Virtual reality effects on dynamic stability of gait in stroke survivors: a study protocol of a randomised controlled trial to assess the role of user experience

Abstract

Introduction

In recent years, virtual reality (VR) training has gained momentum in stroke rehabilitation for reducing falls by enhancing balance and stability.

Objectives

VR interventions offer superior immersion, interactivity and creative engagement compared with conventional therapies. However, the precise mechanisms by which VR influences dynamic stability remain unclear. Moreover, there is currently limited information regarding user experience (UX) in this context.

Methods and analysis

This study protocol outlines a randomised controlled trial designed to evaluate the effects of an immersive VR intervention on the dynamic stability of gait parameters in stroke survivors, incorporating an assessment of UX. Two groups of subjects will be enrolled in this study, including patients with a stroke within the previous 6–12 months who will receive (1) 3 weeks of immersive VR training (intervention) or (2) conventional rehabilitation (control).

The primary outcome of the study will be dynamic stability measures and the UX. The secondary outcomes include anxiety and depression, fatigue, functional mobility, functional walking ability, static balance, fear of falling and adverse effects. All participants will be assessed at baseline, after 3 weeks and after 2 months of the initiation of the study.

Ethics and dissemination

The study ethics were approved by the ethical committee of Tehran University of Medical Sciences (registration No IR.TUMS.NI.REC.1401.091, registered on 8 January 2023). Written consent is going to be taken from all of the participants. The study adheres to the principles outlined in the Declaration of Helsinki. Dissemination of study findings will occur through peer-reviewed publications and knowledge translation activities targeting clinicians, researchers and rehabilitation professionals. Data generated from this study will be anonymised and curated in accordance with institutional and national data protection standards

Trial registration number

IRCT20231031059916N1.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study employs a randomised controlled trial design with concealed allocation, which strengthens causal inference about the effects of virtual reality (VR) on dynamic gait stability compared with conventional therapy.

• The inclusion of validated user-experience questionnaires alongside qualitative semi-structured interviews provides a comprehensive mixed-methods assessment rarely reported in previous VR stroke-rehabilitation trials.

• All VR training sessions are fully supervised and individually tailored in real-time by the same experienced physiotherapist, which ensures optimal exercise intensity and progression for each participant.

• This study is conducted at a single centre (Imam Khomeini Hospital, Tehran), which may reduce the generalisability of findings to other regions or settings; a future multicentre study involving urban and rural sites would address this issue.

• The absence of physiotherapist blinding and the relatively short 3-week intervention with 2-month follow-up restrict conclusions regarding long-term efficacy and retention of gains; extending follow-up to at least 6–12 months in subsequent trials is recommended.

Background

Walking impairment is one of the most prevalent consequences of stroke, affecting about 80% of survivors.¹ This impairment arises from muscle weakness, pain, spasticity and poor balance, which together contribute to an increased risk of falls.² In fact, patients who had a stroke experience falls and related impairments at a higher rate than the general elderly population, leading to significant physical and psychological challenges throughout their post-stroke recovery.³ These challenges can include fractures, soft tissue injuries, immobility and a persistent fear of falling, all of which substantially diminish quality of life.⁴ Consequently, addressing walking impairments and implementing effective fall prevention strategies are among the central goals in stroke rehabilitation.

Several studies indicate that immersive virtual reality (VR) interventions can help reduce falls in individuals with walking impairments by improving stability and balance.^{5 6} In recent years, VR-based interventions have emerged as either a substitute or an adjunct to conventional stroke rehabilitation therapies.⁷ These interventions facilitate repeated practice, deliver multisensory feedback and allow for the progressive increase in task difficulty while maintaining task specificity for individual patients.⁸ Additionally, the gamified elements inherent in VR have been shown to enhance both the therapeutic experience and patient adherence.⁹ However, previous research has primarily focused on using clinical tests or kinematic variability to predict fall risk in patients who had a stroke,¹⁰ and there is still limited understanding of how individuals respond to subtle disturbances throughout the different phases of walking, recognising walking as a dynamic task in which earlier strides can influence subsequent ones.

The present protocol aims to answer the main question, ‘What is the effect of immersive VR on dynamic stability in stroke survivors?’ Additionally, this study seeks to evaluate participants’ motivation and adherence to the treatment. Recent findings indicate that while increased motivation to engage with technology does not inherently confirm its efficacy, the quality of user experience (UX) may have a more significant impact on intervention outcomes.¹¹ Accordingly, the secondary research question is: ‘How does patient experience influence the outcomes in using VR technology?’

To address these questions, this study has established a rigorous protocol with quantifiable measures to compare the impact of immersive VR against conventional functional rehabilitation on dynamic stability and patient experience in stroke survivors. We will assess outcomes at both 3 weeks and 2 months after treatment initiation. In addition to dynamic stability, the study will evaluate the association between the VR intervention and various outcome measures, including anxiety, depression, fatigue, functional mobility, functional walking ability, static balance, fear of falling and any adverse effects. Notably, as far as we know, this is the first study of its kind to examine how immersive VR influences dynamic stability and user experience in stroke survivors, offering a unique opportunity to test the feasibility and integration of immersive VR in neurorehabilitation.

Methods

Recruitment strategy and timeline

The present study is a single-blinded, randomised clinical trial to compare the effects of VR training in two parallel groups. The study ethics was approved by the ethical committee of Tehran University of Medical Sciences (registration No IR.TUMS.NI.REC.1401.091, registered on 8 January 2023), and the study is registered in the Iranian Registry of Clinical Trials (IRCT20231031059916N1).

Participants will be recruited from patients who had a stroke referred to Imam Khomeini Hospital in Tehran, Iran, from January 2026 to April 2026. Patients who experienced a stroke between 6 and 12 months before enrolment will be randomly assigned to one of two training groups: one receiving an immersive VR intervention and the other undergoing conventional functional rehabilitation. All participants will receive 3 weeks of treatment. A flowchart of the experimental process is provided in figure 1, and a timeline following the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) are presented in figure 2.

Figure 1. Flowchart of the study procedure. This study will recruit participants from patients referred to a tertiary medical centre who had experienced at least one stroke within 6–12 months before enrolment. The eligibility criteria will be checked, and informed consent will be obtained before the programmes. Eligible participants will be randomly allocated to the intervention group (undergoing immersive virtual reality training) or control group (undergoing conventional rehabilitation). Both rehabilitation programmes last for 3 weeks, 5 days a week. Assessment of the participants regarding the clinical psychological variables will be undertaken at three time points: baseline (before the rehabilitation programme), at 3 weeks (after programme completion) and at 2 months (follow-up time point). VR, virtual reality.

Figure 2. Participant timeline recommended by Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT). Each X shows the demonstrate the phase that each specific action will be performed in the trial.

Participants

Participants eligible for this study must meet the following inclusion criteria: they must be diagnosed with an ischaemic stroke according to the WHO definition,¹² be between 30 and 70 years of age and be within 6–12 months of their stroke. Additionally, they should have the ability to walk unassisted for at least 10 m (Functional Ambulation Categories ≥4) and sufficient cognitive and communication ability, defined as Mini-Mental State Examination (MMSE) score ≥24/30, able to understand and speak Persian well enough to follow verbal instructions. On the contrary, individuals will be excluded if they are pregnant, have a history of traumatic brain injury, dementia or other neurological diseases affecting cognition or motor function, MMSE <24, severe aphasia (defined as inability to follow two-step commands), suffer from cardiac arrhythmias or require a pacemaker, have a major vascular disease, limb amputation, reduced consciousness or psychiatric disorder requiring regular medication that affects alertness or cooperation, or have severe visual impairments (corrected visual acuity worse than 100/200 or hemianopia with neglect), any orthopaedic or pain condition that limits walking ability.

Sample size calculation

The sample size was calculated using G*Power software based on the Timed Up and Go (TUG) test outcome measure. The calculation was informed by a previous study by Boissoneault et al.¹³ Based on the reported data, an effect size of 0.77, corresponding to a minimum clinically relevant difference of approximately 3.16 s, was considered for the initial power estimation. Using a one-tailed independent-samples t-test with a significance level of α=0.05 and a statistical power of 80% (β=0.20), a planned sample size of 22 participants per group was determined. The obtained sample size will be re-estimated with a dropout rate of 20%.

Setting

Rehabilitation training and evaluation of this study will be conducted in the Rehabilitation Clinic of the Sports Medicine Research Center, Tehran, Iran.

Allocation and masking

Eligible participants will be randomly assigned to one of the two training groups using a computer-generated sequence created by an independent statistician not involved in the study, ensuring equal group allocation. Allocation concealment is done by sequentially numbered, sealed envelopes prepared by the independent statistician. Prior to enrolment, all eligible individuals will be thoroughly informed about the study objectives and will only be included if they voluntarily sign the informed consent form (online supplementary material 1) in adherence with the Declaration of Helsinki. Then, the physiotherapist will open the next consecutively numbered envelope in the presence of the participant to reveal the group assignment. Although the non-pharmacological nature of the interventions prevents blinding of participants and the rehabilitation trainer, outcome assessors and the data analyst will remain blinded to the group allocation. Additionally, participants will be instructed not to disclose their assigned intervention to the attending physician or the outcome assessors.

Unblinding can be carried out by contacting the lead investigator, who has access to the randomisation list, in the event of major adverse events or medical necessity; the date, reason and person making the request will be recorded. There are no scheduled regular interim assessments that would need unblinding.

Interventions

The study interventions adhere to the TIDieR (Template for Intervention Description and Replication) guidelines.¹⁴ To boost adherence, patients will receive session reminders beforehand and are encouraged to reach out to the trial team at any time during treatment. The protocol prohibits the receipt of any additional forms of treatment during the trial. If a participant requests a change in their allocated intervention or receives another treatment, adjustments will be made in accordance with the protocol. Before initiating the study, all subjects were trained to work with VR. The team of experts and examiners also had vast experience working with patients who had a stroke.

VR system

In this study, we will use the Meta Quest 2 VR head-mounted display (HMD) device (Meta Platforms, Menlo Park, California, USA) as a VR device. The weight and dimensions of the HMD are 503 g and 22.4×45.0 cm, respectively. It is a self-contained device with a rechargeable 3640 mAh lithium-ion battery. The main technical features of this device include an LCD of 1832×1920 per eye with a display rate of up to 90 Hz, a Qualcomm Snapdragon XR2 as CPU, 6 GB of RAM, 128 GB of memory, four infrared cameras and built-in speakers. The HMD contains a TDK ICM-42688-P (InvenSense, San Jose, California, USA) inertial measurement unit inside the headset, which tracks 6 df to collect head movements.

In the VR training group, participants will wear an HMD to engage with virtual objects across varied scenarios. Specifically, they will experience four distinct VR-assisted lower limb walking training protocols. These scenarios have been adapted from commercially available models specifically designed for walking rehabilitation, ensuring effective and immersive training to enhance dynamic stability and functional mobility.¹⁵

Safety

Ensuring the safety and well-being of our participants is our top priority throughout this study. In both training groups, each participant will be closely monitored by a dedicated team comprising a physiotherapist, a sports medicine physician and a neurologist. Participants are encouraged to promptly communicate any discomfort or adverse effects they experience, and they are fully informed of their right to withdraw from the study at any time, without any consequences. To maintain rigorous safety standards, the effects of the VR-based rehabilitation will be periodically evaluated, and any adverse events will be immediately reported to the ethics committee. The study protocol will be reviewed and modified as needed to minimise risks further; thus, little harm is anticipated from participation in this trial.

VR training group

The VR training group will engage in 30 min of VR-assisted gait rehabilitation daily for 3 weeks (5 days per week). Participants will receive instructions on the option-selection process, which involves fixating their gaze on a desired option for a brief duration. Once the HMD is donned, participants can freely move their heads to explore the virtual environment and observe features such as the ground and their own feet. Performance will be quantified through training scores based on the successful completion of specified tasks across various training stages. The VR game scenarios contain six progressive task categories with three difficulty levels (easy, medium and hard). The virtual training scenarios can be customised to accommodate individual needs and preferences, and the difficulty level will be adaptively adjusted according to each participant’s lower extremity motor abilities, providing a tailored and personalised VR experience:

Warm-up/cool-down and initial gait assessment (3–5 min): Calm circular or linear path in a quiet garden or a quiet mall environment with minimal distractions.

9-grid training (3 min): 3×3 grid stepping task to improve weight-shift and reactive balance (online supplementary material 2).

Bridge step (3–5 min): The player walks across a virtual bridge made of short, simple sections; after successfully crossing each section, the path length and challenges (gentle slopes, gaps between planks) increase slightly.

Stair climbing training: This game targets leg strength, endurance and balance through repeated stair-stepping activities. User navigate stairs of varying heights or patterns to improve functional mobility and lower-limb control (online supplementary material 3).

Virtual environments include:

• A large calming garden with flowers, small streams, bridges and natural sounds.

• Quiet indoor hospital corridor or laboratory.

• Low-stimulus city park or shopping mall.

Immersion is enhanced by high visual fidelity, spatial audio, real-time visual/auditory feedback (glowing foot markers, voice encouragement) and immediate scoring/rewards (coins, hearts, collectibles). Difficulty progression (speed, obstacle density, number/size of distractions, cognitive load) is adjusted manually every session by the supervising physiotherapist. Each VR session lasts 30 min, three times per week for 3 weeks by the same experienced physiotherapist. An emergency stop button and voice command allow instant pausing if the patient feels discomfort.

Non-VR training group

Participants in the non-VR training group will undertake 30 min of a task-oriented gait and balance programme that is fully matched in dose to the VR group for 3 weeks (3 days per week). The programme follows the 2016 American Heart Association/American Stroke Association Guidelines for Adult Stroke Rehabilitation and Recovery,¹⁶ which recommend intensive, repetitive, task-specific training for mobility and balance in chronic stroke survivors.

• Walking for specified distances (10–30 m overground).

• Turning (360° pivots while walking to simulate community navigation).

• Stair climbing (ascending/descending three to six steps with or without handrail support).

• Synchronising walking to rhythmic auditory cues (overground walking to metronome beats).

• Repetitive practice of functional tasks such as rising to stand and sitting down.

• Balance-specific activities responding to challenges in standing (weight shifting forward–backward or side-to-side, and practice of postural perturbations).

In addition, the same experienced physiotherapist will closely monitor each session, making necessary adjustments to exercise intensity and type based on the individual’s abilities, while also assessing progress, ensuring safety and evaluating the quality of movement.

Outcome measure

Primary outcomes

All participants will be assessed at three time points: at baseline (study start), immediately after the 3-week intervention and at a 2-month follow-up post-treatment. To facilitate complete follow-up assessments, commuting services to the laboratory will be provided. The primary outcome is the dynamic stability of gait, measured as a predictive indicator of fall risk. Non-linear measures, including the Lyapunov exponent and Floquet multipliers (FM), will be employed to assess local and orbital dynamic stability, respectively.¹⁷ These metrics offer insight into the system’s capacity to respond effectively to the small perturbations that occur naturally during normal walking¹⁷ and offer a significant safety advantage, as they allow us to quantify dynamic stability without exposing subjects to large, hazardous external perturbations. For periodic systems, the Lyapunov exponent continuously captures the system’s real-time behaviour in response to small, natural disturbances. In contrast, FM provide a discrete evaluation in aperiodic systems, measuring how the system responds to local perturbations from one cycle to the next at a specific point.¹⁸ Because human walking is characterised by neither strict periodicity nor complete aperiodicity, measuring both non-linear metrics yields a more comprehensive insight into the system’s sensitivity to small perturbations. In addition to dynamic stability, another primary outcome of the study is UX. UX will be assessed using the AttrakDiff questionnaire, a standardised instrument featuring a 7-point semantic differential scale with 28 pairs of opposing adjectives.¹⁶ In this scale, users indicate the point on the 7-point semantic differential that best represents their experience with the corresponding adjective. The AttrakDiff questionnaire evaluates four dimensions of UX: pragmatic quality: assesses how effectively and efficiently the product allows users to manipulate their environment; Hedonic Quality Identity: evaluates the extent to which the product enables users to express themselves and be perceived by others; Hedonic Quality Stimulation: measures the degree of excitement or stimulation the product provides; attractiveness: captures the overall, general impression of the product. Together, these dimensions offer a comprehensive view of the user’s experience with the system.

Secondary outcomes

In addition to the primary outcomes, several secondary measures will be evaluated to capture the comprehensive impact of the intervention. These measures will reflect both immediate benefits and long-term effects across a wide range of rehabilitation and well-being dimensions. Like the primary outcomes, assessments will take place at baseline (study initiation), immediately after the 3-week intervention and again at a 2-month post-treatment follow-up.

Functional mobility: The TUG test will be used to assess participants’ functional mobility.¹⁹ In this test, individuals are instructed to rise from a seated position, walk 3 m (approximately 10 feet), turn around and return to sit down. This tool has demonstrated high efficacy and reliability for evaluating functional activity in stroke survivors.²⁰

Functional walking ability: Functional walking ability will be evaluated using the 6-minute walk test (6MWT), which measures the effectiveness of rehabilitation interventions.²¹ Participants are asked to walk as far as possible within 6 min at a self-selected comfortable pace, along a 40-metre-long corridor marked at 5-metre intervals. Throughout the test, participants receive updates on the elapsed or remaining time at 1-minute intervals, and they are allowed to take breaks or use a chair for resting if needed.

Static balance: Static balance will be evaluated using the Berg Balance Scale (BBS), wherein participants perform 14 predetermined tasks. Each task is scored on a 0–4 scale, with higher cumulative scores indicating better balance.²²

Fear of falling: fear of falling during daily activities and social participation will be assessed using the Falls Efficacy Scale-International (FES-I), a 16-item questionnaire that quantifies concern about falling. FES-I scores range from 16, indicating no concern, to 64, representing severe concern.²³

Anxiety and depression: The Hospital Anxiety and Depression Scale will be employed to assess anxiety and depression among participants.²⁴ This self-assessment tool is widely used in general medical populations and comprises 14 items—7 for anxiety and 7 for depression—taking approximately 2–5 min to complete. Scores below 7 indicate that the individual is not a case for anxiety or depression, scores between 8 and 10 suggest mild symptoms, scores from 11 to 14 indicate moderate symptoms and scores between 15 and 21 reflect severe cases.

Fatigue: The Fatigue Severity Scale (FSS) will be used to measure the severity of fatigue in participants.²⁵ This self-report questionnaire consists of nine statements, each rated on a 7-point Likert scale ranging from 1 (completely disagree) to 7 (completely agree). The overall fatigue score is determined by calculating the mean of these nine responses, providing a concise metric of fatigue severity.

Adverse effects: To evaluate adverse effects induced by immersive VR, the Simulator Sickness Questionnaire (SSQ) will be used.²⁶ The SSQ assesses the severity of simulator sickness by evaluating 16 symptoms, with each symptom rated on a scale from 0 (none) to 3 (severe). This is a comprehensive scoring approach that ensures an accurate assessment of the severity of symptoms experienced during VR exposure.

Data pre-processing

The recorded data requires reprocessing to ensure its suitability for full spectral analysis. To preserve the primary signal information, the data will be filtered within the 0.5–15 Hz range using a Butterworth bandpass filter.²⁷ Local stability will then be assessed using the Lyapunov exponent. As previously described, the Lyapunov exponent measures the average exponential rate of divergence of nearby trajectories in state space. This provides a direct evaluation of the system’s sensitivity to local perturbations, including those affecting gait.^{28 29} It is calculated using the following formula:

$$y\left(i\right)=\frac{1}{\Delta t}\S amp;lt;\mathrm{Ln}{d}_j\left(i\right)\S amp;gt;\mathrm{}$$ Eq.1

In this formula, ${\mathrm{d}}_{\mathrm{j}}\left(\mathrm{i}\right)$ represents the Euclidean distance between the j^th pair of initially nearest neighbours after $i$ discrete-time steps (ie, $\mathrm{i}\mathrm{\Delta }\mathrm{t}$), and 〈〉 represents the average overall $\mathrm{j}$ pairs.^{28 30} A positive local divergence exponent signifies local instability, where higher values of the Lyapunov exponent indicate increased sensitivity and instability to local perturbations. Conversely, negative Lyapunov exponents are indicative of dissipative or non-conservative systems, such as a damped harmonic oscillator. These systems demonstrate asymptotic stability, with greater stability corresponding to more negative exponents.

Orbital stability will be quantified using FM, which estimates the system’s response to small perturbations occurring between consecutive gait cycles.³¹ Poincaré sections were defined for each per cent of the stride cycle, calculated using the following formula:

$$S_{k+1}=F\left(S_k\right)$$ Eq.2

In this formula, $S_k$ is the state of the system at stride k at each given Poincaré section (ie, at each per cent of the stride cycle) and $S_{k+1}$ is the state of the system at the next stride. Fixed points at each Poincaré section were defined from the average trajectory across all strides in a given trial:

$$S^{*}=F\left(S^{*}\right)$$ Eq.3

Orbital stability at each Poincaré section was assessed by analysing the system’s response to small perturbations away from the fixed points using a linearised approximation. The following equation was employed for this linearisation:

$$\left[S_{k+1}-s^{*}\right]=J\left(S^{*}\right)\left[S_k-S^{*}\right]$$ Eq.4

In this formula, $J\left(S^{*}\right)$ is the Jacobian matrix for the system at each Poincaré section. The FM values are the eigenvalues of $J\left(S^{*}\right)$, which quantifies how small perturbations grow or diminish for a given Poincaré section between one cycle and the next cycle. The FM values are the eigenvalues of the Jacobian matrix, and they quantify how small perturbations grow or diminish for a given Poincaré section between one cycle and the next.

The maximum Floquet multiplier (maxFM) was computed at each percentage interval of the gait cycle and subsequently averaged over the entire cycle for each trial. A maxFM value below one suggests that deviations from the fixed point tend to diminish in subsequent strides, indicating orbital stability.³⁰ All signal pre-processing analyses were performed using the Signal Processing Toolbox in MATLAB 2018b.

Data analysis

Statistical analyses will be conducted using IBM SPSS (Statistical Package for Social Sciences, V.22, Chicago, Illinois, USA). Data normality will be assessed initially, and descriptive statistics, including means and SDs, will be presented for all variables at three time points: baseline, 3 weeks and 2 months. The effectiveness of the intervention will be evaluated using repeated measures mixed analysis of variance. This analysis will examine the interaction between time (baseline, 3 weeks and 2 months) and group. Baseline stroke severity, age, time since stroke and baseline values of the outcome measures will be included as covariates in the model as these variables may influence treatment response or differ between groups at baseline. Bonferroni’s post hoc test will be applied for within-group comparisons. Additionally, an intention-to-treat analysis will be performed to address any missing follow-up data. Results will be considered statistically significant at p<0.05.

Trial status

We expect the study to span 18–24 months. In preparation for the main trial, a pilot study was conducted with one patient to refine key aspects, including the selection of appropriate outcome measures, the evaluation of the data collection device and the assessment of intervention acceptance. The dynamic stability measurements obtained during this pilot are presented in figure 3. Based on the insights gained, the main trial is going to start in January 2026.

Figure 3. The primary findings of dynamic stability in one patient. (A) representative example of the Lyapunov exponent (Panel A) and Floquet multipliers (Panel B) in our pilot data. In both figures, the X-axis represents the stride numbers. The Y-axis in the first and second panels show measures of Lyapunov exponent and Floquet multipliers, respectively. Data are shown in three axes, namely frontal, longitudinal and sagittal. maxFM, maximum Floquet multipliers.

Patient and public involvement

Stroke survivors and caregivers contributed to the design and refinement of trial materials, including patient-facing instructions and VR training scenarios, to ensure clarity, safety and acceptability. Their feedback informed adjustments to session length, difficulty progression and support services. We will share study results with participants and caregivers in accessible formats and involve them in shaping dissemination methods.

Discussion

This study proposes a protocol to rigorously evaluate the impact of immersive VR technology on walking stability in patients who had a stroke. Our primary objective is to assess the effects of immersive VR on dynamic gait stability after 3 weeks of intervention, with follow-up assessments at 2 months to determine the longevity of any improvements. In addition to the primary outcome, the protocol is designed to capture a wide range of secondary outcomes. These include subjective patient experience, functional mobility, functional walking ability, static balance, anxiety and depression, fatigue severity, fear of falling and any VR-related adverse effects. This comprehensive approach ensures that both clinical efficacy and patient-centred outcomes are measured. Preliminary data from two subjects confirm our approach and the robustness of our methodology. The dynamic stability measurements validate our protocol and suggest that immersive VR may serve as a promising adjunct to conventional stroke rehabilitation, with further findings expected to clarify its role in enhancing gait stability and overall outcomes.

One of the key features of our VR protocol study is the integration of non-linear dynamics as an objective, data-driven approach to assess the locomotion of stroke survivors. In this context, the Lyapunov exponent and FM quantify individuals’ responses to local and orbit perturbations, respectively.^{32 33} These measures have been demonstrated to differentiate elderly fallers from non-fallers as well as older adults from younger controls, thereby providing an accurate representation of real-life gait status.³⁴ Unlike previous VR interventions in stroke rehabilitation that relied solely on conventional linear measures, our study combines both non-linear dynamic analyses and traditional metrics.¹⁰ We anticipate that integrating secondary gait-related outcomes, such as the TUG test, the 6MWT and the BBS, with non-linear dynamic assessments will offer a more comprehensive understanding of a patient’s gait profile and the overall effectiveness of the intervention.

Another innovation of this study is the comprehensive UX assessment embedded within our VR-based intervention. UX is frequently overlooked in the development of digital health products and clinical trials,³⁵ yet it is critical for understanding patient acceptance and sustained use. In our study, we will explore the association between signal analysis measures and UX items, thereby evaluating the clinical and practical relevance of our approach. Prior research indicates that the UX of institutionalised elderly patients significantly influences their acceptance of immersive VR, promoting continued usage.³⁶ In addition to administering the AttrakDiff questionnaire for UX evaluation, we will systematically record any adverse effects experienced during the trial. Unlike some previous VR-based stroke interventions, which primarily focused on falls and cardiovascular responses, our protocol includes a detailed assessment of adverse effects using the SSQ to quantify the severity of simulator sickness.³⁷ We anticipate that adverse effects may occur in both study arms and may correlate with UX ratings or gait-related outcomes at the end of the study. This dual focus on UX and adverse effects promises to yield valuable insights into the practical application of immersive VR in stroke rehabilitation.

In this study, we will include stroke survivors diagnosed within 6–12 months, in contrast to many previous studies.^{38 39} Recruiting patients after the acute phase (ie, after 6 months) may mitigate adherence challenges associated with novel rehabilitation tools, as post-stroke distress in the early phases has been shown to deteriorate quality of life and impair rehabilitation adherence.³³ Additionally, neuroimaging findings indicate that spontaneous neuroplasticity begins shortly after a stroke and continues for at least 3–6 months.⁴⁰ We anticipate that the immersive VR intervention will extend this period of enhanced neuroplasticity, thereby improving gait and motor symptoms. Finally, implementing immersive VR technology for clinical purposes is particularly novel in developing countries such as Iran, where patients’ experiences, expectations and cultural factors may differ significantly from those observed in other contexts.

Using immersive VR technology to assess gait improvement and UX in patients who had a stroke poses several challenges that we have proactively addressed. For instance, patients who had a stroke may experience visual or cognitive impairments that hinder smooth participation in the trial. To mitigate this, we will screen all candidates to ensure they possess the necessary capacity to engage in the study. Additionally, given the innovative nature of VR, there is a risk of higher dropout rates; therefore, we plan to increase our sample size by 20% to account for potential attrition. To further minimise selection bias, participants will be randomly assigned to intervention groups. We will also conduct systematic follow-ups to evaluate both immediate and long-term outcomes, providing insight into the sustained impact of the intervention. To maximise safety, all sessions will be conducted under the vigilant supervision of a multidisciplinary team comprising a physiotherapist, a sports medicine physician and a neurologist. This dedicated approach ensures that any acute adverse effects are promptly addressed while upholding the highest safety standards throughout the trial.

To sum up, our proposed VR training protocol has the potential to offer valuable preliminary evidence regarding the role of immersive VR in improving dynamic gait stability, user experience after stroke and broaden its application to both rehabilitation and telerehabilitation of various neurovascular disorders.