This study consisted of two phases: an exploratory survey (Delphi method) conducted online to inform platform design and a subsequent in-person co-design workshop for usability evaluation.

In phase 1, an exploratory Delphi survey was conducted online, where a convenience sampling method was used to recruit individuals with lived experience of FND as “experts by experience” from the United Kingdom Royal Preston Hospital’s FND service team’s networks led by the PPIE (Patient and Public Involvement and Engagement) leads. In total, 20 individuals (experts by experience) with FND participated in the initial round of the Delphi survey. Participants provided feedback via an online questionnaire. The survey collected both quantitative and qualitative data on several topics: familiarity with VR and haptic technologies, perceptions of comfort and ease of use, anticipated relevance and impact of an XR-based therapy for FND, and potential barriers to adoption (such as, cost, access to equipment, technical support, and side effects). Responses were analyzed to extract common themes and requirements that the PPIE lead presented at the National Rehabilitation Centre (NRC) Rehabilitation Technologies Conference 2024 [21] (NRC Rehabilitation Technologies Conference 2024 poster and slides in Multimedia Appendix 1). Based on the survey findings, we codeveloped with the PPIE leads and industry partners (Human Robotix Ltd and Nudge Reality Ltd) a prototype XR neurotechnology platform. We selected the Human Robotix HRX-1 upper-limb haptic system (a portable robotic device providing force feedback) and Nudge Reality’s “Hoop Hustle” XR game as the core components for our platform, as these were judged by the PPIE leads to best meet the identified needs (detailed specifications of the hardware and game options are available in Multimedia Appendix 2). The HRX-1 device can assist or resist arm movements with precise torque control, while Hoop Hustle is a VR game that can be adapted for therapeutic exercises.

For the phase 2 co-design and usability testing workshop, a purposive sampling approach was used to recruit participants specifically from the United Kingdom Royal Preston Hospital’s FND Service team, including PPIE leads. We then conducted an in-person workshop involving 6 participants drawn from the FND service community: 3 FND patient representatives (1 female and 2 male) and 3 health care professionals (2 physiotherapists and 1 neurologist; 2 female and 1 male). All 6 participants are coauthors of this paper for the participatory design approach. Before the workshop, participants provided informed consent. The session took place in a rehabilitation clinic setting and lasted about half a day.

As this work was part of a patient engagement and technology co-design project, it was conducted with institutional review board notification but was determined to be a service development and quality improvement activity not requiring full National Health Service (NHS) Research Ethics Committee review. All participants gave written informed consent for their involvement and for publication of deidentified feedback. The study was carried out in accordance with the Declaration of Helsinki principles of ethical research.

Given the selection of Human Robotix’s HRX-1 system for upper limb rehabilitation (Human Robotix’s HRX-1 system in Multimedia Appendix 3) and Nudge Reality’s “Hoop Hustle” game (Nudge Reality’s XR games in Multimedia Appendix 4) by PPIE leads, efforts were focused on adapting these technologies to test 3 conditions: VR relaxation, XR positional feedback, and XR force feedback. During the co-design workshop, the prototype XR platform was introduced, and participants were guided through 3 interactive training tasks, each representing a different mode of biofeedback.

After completing all 3 tasks, participants filled out the system usability scale (SUS) questionnaire for each task. The SUS is a 10-item questionnaire yielding a score from 0 to 100, where higher scores indicate better perceived usability. We chose the SUS because it is a well-established, quick tool for usability assessment, suitable even for small samples [26]. Participants also provided written free-text feedback on their experience with each task and the overall platform. These responses were collected on paper forms and later transcribed. In addition, the workshop concluded with a short group discussion, allowing participants to collectively reflect on what aspects of the platform worked well and what improvements they would prioritize. The discussion was later summarized in notes.

Quantitative data from the SUS were summarized using descriptive statistics, given the small sample size. We report individual SUS scores per participant and per task, as well as the range and median for each task’s scores. Following convention [27], we interpret SUS scores using an adjective rating scale for context: scores above~85 are considered “excellent,” around 70‐85 “good,”~50‐69 “okay,” and below 50 “poor” in terms of usability perception. We did not perform inferential statistical tests due to the exploratory nature of this pilot and the limited number of subjects. Qualitative data (written feedback and facilitator notes) were analyzed thematically. Two researchers (1 patient representative and 1 study investigator) independently reviewed the feedback to identify recurring themes. Using NVivo 12 (QSR International), feedback comments were coded with initial labels corresponding to aspects of user experience (eg, “hardware discomfort,” “audio feedback,” and “game difficulty”). These codes were then grouped into higher-level themes through discussion and consensus. Representative participant quotes were extracted to illustrate each theme in the Results.

We present the results from our first round of the Delphi survey according to the Checklist for Reporting Results of Internet E-Surveys (CHERRIES) [28], aimed at the translation of the VR haptics technology for biofeedback training in FND. The survey gathered online feedback from 20 (N=20) individuals with lived experience of FND, considered experts by experience for technology translation.

The survey was hosted on a General Data Protection Regulation (GDPR)-compliant, secure online platform (MS Forms is part of Microsoft 365, which adheres to GDPR, Health Insurance Portability and Accountability Act (HIPAA), and ISO 27001 security standards).

Anonymous participation was allowed; no email registration was required, and no IP tracking or cookies were used.

The question structure of the survey included a combination of question types, as listed in Textbox 1.

This study has two limitations. The first is the small sample size (N=20); the results are preliminary and not generalizable to all patients with FND. The second is the use of the single-round Delphi survey; the findings require further validation through additional rounds or larger-scale studies.

The first round of the Delphi survey provided key insights into the usability, expectations, and barriers associated with XR haptics biofeedback training for FND rehabilitation.

Participants perceived high potential benefits but highlighted cost, accessibility, and usability concerns. There was a strong interest in real-time feedback and customization to tailor the technology to individual needs. Concerns about motion sickness, equipment comfort, and NHS availability need to be addressed for successful adoption.

Refining usability features based on patient feedback in phase 2 co-design and usability testing. Further stakeholder engagement with clinicians, patient organizations, and industry partners in phase 2 co-design and usability testing. Scaling the study to validate findings with a larger sample and additional Delphi rounds following in phase 2 co-design and usability testing.

Basic usability testing typically benefits from the purposive selection of 5‐10 participants [29]. Here, all 6 workshop participants completed the XR position feedback and XR force feedback tasks, and 5 completed the VR relaxation task (1 health care professional was unable to try the VR relaxation due to time constraints). Table 1 presents the SUS scores given by each participant for each task. Overall, the XR position feedback game received the highest ratings with a median score of 91.3, and all participants rated it above 70. The VR relaxation task had a bimodal distribution of scores—2 participants rated it very highly (~88‐90) while 3 participants gave it scores below 50, indicating poor usability for those individuals. The XR force feedback task had generally positive scores from 4 participants (range, 80.0‐95.0), but 1 participant (Participant 3) gave a very low score (27.5). According to Bangor et al’s [27] adjective rating scale for SUS, the low scores in the 40s for the VR relaxation task correspond to a “poor” usability experience, despite the same task being rated as “excellent” by others. Similarly, the force feedback task’s scores suggest mostly “good” to “excellent” usability, with one clear outlier in the “poor” range. In contrast, the XR position feedback task’s scores correspond to “good” or “excellent” usability across all users. These results highlight a high degree of variability in user experience for the more complex or condition-sensitive tasks (VR relaxation and force feedback), compared to the consistently positive experience with the position feedback game (XR system usability testing script and XR system usability testing results in Multimedia Appendices 5-8).

Qualitative analysis of the feedback revealed several key themes regarding the user experience and suggestions for improvement. Participants provided free-text responses regarding their VR relaxation task experience, which were analyzed for future technology improvement.

This study is, to our knowledge, the first mixed methods evaluation of an XR-based biofeedback training platform co-designed for individuals with motor FND. Through a 2-phase coproduction approach, we obtained rich stakeholder input and preliminary evidence of usability. Our key finding is that the XR position feedback game was the most well-received component of the platform, with consistently high SUS usability scores and positive feedback from users. This task, dubbed “Hoop Hustle” in the prototype, required participants to perform wrist movements to control a VR interface with accompanying visual feedback. The strong performance of this task suggests that combining visual feedback in an intuitive pointing game can be highly engaging and easy to use for people with functional weakness. Participants likely benefited from the clear, immediate cause-and-effect in this game, which may have contributed to a sense of accomplishment and control.

In contrast, the VR relaxation module yielded a polarized reaction: some individuals felt deeply relaxed and enjoyed the experience (reflected in very high SUS scores), while others struggled with aspects of the VR environment (leading to poor usability ratings). These divergent outcomes highlight that a one-size-fits-all relaxation experience may not suit everyone; factors such as susceptibility to motion sickness, comfort with wearing a VR headset, and personal preference for meditation-style activities can greatly influence one’s experience. The XR force feedback task showed intermediate and more variable usability. Most participants handled the force-feedback task moderately well (SUS~80‐95 for 4 participants), indicating that they understood the task and could perform it, but one participant (P3) had an extremely negative experience (SUS 27.5). P3’s case is particularly informative: this participant has functional dystonia (a subtype of FND causing involuntary muscle contractions), which likely made it difficult to perform the steady force output required by the task. This resulted in frustration or fatigue, as reflected in both the low usability rating and the participant’s comments describing the force task as “hard to manage” and “tiring.” This finding underscores that individual clinical differences (such as the type of motor symptoms) can dramatically affect the usability of specific training tasks. Notably, the same participant (P3) rated the XR position task very highly (92.5), much higher than they rated the other 2 tasks. We interpret this to mean that while the force feedback task was not well-tolerated by P3, the position feedback game was accessible and enjoyable even for someone with dystonia. It is possible that the position task’s design—emphasizing range of motion and coordination rather than sustained force—was better aligned with this participant’s abilities. This suggests a need for personalized task selection or customization: users might benefit from having multiple training task options and skipping or modifying those that aggravate their symptoms.

Across all tasks, the qualitative feedback provided further insight into these quantitative results. For instance, participants who gave lower SUS scores often cited specific issues that explained their discomfort. Those who rated the VR relaxation poorly mentioned problems like visual graininess and a sense of disorientation when the virtual scene “breaks” (one user described a loss of immersion at a certain transition, eg, reaching a virtual staircase where the illusion was not convincing). On the other hand, participants who enjoyed the relaxation task commented on feeling calm and appreciating the break from active gameplay, which may reflect personal differences in how individuals prefer to engage (active interaction vs passive relaxation). Similarly, mixed feedback on the force task corresponded with whether users felt the haptic feedback was appropriate; some found it novel and motivating, while others found it confusing or difficult to calibrate their strength. We can summarize the qualitative feedback themes as follows.

Our findings align with existing literature emphasizing user-centered design for health technologies. Previous studies have noted that even small samples (5‐10 users) can uncover the majority of usability issues in a system [29]. In our case, 6 users were sufficient to highlight distinct strengths and weaknesses of the platform. The variability in VR relaxation feedback is reminiscent of observations in broader VR applications: while VR can provide immersive therapeutic experiences, factors like motion sickness and comfort remain challenges to address. The need for personalization in rehabilitation technology is well-documented; for instance, usability studies of other rehab games have found that adaptive difficulty can significantly improve user engagement and outcomes. Our results specifically extend this understanding to FND, suggesting that personalization may not only improve engagement but might be necessary to accommodate neurological symptoms like dystonia or fatigue. From a neurological perspective, the concept of using haptic feedback and VR to retrain the perception-action cycle in FND draws on theories of sensory attenuation and agency in functional movement disorders. By providing congruent visual and haptic inputs corresponding to the user’s intended movements, the platform aims to reinforce the association between effort and sensory feedback, potentially strengthening the efference copy mechanism that is hypothesized to be underactive in patients with FND [14,30,31]. While our study did not directly measure clinical outcomes or neurophysiological changes, the positive usability of the position and force feedback tasks is a critical first step toward implementing such therapeutic concepts in practice. A recent review by [12] also emphasized VR’s promise for addressing mechanisms of agency and attention in FND; our practical findings complement this by showing that patients are willing to engage with VR or haptic systems, provided they are comfortable and accessible.

This study has several limitations. First, the sample size was small (5‐6 participants for usability testing), and all participants were from a single clinical center and also coauthors, which could introduce some bias or limit critical feedback. The findings should be interpreted as preliminary and exploratory; a larger, independent sample will be needed to validate and generalize the usability results. Second, participants’ familiarity with XR technology varied, and those with previous VR or gaming experience might have found the system easier to use, potentially influencing their SUS scores. We did not formally quantify each participant’s XR technology background, which is a confounding factor that future studies should measure. Third, we focused on 3 specific XR tasks (VR relaxation, XR position feedback, and XR force feedback). Other functionalities (eg, bilateral training or cognitive tasks in XR) were not included and could present additional usability challenges or benefits not captured here. Fourth, the reliance on subjective SUS scores introduces potential bias, as individual expectations or novelty effects can influence ratings. We mitigated this by collecting detailed qualitative feedback, but objective performance metrics were not analyzed in this pilot. Fifth, as an initial co-design and usability study, we did not assess clinical efficacy, for instance, whether using the platform yields improvements in motor function or FND symptoms. Such outcomes will need evaluation in subsequent trials. Finally, our personalization of the tasks was done manually by the facilitators rather than through built-in adaptive algorithms. This limits the consistency of the user experience; an automated personalization mechanism would be ideal to ensure each user gets an optimally challenging experience. Despite these limitations, the study provides valuable insights into the user experience of an XR neurotechnology platform tailored for FND. To our knowledge, this is one of the first studies to report detailed usability data for an XR haptics platform in FND rehabilitation. The co-design approach proved effective in identifying user priorities and potential pitfalls early in the development process.

The next steps following this study will address the identified issues and test the platform on a broader scale at home [32]. FMD involves involuntary-feeling but voluntary-appearing movements, linked to a disrupted sense of agency due to impaired sensory attenuation, that is, the brain’s predictive suppression of self-generated sensory feedback [33]. This impairment, involving brain regions like the primary motor cortex, cerebellum, and right temporo-parietal junction, leads to difficulties in distinguishing self-initiated actions from external stimuli [34]. Conversely, sensory amplification, mediated by the posterior parietal cortex, heightens sensory perception through attention. XR presents therapeutic potential by balancing sensory attenuation and amplification [6]. VR allows controlled manipulation of predictive coding, helping recalibrate agency and sensory processing in FMD. The comparator model (refer to Figure 3) suggests that agency arises when predicted sensory outcomes align with actual feedback, which can be reinforced through haptic feedback in XR. Here, linking active inference in motor control lies in its ability to explain and address motor dysfunctions [35]. By recognizing that the brain updates perceptions and modifies actions to minimize prediction errors, this framework offers insights into abnormal motor control, where disrupted sensory prediction leads to impaired agency and movement errors. In rehabilitation, this perspective supports the development of XR biofeedback interventions, where haptic and visual feedback can help recalibrate faulty sensorimotor predictions. By reinforcing accurate sensory-motor associations, these technologies may restore agency and improve motor function, offering a novel, personalized approach to therapy. Indeed, XR technologies have been shown to enhance sensorimotor processing, but usability for patients with FMD must be assessed. Early user involvement, particularly in conditions like functional dystonia, is critical to refining XR rehabilitation design. Our study engaged stakeholders from academia, industry, and health care (NHS, England) to classify technological needs into incremental or revolutionary advancements. Notably, no commercial or research-based XR biofeedback systems currently exist specifically for FMD rehabilitation.

In response to user feedback, we are working with the developers to improve the VR relaxation module (enhancing graphics, refining the audio guidance, and possibly adding options for different scenes or background music) [32]. We are also implementing in-software settings that allow end-users or therapists to easily adjust game difficulty, visual or auditory feedback levels, and force feedback intensity. In addition, we plan to incorporate a brief calibration or tutorial at the start of a session, where the system can gauge a user’s comfortable range of motion and strength, and automatically set initial task parameters accordingly. These changes aim to embed personalization directly into the platform. A follow-up study is being designed to involve a larger cohort of patients with FND in a multisession at-home trial with the refined platform [32]. That study will evaluate not just usability, but also short-term effects on motor function and symptoms, using clinical scales and objective performance metrics within the game. We will also examine learning effects–whether repeated use leads to improved user proficiency or changes in feedback preferences–to understand how usability evolves over time. An important future direction is to explore remote or home usability of this platform [32]. Given the interest in home-based rehabilitation (and lessons learned during the COVID-19 pandemic), we aim to test whether patients can effectively use the XR system at home with minimal supervision. This will involve developing comprehensive user guides, integrating remote monitoring capabilities (so therapists can track usage and progress), and ensuring robust technical support is available. Addressing these factors will be essential for translating this coproduced XR platform into a scalable, real-world therapeutic option for individuals with FND.

Through a collaborative coproduction approach, we developed and pilot-tested a novel XR (VR+ haptic) biofeedback training platform for patients with functional upper limb weakness due to FND. Our usability findings are encouraging: an interactive XR position feedback game was rated highly usable by all participants, and a VR relaxation experience received very positive feedback from some users. At the same time, the variability in responses, particularly the challenges faced by one participant during the force feedback task, highlights the necessity of a flexible, user-tailored design in such neurotechnologies. One-size-fits-all solutions are unlikely to succeed in the FND population given the diversity of symptoms and user preferences. By systematically incorporating user feedback, we identified concrete areas for improvement (such as hardware comfort and software adaptability) that will guide the next iteration of the platform. This study demonstrates that patients with FND and clinicians are not only capable of providing meaningful input into technology design but are eager to do so when the goal is to enhance therapy. With further refinement and larger-scale testing, the XR platform has the potential to become a valuable tool in FND rehabilitation, offering engaging, at-home training that reinforces patients’ agency and motor function in a way that is enjoyable and customized to their needs.