverse event tracking			✓	✓	✓	✓
Postintervention questionnaire					✓

Primary outcomes: composite EF score

Executive functions will be assessed with a composite EF score by summing standardized scores on three different tasks: the DSST [41], flanker interference [42,43] and WWT [44] tasks. During the DSST, participants are presented with a series of nine digit-symbol pairs followed by a series of digits. Participants are asked to write down as many corresponding symbols they can within 120 s, and the number of correct trials during this time, pre- and post-intervention, is of primary interest (higher values reflect a better outcome). DSST engages both EF and processing speed [45], has been linked to aging [46], dementia [47] and mortality [48,49], and is also associated with physical functions [50].

During each trial on the flanker interference task, participants are presented with a line of 5 arrows that point either to the right [> > > > >] or left [< < < < <], and they are asked to report the orientation of the central (middle) arrow with a right or left mouse press, while ignoring the orientation of the peripheral or flanking arrows. The central arrow is either congruent with the flanking arrows ([> > > > >]; [< < < < <]; congruous trials) or incongruent with the flanking arrows ([> > < > >]; [< < > < <]; incongruent trials), and the difference in response time to incongruous and congruous trials before and after the interventions is of primary interest (lower values reflect a better outcome). Each participant completes a total of 60 congruent and 60 incongruent trials presented in a random order over the course of two blocks, separated by a short resting period. Flanker interference tasks taps into interference resolution, an EF that involves selecting appropriate responses in the presence of interfering stimuli, is a component of the NIH toolbox [51], and is sensitive to aerobic exercise [6].

During the WWT task, participants are asked to walk at their normal pace (single-task walking), stand while reciting alternate letters of the alphabet (single-task talking), and walk while reciting alternate letters of the alphabet (WWT). The difference in gait velocity (cm/s) during WWT (dual-task walking) and walking alone (single-task walking), pre- and post-intervention, is of primary interest (lower values reflect a better outcome). Gait velocity is quantified from a 20-feet instrumented walkway (GAITRite^® electronic walkway system; NJ, USA), and additional quantitative gait parameters are also obtained and available for analysis (Table 2). WWT necessitates attention allocation to competing task demands, and has been linked to falls, frailty, disability, mortality and dementia [44,52,53].

Secondary outcomes: neuroplasticity

Neuroplasticity will be assessed using functional MRI (fMRI) during fMRI-adapted versions of the DSST, flanker interference, and WWT tasks. During each trial of the fMRI adapted version of the DSST task, participants are presented with a code table consisting of nine digit-symbol pairs on the top of the screen, and a digit-symbol pair at the bottom of the screen. Participant are asked to decide (with a button press) if the digit-symbol pair at the bottom of the screen corresponds to the code on the top of the screen, or if it does not correspond to the code on the top of the screen. A total of five testing blocks, with 18 randomly presented items each are completed first outside of the scanner, and then inside of the scanner. This digit-symbol substitution task has been previously validated in older adults [54]. The blood-oxygen-level-dependent (BOLD) signal (activation/deactivation pattern) during digit-symbol decisions relative to rest, pre- and post-intervention, is of primary interest. The fMRI version of the flanker interference task was identical to the one completed outside of the scanner (see primary outcomes), except for that responses were made with a left or right button press rather than a left or right mouse press. The BOLD signal during incongruous relative to congruous trials, pre- and post-intervention, is of primary interest. Finally, we will examine functional activation and deactivation patterns with an fMRI or motor imagery version of the WWT task that was developed and validated by our group [55]. There are three phases to this task. During the first phase, which takes place outside of the MRI, participants are asked to walk around a 4 × 14 foot Zeno electronic walkway using ProtoKinetic Movement Analysis Software (PKMAS; Zenometrics LLC, NY, USA). Participants are asked to walk around this walkway three times at their normal pace (W) and three times while reciting alternate letters of the alphabet (WWT). They are also asked to stand still on this walkway while reciting alternate letter of the alphabet (T) for 30 s. During the second phase, which also takes place outside of the MRI, participants undergo imagery training. First, they are asked to execute and then imagine fairly simple motor movements, such as raising their arm and tapping their foot, as part of the Vividness of Imagery Questionnaire [56]. They are instructed to close their eyes during imagery and rate the visual and kinesthetic qualities of their images on a scale from 1 (no image; no sensation) to 5 (image as clear as seeing; as intense as executing the action). Then, seated at a desk, they are asked to close their eyes and imagine or envision themselves walking around the walkway (iW), walking around the walkway while reciting alternate letters of the alphabet (iWWT), and standing on the walkway while reciting alternate letters of the alphabet (iT). During the first trial, instructions are detailed (i.e., “Imagine Walking: At the start of the next tone, close your eyes and imagine or envision yourself walking on the mat, at the start of the following tone, stop and wait for further instructions”). During subsequent trials, participants are simply prompted to ‘Imagine walking’. Following each trial, participants rate the visual and kinesthetic qualities of their images on a scale from 1 to 5. Long and short imagery instructions are presented visually, and the beginning and end of imagery is initiated with a tone. One imagery event lasts for 16 s, and the whole imagery training procedure lasts for approximately 15 min. During the third phase of the imagery of WWT task, participants are prompted to iW, iWWT and iT during fMRI scanning. Each imagery event lasts for 16 s and is repeated six times, separated over three blocks, with a short break in between. After each block, participants are asked to rate the overall visual and kinesthetic qualities of their images on a scale from 1 to 5. The BOLD signal during iW, iT and iWWT trials, pre- and post-intervention, is of primary interest.

MRI scanning will be performed locally with a Philips 3T system (Ingenia Elition) at our institution. All BOLD (T2*-weighted) images [57,58] will be acquired with multi-band echo planar imaging (EPI) using a whole brain gradient over a 224 × 224 × 128 mm field of view (FOV) on a 112 × 112 acquisition matrix, 2 mm slice thickness (no gap); TE = 28 ms, TR = 2000 ms, flip angle = 90 degrees and 64 trans-axial slices per volume. A T1-weighted whole head structural image is also acquired using axial 3D-MP-RAGE parameters over a 240 × 188 × 220 mm FOV and 1.0 mm isotropic resolution, TE = 4.6 ms, TR = 9.9 ms, α = 80, with SENSE factor 2.6. Our fMRI adapted DSST, Flanker interference and imagery of WWT protocols are written in E-Prime 2.0 (Psychology Software Tools, Inc.) and presented with an InVivo Eloquence fMRI system. Additional MRI images are also acquired as potential covariates and tertiary outcomes, including diffusion-weighted and Fluid Attenuated Inversion Recovery (FLAIR) images (Table 2). The MRI scan lasts for approximately 1 h and 15 min, including screening, participant preparation, and actual scanning time (50 min, 42 s).

Study interventions: social ballroom dancing & treadmill walking

Participants assigned to the social ballroom dancing arm will complete 90-min dance sessions twice a week for 6 months. Each dance session includes warm-up (10 min of gentle stretching to music), low intensity dance instruction (30 min), break (10 min), higher intensity dance instruction (30 min) and cool down (10 min). Low intensity dances include foxtrot and waltz, and higher intensity dances include salsa and east coast swing. Heart rate and blood pressure is checked before warm-up, during the break, and after cool down. A participant cannot start a session (and a session is stopped) if systolic blood pressure is >150 mmHg. Participants are also monitored with the subjective Borg scale [59] at 2 points during each session (in the middle and at the end of the sessions) to monitor level of exertion. Dance sessions are led by instructors who are experienced in teaching older adults (www.rbcares.org). Heart rate and blood pressure during the sessions is monitored by trained research personnel. Dance instructors ensure that the progression to new dances is gradual, and take into account balance, cardiac and cognitive demands. They also ensure that participant rotate partners during dance sessions. Each dance is first demonstrated by the instructor. The instructor then repeats the dance sequence several times, and the participants tries to reproduce it along with the instructor, until the sequence has been learned. Finally, participants reproduce the sequence along with the music. Note that a recent review of dance interventions in older adults for cognitive, and other health outcomes, recommend a minimum of one 45-min session per week for 6 weeks [60], which is lower than the duration of the current study, but this review did not specifically address dance dose for potential neuroplasticity effects. In a recent RCT conducted in Japan, we have also shown cognitive benefits of dancing for 1 h once a week for 9 months in older adults with MCI, which is lower than the current dose [20]. Finally, note that the exercise dose in both the social dancing arm and treadmill walking are higher than current CDC exercise recommendations [13].

Participants assigned to the treadmill walking arm will complete 90-min treadmill walking sessions twice a week for 6 months. Analogous to the dance sessions, each 90-min treadmill walking session includes warm-up (10 min of gentle stretching to music), low-intensity walking (30 min), break (10 min), higher intensity walking (30 min) and cool down (10 min). Treadmill sessions are monitored by trained research personnel who ensure that walking speed begins at 1.0 mile per hour, and is then gradually increased to a moderate pace where participants is still able to carry on a conversation, but does not exceed 3.0 miles per hour. As in the dance arm, participant's level of exertion is monitored with the subjective Borg scale [59] in the middle and at the end of the sessions. Trained research personnel also monitor heart rate and blood pressure before warm-up, during the break and after cool down. A participant cannot start a session (and a session is stopped) if systolic blood pressure is >150 mmHg. Treadmill sessions are done in groups to mimic the social interactions of the dance arm. Moreover, music from the same playlists used during dance sessions are played during treadmill walking sessions to further reduce the variability between the two arms.

Study interventions: performance & safety monitoring & feasibility

The social dancing and treadmill walking interventions will be administered by trained personnel under controlled conditions in order to protect the internal validity of the study, and to ensure compliance with the protocol. Heart rate and blood pressure is monitored prior, during, and after each intervention session to ensure safety and to permit comparison of aerobic exertion between intervention arms. Adverse events are tracked and logged for each participant from enrolment until their final study visit. Trained research assistants ask participants at each visit about falls and changes in their health since their prior visit. If participants miss any intervention session or study visit, research assistants will reach out to them to find out the reason. If the research assistants become aware of any serious and/or potentially related adverse events at any point during the study the safety officer and NIH representative will be notified. All adverse events are logged and reported to the Einstein IRB on an annual basis and to the safety officer on a biannual basis. Interim administration of primary outcomes (2 and 4 months after the start of the intervention) permit examination of the trajectories and asymptote of potential cognitive benefits of our intervention in order to inform the design of future (large-scale) dance and treadmill interventions. Administration of primary outcomes 3 months after intervention completion permit us to explore the longevity of potential cognitive benefits. A number of feasibility metrics will be examined, including recruitment sources, implementation (i.e., intervention session completion rates), retention rates, acceptability (post-study interviews) and safety (adverse events monitoring). Reporting of study design and results will be done in accordance with CONSORT guidelines for pilot feasibility study [61].

Statistical plan: primary outcomes

The primary statistical objective of this pilot study is to obtain estimates of intervention effects on EF for the purpose of designing a full-scale study in the future. Linear mixed effects models will be used to evaluate and compare trajectories of the composite EF scores as a function of intervention arm (social ballroom dancing vs treadmill walking) and time (baseline/pre-intervention, 2 months, 4 months, 6 months/post-intervention, 9 months/3-months post-intervention). Estimates with 95% confidence intervals (CI) will be reported for the mean EF score at each time point and the slope of change across time within each group, as well as the differences between groups. Models will be adjusted for age and sex; further adjusting for education and chronic illnesses will also be considered. Unadjusted models will also be explored. Baseline distribution of covariates will be compared with assess adequacy of randomization. Analyses will be performed with the intention to treat such that all randomized participants are considered in the analyses – regardless of adherence, whether the intervention was received and subsequently withdrawn, or deviated from the protocol. Exploratory analyses will also be performed on individual EF tests, physical intensity (heart rate and Borg Scale), and dose (number of sessions completed).

Statistical plan: secondary outcomes

Pre-processing and first-level analyses will be performed with statistical parametric mapping (SPM12) for each fMRI-adapted EF task separately. In short, each participant's EPI dataset will be: realigned to the first volume to correct for motion, temporally shifted to correct for the order of slice acquisition, co-registered to the T1-weighted/structural image, spatially normalized in Montreal Neurologic Institute space [62] and spatially smoothed with an isotropic Gaussian kernel, full-width at half-maximum = 6 mm. During first-level analyses, each pre-processed EPI time series is modeled with regressors that represent the expected BOLD response (implicitly relative to blanks) for each trial type in each task (i.e., digit symbol decision and rest trials, incongruous and congruous trials, iW, iT and iWWT trials). The contrast maps for each trial type that is generated during first-level analyses are then used in second-level or group-level covariance analyses. Ordinal trend covariance analyses (OrT-CVA) implemented with the generalized covariance analysis suite (www.nitrc.org/projects/gcva_pca [63,64]) will be used to identify patterns of BOLD signal increases (activation) and decreases (deactivation) as a function of trial type during each MRI visit (baseline/pre-intervention and 6 months/post-intervention) for each intervention arm (social ballroom dancing and treadmill walking). OrT-CVA employs a PCA to the data matrix that is then transformed to a matrix of the experimental design (e.g., iW, iT and iWWT). Linear regression is then applied to detect a covariance pattern (ordinal trend) in the BOLD signal as a function of task conditions (e.g., trial type) that is based on a linear combination of a small set of principal components. An ordinal trend is a monotonic change in pattern expression as a function of task conditions. The expression of an ordinal trend is quantified in terms of a participant-specific expression (or factor) score that is derived by projecting the covariance pattern onto a participant's scan for each task condition.

Power: primary outcomes

The pilot sample size was estimated based on the pragmatics of recruitment, and necessities for examining feasibility [65]. We recognize, however, that this pilot study may be underpowered to detect our hypothesized effect, and therefore our power analyses reflect the effect that we can detect given our sample size, rather than our expected effects. We expect a correlation of at least 0.80 between repeated composite EF scores. With 16 participants in each intervention arm, and assuming a 20% dropout rate at post intervention, we can detect a difference of 0.0075 standard deviation (SD) per month in EF slope, using baseline/pre-intervention, 2, 4 and 6 months/post-intervention between groups, with 80% power, using a two-sided tests with an alpha level of 0.05.

Power: secondary outcomes

The sample size of 16 per group is within the number recommended for neuroimaging pilot feasibility study [66–68], and will be used to guide study design and modeling strategy in future (large-scale) dance and treadmill interventions.

Data management & quality control

Trained research personnel will collect and enter behavioral data in a secure REDCap (Research Electronic Data Capture [69]) database during each study visit and intervention session. REDCap databases provides central storage and back-up of data, real-time data validation and integrity checks, and export functions into different statistical software. MRI data will be stored, pre-processed and analyzed on a secure Linux server. Only trained research personnel will be assigned data viewing, entry, or login privileges in these databases. Data monitoring reports will be provided to the safety officer on a biannual basis.

Discussion

This pilot RCT aims to establish the feasibility, and explore the relative efficacy, of a 6-month social dancing intervention and a 6-month treadmill walking to music intervention in older adults at increased risk for AD and related dementias (ADRD). Our dancing intervention builds upon emerging evidence, which suggests dancing may benefit mobility and cognitive functions in older adults [19–28] – and specifically targets a group of older adults at increased risk for future cognitive decline, AD and related dementias. This pilot is the first to explore the relative efficacy of social dancing and treadmill walking with a composite EF score, which overcomes the limitations associated with using any single EF task [70–72]. This composite EF score is derived from standardized, computerized and cognitive-motor tasks that have been previously linked to aging, dementia and/or physical functions [6,46,47,50,53]. These tests have also been incorporated or adapted to explore functional brain changes following social dancing and treadmill walking using fMRI [54,55], an issue that has not been previously explored. In fact, very limited evidence is available regarding the potential for neuroplasticity following dancing [29]. Our focus here is on functional brain changes because they may be more sensitive biomarkers of ADRD than structural brain changes [73]. The use of parallel tests outside and inside of the scanner also permit potential behavioral improvements to be ‘more directly’ linked to functional brain changes. We hypothesize that the social, cognitive and physical demands of dancing will lead to greater improvements and functional neuroplasticity than treadmill walking, particularly in supplementary motor, anterior cingulate and PFC regions.

The public health impact of ADRD is substantial and rising [74–76] – and there are currently no pharmacological or non-pharmacological cures or effective treatments. Identifying novel, noninvasive and cost-effective interventions that can be used to reduce the incidence and slow the progression of ADRD is therefore of utmost importance. Aerobic exercise interventions generate modest improvements in cognition and induce neuroplasticity in relatively healthy older adults [3–8], but few older adults exercise at recommended levels [13]. Social dancing is potentially a more enjoyable physical, cognitive and social activity that can be used to improve cognitive function, particularly EF, and to induce neuroplasticity in older adults. This pilot RCT targeted to older adults at increased risk for AD and related dementias will establish feasibility, and provide preliminary data of the relative efficacy of dancing and walking on EF. The results of this pilot will be used to develop a more definitive large-scale RCT to support or refute the use of social dancing to improve cognitive functions in older adults at high risk for AD and related dementias.