The use of wearables in rehabilitation for functional neurological disorder (FND) has expanded from experimental tools in research settings to increasingly common adjuncts in routine clinical care. Multidisciplinary FND teams now integrate consumer-grade and medical-grade devices to collect continuous data on movement, autonomic function, sleep, and daily activities, with the aim of linking objective patterns to symptom fluctuations and treatment response. This shift reflects a broader trend in rehabilitation medicine toward data-informed, patient-centered care that extends beyond the clinic and into everyday environments where functional symptoms are often most disabling.
Clinicians working with FND primarily use wearables for three purposes: monitoring, therapeutic biofeedback, and enhancing engagement with rehabilitation. Monitoring applications focus on tracking gait quality, tremor frequency, balance, and overall activity levels as patients participate in physical therapy, occupational therapy, or self-directed exercise. In motor FND, inertial measurement units embedded in wristbands, ankle sensors, or smartphones can quantify step counts, walking speed, stride variability, and postural sway. These metrics provide a more nuanced picture than self-report alone, particularly for patients who have difficulty recognizing fluctuations in their symptoms or who underestimate their functional abilities.
Therapeutic biofeedback is emerging as a central use case, with wearables offering real-time information on physiological processes relevant to FND, such as heart rate, heart rate variability, breathing patterns, and muscle tension. Many patients with FND experience heightened autonomic arousal, persistent hypervigilance, or difficulty downregulating stress responses, which can exacerbate motor and sensory symptoms. Devices capable of providing live feedback on arousal states allow patients to practice grounding techniques, paced breathing, and other regulation strategies while immediately seeing their effect on physiological parameters. Over time, this can support better interoceptive awareness and a greater sense of control over symptoms.
Another growing application is the use of activity tracking to structure graded increases in movement and participation. Many FND rehabilitation programs emphasize exposure to feared activities, reduction of safety behaviors, and a shift from symptom-focused monitoring to function-focused goals. Wearables that log steps, standing time, and task completion can help clinicians collaborate with patients to set realistic, measurable targets, adjust pacing, and identify patterns of boom-and-bust activity that might otherwise remain unclear. This objective data supports more precise titration of rehabilitation intensity and can improve transparency when discussing progress and setbacks.
On the technological side, the current landscape includes a mix of general-purpose consumer devices and specialized systems. Popular smartwatches and fitness bands provide basic metrics such as heart rate, step count, sleep duration, and sometimes heart rate variability, making them accessible options in clinics with limited resources. At the same time, research groups and some tertiary centers employ medical-grade accelerometers, electromyography sensors, and multi-sensor platforms that can capture more detailed biomechanical and autonomic data. Smartphone-based tools, using the built-in accelerometer and gyroscope, are increasingly used as a low-cost middle ground that reduces the need for additional hardware.
The integration of these technologies into routine FND care remains heterogeneous across settings and countries. Some specialized centers have embedded wearables into standard assessment protocols, using short-term monitoring to help distinguish functional movement patterns from other neurological conditions and to establish baseline functioning before rehabilitation. Others focus on longitudinal tracking, issuing devices for weeks or months to observe how changes in therapy, stressors, sleep, or medication relate to symptom intensity and functional outcomes. Community and outpatient services may rely more heavily on patient-owned devices, asking individuals to share downloaded reports or screenshots during follow-up visits or telehealth sessions.
Patient engagement and motivation are key drivers behind the adoption of wearables in FND rehabilitation. Many individuals appreciate having concrete indicators of their efforts, especially when subjective improvements feel slow or inconsistent. Visual feedback in the form of daily activity rings, progress graphs, or biofeedback dashboards can validate small but meaningful gains, reinforce adherence to home exercises, and provide a sense of agency during a disorder that often feels unpredictable and uncontrollable. For younger patients and those already accustomed to digital health apps, the integration of wearables into therapy can feel intuitive and acceptable.
In addition to individual clinical use, wearables are increasingly central to FND research, where they enable large-scale, ecologically valid data collection. Continuous monitoring across days or weeks allows investigators to map relationships between stress exposure, sleep disruption, and symptom exacerbations, or to identify digital biomarkers that distinguish FND from other neurological or psychiatric conditions. Studies are beginning to explore whether specific signatures in gait, tremor patterns, autonomic responses, or variability in activity levels can predict treatment response or risk of relapse. These efforts lay the groundwork for stratified rehabilitation approaches that match interventions to patients based on objective data rather than solely on clinical impression.
Despite these advances, use of wearables in FND rehabilitation is still in an early phase compared to their adoption in fields such as cardiology or diabetes care. Many protocols remain pilot projects, and there is significant variability in how data are interpreted and acted upon. Some clinicians use device readouts as supportive information alongside standard assessments, while others incorporate them into structured pathways that define thresholds for adjusting therapy plans. Standardization of definitions, metrics, and reporting practices is still developing, and cross-study comparisons can be challenging. Nonetheless, there is broad recognition that continuous, real-world data can complement traditional episodic assessments and provide a more comprehensive view of functional capacity.
An important trend is the move from passive data collection toward interactive, adaptive systems. Rather than simply recording metrics for later review, newer platforms aim to deliver context-sensitive prompts, reminders, or micro-interventions based on real-time data. For example, a device might detect prolonged inactivity following a reported symptom flare and suggest a brief, achievable movement task, or notice rising physiological arousal and prompt a breathing exercise. In FND rehabilitation, such just-in-time support holds promise for reinforcing therapy principles between sessions and helping patients generalize skills to their everyday lives, where triggers and challenges are most salient.
Collaboration between neurologists, physiotherapists, psychologists, occupational therapists, and digital health teams is shaping how wearables are selected, configured, and integrated into care. Successful implementation often requires aligning device capabilities with specific therapeutic goals, such as focusing on gait retraining, autonomic regulation, or building tolerance for daily tasks. It also involves planning for practical issues, including battery life, data storage, connectivity, and how information will be communicated back to patients in a clear and supportive manner. As experience grows, many teams are moving away from ad hoc device use and toward more structured protocols that define the timing, duration, and purpose of wearable-based monitoring within the overall rehabilitation plan.
The current landscape is characterized by rapid innovation, uneven adoption, and growing evidence that thoughtfully deployed wearables and biofeedback can enrich FND rehabilitation. While definitive guidelines are still emerging, clinical practice is increasingly influenced by the possibilities of continuous, objective measurement and patient-facing feedback, setting the stage for more individualized and responsive rehabilitation pathways that are grounded in real-world data and functional outcomes.
Physiological signals and biofeedback modalities
Physiological monitoring in FND rehabilitation typically centers on a core group of signals that relate to arousal regulation, movement control, and interoceptive awareness. Wearables collect these signals continuously or during targeted exercises, transforming otherwise opaque internal states into data that can be visualized, interpreted, and used for therapeutic learning. The goal is not to “normalize” numbers in isolation, but to help patients understand how their bodies respond to stress, attention shifts, and movement demands, and to support more adaptive responses in real time.
Autonomic nervous system activity is a primary focus, given its strong association with symptom onset and fluctuation in FND. Basic heart rate monitoring is widely available in consumer devices and can reveal patterns of persistent tachycardia, exaggerated responses to mild exertion, or delayed recovery after stress. More advanced analysis of heart rate variability, particularly time-domain and frequency-domain measures, provides a window into the balance between sympathetic and parasympathetic influences. In rehabilitation, clinicians may use trends in heart rate variability to infer when a patient is spending much of the day in a high-arousal state, potentially linked to hypervigilance, pain, or fatigue, and then pair this information with training in grounding, paced breathing, or body scan exercises.
Biofeedback protocols that incorporate heart rate or heart rate variability often rely on simple visual or auditory cues to help patients learn how specific techniques affect their physiology. For instance, a patient may practice slow diaphragmatic breathing while watching a graph that shows moment-to-moment changes in heart rate, or follow a paced-breathing animation that synchronizes with their pulse wave. The wearable device translates subtle cardiovascular shifts into easily recognizable patterns, providing immediate reinforcement when regulation strategies are effective. Over sessions, patients can experiment with different approaches—such as imagery, progressive muscle relaxation, or posture adjustments—and discover which ones reliably reduce physiological arousal or stabilize heart rate variability.
Respiratory signals are another important target, especially in patients whose symptoms are closely linked to sensations of breathlessness, chest tightness, or panic. Some wearables use chest straps, respiratory belts, or strain sensors to monitor respiratory rate and effort, while others infer breathing patterns from variations in heart rate or accelerometer data. These inputs can power breathing-focused biofeedback, where patients learn to recognize and counteract rapid, shallow breathing or breath-holding that may occur during pain spikes, dissociative episodes, or functional attacks. Therapists may design exercises where patients practice lengthening the exhalation phase, coordinating movement with exhalation, or maintaining a steady respiratory rhythm during graded exposure to symptom triggers.
Electrodermal activity, often captured via skin conductance sensors on the wrist or fingers, is used to index changes in sympathetic arousal. Although frequently associated with research settings, electrodermal biofeedback is increasingly available in commercially produced wearables. In FND rehabilitation, heightened skin conductance responses during everyday tasks can highlight situations where threat appraisal, worry, or anticipatory anxiety is particularly strong, even when the patient reports feeling “numb” or detached. Clinicians may use this information to guide psychoeducation about the link between bodily arousal and functional symptoms, and to structure in-session exercises where patients practice grounding or cognitive reframing while observing how their skin conductance rises and falls.
Muscle activity, measured through surface electromyography (EMG), plays a central role in many biofeedback interventions for functional motor symptoms. EMG sensors placed over specific muscle groups can detect inappropriate co-contraction, excessive tension, or paradoxical activation patterns common in functional gait disturbance, tremor, or dystonia-like postures. During therapy, the patient may see a real-time bar graph or line trace representing activation levels in agonist and antagonist muscles while performing simple tasks such as standing, stepping, or reaching. The clinician then guides the patient to adjust attention, intention, and movement strategy—for example, focusing on a goal-directed task rather than symptom monitoring—while using EMG feedback to reinforce more efficient, less effortful patterns of activation.
In functional tremor, EMG and accelerometer data can be combined to provide feedback on tremor frequency, amplitude, and variability. Patients may be encouraged to perform dual-task exercises, rhythmic movements, or distraction maneuvers while observing how their tremor pattern changes. This can help challenge fixed beliefs about symptom uncontrollability and highlight the influence of cognitive and emotional context on motor output. Over time, the aim is for patients to internalize a sense of agency and to reproduce more stable, voluntary movement patterns without needing continuous visual feedback.
Movement and posture signals derived from accelerometers and gyroscopes underpin many rehabilitation protocols focused on mobility, balance, and coordination. Activity tracking on wearables is often configured to capture step counts, cadence, stride length surrogates, and postural transitions (such as sit-to-stand events). Beyond simple metrics, algorithms can detect gait asymmetry, freezing episodes, or irregular stride timing during real-world walking. These data enable targeted biofeedback in which patients receive in-the-moment cues—for instance, auditory metronome beats to encourage steady cadence, or vibration prompts to shift weight evenly across both legs. When integrated into task-specific training, such feedback can gradually reduce compensatory behaviors, such as overreliance on one limb or excessive visual checking of the ground.
Postural control biofeedback may make use of trunk or waist-mounted sensors to quantify sway and center-of-mass movement during standing tasks. Patients can practice static and dynamic balance exercises while seeing a simplified representation of their sway pattern, learning how minor adjustments in foot placement, knee flexion, or gaze direction influence stability. For individuals with functional dizziness or visually induced unsteadiness, pairing vestibular habituation exercises with real-time sway feedback can help decouple sensations of instability from catastrophic interpretations, reinforcing the concept that apparent “loss of balance” often occurs in the context of intact protective responses.
Some systems also track upper limb movement quality, including smoothness, range of motion, and movement variability during reach-and-grasp tasks or activities of daily living. In the context of functional limb weakness or abnormal posturing, biofeedback may involve encouraging fluid, continuous motion while minimizing start–stop patterns or excessive co-contraction. Graphical displays that show movement trajectories or speed profiles help patients differentiate between effortful, guarded movement and more automatic, goal-oriented patterns, aligning with principles of redirecting attention away from the affected limb and toward the functional task.
Sleep and rest-activity cycles, although often considered secondary outcomes, are increasingly integrated into FND rehabilitation via wearable-based actigraphy. Devices estimate sleep onset, duration, wake after sleep onset, and daytime inactivity. For patients with FND, disrupted sleep, irregular routines, and frequent napping can exacerbate fatigue, cognitive fog, and symptom flare-ups. Visualizing sleep–wake patterns over days or weeks helps both patients and clinicians identify cycles of overexertion and collapse, aligning behavioral interventions with evidence from activity tracking. Biofeedback in this context is less about moment-to-moment signals and more about giving patients longitudinal insight into how consistent routines, light exposure, and graded activity schedules influence symptoms and daytime functioning.
Multimodal biofeedback, which combines several physiological signals, is emerging as a promising approach for complex symptom presentations. For example, a protocol might integrate heart rate, respiratory rate, and accelerometry to deliver context-sensitive prompts: a rise in heart rate disproportionate to movement level, plus an increase in respiratory rate, could trigger an on-screen suggestion to pause, ground, and employ a preferred regulation technique. Similarly, a combination of EMG and accelerometry might detect a pattern characteristic of the onset of a functional tremor episode, leading the system to guide the patient through a sequence of voluntary movements that have previously helped interrupt or modify the episode.
The choice of biofeedback modality is shaped by clinical goals, patient preferences, and practical considerations such as comfort and wearability. Some individuals are highly motivated by clear numerical targets, graphs, or “scores,” finding that these outputs foster a sense of progress and mastery. Others may benefit more from simplified cues such as color changes, tones, or tactile feedback that signal whether they are moving toward or away from a desired physiological state. Clinicians must also consider cognitive load: complex dashboards may be overwhelming for patients with fatigue, dissociation, or attentional difficulties, so interfaces are often simplified to emphasize only the most relevant signals for a given stage of rehabilitation.
Importantly, the therapeutic use of physiological data in FND hinges on careful framing. Rather than presenting readings as evidence of “damage” or deficits, clinicians typically use them to support a biopsychosocial understanding of symptoms: the signals represent dynamic patterns of brain–body interaction that can change with practice, context, and learning. When used in this way, biofeedback can reinforce the core rehabilitation message that functional symptoms are real, but also modifiable, and that improvements in physiological regulation and movement patterns are meaningful contributors to better functional outcomes in daily life.
Clinical applications and case examples
Clinical use of wearables and biofeedback in FND spans inpatient, outpatient, and community-based rehabilitation, often embedded within multidisciplinary programs that address motor, sensory, cognitive, and dissociative symptoms. In specialized centers, devices are not used as stand-alone treatments but as tools that enhance core therapeutic approaches such as motor retraining, graded exposure, cognitive-behavioral interventions, and autonomic regulation training. The following examples illustrate how these technologies are integrated into everyday clinical practice and how they can influence engagement and outcomes.
One common scenario involves individuals with functional gait disturbance admitted to intensive rehabilitation units. At the outset, physiotherapists may attach inertial sensors to the ankles or use a smartwatch to quantify baseline walking speed, step length surrogates, and variability. During the first days, many patients demonstrate slow, cautious gait with prominent asymmetry and frequent reliance on mobility aids, despite preserved strength on examination. Therapists introduce task-oriented gait retraining—such as walking to targets, dual-task exercises, or treadmill walking—while continuously recording movement parameters. As patients shift attention from symptom monitoring to functional goals, wearables often show rapid improvements in cadence and stride regularity, even before patients subjectively feel “more confident.” Reviewing simple graphs of activity tracking data helps patients recognize these early gains and reinforces the message that their nervous system retains the capacity for efficient movement.
In some cases, feedback from wearable sensors is used directly during gait training. For example, a patient with inconsistent foot drag and variable step timing may wear an ankle sensor linked to a tablet that displays real-time cadence and symmetry metrics. The physiotherapist sets a target cadence range based on the patient’s unimpaired capacity, and the display changes color or emits a tone when steps fall within this zone. The patient practices walking while focusing on reaching and maintaining the target cadence rather than thinking about leg weakness. Over several sessions, the need for real-time visual feedback decreases as the more automatic gait pattern consolidates, and the device is then used primarily for periodic progress checks.
Functional tremor provides another context in which clinical teams leverage wearable technology. Patients often present with highly variable tremor that worsens under observation but may diminish with distraction. Neurologists and physiotherapists can attach accelerometers or combined accelerometer–EMG sensors to the affected limb during standardized tasks such as rest, posture holding, and goal-directed movement. Data collected across conditions frequently demonstrate entrainment to external rhythms or marked changes with cognitive distraction, supporting the functional diagnosis and providing a concrete way to explain to patients how attention and expectation shape motor output. During subsequent therapy sessions, the same sensors can be used as biofeedback tools: patients perform tasks like rhythm tapping, bilateral movements, or object manipulation while viewing a simplified display of tremor amplitude. As they practice focusing on the task rather than the tremor, they can see real-time reductions in amplitude, which underscores their capacity to influence symptoms.
Patients with functional limb weakness or non-epileptic attacks often experience heightened autonomic arousal and poor interoceptive awareness. In outpatient settings, clinicians may provide a wrist-worn device or chest strap that captures heart rate and, when possible, heart rate variability during daily life. Over one or two weeks, patients keep a brief electronic diary noting perceived stressors, fatigue, and symptom episodes. When reviewing the data together, patterns frequently emerge: episodes of functional weakness may cluster after nights of reduced sleep, during prolonged sedentary periods, or following interpersonal conflicts, accompanied by sustained elevations in heart rate or reduced heart rate variability. This shared review supports psychoeducation about the interplay between autonomic regulation, stress, and symptom expression, and it sets the stage for targeted interventions such as paced breathing, scheduling of rest, and graded activity.
Once psychoeducation is established, clinicians can use real-time biofeedback to practice regulation skills. For instance, in clinic or via telehealth, a psychologist guides a patient through slow diaphragmatic breathing or grounding exercises while the wearable streams heart rate data to a simple on-screen display. The patient observes how small changes in breathing pattern or posture affect cardiovascular arousal, often noticing delayed reductions in heart rate after sustained practice. When heart rate variability metrics are available, clinicians may use them qualitatively, highlighting trends rather than focusing on exact numerical targets. The emphasis is on learning that physiological arousal is malleable and that consistent practice of regulation strategies influences symptoms indirectly by stabilizing the underlying autonomic context.
In individuals with frequent functional seizures or dissociative episodes, wearables are sometimes used as part of safety and self-management plans. Smartwatches that detect abrupt changes in movement or heart rate can be configured to alert the patient (for example, via vibration) when patterns associated with early stages of an episode are detected. This early warning may prompt the patient to initiate pre-planned coping strategies such as moving to a safe space, engaging support persons, or practicing grounding techniques. Clinicians and patients may later review logs of episodes, combining event markers from the device with self-reported triggers and recovery strategies. Over time, this can clarify which antecedents are most relevant, which coping responses shorten or lessen episodes, and how broader lifestyle factors such as sleep and workload influence episode frequency.
Day-to-day activity tracking is frequently incorporated into graded functional programs, particularly for patients whose symptoms are maintained by cycles of overexertion and subsequent prolonged rest. Occupational therapists might ask a patient to wear a fitness band continuously for two weeks while maintaining a simple diary of major daily tasks. Initial data often show irregular patterns, with days of high step count and numerous errands followed by near-complete inactivity. Using this information, the therapist collaborates with the patient to design a pacing plan that sets modest, achievable daily activity targets, gradually increasing step count or standing time while ensuring scheduled rest breaks. Follow-up graphs from wearables allow both parties to see whether activity has become more consistent and whether symptom flare-ups decrease in intensity or duration as pacing improves.
For some patients, particularly adolescents or young adults, gamified activity tracking elements are purposefully harnessed. Clinicians may align rehabilitation goals with features such as daily “move rings,” streaks, or point systems. For example, a teenager with functional limb weakness working on returning to school might agree on goals related to standing time and number of classroom transitions, monitored through a smartwatch. Achieving these function-focused targets, rather than symptom disappearance, becomes a visible source of mastery. Parents and school staff can be involved in reviewing the data at predetermined intervals, helping to coordinate accommodations while avoiding excessive monitoring that could reinforce symptom preoccupation.
EMG-based biofeedback is often used in focused blocks of therapy for complex motor patterns such as fixed postures, gait disturbances with knee buckling, or functional dystonia-like symptoms. In a typical outpatient block, the physiotherapist places EMG electrodes over relevant agonist and antagonist muscles, then guides the patient through a sequence of functional tasks—such as sit-to-stand transfers, step-ups, or reaching while standing. The patient watches a simple display indicating relative muscle activation, learning to reduce excessive co-contraction and to recruit muscles in smoother, more graded ways. For example, a patient with recurrent knee “giving way” may discover that they habitually overactivate the quadriceps while underutilizing hip and trunk muscles, leading to a stiff, unstable pattern. With practice, and supported by EMG feedback, they experiment with alternative strategies that produce lower overall activation but greater stability. Follow-up without real-time feedback assesses whether new patterns generalize beyond the therapy setting.
Sleep and fatigue management represent another area where clinical teams use wearable data. Many individuals with FND report non-restorative sleep, frequent nighttime awakenings, and excessive daytime fatigue, all of which can worsen motor and cognitive symptoms. Actigraphy from wrist-worn devices offers an accessible way to approximate sleep timing, duration, and fragmentation across several weeks. Sleep medicine specialists or rehabilitation physicians can review actigraphy reports alongside symptom logs, often identifying patterns such as delayed sleep phase, irregular bedtimes across the week, or extended time in bed without consolidated sleep. Guided by these findings, clinicians implement behavioral sleep interventions—standardizing wake times, limiting daytime naps, and aligning graded activity with periods of greatest alertness—while continuing to monitor via the wearable to evaluate changes over time.
Case examples from community rehabilitation programs highlight how low-cost, patient-owned devices can extend care beyond specialist centers. A person living in a rural area with limited access to FND expertise might work with a telehealth team that uses their existing smartwatch and smartphone. The patient shares weekly screenshots of step counts, resting heart rate trends, and sleep estimates during virtual visits. Together, they set individualized goals, such as walking for brief periods three times per day or practicing breathing exercises when nocturnal awakenings are detected. Over several months, incremental changes in activity tracking and self-reported function can guide decisions about whether to intensify, maintain, or taper formal therapy, even in the absence of frequent in-person assessments.
Interdisciplinary pain and FND services sometimes combine multimodal wearables with structured group interventions. In such programs, participants may attend education and skills groups while wearing devices that monitor activity, heart rate, and occasionally electrodermal activity during sessions and at home. Group sessions include modules on stress physiology, pacing, and the role of attention in symptom amplification, with instructors periodically displaying anonymized group-level data patterns. For instance, participants might see a composite graph showing how average heart rate decreased over the course of a relaxation exercise or how activity levels increased slightly but steadily during a four-week graded exercise plan. These visualizations normalize variability between individuals while underscoring the broader principle that, as behaviors shift, physiology and symptoms can change as well.
Emerging clinical pilots demonstrate the potential of adaptive interventions driven by wearable data. In one model, a smartphone app linked to a smartwatch continuously monitors movement and heart rate, delivering brief prompts when certain thresholds are crossed. If the system detects extended inactivity during daytime hours in a patient whose rehab plan emphasizes gradual re-engagement, it may send a simple suggestion to stand, stretch, or complete a pre-agreed functional task. Conversely, if physiological markers of high arousal appear alongside reported stressors, the app may propose a short regulation exercise, echoing skills learned in therapy. Clinicians then review summary logs during appointments, adjusting thresholds and suggested activities based on the patient’s response and priorities.
Across these varied applications, a consistent theme is the need to integrate wearable-derived information into a coherent therapeutic narrative rather than treating numbers as definitive judgments of progress or failure. Clinicians often frame data as one piece of evidence among many, to be interpreted within the context of the patient’s subjective experience, goals, and life circumstances. When used collaboratively, wearables and biofeedback can enhance patients’ understanding of how thoughts, emotions, physiology, and behavior interact in FND, while offering concrete, trackable markers of change that make abstract rehabilitation concepts more tangible in everyday life.
Implementation challenges and ethical considerations
Integrating wearables and biofeedback into FND rehabilitation raises a number of practical and ethical challenges that must be addressed deliberately if these tools are to enhance, rather than complicate, care. On the implementation side, services face decisions about hardware selection, software platforms, data workflows, and how to embed device use into existing pathways without overwhelming clinicians or patients. Ethically, issues of privacy, consent, data ownership, equity, and the potential for unintended psychological effects require particular attention, especially in a population already vulnerable to stigma and misinterpretation of symptoms.
One major practical challenge is infrastructure. Even when consumer devices are relatively inexpensive, deploying them at scale demands systems for device provisioning, charging, pairing, and troubleshooting. Clinics need reliable procedures to ensure that sensors are correctly placed, synchronized, and calibrated, and that data transfer from wearables to clinical or research systems is secure and consistent. Fragmentation across proprietary platforms complicates this process: different manufacturers offer distinct data formats, dashboards, and analytic tools, which may not integrate well with electronic health records. Without deliberate planning, clinicians are left logging into multiple portals, manually exporting files, or relying on screenshots from patients’ phones, a workflow that is rarely sustainable in busy services.
Data overload compounds these logistical issues. Continuous activity tracking and physiological monitoring can generate thousands of data points per day per patient. Translating this volume into clinically meaningful insights is nontrivial. Rehabilitation clinicians may be presented with graphs of step counts, heart rate variability indices, sleep estimates, and symptom logs without clear guidance on how to interpret or act on specific patterns. In the absence of standardized thresholds linked to outcomes, there is a risk of either underutilizing the data—ignoring potentially useful information—or overinterpreting noise, such as normal fluctuations in heart rate or sleep. Implementation efforts therefore need to include decision-support tools, simple summary reports, and training that help clinicians integrate data into treatment planning without becoming de facto data analysts.
Training and role clarity pose further challenges. Neurologists, physiotherapists, psychologists, occupational therapists, and nurses may all interact with wearable-derived information, but their comfort with digital health tools and quantitative metrics varies widely. If expectations for device use are not well defined, some clinicians may feel pressured to incorporate technology in ways that do not align with their expertise or therapeutic style, while others may avoid using the tools altogether. Clear protocols that specify who is responsible for reviewing which data, at what intervals, and for what clinical decisions can prevent duplication of effort and reduce the sense that wearables add unstructured work to already busy caseloads.
Clinician trust in digital measures is also a key issue. Many FND practitioners are accustomed to basing decisions on detailed clinical interviews, behavioral observation, and standardized assessments. They may question the validity of consumer-grade sensors, particularly when readings appear inconsistent with observed function or patient report. Inaccurate heart rate readings, missed steps, or misclassified sleep can erode confidence quickly. Implementation strategies therefore benefit from transparent discussion of device limitations, realistic expectations about accuracy, and pilot phases where teams compare digital metrics with established measures, allowing them to understand typical discrepancies and their implications.
From the patient perspective, digital literacy, comfort with technology, and access to compatible devices vary considerably. While some people with FND are enthusiastic early adopters, others have limited experience with smartphones, data plans, or apps. Language barriers, cognitive symptoms, and fatigue can make onboarding and daily use of devices more difficult. If participation in wearable-based interventions implicitly assumes access to high-end phones or stable internet connections, there is a risk of excluding individuals from lower socioeconomic backgrounds or those living in rural areas. Equitable implementation requires providing loan devices when possible, designing low-bandwidth solutions, simplifying user interfaces, and offering support for setup and troubleshooting that does not rely solely on written instructions.
Equity concerns extend beyond access to technology. The algorithms underlying activity classification, heart rate variability estimation, or seizure detection are often developed and validated in populations that do not reflect the diversity of people with FND in terms of age, gender, skin tone, comorbidities, and movement patterns. For instance, optical sensors may perform less accurately in individuals with darker skin tones or extensive tattoos, and gait algorithms trained on typical walking patterns may misclassify or fail to capture the complex, variable movements seen in motor FND. This raises the ethical question of whether certain subgroups systematically receive less accurate feedback or less reliable alerts, potentially reinforcing health disparities. Transparency around algorithm limitations, efforts to validate tools in FND-specific cohorts, and inclusive research recruitment are important countermeasures.
Privacy and data security are central ethical considerations. Wearables collect granular data about movement, location, physiological states, and daily routines, creating detailed behavioral profiles. Storing and transmitting these data through commercial platforms introduces multiple points of vulnerability, including potential access by vendors, third-party analytics partners, or insurers. Patients may not realize that accepting a device’s default terms of service authorizes broader use of their data than is typical in healthcare. Rehabilitation services that incorporate commercial apps or cloud services need to evaluate vendor privacy policies carefully, ensure data are encrypted in transit and at rest, and clarify who can access raw and processed data. Where possible, using platforms that allow local data storage or health-system-controlled cloud environments can offer greater protection.
Informed consent for wearable use must go beyond a simple explanation that “we will track your steps” or “we will monitor your heart rate.” Patients should understand what types of data will be collected, how long they will be stored, who will have access, and for what purposes (clinical care, service evaluation, research, or product development). They should also be informed about foreseeable risks, such as data breaches, misinterpretation of metrics, or emotional distress from monitoring, and about their right to opt out without jeopardizing other aspects of their care. For individuals with FND, who may be particularly sensitive to perceived surveillance or invalidation, consent discussions should explicitly address how data will and will not be used—for example, that activity logs will not be used to “prove” or “disprove” the reality of symptoms, but rather to support functional goals.
Data ownership and secondary use present additional ethical dilemmas. In many digital health ecosystems, vendors reserve broad rights to use de-identified or even identifiable data for algorithm development, marketing, or commercial partnerships. Patients and clinicians may not fully appreciate the extent of these rights when they first adopt the technology. In rehabilitation contexts, this can be especially problematic when sensitive information about mental health, trauma history indirectly inferred from behavior patterns, or frequent functional attacks is used in ways that are misaligned with patient preferences. Clear institutional policies that define who owns the data, under what conditions they can be shared, and how de-identification is managed are critical, as is giving patients meaningful choices about participation in optional data-sharing for research or product improvement.
Ethical concerns also arise around the potential for wearables to inadvertently increase symptom focus, anxiety, or self-criticism. While many individuals experience biofeedback and activity tracking as empowering, others may become preoccupied with numbers, checking metrics repeatedly and interpreting normal variability as signs of deterioration or personal failure. A minor dip in daily steps or a transient reduction in heart rate variability, for example, can be misread as evidence that a coping strategy “isn’t working” or that they are “going backwards,” reinforcing catastrophizing and perfectionism. For some, alerts about elevated heart rate or increased electrodermal activity may actually trigger or intensify panic or dissociation, particularly if they are framed as “stress alarms” rather than neutral physiological information.
To mitigate these risks, clinicians need to carefully frame the role of data in therapy, emphasizing that metrics are tools for learning patterns over time rather than strict targets to be met every day. Discussions should normalize fluctuations, highlight broad trends instead of single data points, and explicitly explore emotional responses to the numbers. In some cases, it may be appropriate to limit the degree of real-time feedback available to the patient—such as hiding detailed dashboards and focusing on simple traffic-light indicators, or restricting access to raw data while clinicians review and summarize findings collaboratively during sessions. Screening for tendencies toward obsessive tracking, health anxiety, or perfectionism can guide decisions about whether a particular device or feature set is suitable for a given person.
The possibility of surveillance or coercive use of wearable data is another important ethical concern. Families, employers, insurers, or legal representatives may seek access to activity logs or physiological records in order to “verify” symptoms, monitor adherence, or challenge disability claims. When such pressures exist, patients may feel compelled to share data or to consent to monitoring as a condition of support, blurring the line between therapeutic and surveillance-oriented uses. Clinicians should be explicit that their primary obligation is to the patient’s well-being and confidentiality, and that data collected in the context of rehabilitation will not be shared with third parties without clear, voluntary consent except where mandated by law. Establishing boundaries around use of wearable data in medico-legal contexts, and discussing these openly with patients, can help maintain trust.
Algorithmic decision-making and automated alerts introduce their own set of ethical questions. Systems that aim to predict functional seizure episodes, detect “non-compliance” with activity goals, or classify mood states based on physiological proxies may produce false positives and false negatives with meaningful consequences. A missed alert could fail to warn a patient before a high-risk episode, while a false alarm might trigger unnecessary worry, emergency responses, or avoidant behavior. If systems are programmed to adjust therapy intensity or send motivational messages based on thresholds, they may inadvertently encourage overexertion in patients who are already prone to pushing themselves too hard, or they may misinterpret purposeful rest days as setbacks. Transparent communication about the fallibility of algorithms, and maintaining human oversight for significant decisions, are essential safeguards.
Responsibility for acting on data represents another gray area. Continuous monitoring can create an implicit expectation that clinicians are “watching” patients between sessions and will intervene if concerning patterns emerge. In reality, many services lack the capacity to review real-time data streams or to respond promptly outside scheduled contact. If boundaries around monitoring responsibilities are not clearly established, patients may assume that absence of outreach means everything is fine, or conversely feel abandoned when device flags are not followed by immediate action. Implementation protocols should therefore specify how frequently data are reviewed, what constitutes an actionable alert, and when patients should seek help through standard channels rather than relying on wearable systems.
In research contexts, additional ethical challenges arise when wearable data are used to derive biomarkers or predictive models for diagnosis and prognosis. Efforts to distinguish FND from other conditions, or to predict who will benefit most from certain interventions, may be valuable at a population level but can be distressing at the individual level if framed as definitive predictions. For example, if an algorithm suggests that a patient’s pattern of activity or heart rate variability is associated with “poor outcome” groups, there is a risk of self-fulfilling pessimism, therapeutic nihilism, or reduced investment in rehabilitation. Researchers and clinicians must take care in how probabilistic findings are communicated, emphasize uncertainty and modifiability, and avoid using early-stage digital markers as rigid gatekeepers for access to services.
Implementation also needs to grapple with regulatory and legal frameworks that have not fully caught up with the complexity of digital health tools. Some devices used for biofeedback or episode detection are regulated as medical devices, while others are marketed as wellness gadgets with less oversight. Clinicians must navigate questions about liability when recommending or relying on consumer products: what happens if a seizure-detection app fails to warn the patient, or if inaccurate activity tracking leads to inappropriate changes in rehabilitation intensity? Clear institutional guidance, documentation of shared decision-making about device use, and preference for tools with at least some evidence base in FND or related populations can reduce risk.
Resource allocation is an additional consideration. Investing in wearables, software licenses, and staff training may divert funds from other components of rehabilitation, such as face-to-face therapy time, group programs, or community-based supports. Services need to evaluate whether and how digital tools add value relative to their cost, not just in terms of clinical outcomes but also patient experience, staff workload, and long-term sustainability. Pilot programs with explicit evaluation metrics can inform decisions about scaling up or modifying approaches, rather than assuming that new technology is inherently beneficial.
A further challenge lies in aligning digital tools with the conceptual framework of FND rehabilitation. Models that emphasize re-establishing automatic, effortless movement and reducing excessive self-monitoring can appear at odds with interventions that encourage frequent checking of physiological signals or step counts. If not carefully integrated, wearables may inadvertently reinforce the very hypervigilance to bodily sensations that rehabilitation aims to reduce. Ethical implementation therefore requires congruence between the technology and therapeutic principles: devices should be used to shift attention toward functional goals, environmental engagement, and mastery of regulation skills, rather than fostering continuous symptom scanning.
Meaningful patient involvement in the design and deployment of wearable-based interventions can help address many of these challenges. Co-design approaches that invite people with FND, carers, and advocacy groups to shape device selection, interface design, feedback frequency, and data-sharing policies are more likely to surface concerns about stigma, burden, or misinterpretation early in the process. Patients can highlight what types of feedback feel empowering versus intrusive, which metrics they find understandable, and how often they wish to receive prompts or summaries. Incorporating this input into protocols and consent materials supports a more ethical, person-centered use of technology.
Across all these domains, a common thread is the need for explicit, ongoing conversation about the role of wearables and biofeedback in each person’s rehabilitation journey. Implementation is not a one-time technical deployment but a continuous ethical practice that adapts to changing circumstances, treatment phases, and preferences. Regularly revisiting whether the technology is helping or hindering, pausing or modifying use when it appears to increase distress or preoccupation, and ensuring that digital data are always interpreted in dialogue with the patient and the broader clinical picture are key strategies for navigating the complex terrain of digital FND rehabilitation responsibly.
Future directions and research priorities
Future work with wearables in FND rehabilitation is likely to revolve around developing more precise, clinically meaningful digital biomarkers that can be linked to specific mechanisms and treatment targets. Current measures such as step count, basic activity tracking, and average heart rate provide broad indicators of function, but they rarely capture the nuanced shifts in attention, intention, and arousal that characterize FND. Research priorities include defining composite indices that integrate movement features (for example, gait variability or tremor coherence) with autonomic markers such as heart rate variability and electrodermal activity, and then validating how these indices relate to symptom severity, functional capacity, and response to particular rehabilitation strategies.
To move in this direction, large-scale, multicenter cohorts will be essential. Single-center pilot studies have demonstrated feasibility, but they are typically underpowered to identify robust patterns or to build generalizable models across diverse FND presentations. Collaborative networks can standardize sensor configurations, sampling frequencies, and core outcome sets, enabling pooled analyses that test whether specific digital signatures are reproducible across settings and populations. For example, projects could examine whether certain patterns of daytime activity fragmentation, nocturnal restlessness, and low heart rate variability consistently predict poorer rehabilitation outcomes, or whether distinct movement and autonomic profiles differentiate subgroups that benefit most from intensive motor retraining versus those who respond better to autonomic regulation–focused interventions.
The field also needs methodologically rigorous trials that evaluate wearables and biofeedback not simply as add-ons, but as structured components of treatment packages. Many existing studies treat devices as background measurement tools, without clearly specifying how data influence therapeutic decisions. Future randomized and adaptive trials could compare standard FND rehabilitation with and without integrated wearable-informed decision support, examining whether data-guided titration of activity, real-time regulation prompts, or feedback on functional goals produce superior functional outcomes, reduced healthcare utilization, or more durable gains. Such designs should include long-term follow-up to assess whether benefits persist after device use is tapered or discontinued.
Another priority is the development and validation of adaptive, just-in-time interventions that respond dynamically to momentary states rather than delivering generic prompts. This requires advances in real-time signal processing and context modeling. Algorithms must distinguish, for example, between elevated heart rate due to normal exertion and similar values arising from escalating anxiety or pre-episode arousal, and then select appropriate micro-interventions accordingly. Research programs can experiment with combining passive sensing (movement, autonomic signals) with brief ecological momentary assessments to infer context, testing different rules for when and how to deliver suggestions for movement, rest, grounding, or cognitive reframing. Measuring not only symptom change but also acceptability, perceived intrusiveness, and adherence will be critical to refining these systems.
Personalization is likely to be a central theme in future directions. FND is heterogeneous, and the same digital feedback that empowers one person may overwhelm another. Rather than relying on one-size-fits-all protocols, research should investigate how to tailor device choice, feedback modality, and intensity of monitoring to individual profiles. This could involve building preference-sensitive algorithms that learn over time which prompts a given person tends to accept or ignore, which sensor-derived patterns precede their symptom flares, and what types of feedback (graphical, auditory, haptic, or text-based) they find most helpful. Studies can compare fixed versus adaptive personalization strategies, assessing not only symptom outcomes but also engagement trajectories and time-on-task with home exercises.
There is also a need to expand the range of clinically relevant variables captured by wearables beyond gross motor and basic autonomic measures. Cognitive and affective dimensions—such as sustained attention, cognitive load, and mood reactivity—play key roles in FND but are rarely assessed continuously. Future devices and analytic methods may infer proxies for these states through combinations of keystroke dynamics, speech features, micro-movement patterns, and fluctuations in heart rate variability. Feasibility studies can explore whether such multimodal sensing provides added value for identifying periods of high vulnerability to dissociation, functional seizures, or motor exacerbations, and whether timely interventions during those windows alter trajectories.
At the same time, the field must refine how outcomes are defined and measured in trials involving digital tools. Traditional endpoints such as symptom frequency or global impression of change are important but may not fully capture the functional gains and shifts in self-efficacy that clinicians observe when biofeedback and activity tracking are well integrated. Consensus-building efforts should prioritize outcome sets that include objective functional indices (for example, work participation, school attendance, or community mobility), patient-reported measures of agency and confidence in self-management, and digital metrics such as stability of daily routines or reduction in extreme peaks and troughs of activity. Harmonizing these outcomes across studies will facilitate meta-analyses and translational work from research to routine services.
A related research agenda involves understanding mechanisms of change in digital-supported FND rehabilitation. It is still unclear, for instance, whether improvements associated with wearable use are primarily driven by increased awareness of functional capacity, by reinforcement of graded exposure principles, by enhanced regulation of autonomic arousal, or by nonspecific factors such as novelty and attention from clinicians. Mechanistic studies that combine qualitative interviews with quantitative modeling can examine how patients interpret and use feedback, how their narratives about symptoms and control evolve over time, and how these shifts correlate with changes in sensor-derived variables. Experimental designs might manipulate specific components—for example, comparing real-time physiological biofeedback with sham or delayed feedback—to isolate the role of contingent information versus general engagement.
Another important direction is the co-development of FND-specific algorithms rather than repurposing tools built for other conditions. Activity classifiers, seizure detectors, and fall-detection models are often trained on populations with epilepsy, Parkinson’s disease, or healthy volunteers, which may not capture the distinctive variability and context dependence of functional symptoms. Partnerships between clinicians, data scientists, and people with FND can inform labeling schemes that make sense for this population—for instance, differentiating functional attacks from panic episodes, or distinguishing helpful rest from avoidance-related inactivity. Dataset curation must ensure that labels are grounded in careful clinical assessment, ideally with multimodal confirmation (video, EEG where relevant, therapist observation) to avoid baking diagnostic uncertainty into models.
Looking ahead, integration of wearable data with neuroimaging, neurophysiological measures, and psychometric assessments offers an avenue for more comprehensive, multi-level models of FND. Longitudinal studies could investigate how changes in resting-state connectivity, cortical excitability, or task-related activation patterns relate to evolving daily-life patterns captured by wearables. For example, improvements in automaticity of movement observed in functional MRI paradigms might correspond with reduced variability in gait parameters and less reliance on visual checking as detected by inertial sensors. Such convergent evidence can strengthen theoretical frameworks linking brain network dynamics, bodily regulation, and behavior, ultimately informing refinements to both digital and non-digital aspects of rehabilitation.
Technological innovation will also need to focus on improving comfort, aesthetics, and integration into everyday life. Bulky or conspicuous sensors can interfere with social participation, draw unwanted attention, or remind patients of illness at times when they are trying to shift focus toward valued activities. Research into skin-integrated electronics, textile-based sensors, and low-profile form factors could reduce these barriers. Parallel efforts in user experience design should explore interfaces that minimize cognitive load and avoid medicalized imagery, perhaps embedding feedback within general wellness or productivity apps in ways that feel less stigmatizing while still delivering clinically meaningful information.
Interoperability and data standards represent additional areas for development. At present, many rehabilitation services are constrained by ecosystems in which each device produces proprietary output, limiting cross-platform comparisons and long-term archiving. Future work should prioritize the creation and adoption of open data standards specific to rehabilitation and FND, allowing raw and derived metrics to be stored, shared (with appropriate safeguards), and reanalyzed as analytic methods evolve. Collaborative consortia could maintain shared repositories of anonymized, well-annotated datasets that support benchmark challenges—for instance, competitions to develop the most accurate models for predicting functional seizure onset or for classifying activity patterns associated with successful pacing—accelerating innovation while maintaining transparency about performance.
On the service-delivery side, research should evaluate scalable implementation models that can be adapted to varied resource settings, from tertiary academic centers to community clinics and telehealth-only programs. Implementation science frameworks can guide studies that test different strategies for training staff, integrating digital information into multidisciplinary meetings, and clarifying responsibilities for monitoring and responding to data. Comparative evaluations might explore, for example, whether centralized data review by a small digital health team provides better support and clinician satisfaction than decentralized review by individual therapists, or whether automated summary reports embedded in electronic health records are more acceptable than standalone dashboards.
The interplay between digital equity and FND rehabilitation warrants ongoing investigation. Future studies should systematically assess who is and is not benefiting from wearable-based interventions, analyzing differences by socioeconomic status, race and ethnicity, age, rural versus urban residence, and comorbid conditions. Mixed-methods research can identify structural barriers (such as limited connectivity or competing caregiving responsibilities) and inform policy-level solutions like provision of loan devices, subsidized data plans, or community-based digital literacy programs. Ensuring that wearables do not widen existing disparities but instead support more inclusive, accessible care is likely to be a key determinant of their long-term role in FND services.
There is a need for sustained, participatory ethics research alongside technological and clinical innovation. Questions about acceptable levels of monitoring, data retention durations, and use of predictive models in triage or resource allocation cannot be resolved solely through top-down guidelines. Deliberative forums, citizen juries, and ongoing advisory panels that include people with FND, clinicians, ethicists, legal experts, and technologists can help shape norms around what counts as respectful, non-coercive digital care in this context. Empirical studies that track how patients’ attitudes toward monitoring change over time, what kinds of consent models they prefer, and how they weigh trade-offs between convenience, privacy, and perceived safety will provide a crucial evidence base for refining both practice and policy as the role of wearables in FND rehabilitation continues to evolve.
