A Pilot Study of Home-Based Rhythmic Auditory Stimulation to Evaluate Mobility, Balance, and Patient-Reported Outcomes After Traumatic Brain Injury

1. PURPOSE, BACKGROUND, SCIENTIFIC RATIONALE Traumatic brain injury (TBI) affects approximately 2.8 million people annually in the United States and remains a major public health concern due to its high rates of disability and long-term functional impairment.1 TBI commonly results from falls, motor vehicle collisions, sports -related injuries, and assaults, and can lead to persistent deficits in mobility, balance, cognition, and overall functional independence.2 Even after acute recovery, many individuals experience prolonged limitations that interfere with community participation and quality of life.3,4 As survival rates improve, optimizing rehabilitation strategies that support functional recovery has become increasingly important. Impaired mobility and balance are two common long-term sequelae of TBI5-7 that have substantial negative life impacts and often persist for years post injury.8 Individuals with TBI walk more slowly,9-13 demonstrate greater imbalance,11-13 and have reduced endurance14 compared to nondisabled counterparts; evidence suggests that improving mobility and balance after TBI is associated with improved quality of life and community participation.15 Despite clinical guidelines emphasizing the importance of ongoing, task-specific rehabilitation to promote neuroplasticity, access to sustained outpatient therapy for individuals with chronic TBI is limited by cost, transportation barriers, workforce constraints, and insurance coverage.16 As a result, there is a critical need for scalable, home-based rehabilitation approaches that can safely deliver high-quality, progressive mobility training while supporting adherence and engagement over time. Rhythmic auditory stimulation (RAS) is a well established, evidence based neuromodulation approach that uses rhythmic auditory cues, such as music, to synchronize and support motor activity through auditory-motor entrainment..17,18 This synchronization can enhance gait neuromotor control by activating intact motor networks to facilitate recovery after neurologic injury.19,20 Stroke literature demonstrates that RAS improves walking speed, cadence, symmetry, endurance, and functional mobility, with emerging evidence that music-based cueing may also enhance motivation, enjoyment, and engagement relative to metronome-only cueing.17 A growing body of literature suggests that music-based interventions may offer multi-domain benefits for individuals with TBI, extending beyond motor recovery to cognitive, emotional, and psychosocial outcomes.21,22 Randomized controlled trials of neurological music therapy in moderate-to-severe TBI have demonstrated improvements in executive function, behavioral regulation, and attentional control.22,23 Complementary neuroimaging analyses provide preliminary evidence that these functional gains may be supported by music-induced neuroplasticity, including structural and network-level changes within prefrontal and frontotemporal circuits implicated in cognitive control and emotional regulation. Systematic reviews further indicate that music-based interventions are associated with improvements in mood, agitation, stress, social interaction, and sleep quality in TBI populations, supporting their potential relevance for quality-of-life outcomes. However, despite these promising findings, the evidence base remains limited by small sample sizes, heterogeneous intervention protocols, variable outcome measures, and inconsistent reporting of dose, timing, and treatment fidelity. Recent reviews emphasize the need for larger, methodologically rigorous trials with standardized outcomes and clearer mechanistic frameworks to support translation into routine neurorehabilitation practice. MedRhythms' MR-001 represents a significant advancement in the delivery of RAS. The system integrates wearable inertial sensors with proprietary, closed-loop algorithms to autonomously deliver individualized, progressive, music-based rhythmic cueing during walking in the home environment. Specifically, the MR-001 neurorehabilitation system consists of three components: two shoe-worn wearable inertial sensors that measure walking patterns and gait parameters in real time, a bluetooth headset to deliver RAS via music, and a touchscreen control unit preloaded with intervention software to deliver personalized therapy. The MR-001 neurorehabilitation system continuously assesses the user's entrainment to the target tempo and evaluates gait symmetry and variability, while safely and autonomously adjusting the target walking speed without direct clinician input. Unlike traditional RAS approaches that require in-person clinician oversight, this technology adapts tempo and cueing in real time based on gait quality and entrainment, enabling safe, unsupervised home use. A large-scale study (n=204) in the stroke population provides strong preliminary evidence supporting the feasibility, safety, and effectiveness of this approach. In the OrcHESTRAS single arm, pragmatic home-based trial, over 81.9% (95% confidence interval \[CI\] 0.76-0.87) achieved moderate to high weekly use \> 4 weeks meeting the primary endpoint (p\<0.001) over 12 weeks, exceeding the predefined feasibility benchmark of 60%. Additionally, participants demonstrated significant improvements in walking endurance (6-minute walk test \[6minWT\] 26.1 ± 5.6m; 95% CI 14.99, 37.22) and functional mobility (Timed Up and Go \[TUG\] -1.45 ± 0.31s; 95% CI -2.06, -0.84 p\< 0.001). Weekly use influenced effectiveness with each additional week of use predicting a 5.82m greater gain in the 6minWT (standard error \[SE\] = 2.05; 95% CI 1.77, 9.87\], p\<0.005. Importantly, the intervention exhibited an acceptable safety profile across more than 100,000 minutes of autonomous use, with adverse events and falls largely consistent with baseline risk in chronic stroke populations (preprint doi.org/10.64898/2026.03.13.26348352) Brain derived neurotrophic factor (BDNF) is a growth factor known to be important for both neuronal and cognitive plasticity. Increased BDNF levels improves cell signaling promoting synaptic plasticity and neurogenesis .37,38 High intensity interval training and continuous moderate to high intensity exercise are both associated with increased BDNF levels in animal models,39 but the impact of high-intensity training on BDNF has not been tested in individuals with moderate to severe TBI. We propose to evaluate changes in BDNF levels after 6 weeks of training interventions (including high-intensity training) in individuals with moderate to severe TBI. Although a growing body of literature supports RAS in brain injury, the neural mechanisms underlying RAS remain poorly characterized. Functional near-infrared spectroscopy (fNIRS) offers a portable, ecologically valid method for assessing cortical activation during auditory and motor tasks and is particularly well suited for rehabilitation research involving movement. Emerging fNIRS evidence in acquired brain injury demonstrates altered and exaggerated prefrontal cortical hemodynamic responses to musical and rhythmic auditory stimuli, suggesting inefficient or compensatory neural processing during auditory-cognitive integration tasks.24 These findings support the feasibility of using fNIRS to objectively quantify brain responses to rhythmic auditory input after brain injury and highlight a critical gap in understanding how RAS modulates cortical networks during motor rehabilitation in individuals with TBI. Additionally, while RAS is supported by a robust evidence base in stroke, comparable data are lacking in individuals with TBI despite shared mechanisms of motor control impairment and neuroplastic potential. Individuals with TBI may demonstrate even greater benefit from music-based, home-delivered interventions that reduce cognitive load, enhance enjoyment, and support sustained engagement in rehabilitation. However, foundational data are needed to establish feasibility, adherence, safety, and preliminary signal of benefit before larger efficacy trials in this population are warranted. Study Objective and Design The purpose of this pilot randomized controlled trial with a waitlist control study is to determine the safety, feasibility, and adherence of a home-based, music-guided RAS intervention using the MR-001 for individuals with TBI, and to explore preliminary effects on mobility, cognition, mood, enjoyment, perceived change, and cortical excitation. This work addresses critical barriers to long-term rehabilitation by evaluating a home-based approach to gait recovery in the chronic TBI population. Aim 1: To determine the safety, feasibility, and adherence of a home-based, music-guided RAS intervention delivered via the MR-001 in individuals with TBI. Hypothesis 1: The MR-001 intervention will be safe and feasible for individuals with chronic TBI (\>6 months). Safety will be demonstrated by the absence of serious device-related adverse events. Feasibility will be demonstrated by the ability to enroll and retain 30 individuals with TBI; and adherence will be defined as enrolled individuals will complete \> 60% of scheduled training sessions (30-minute sessions three times per week) over the 8-week intervention period. Aim 2: Evaluate the preliminary effects of using home-based RAS (MR-001) on the change in walking endurance, gait speed, and balance, both between groups and within the waitlist control group. Hypothesis 2a: Participants will demonstrate clinically meaningful improvements in walking endurance (6-Minute Walk Test \[6minWT\]), gait speed (10-Meter Walk Test \[10MWT\]), and balance (Berg Balance Scale \[BBS\]; Timed Up and Go \[TUG\]; Functional Gait Assessment \[FGA\]; High Level Mobility Assessment Tool \[HiMAT\]) over the intervention period (eight weeks) compared to the waitlist control group. Hypothesis 2b: Participants within the waitlist control group will demonstrate clinically meaningful improvements in walking endurance, gait speed, and balance over the intervention period (eight weeks) compared to the eight-week waitlist control period. Aim 3: Evaluate the preliminary effects of using home-based RAS (MR-001) on the change in cognition, mood, fatigue, and impression of change, both between groups and within the waitlist control group while also assessing activity enjoyment in both groups after training. Hypothesis 3a: Participants will demonstrate clinically meaningful improvements in cognition (Brief Test of Adult Cognition by Telephone \[BTACT\]), mood (Brief symptom Inventory, \[BSI-18\]), fatigue (Modified Fatigue Impact Scale \[MFIS\]), and participant global impression of change (PGIC) over the intervention period (eight weeks) compared to the waitlist control group. Hypothesis 3b: Participants within the waitlist control group will demonstrate clinically meaningful improvements in cognition, mood, fatigue, and impression of change over the intervention period (eight weeks) compared to the eight-week waitlist control period. Hypothesis 3C: Participants in both groups will report greater enjoyment (PACES) with training using the MR-001. Exploratory Aim: Explore within participant changes in global cortical excitation (fNIRS) when walking overground with MR-001 compared to walking overground without MR-001 PARTICIPANT SELECTION CRITERIA Inclusion and exclusion criteria for participation are provided in Table 1 below: Table 1. Eligibility criteria for participants Inclusion Exclusion 1) have a moderate to severe TBI that required inpatient rehabilitation; 2) able to ambulate independently, with or without an assistive device.; 3) ambulating \< 1.4 m/s (comfortable walking speed) 4) demonstrate a reciprocal gait pattern; 5) ability to follow directions/standardized instructions; 6) minimum 18 years of age at consent. 1) Uncontrolled cardiopulmonary, metabolic, or infectious disorder; 2) history of orthopedic or additional neurological disorder that limited motor function before TBI; 3) any reason that, in the opinion of the study investigators or medical team, would interfere with completing the study protocol such as behavioral concerns; 4) uncontrolled seizure disorder; 5) participation in any gait interventional trials in last eight weeks; 6) hearing impairment limiting perception of rhythmic cues. 2. RECRUITMENT / SCREENING OF PARTICIPANTS We will enroll 33 (accounting for 10% attrition to yield a total sample size of 30) individuals with moderate to severe TBI who have residual gait impairment and have been discharged from inpatient rehabilitation. Participants will be recruited from eligible individuals who reside within 50 miles of the Denver metro area and/or report they can attend in-person training sessions and assessments when needed. Study personnel will recruit from Craig's outpatient rehabilitation program, our community fitness center (The PEAK), and the TBIMS database. 3. STUDY PROCEDURES After completing the informed consent process, all participants will complete baseline testing to include all outcome measures except fNIRS (cortical excitation) and PACES. After baseline testing, individuals will be randomized to immediate treatment or a waitlist control with randomization stratified by baseline walking speed of 0.8m/s. Immediate treatment group: Those randomized to immediate treatment will receive their MR-001 home device during baseline testing. Participants and caregivers will demonstrate competency in device set up and application before taking the system home. Individuals will be instructed to train with the in-home device for at least 30 minutes 3x/week X eight weeks. A research coordinator will check in with them weekly over the phone to problem solve any issues that may arise with training in the home throughout the course of the study and assess safety, and training adherence. Post treatment assessments will be completed after eight weeks of training. Waitlist Control Group: Those randomized to the waitlist control group will complete baseline testing (except fNIRS and PACES), upon study enrollment and at eight weeks before starting active treatment. Treatment intervention will mirror the immediate treatment group. All study therapists will receive training and demonstrate competency on the use of the MR-001 device. Two additional physical therapists masked to group allocation, will demonstrate competency assessing walking, balance, and cognitive outcome assessments. Participants in both groups will complete the PACES at the end of their treatment phase to assess enjoyment with training. Participants in both groups will complete two additional training sessions, one at the start of their treatment phase and another one at eight weeks (post-treatment) designed to assess cortical excitation using fNIRS. Using a randomized allocation, participants will complete 6 minutes of walking overground alone followed by 6 minutes of walking overground augmented with MR-001, or the reverse sequence (6 minutes of walking overground augmented with MR-001 followed by 6 minutes of walking overground alone). Following signal optimization and signal quality checks, baseline fNIRS data will be obtained during 60-second periods of quiet, static sitting and standing. Task dependent fNIRS data will be acquired during specific epochs - three at the early, middle, and late segments of both training conditions. Epochs will be generated using stimulus presentation software (PsychoPy2),45 which interfaces with Aurora via a lab-streaming layer. The research team member will initiate the start of each epoch which will generate a trial marker for early, middle, and late epochs in the fNIRS data file. These trial markers will be used to average brain responses for early, middle, and late activity; Complete procedures for fNIRS set up and data acquisition can be reviewed in Stephens et al.44. A fasting blood draw will be collected from participants during their first week of study participation and during the last week of their study participation. The sample will be processed and stored at -80C and shipped to University of Colorado Human Immunology and Immunotherapy Initiative (HI3) subdivision of Human Immune Monitoring Share Resource Core Lab (HIMSR) who will provide analysis of BDNF concentration levels that will be compared between the two timepoints of collection. Dr. Kimberly Jordan, PhD and colleagues with experience in BDNF analysis, validation, and measurement will be performing the analysis. Outcome Measurement Plan Walking Function and Balance The 10-Meter Walk Test (10MWT) is a standardized measure of gait speed. This test demonstrates excellent test-retest and interrater reliability, with interclass correlation coefficients (ICC) of 0.95 and 0.99 , respectively in individuals with TBI.25,26 An MCID of 0.15 meters/second in comfortable walking speed has been reported.27 The 6-Minute Walk Test (6MWT)28 measures distance walked (in meters) over 6 minutes as a sub-maximal test of aerobic capacity and endurance. This test shows excellent test re-test reliability (ICC = 0.94) for the TBI population.29 The MCID is 34.4 meters for the stroke population.30 The Berg Balance Scale (BBS)31 is a 14-task performance measure that objectively assesses static and dynamic balance and fall risk in adults. Item-level scores range from 0-4 with higher total scores indicating better balance. BBS has excellent test-retest reliability (ICC = 0.99) in TBI and a standard error of measurement of 1.65.32 The Timed Up and Go (TUG) is designed to quickly assess mobility, balance, walking ability, and fall risk. PSYCHOMETRICS and REF\*\* The Functional Gait Assessment (FGA) is a ten-task dynamic balance assessment with item-level scores ranging from 0-3 and higher scores indicating greater balance. The FGA has excellent intra-rater (ICC = 0.94) and interrater reliability (ICC = 0.97).33 The High-Level Mobility Assessment Tool (HiMAT) is a validated performance-based measure of high-level mobility designed to assess advanced gait and balance skills in individuals with TBI. The HiMAT has demonstrated strong reliability, validity, and responsiveness in the TBI populatiosn.34,35 Cognition The Brief Test of Adult Cognition by Telephone (BTACT) assesses multiple cognitive dimensions central for effective functioning across adulthood: episodic memory, working memory, reasoning, verbal fluency, and executive function. Although originally designed for telephone use, evidence establishing construct validity has been published in a subsample of individuals with TBI who were tested in person.36-37 Global Impression of Change Participant Global Impression of Change (PGIC) scales have been recommended for use as a core outcome measure in rehabilitation studies38-40 to improve the applicability of information from clinical trials to clinical practice. We will utilize a 7-point scale: "With respect to walking and balance function, how would you describe yourself compared to before the treatment began?" with -3 representing very much worse, 0 representing unchanged, and +3 representing very much better. Activity Enjoyment The Physical Activity Enjoyment Scale (PACES) is a validated self-report measure designed to assess an individual's affective enjoyment of physical activity, a key determinant of adherence and sustained engagement. The original 18-item scale and its validated short forms demonstrate strong internal consistency, construct validity, and test-retest reliability across adult and clinical populations and have been widely used in rehabilitation and exercise research.41 Mood/Fatigue The Brief Symptom Inventory-18 (BSI 18) is a brief self-report measure of psychological distress assessing somatization, depression, and anxiety, with a Global Severity Index (GSI) as the primary outcome. It has low burden (\~3-5 minutes) and strong psychometric support. Across populations, the GSI demonstrates excellent reliability (α≈0.91), with acceptable subscale reliability. In traumatic brain injury (TBI), the GSI shows strong internal consistency (α=0.84-0.91) and good construct validity.461 The Modified Fatigue Impact Scale (MFIS) has 21 items rated on a 5-point scale with total scores ranging from 0-84 (higher score indicates greater fatigue). The mFIS assesses the perceived impact of fatigue on physical, cognitive and psychosocial functioning over the past 4 weeks. It has excellent internal consistency, good test-rest reliability and a minimally important difference of 4.00.2 This pencil and paper survey takes 2-10 minutes to complete. Sleep We will capture a self-report measurement of sleep using the Pittsburgh Sleep Quality Index (PSQI), a 19-item self-report questionnaire, to assess sleep quality and disturbance.44 The specific aspects of sleep assessed include: subjective sleep quality, efficiency, and latency, use of sleep medications, sleep disturbances, and daytime dysfunction. PSQI scores range from 0 to 21 points (0 = no difficulty in any area of sleep; 21 = severe difficulties in all areas of sleep). Scores \>5 are indicative of poor sleep quality.45 The Insomnia Severity Index (ISI) is a seven-item questionnaire validated to measure the nature and symptoms of their sleep problems.46 These pencil and paper surveys take 5 minutes each to complete and will be administered at the three assessment time points (baseline, post treatment, and follow-up). Biomarkers Blood-based biomarker BDNF will be collected from participants via a fasting blood draw on the day of their first training session and the day after their last training session. The sample will be processed and stored at -80C and shipped to University of Colorado Human Immunology and Immunotherapy Initiative (HI3) subdivision of Human Immune Monitoring Share Resource Core Lab (HIMSR) who will provide analysis of BDNF concentration levels that will be compared between the two timepoints of collection. Dr. Kimberly Jordan, PhD and colleagues with experience in BDNF analysis, validation, and measurement will be performing the analysis. Cortical Excitation Functional near-infrared spectroscopy (fNIRS) is a neuroimaging technique that uses near-infrared light to evaluate changes in brain activity via proxy measures of oxygenated (HbO) and deoxygenated hemoglobin (HbR). Specifically, following neural activity, oxygen is pulled from hemoglobin resulting in an immediate increase of HbR. Next, the reduction in oxygen (and glucose) elicits an increase in localized cerebral blood flow (CBF), a mechanism called neurovascular coupling. As CBF increases, a greater concentration of HbO can be detected. Both HbR and HbO have optical properties, meaning that their concentrations can be measured with near-infrared light. Thus, fNIRS is able to detect localized neural activity by evaluating HbR and HbO, and increases in HbO concentrations reflect site-specific increases in neural activity.43 Our team has recently used fNIRS to measure cortical activation in different walking conditions44 and we will use fNIRS to evaluate differences in brain activation in a single session when walking with and without MR-001. Data Analysis Plan Aim 1: Safety, feasibility, and adherence for the active intervention MR-001 walking will be analyzed descriptively. Frequency counts and percentages will be used to summarize the number of adverse events, and participant's session completion and dropout rate. Aim 2: The 10MWT, 6minWT, FGA, BBS, and HiMAT will be described using means, standard deviations (SDs) and summary statistics (Minimum, 25th percentile, Median, 75th percentile, Maximum) for the overall sample and by treatment group. 1. Between group differences will be evaluated using statistical testing, effect sizes (ESs), and clinically meaningful change (Baseline - Post-treatment). Statistical testing will be performed using 2-sided t-tests and Mann-Whitney U tests for each measure. ESs will be calculated using Cohen's D. Clinically meaningful change will be determined using the MCIDs (Minimally Clinically Important Differences) for each respective measure (10MWT = 0.15m/s, 6minWT = 34.4m, BBS = 6 points, FGA = 4 points). These results will be summarized using frequency counts and percentages. 2. The within group differences will be evaluated similarly to 2a. Calculations for ESs and clinically meaningful change will be identical. Statistical testing will be performed using Paired t-tests and Wilcoxon Signed-Rank tests. Aim 3: The BTACT, BSI-18, MFIS, and PGIC will be described using means, SDs and summary statistics for the overall sample and by treatment group. The PGIC will also be summarized using frequency counts and percentages. 1. Between group differences will be evaluated using statistical testing, ESs, and clinically meaningful change. Statistical testing will be performed using 2-sided t-tests and Mann-Whitney U tests for the continuous measures, and chi-square tests for the categorical. ESs will be calculated using Cohen's D. These results will be summarized using frequency counts and percentages. 2. The within group differences will be evaluated similarly to 3a. Calculations for ESs and clinically meaningful change will be identical. Statistical testing for the continuous measures will be performed using Paired t-tests and Wilcoxon Signed-Rank tests. C) Post-treatment PACES scores will be summarized with means and SDs. Exploratory Aim: fNIRS processing steps of conversion and spatial registration will be completed automatically by Satori software. Specifically, in the conversion step, the raw light intensity data will be converted to optical density values and then converted to total hemoglobin (HbT), HbO and HbR values using the Modified Beer-Lambert Law. Then, all data will be spatially registered to the fNIRS head probe and visually inspected. Following these automated steps, the data will be segmented into 'events,' which represent the three epochs of data collection and their durations. Next, temporal pre-processing steps of motion artifact detection and correction will be completed with a spike removal procedure that uses the Satori default parameters (10 interactions, 5s lag, 3.5 threshold, 0.5 influence). Given that participants will be moving during data acquisition, this step - motion artifact detection is the most important step for assessing success in fNIRS data analysis. Dr. Stephens will compare original data files to analyzed data files to assess any instances of un-corrected motion artifacts. Successful data analysis includes successful conversion, spatial registration, and motion artifact detection and correction for all motion artifacts. Within subject global cortical excitation will be compared between the two conditions.