Trial design

This protocol is a single-blinded, randomized controlled trial. The study received approval from a joint Oregon Health & Science University and Veterans Affairs Portland Health Care System Institutional Review Board (eIRB #25890) and will be executed in compliance with the Declaration of Helsinki. The trial is registered on ClinicalTrials.gov (Identifier: NCT06381674; date of registration: 2024-04-04). This protocol adheres to the Standard Protocol Items: Recommendations for Interventional Trials schedule (Fig 1). The study design is illustrated in Fig 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Participant timeline: Schedule of enrollment, interventions, and assessments.

https://doi.org/10.1371/journal.pone.0340867.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Study design.

https://doi.org/10.1371/journal.pone.0340867.g002

Participants and eligibility criteria

This study will involve 100 participants experiencing persistent symptoms from mTBI and 40 healthy controls that are active-duty service members. Participants will be recruited between June 2024 to January 2027. The data collection will be expected to be completed by April 2027, and the results are expected January 2028. Testing and rehabilitation (Aim I & II) for civilians will be conducted at Oregon Health & Science University (OHSU), Veterans Affairs (VA) Portland Health Care System, and University of Utah (UU). Home monitoring with wearable sensors (Aim III) will include a subset of participants at OHSU and UU as well as active-duty military service members at Joint Base San Antonio- Fort Sam Houston (FSH), in San Antonio, Texas.

Recruited participants must: 1) have been diagnosed with mTBI according to VA and Department of Defense criteria [34,35], 2) be 18−60 years old, 3) be capable of standing unassisted for at least 10 minutes, 4) be outside of the acute stage (> 2 weeks post-mTBI) but within 6 months of their most recent mTBI and still reporting symptoms, 5) demonstrate some level of measurable V/O deficits based on the Concussion Clinical Profile Screening (details provided below), 6) have sufficient vision (corrected or uncorrected) for unassisted reading and performance of daily personal tasks and independent community ambulation, and 7) have adequate hearing (without amplification) for engaging in close-range personal or telephone conversations. For Aim III, active-duty military service members must have no recent history (< 3 years) of mTBI and no residual symptoms from any previous mTBI.

Exclusion criteria for all aims include participants that: 1) have a current or past history of any other injury, medical, or neurological illness that could potentially explain balance or vision deficits (e.g., central nervous system disease, stroke, epilepsy, greater than mild TBI, Meniere’s disease, bilateral vestibular loss, recent lower extremity or spine orthopedic injury requiring activity modifications (or duty limitations for service members)), 2) meet criteria for moderate to severe substance-use disorder within the past month as outlined by Diagnostic and Statistical Manual of Mental Disorders-V, 3) display behavior that would significantly interfere with validity of data collection or safety during the study, 4) are in significant pain during evaluation (> 7/10 by patient subjective report), 5) are pregnant (due to balance considerations), 6) have been hospitalized for any brain injuries (separate from emergency department visits), 7) have significant joint pain or recent musculoskeletal injury that limits walking or mobility, 8) have had any major surgeries in the past year or amputation, 9) use an assistive device, 10) are unable to stand barefoot, 11) currently are in or have received rehabilitation services or physical therapy for symptoms related to their mTBI within 1 month of the screening date.

Recruitment, randomization, and allocation

Individuals with mTBI will primarily be recruited from the OHSU Concussion Clinic, OHSU Hospital System, the UU Concussion and Brain Injury Program, and the UU Health System. Active-duty military service members will be recruited from FSH, primarily from personnel at the U.S. Army Medical Center of Excellence. Recruitment efforts will include word-of-mouth referrals, distribution of flyers and posters within these facilities, and briefings for assigned personnel.

Participants with mTBI will be randomized into multimodal rehabilitation intervention consisting of either 1) vestibular rehabilitation therapy or 2) vestibular rehabilitation therapy with augmented real-time biofeedback using wearable sensors. Following the baseline assessments, participants will be randomized with an automated Research Electronic Data Capture system database algorithm that is site controlled and includes phase of recovery based on time since injury and V/O HI or LO classification at baseline. The allocation sequence is concealed until participants are enrolled and assigned to an intervention. Given the significance of V/O symptoms in our hypotheses, we will employ a stratified block randomization method. This approach will stratify based on V/O HI and V/O LO to ensure that both rehabilitation interventions receive an approximately equal number of participants with similar severity of V/O symptoms. Block randomization will assist in ensuring balanced allocation to the two intervention groups, with block sizes (2, 4, 6, or 8) varied to mitigate potential selection bias.

Blinding

Participants will remain unaware of the favored intervention. The researchers delivering the intervention will be aware of participants’ assigned intervention. The researchers conducting the pre- and post-tests and analyzing the data will be unaware of the group assignments, ensuring blinding throughout the study.

Assessment procedures

The pre-screening for inclusion and exclusion criteria will be conducted via phone. All eligible participants with mTBI will come into either OHSU or UU for the informed consent process. An investigator will verbally explain the consent forms to the participants and allow ample time to read through the consent forms (S2). Participants will confirm their consent to participate by signing the consent forms.

Two days of baseline testing (approximately six total hours) will occur in the Balance Disorder Laboratory at OHSU, National Center for Rehabilitative Auditory Research at VA Portland Health Care System, Neuromechanics and Applied Locomotion Laboratory, and Balance and Mobility Clinic at UU. Participants will complete validated questionnaires; motor assessments of standing and reactive balance, gait, and turning; standard vestibular and ocular-motor tests; and a computerized neurocognitive test (see Tables 2 and 3 for detailed lists of tests). Following baseline assessments, a convenience sample of 50 individuals with mTBI, with approximately equal number of participants with V/O HI and LO, will undergo 7 days of daily-life mobility monitoring. Subsequently, participants will receive a 6-week intervention, detailed below. Post-testing will be conducted after completing the intervention sessions. Additionally, 40 active-duty service members will undergo 7 days of daily-life mobility monitoring at FSH. The overview of the study is shown in Fig 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Patient-reported subjective questionnaires.

https://doi.org/10.1371/journal.pone.0340867.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Clinical and instrumented assessments.

https://doi.org/10.1371/journal.pone.0340867.t003

Daily-life mobility monitoring

Participants will utilize Opal Instrumented Socks (APDM Wearable Technologies, a Clario company, USA), consisting of a thin elastic fabric embedded with inertial sensors, to monitor their mobility in daily activities (Fig 3). The sensors will be positioned on the dorsum of the foot with the battery situated above the lateral malleolus [30,31]. Additionally, participants will wear an Opal sensor on a belt, positioned over the lumbar area. These sensors will be worn throughout the day for approximately 8–10 hours over the course of 7 days, with nightly charging. At the end of the 7-day period, participants will return the sensors and the journal, which records the times that the sensors were put on and remove, either by mail or in-person. The quality and quantity of turns measured by wearable sensors will be calculated [29–31].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Instrumented socks with inertial sensors.

(A) Inertial sensors placed on the top of foot and battery on the side (B) Inertial sensors on the posterior side of the waist, near the lumbar region. The individual shown in Fig 3 has given written informed consent (as outlined in PLOS consent form) to publish these images.

https://doi.org/10.1371/journal.pone.0340867.g003

Intervention

The intervention will consist of twelve sessions, each lasting approximately 60 minutes, delivered twice per week for 6 weeks. The intervention will be administered and/or supervised by a licensed physical therapist. The physical therapy program is based on the Clinical Practice Guideline recommendations [2,76], which advocate for a multimodal physical therapy approach tailored to the participant’s impairments and symptoms and progressed to their tolerance. During the first rehabilitation session, participants will be screened for benign paroxysmal positional vertigo with Dix-Hallpike and supine roll tests, and positive findings will be treated according to the benign paroxysmal positional vertigo Clinical Practice Guideline recommendations [77]. Additionally, participants will be assessed for cervical dysfunction with the Numeric Pain Rating Scale, Active Range of Motion (ROM) testing, Head-Neck Differentiation Test [78], and the Cervical Flexion Rotation Test [79]. If cervical dysfunction is present, it will be treated with a home exercise program. Our intervention includes six exercise categories: ocular-motor, standing balance, perturbations, gaze stabilization, visual motion sensitivity, and dynamic balance related to walking tasks (Table 4). Exercises will be advanced in each category through changes in surface conditions (i.e., firm, foam, Both Sides Utilized (BOSU) [BOSU® Fitness, LLC, USA], or rocker board), visual conditions (i.e., blank background, high contrast checkerboard, or moving lights), and the addition of cognitive or motor dual-tasking (Table 4) [80].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Multimodal rehabilitation sessions.

https://doi.org/10.1371/journal.pone.0340867.t004

The biofeedback group will receive biofeedback on various parameters including head velocity or ROM, sternum velocity or ROM, and postural sway (Fig 4). Target parameters for each exercise will be based on participants’ performance and previously collected data from healthy controls as target zones. Biofeedback will be provided through audio channels and can be adjusted to provide either continuous or intermittent feedback to maximize motor learning. The standard of care group will complete the same exercise protocol while wearing the sensors but will not receive any biofeedback. For both study groups, surface progression will be standardized so that all participants perform exercises on firm ground for sessions 1–4, foam for sessions 5–8, and BOSU or rocker board for sessions 9–12. If for any reason a participant cannot tolerate the progression, they will remain at the previous level until they can be safely progressed. Visual challenges will be individualized to the participant for each exercise for both groups.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Exercise stations of the multimodal rehabilitation intervention.

The individuals shown in Fig 4 have given written informed consent (as outlined in PLOS consent form) to publish these images.

https://doi.org/10.1371/journal.pone.0340867.g004

All participants will receive a home exercise program (HEP) consisting of three exercises. These exercises will be selected from five categories: ocular-motor, static balance, gaze stabilization, visual motion sensitivity, and dynamic balance. The selection will include exercises tailored to individual needs and will be advanced based on the participant’s tolerance, including changes in surface condition and visual stimuli. Participants who score ≥2 on the cervical modifier on the pre-rehabilitation intervention CP Screen or those for whom the physical therapist determines that the cervical spine limits their ability to perform the intervention will be educated on an additional cervical exercise component for their HEP. Compliance with the HEP will be tracked through journals collected at each session.

Community based participatory research

Our study team works with a Community Advisory Board (CAB) including organizational partners, clinicians, and lived experience consultants. This board will ensure a Community-Based Participatory Research approach throughout the planning, design, execution, analysis and dissemination of this research. The input from all members will come from moderated discussions (quarterly) and will be included during the entirety of the study. Currently, the CAB for this study includes physical therapists, physical therapy students, athletic trainers, occupational therapists, Veterans, and lived experience consultants (i.e., individuals who have sustained a mTBI in the past). The inputs from our CAB have guided development of the rehabilitation program, treatment tolerance expectations with the planned rehabilitation approach, symptom education and mitigation strategies, techniques to address barriers to care, and additional topics for this project.

Safety issues

Any adverse events occurring during either physical therapy sessions or data collection will be documented and reported to the Institutional Review Board.

Data management and monitoring

Team members will manage data entry tasks, inputting data into both Research Electronic Data Capture system and the OHSU Balance Disorders Laboratory database. Physical copies of records will be securely stored in a locked office cabinet at OHSU and UU. Instrumented de-identified data will be gathered on a laptop computer with password protection and data encryption at OHSU and UU. After completing the protocol, this data will be uploaded to an OHSU secure server, where the Balance Disorders database resides. To ensure accuracy, all data will undergo quality-checking quarterly by an individual not directly involved in the study’s day-to-day operations.

The OHSU Institutional Review Board conducts annual reviews to ensure participant safety and data integrity, including monitoring of adverse events, protocol deviations, and general procedures. Quarterly and annual reports are submitted to the Department of Defense, providing specific updates on study proceedings and progress on the scope of work. All annual reviews are also submitted to Office of Human Research Oversight, to ensure that important protocol modifications are communicated to relevant parties.

Stopping the trial

The study will be stopped prior to its completion if (1) the intervention is associated with adverse effects that question the safety of the intervention, (2) new information becomes available during the trial that necessitates stopping the trial, or (3) there are other situations that warrant stopping the trial.

Sample size calculation

Sample size calculations for Aim I and II were based on pilot data on participants with mTBI who received rehabilitation during the development phase of the biofeedback system (Grant No. W81XWH-17-1-0424). A group of 43 participants with mTBI were randomized into standard vestibular rehabilitation therapy or vestibular rehabilitation therapy while using wearable sensors at home (unsupervised). The wearable sensors measured movement during home exercises and that information was discussed with participants during subsequent therapy sessions. For those using wearable sensors, feedback was not provided in real-time but was instead delivered during the subsequent physical therapy session after the sensors had been used during home exercise. Participants rated their perception of change using the PGIC at the end of rehabilitation. In the pilot study, participants who provided sensor data had a mean PGIC score 0.5 ± 1.3 points higher than the standard vestibular rehabilitation group. For Aim I, we expect a larger mean difference (i.e., 0.75) because the sensor group will receive immediate feedback. Also, we anticipate a smaller standard deviation because of more consistent therapeutic results (i.e., SD = 1). We will recruit 50 participants per group. After accounting for an anticipated dropout rate of 25%, we will have 38 participants per group. This sample size could achieve 90% power based on the two-sample t-test to detect the difference of 0.75 ± 1.00 using a two-sided test, alpha at 0.05.

For Aim III, we used pilot data collected from eight participants on objective daily life mobility metrics. Participants were categorized as V/O HI (n = 3) or as V/O LO (n = 5) based on their vestibular testing. From the pilot data, mean differences and common standard deviation between the two groups for turn angle was 4.48 ± 4.14^o, turn angle variability was 0.018 ± 0.016, and peak turn velocity was 7.95 ± 4.47^o/s. Additionally, these three metrics had very strong effect sizes (1.1, 1.1, and 1.8, respectively). Therefore, we conservatively determined the sample size calculation with an effect size of 0.90 when comparing two groups. To gain at least 80% power to detect group difference, we need at least 21 subjects for each group (a total of 42 participants), alpha at 0.05. Given an estimated completion rate greater than 85%, we plan to enroll 25 subjects per group (n = 50).

Statistical analysis

For Aim I, responsiveness to rehabilitation will be determined by the participants’ score on the PGIC after rehabilitation – the primary outcome. We will first conduct a Mann-Whitney U test to assess group differences. Next, PGIC scores will be dichotomized by “responder” or “non-responder”. We will classify participants who score 6 or above on the PGIC as “responders” and those who score 5 or below as “non-responders” [42]. To compare the groups, the Chi-square test will be used to compare responders vs. non-responders in PGIC scores between groups. Then we will conduct a logistic regression on responder vs. non-responder as the binary outcome to assess differences between groups accounting for covariates (e.g., age, sex, Post-Traumatic Stress Disorder, etc.). Additionally, we will conduct ordinal logistic regression on PGIC score levels to examine group differences accounting for covariate effects. Secondary outcomes (e.g., patient-reported symptoms, clinical/instrumented objective assessments) will be measured at baseline and post-rehabilitation, and pre-post change will be calculated for each participant. To compare the differences in pre-post changes in secondary outcomes between the real-time biofeedback group and the standard vestibular rehabilitation group, we will conduct a two-sample t-test or Mann-Whitney U test for each secondary outcome. We will also employ mixed effects models to examine the differences in change in secondary outcomes between the real-time biofeedback and standard vestibular rehabilitation group. Two fixed effects will be included in the model: 1) group effect (real-time biofeedback vs. standard vestibular rehabilitation) and 2) time effect (baseline and post-rehabilitation). The interaction between group and time (group x time) will be included to assess the differences in pre-post change between groups. Random intercepts will be included to account for the clustering effect within subjects over time, and covariates will be included in the fully adjusted model. We will use an inverse probability weighting approach to account for participant attrition in the study and perform sensitivity analyses for missing data [81].

For Aim II, we first will examine the distribution of PGIC scores and the proportion of responders vs. non-responders in V/O HI group and V/O LO group, regardless of randomization group assignment. Then, we will assess the PGIC as an ordinal outcome variable and test group differences (V/O HI vs. V/O LO) in responsiveness using ordinal logistic regression. Additionally, we will conduct a logistic regression on responder vs. non-responder as the binary outcome to assess differences between V/O HI group and V/O LO group. Both regression models will allow us to control for covariates (e.g., age, sex, days since injury, intervention arm) in the models. We will use the data to obtain estimates to inform future studies and power calculations. Finally, to explore the strength of associations between the patient-reported and clinical/instrumented V/O measures, we will measure the associations using Pearson correlation coefficients for continuous variables and Spearman correlation coefficients for ordinal variables.

For Aim III, participants will be grouped according to their level of severity of V/O symptoms (V/O HI or V/O LO symptoms), regardless of intervention group. To test whether daily life mobility measures and quantitative activity measures differ between V/O HI and V/O LO groups, we will test the metrics for group differences using a two-sample t-test or a Mann-Whitney U test. In addition, we will conduct subsequent analyses to compare metrics between the two groups using the multivariate generalized linear modeling approach. The model will be constructed for each metric separately and will be adjusted for covariates. Other quality and quantity metrics will be examined using the same statistical approach. Data from the healthy military normative group will be collected and examined to provide preliminary data for our next studies exploring daily life mobility in those cleared to return-to-duty after mTBI. We will provide both statistical and graphical evaluation for the distributions of turn quality metrics and quantitative activity measures. The descriptive statistics of these metrics will be reported as mean (standard deviation).