Extended Reality in Rehabilitation Medicine: A Systematic Review Authors: Jie Hao, Yao Yao, Ka‐Chun Siu Published: 2026 ABSTRACT ABSTRACTBackground and AimsThis systematic review synthesizes current evidence on the application of extended reality (XR) technologies, comprising virtual reality (VR), augmented reality (AR), and mixed reality (MR), in rehabilitation medicine, assessing their effects across diverse rehabilitation domains and clinical outcomes.MethodsA comprehensive search of three biomedical databases was conducted, yielding 58 randomized controlled trials that examined XR interventions in geriatric, neurological, orthopedic, and pediatric rehabilitation. Methodological quality was assessed using the Physiotherapy Evidence Database scale. Results were qualitatively synthesized according to intervention characteristics and outcome measures.ResultsThe findings indicate that XR technologies can enhance rehabilitation outcomes, including improvements in motor and cognitive function, pain management, psychological well‐being, and quality of life. XR interventions were shown to improve balance, strength, and cognitive and psychological function in older adults; facilitate motor recovery and reduce neuropathic pain in neurological populations; support postoperative rehabilitation and conservative interventions in orthopedic patients; and enhance engagement and therapeutic goals attainment in pediatric patients.ConclusionXR technologies present a promising and innovative approach to rehabilitation, offering personalized and engaging therapeutic experiences. Clinicians should leverage XR technologies to empower patients and optimize clinical outcomes.Keywords: function, health, immersion, physiatry, simulation, therapy Introduction The rapid evolution of science, technology, and engineering has revolutionized human society and catalyzed significant innovations in healthcare. Artificial intelligence and simulation technologies are increasingly being developed and adopted in clinical research and practice to improve access and quality of healthcare. Among them, extended reality (XR) is a burgeoning field. XR is defined as the spectrum of computer‐generated immersive technologies, which encompasses virtual reality (VR), augmented reality (AR), and mixed reality (MR) [1]. VR allows users to be completely immersed in a simulated environment by using an interactive device and a head‐mounted display. AR superimposes digital components to the real world, and users can view and interact with virtual objects and information as if they were in the real‐world. MR connects the physical and virtual worlds and allows digital objects to interact with their surroundings. According to a report by the International Market Analysis Research and Consulting Group [2], the global market size of XR in healthcare reached USD 6.0 billion in 2023 and is expected to have a compound annual growth rate of 29.02% from 2024 to 2032. Rehabilitation medicine aims to restore function, reduce disability, and improve the quality of life in individuals with various health conditions. People with cardiovascular, musculoskeletal, neurological, or other chronic diseases may suffer from short‐ to long‐term impairments and functional deficits that impede independence in personal life and social participation. Rehabilitation is an essential component of healthcare, and it was estimated that 2.4 billion people with disabilities have demands for rehabilitation [3]. In addition, in light of the global aging population with an increased burden of disability related to chronic diseases, the need for rehabilitation is projected to grow in the next decades [4]. Given the growing and unmet needs for rehabilitation [5], it is imperative to incorporate the advances in the state‐of‐the‐art technologies into rehabilitation medicine to optimize function and promote the well‐being of people with disabilities. In rehabilitation medicine, XR technologies provide a unique platform for enhancing physical and cognitive recovery. Conventional rehabilitation relies on intensive physical, occupational, and speech therapies which involve repetitive therapeutic exercises or task‐specific training delivered by therapists. For many patients and their families, participating in rehabilitation programs can be a long‐term and tedious process. XR holds the promise to transform conventional rehabilitation paradigms by enabling patients to perform a variety of interactive activities that are customized to patients’ needs in an immersive environment [6]. Patients can receive real‐time multimodal feedback on their performance, facilitating skill acquisition and functional recovery. By creating a novel rehabilitation experience, XR can increase patients’ motivation and engagement, improving adherence to therapeutic regimens [7]. Therapists can leverage XR's capacity to produce adaptable and patient‐specific environments to foster the efficiency and effectiveness of rehabilitation interventions. Besides, XR plays an instrumental role in pain management by diverting patients’ attention from painful movements or procedures to relaxing and aesthetic environments [8], improving patients’ comfort and tolerance to rehabilitation. Although VR, AR, and MR differ in their levels of immersion, interaction with the physical environment, and technological implementation, they can be conceptualized as part of the broader XR continuum. These technologies share common therapeutic principles relevant to rehabilitation, including immersive visualization, interactive task‐based training, and the provision of real‐time multimodal feedback to facilitate motor and cognitive recovery. Synthesizing VR, AR, and MR within a unified XR framework allows for a more comprehensive understanding of how immersive technologies are applied across diverse rehabilitation contexts. Recent literature has reviewed the application of XR to various medical specialties, including cardiology [9], spine medicine [10], radiology [11], ophthalmology [12], orthopedic surgery [13], neurosurgery [14], and primary care [15]. However, no reviews have systematically investigated how XR can be used in rehabilitation medicine. This systematic review aims to identify, critically appraise, and synthesize current evidence regarding the application of XR technologies in rehabilitation medicine, with a focus on clinical outcomes, such as motor and cognitive function, mood, pain reduction, functional recovery, and quality of life. By synthesizing the available literature, this review can provide insights into the therapeutic potential of XR, and inform future development and implementation of XR in rehabilitation medicine. Methods Reporting of this systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta‐Analysis (PRISMA) guidelines [16]. The protocol of this systematic review was registered at the International prospective register of systematic reviews (PROSPERO): CRD42024604134. Three bibliographic databases, PubMed, Embase, and Scopus, were searched using strategies designed for each database. The search strategies of PubMed are described in the Appendix. Databases were searched from inception to December 30, 2024. The search strategies combined controlled vocabularies and title/abstract keywords on search themes of XR, VR, AR, MR, and rehabilitation. The literature search was limited to human studies reported in English in peer‐reviewed journal articles. Records retrieved from each database were then exported to Covidence software (Veritas Health Innovation Ltd., Melbourne, Australia) for duplication removal. The inclusion criteria of this systematic review were generated based on the PICOS framework [17]: (1) Population: patients with a disability or functional deficit who are candidates for rehabilitation services; there were no restrictions on age, gender, or diagnosis; (2) Intervention: XR technologies (VR, AR, or MR) were used for rehabilitation; (3) Comparison: conventional rehabilitation, usual care, or waitlist control; (4) Outcome: performance‐based or patient‐reported outcome measures related to motor, sensory, cognitive, or perceptual function; psychological outcomes; activities of daily living ability and functional independence; disability; quality of life; pain; or biomechanical parameters; and (5) Study design: longitudinal experimental studies designed as randomized controlled trials. Studies were excluded in this systematic review if they met any of the following criteria: (1) they were evaluated as fair or poor in quality assessment; (2) study design was a qualitative study, case study, cross‐sectional laboratory‐based study, or research protocol; or (3) study was conference proceeding or abstract‐only. Two reviewers independently screened titles and abstracts for potential eligibility and reviewed full‐text articles to determine final eligibility. For each included study, they extracted data on participant characteristics, sample size, intervention in the experimental and control groups, outcome measures, and main findings. Throughout this process, any discrepancies were resolved through group discussion with a third senior reviewer until a consensus was reached. For each randomized controlled trial, the risk of bias was assessed by the Physiotherapy Evidence Database (PEDro) scale, a valid and reliable tool used to evaluate the methodological quality of clinical trials in rehabilitation medicine literature to inform clinical decision‐making [18]. There are 11 items on this scale: Item 1 is classified as “yes or no,” and items 2–11 each count for one point, resulting in a total score range of 0–10. Higher scores indicate better study quality, with scores of 9–10 commonly interpreted as reflecting excellent, 6–8 as good, 4–5 as fair, and 0–3 as poor quality [19]. Two reviewers independently assessed the quality of each included study and reached a consensus with the third reviewer on any identified discrepancies in their rating. We decided to exclude studies with low‐quality (fair or poor as assessed by the PEDro) from data synthesis in priori to avoid the risk of bias. We qualitatively evaluated clinical and methodological heterogeneity among included studies and the risk of biases that could impact the interpretation of the evidence. We qualitatively synthesized included studies according to domains of rehabilitation medicine (geriatric, neurological, orthopedic, and pediatric rehabilitation) and types of outcome measures to summarize the effects of XR interventions on clinical outcomes in different populations. We decided a priori not to perform a meta‐analysis due to substantial heterogeneity across studies. Specifically, heterogeneity was observed in XR intervention characteristics (e.g., type of technology, hardware platforms, and training protocols), intervention dosage and duration, patient populations with diverse clinical conditions and severity levels, and outcome measures spanning motor, cognitive, psychological, and functional domains. Results A total of 6175 records were identified from the three bibliographic databases. After removing duplicates, 4499 remaining records were screened by titles and abstracts. Subsequently, 164 full‐text articles were assessed for eligibility. Of these, 106 studies were excluded due to wrong intervention (n = 85), wrong outcomes (n = 7), wrong study design (n = 3), and inadequate methodological quality (n = 11). Ultimately, 58 articles were included in this systematic review. The PRISMA flow diagram (Figure 1) delineates the study identification process and reasons for excluding articles. PRISMA 2020 flow diagram for new systematic reviews which included searches of databases and registers only. The 58 included randomized controlled trials report the application of XR in geriatric, neurological, orthopedic, and pediatric rehabilitation. In particular, neurological conditions include Parkinson's disease, spinal cord injury, stroke, and vestibular dysfunction. Orthopedic conditions include neck pain and post‐operative rehabilitation. Among the 58 studies, 42 used VR, 13 used AR, and 3 used MR. XR technologies were incorporated into rehabilitation programs for upper extremity, lower extremity, balance and postural control, mobility and gait, kinematic, cognitive, perceptual, and vestibular‐specific training. For outcome measures, performance‐based outcomes included measures such as range of motion, balance ability, postural and gait parameters, and cognitive and perceptual performance, whereas patient‐reported outcomes included pain, psychological distress, self‐efficacy, balance confidence, and self‐reported health and functional status. Half of these studies assessed the occurrence of adverse events, and 10 reported adverse events related to XR, including headache, dizziness, motion sickness, fatigue, and eye symptoms. Lee et al. [20] reported a participant with an ischemic stroke due to vertebral artery had a transient ischemic attack in the AR group after 3 weeks of training. The characteristics of the included studies are summarized in Table 1. Table 1Characteristics of the included studies.ArticleConditionType of XRInterventionComparisonOutcome measuresMain findingsAdverse eventsGeriatric rehabilitationAppel [21]Patients with dementia in acute careVRVR films to promote nature visualization and well‐beingStandard of care (medical and nursing management)Behavioral and psychological symptom of dementiaVR demonstrates significant effects on aggressive behaviors (physical aggression and loud vociferation) in patients with dementia in acute care.No adverse events.Falls n = 69Oculus Go deviceUp to 20 min every 1–3 daysLength of stayQuality of lifeCampo‐Prieto [22]Institutionalized older adultsVRVR exergamesRoutine care onlyBalance and gaitVR is a feasible method to approach a personalized exercise program and an effective way to improve physical function.No adverse events.Mobility and lower extremity strength n = 24HTC Vive6 min × 3 sessions × 10 weeksGrip strengthQuality of lifeCybersicknessCampo‐Prieto [23]Nonagenarian women in a senior centerVRVR exergamesRoutine therapy programsTinetti TestVR is feasible and effective to enhance balance and to reduce the risk of falls in community‐dwelling nonagenarian women.No adverse events.Timed UP and Go Test n = 12HTC Vive6 min × 3 sessions × 10 weeks45 min × 5 sessions/weekSimulator Sickness QuestionnaireRoutine therapy programsSystem Usability Scale45 min × 5 sessions/weekJeon [24]Elderly womenARAR‐based exercise programWaitlist controlSenior Fitness TestThe AR‐based exercise program is effective increasing the sustainability of exercise, thus preventing muscle reduction in the elderly.Not reported. n = 2730 min × 5 sessions × 12 weeksExercise Self‐efficacy ScaleKu [25]Community‐dwelling older adultsARAR‐based exercise programConventional physical fitness programBalance and mobilityThe AR‐based exercise program can improve balance in the elderly more effectively.Not reported.Modified Barthel Index n = 363D‐ARS system30 min × 3 sessions × 4 weeks30 min × 3 sessions × 4 weeksFugl‐Meyer Assessment of lower extremityPosturography assessmentLima [26]Older adults with balance deficitsVRVR exergames for balanceConventional balance trainingDynamic Gait IndexVR training is effective for balance‐related outcomes, although it is not superior to conventional therapy.No adverse events.Clinical Test of Sensory Interaction and BalanceTimed Up and Go Test n = 37Oculus RiftBoxVR, Baskhead, InCell, and Thrills and Chills Roller Coasters50 min × 2 sessions × 8 weeksFunctional Reach Test50 min × 2 sessions × 8 weeksDizziness Handicap InventoryOptale [27]Older adults with memory deficitsVRVR memory trainingMusic therapyGlobal cognitionVR demonstrates significant improvements in memory tests, especially in long‐term recall, and in several other aspects of cognition.Not reported.Verbal memoryExecutive function n = 3630 min × 2–3 sessions/week × 3 months30 min × 2–3 sessions/week × 3 monthsVisuospatial processingActivities of daily living, depressionPark [28]Older adults with mild cognitive impairmentMRMR‐based cognitive trainingComputer‐assisted cognitive trainingCognitive functionMR‐based cognitive training system can be used as a cognitive training tool to improve visuospatial working memory.Not reported.Depression n = 2130 min × 3 sessions × 6 weeks30 min × 3 sessions × 6 weeksActivities of daily livingQiu [29]Community‐dwelling older adultsVRVR‐enhanced Tai ChiTraditional Tai ChiGeriatric Depression ScaleVR group exhibited a more substantial reduction in depression compared to the traditional Tai Chi group.Not reported. n = 300Oculus Quest 2One session/week × 6 monthsOne session/week × 6 monthsZhu [30]Community‐dwelling older adultsARAR memory trainingHealth science program watchingWorking memory assessmentsAR memory training could improve working memory of Chinese older adults, showing potential for wider promotion and adoption.Not reported.Digit Span Test30 min × 3 sessions × 4 weeks60 min × 3 sessions × 4 weeksCorsi Block‐Tapping Task n = 180Health science program watchingStroop Color and Word Test30 min × 3 sessions × 4 weeksNeurological rehabilitation: Parkinson'sGulcan 2022Patients with Parkinson's diseaseARAR gait and balance trainingConventional physical therapyMotor symptomsAR gait training provides the opportunity to practice walking with different tasks in increasingly difficult environments, thus improving balance and walking by facilitating motor learning.Not reported.Berg Balance Scale, posturography n = 30Motek C‐Mill3 sessions × 6 weeks3 sessions × 6 weeksActivity‐specific balance confidence scaleSpatiotemporal gait analysisTimed Up and Go TestSchuch [31]Patients with Parkinson's diseaseVRVR balance trainingPsychoeducationTimed Up and Go TestVR‐based balance training is a feasible and safe rehabilitation activity. However, it was insufficient to promote improvement in mobility smoothness and cognitive function.No adverse events.10‐meter walk test n = 2320 min × 2 sessions × 5 weeks20 min × 2 sessions × 5 weeksRivermead Behavioral Memory TestState‐trait anxiety inventoryNeurological rehabilitation: Spinal cord injuryAn [32]Patients with incomplete spinal cord injuryVRVR soccer game while sitting on a wheelchairSeated lower extremity exercises on a wheelchairMuscle strength, range of motionVR rehabilitation improves both dynamic and static balance and reduces the risk of falls.Not reported.Chair Stand Test n = 4030 min × 3 sessions × 4 weeks30 min × 3 sessions × 4 weeksTimed Up and Go Test10‐meter walk testAustin 2020Men with spinal cord injury and neuropathic painVRVR experience: Nature TrekSame scene shown on a laptop screenPain intensityVR demonstrates significantly reduced pain intensity and higher perceived levels of presence compared to a 2D screen.No adverse events.Depression Anxiety Stress Scale n = 16Oculus Rift15 min15 miniGroup Presence QuestionnaireGoel 2021Patients with incomplete spinal cord injuryVRVR exergames focusing on trunk movementsFunctional electrical stimulation on erector spinae and rectus abdominisModified functional reach testVR as an adjunct to conventional physical therapy is proved to be an effective treatment to improve balance among individuals with incomplete paraplegia.Not reported.Function in sitting test n = 1845 min × 5 sessions × 4 weeks45 min × 5 sessions × 4 weeksSpinal Cord Independence Measure IIIConventional physical therapyConventional physical therapyLim [33] Patients with incomplete spinal cord injury VR VR upper extremity trainingConventional occupational therapyMuscle strength test of the upper extremityVR training of upper limb function can provide an acceptable adjunctive rehabilitation method without significant adverse effects.No adverse events.Hand Strength Test30 min × 3 sessions × 4 weeksBox and Block Test n = 20HTC ViveConventional occupational therapy60 min × 3 sessions × 4 weeksNine‐Hold Peg Test30 min × 3 sessions × 4 weeksAction Research Arm TestNeurological rehabilitation: StrokeAmin [34]Patients with subacute strokeVRVR hand exercisesConventional physical therapyFugl‐Meyer Assessment ‐ Upper ExtremityVR‐based games intervention is effective in substantially sustaining the motor recovery of subacute patients.Not reported.Action Research Arm TestBox and Block Test n = 52Oculus Quest 2Conventional physical therapy 4 sessions × 6 weeks4 sessions × 6 weeksModified Barthel IndexQuality of lifeChoi [35]Subacute stroke patients with unilateral spatial neglectVRVR spatial neglect trainingConventional rehabilitationLine bisection testVR demonstrates significantly greater improvements in the line bisection test and visual perception tasks.Not reported.Catherine Bergego ScaleModified Barthel Index n = 24Oculus Rift DK230 min × 3 sessions × 4 weeks30 min × 3 sessions × 4 weeksVisual perceptual tasksHorizontal head movementsCrosbie [36]Patients with chronic strokeVRVR upper extremity rehabilitationConventional rehabilitationMotricity IndexBoth groups demonstrate small, but non‐significant, changes to their arm impairment and activity levels.No adverse events. n = 1830–45 min × 3 sessions × 3 weeks30–45 min × 3 sessions × 3 weeksAction Research Arm TestDeff 2013Patients with chronic strokeMRMR upper extremity rehabilitationConventional rehabilitationKinematic assessmentsBoth groups demonstrate significant improvements in upper extremity functional scales; only the MR group demonstrates consistent improvement in kinematic measurements.Not reported.Motor Activity Log n = 21Adaptive Mixed Reality Rehabilitation system1 h × 3 sessions × 4 weeks1 h × 3 sessions × 4 weeksStroke Impact ScaleWolf Motor Function TestFugl Meyer Assessment—Upper ExtremityEvron [37]Patients with subacute strokeARAR mobility exercisesConventional therapyTimed Up and Go TestAR group demonstrates better score on the Timed Up and Go test compared to the control group. Participants reported high satisfaction with the AR system.Not reported.10‐meter walk test6 sessions15 sessionsDynamic Gait Index n = 25Selfit systemConventional therapyActivities‐specific balance confidence scale9 sessionsHegazy [38]Patients with subacute strokeVRVR upper extremity rehabilitationTask‐oriented trainingUpper extremity functional indexVR group demonstrates significantly improved upper limb function and grip strength compared to the control.Not reported.10–15 min × 3 sessions × 6 weeks1 h × 3 sessions × 6 weeksGrip strength n = 20Task‐oriented training1 h × 3 sessions × 6 weeksHsu [39]Patients with chronic strokeVRVR mirror therapy for hand functionGroup 1: mirror therapyFugl‐Meyer Assessment—Upper ExtremityThe results partially support the VR mirror therapy has better effects on enhancing the motor function of the affected upper limb in stroke participants.No adverse events.Motor activity logBox and Block Test n = 5430 min × 2 sessions × 9 weeksGroup 2: conventional therapyModified Ashworth Scale30 min × 2 sessions × 9 weeksMonofilament TestHuang [40]Patients with chronic strokeVRVR upper extremity rehabilitationConventional rehabilitationFugl‐Meyer Assessment—Upper ExtremitySignificant improvements in overall upper limb function and range of motion were found in both groups; only elbow extension and forearm pronation were significantly different between groups.The most common adverse symptoms were eye strain (46.67%) and sweating (46.67%). Most symptoms were mild, except for eye strain and blurred vision, which were reported to cause moderate discomfort.Upper extremity range of motion n = 30HTC Vive1 h × 16 sessions1 h × 16 sessionsSerum biomarkersJo 2024Patients with subacute strokeVRVR mirror therapyGroup 1: Mirror therapyFugl‐Meyer Assessment—Upper ExtremityVR mirror therapy demonstrates significantly greater improvements in upper extremity function compared to mirror therapy and conventional therapy.No adverse events.Manual function Test30 min × 3 sessions × 4 weeksBox and Block Test30 min × 3 sessions × 4 weeksConventional therapyPatients’ experience and satisfaction n = 45Group 2: Conventional therapy onlyConventional therapy60 min × 5 sessions × 4 weeks60 min × 5 sessions × 4 weeksKang 2011Patients with chronic strokeVRVR optic flow during treadmill trainingTreadmill trainingTimed Up and Go TestVR speed‐modulated optic flow treadmill training significantly improves balance and gait abilities compared to treadmill training and conventional therapy.Not reported.Functional Reach Test30 min × 3 sessions × 4 weeks10‐meter walk test n = 3030 min × 3 sessions × 4 weeksConventional therapy6‐min walk test30 min × 3 sessions × 4 weeksLee [41]Patients with chronic strokeARAR postural control trainingConventional therapyBerg Balance ScaleAR postural control training demonstrates significant improvements in several gait variables (gait velocity, step length, and stride length) compared with conventional therapy only.Not reported.Timed Up and Go Test30 min × 3 sessions × 4 weeks30 min × 5 sessions × 4 weeksSpatiotemporal gait analysis n = 21Conventional therapy30 min × 5 sessions × 4 weeksLee [20]Patients with strokeARAR home exercise programWritten and pictorial home exercise programTimed Up and Go TestAR‐based home exercise program might be effective to improve balance in patients with stroke. However, more careful precautions are also required.A participant with ischemic stroke due to vertebral artery dissection (enrolled 6 days after stroke onset) had a transient ischemic attack in the AR group after 3 weeks.Berg Balance ScaleTinetti Test n = 68UINCARE Home4 weeks4 weeksFall efficacyQuality of lifeLi 2021Patients with subacute strokeARAR motor and cognitive trainingConventional occupational therapyFugl‐Meyer Assessment—Upper ExtremityCombined AR and conventional OT rehabilitation is more effective in improving both upper limb function and cognitive function.Pain and fatigue in the shoulder (n = 3) and elbow (n = 2)Action Research Arm TestModified Barthel Index n = 30Cellphone AR rehabilitation system30 min × 5 days × 2 weeks60 min × 5 days × 2 weeksCognitive functionConventional occupational therapyUser Satisfaction Evaluation Questionnaire30 min × 5 days × 2 weeksLiu 2023Patients with chronic strokeVRVR puzzle gamesCognitive trainingMontreal cognitive assessmentVR puzzle games are a promising approach to improve post‐stroke cognitive function; especially executive function visual–spatial attention.Dizziness (n = 2), dry eyes (n = 2), and eye fatigue (n = 3)Trail‐making test‐A, digit symbol substitution test, digital span test, verbal fluency test n = 3015 min × 6 sessions × 6 weeks15 min × 6 sessions × 6 weeksModified Barthel IndexMekbib [42]Patients with strokeVRVR upper extremity rehabilitationConventional occupational therapyFugl‐Meyer Assessment—Upper Extremity Both groups significantly improved Barthel Index. The VR group demonstrated significant improvements on upper extremity motor function than the control group. Not reported.1 h × 4 days × 2 weeks2 h × 4 days × 2 weeksBarthel Index n = 23HTC ViveConventional occupational therapy1 h × 4 days × 2 weeksOgun [43]Patients with chronic strokeVRVR upper extremity rehabilitationSham VRFugl‐Meyer Assessment—Upper ExtremityVR rehabilitation appeared to be effective in improving upper extremity function and self‐care skills, but it did not improve functional independence.Not reported.15 min × 3 sessions × 6 weeksAction Research Arm TestConventional therapy15 min × 3 sessions × 6 weeksFunctional Independence Measure n = 6545 min × 3 sessions × 6 weeksConventional therapyPerformance Assessment of Self‐Care Skills45 min × 3 sessions × 6 weeksPelaez [44]Patients with chronic strokeVRVR exergamesConventional rehabilitationBerg Balance ScaleVR demonstrates significantly greater improvements in balance and mobility, functionality, trunk control, and ambulation compared to the control.Not reported.Tinetti TestMotricity Index n = 24Oculus Quest 21 h × 3 sessions × 6 weeks1 h × 5 sessions × 6 weeksMuscle strength and spasticityFunctional AmbulationConventional rehabilitationClassification1 h × 5 sessions × 6 weeksShin [45]Patients with unilateral spatial neglectVRVR visual exploration therapyConventional rehabilitationLine Bisection TestVR visual exploration therapy significantly improves unilateral spatial neglect.No adverse events.Catherine Bergego Scale n = 11Oculus Rift DK15 sessions × 4 weeksStar Cancellation TestConventional rehabilitationStammler [46]Patients with unilateral spatial neglectARAR active exploration therapyStandard neglect therapyLine Bisection TestAR training demonstrated significantly better outcomes than standard neglect therapy; the effects remained stable 1–2 months after the end of treatment.Not reported.Bells Test n = 2025 min × 5 sessions × 2 weeksSmooth pursuit eye movement trainingLetter Cancellation TestVisual exploration therapyCopying taskYang [47]Patients with subacute strokeARAR gait adaptation trainingTreadmill gait training10‐meter walk testBoth groups improved mobility outcomes; AR gait adaptation demonstrated greater improvement in obstacle avoidance, balance, turning, and stand‐to‐sit.No adverse events.Obstacle avoidanceBerg Balance Scale n = 40Motek C‐Mill35 min × 3 sessions × 5 weeks35 min × 3 sessions × 5 weeksTimed Up and Go TestFall incidenceNeurological rehabilitation: Vestibular dysfunctionMiracarelli 2017Patients with unilateral vestibular hypofunctionVRVR vestibular exercises, a combination of adaptation, substitution, habituation, and balance exercisesVestibular rehabilitationVideo head impulse testingVR could be proposed as an innovative, ecological, self‐assessing, low‐cost, and compliant tool, useful in maximizing vestibular rehabilitation outcomes.Not reported.Static posturography testingDizziness Handicap Inventory n = 4730–40 min daily × 4 weeks30–45 min × 2 sessions × 4 weeksActivities‐specific Balance Confidence ScaleVestibular rehabilitationDynamic Gait Index30–45 min × 2 sessions × 4 weeksOzdil [48]Patients with Benign Paroxysmal Positional VertigoVRVR vestibular exercisesCanalith repositioning maneuver only10‐meter walk testVR vestibular exercises are effective in improving gait speed, balance, processing speed and subjective complaints than canalith repositioning maneuver only.The majority of individuals in the VR group reported symptoms of nausea, vertigo, and lightheadedness during the initial 1–6 sessions.Gait speed with vertical and horizontal head turnsDynamic Gait Index n = 22PlayStation 4VR50 min × 3 sessions × 8 weeksProcessing Time TestCanalith repositioning maneuverFullerton Advanced Balance ScaleStankiewicz [49]Patients with unilateral vestibular hypofunctionVRVR Cawthorne‐Cooksey exercisesConventional Cawthorne‐Cooksey exercisesVertigo Symptom Scale—Short FormVR vestibular rehabilitation is as effective as conventional rehabilitation, with significantly increased levels of patient satisfaction.No adverse events. n = 205 sessions in 5 consecutive days5 sessions in 5 consecutive daysPatient satisfactionViziano [50]Patients with unilateral vestibular hypofunctionVRHome‐based VR vestibular exercises, a combination of adaptation, substitution, habituation, and balance exercisesVestibular rehabilitationVideo head impulse testingVR home exercises are viable, effective, additional to improve long‐term vestibular rehabilitation outcomes.Not reported.Static posturography testingDizziness Handicap Inventory n = 4730–40 min daily × 4 weeks30–45 min × 2 sessions × 4 weeksActivities‐specific Balance Confidence scaleVestibular rehabilitationDynamic Gait Index30–45 min × 2 sessions × 4 weeksYan [51]Patients with Benign Paroxysmal Positional VertigoVRVR vestibular exercisesGroup 1: No interventionSeverity of vertigo and residual symptomsVR‐based vestibular exercises can cure the residual symptoms after canalith repositioning procedure and restore balance ability.Not reported.Vestibular evoked myogenic potentials n = 12415–20 min × twice daily × 4 weeksGroup 2: Cawthorne‐Cooksey exerciseBerg Balance ScaleGroup 3: Brandt‐Daroff exerciseAnxiety and depressionOrthopedic rehabilitation: Neck painBahat [52]Patients with chronic neck painVRVR‐based kinematic trainingKinematic training with a laser pointerNeck Disability IndexBoth groups demonstrated significant improvements in neck disability, cervical rotation range of motion, velocity, and step test.Four patients experienced motion sickness.Cervical range of motion and kinematics Kinesiophobia n = 26Wrap 1200VR systemHome exercise programHome exercise programStatic balance30 min × 4–6 sessions30 min × 4–6 sessionsBahat [53]Patients with chronic neck painVRHome program, VR‐based kinematic trainingKinematic training with a laser pointerNeck Disability IndexVirtual reality group demonstrated greater improvements in velocity, pain intensity, motion accuracy, and quality of life.Five patients experienced motion sickness.Cervical range of motion and kinematicsGlobal perceived benefit n = 51Oculus Rift DK120 min × 4 days × 4 weeks20 min × 4 days × 4 weeksQuality of lifeKinesiophobiaBahat [54]Pilots with Flight‐associate neck painVRSelf‐VR kinematic trainingPhysical therapy onlyNeck Disability IndexThis self‐VR kinematic training promoted kinematic performance, but was ineffective in engaging the pilots to exercise, and consequently did not improve pain and disability.Not reported.Cervical kinematics n = 45Oculus Rift DK15 min × 4 days × 4 weeksQuality of lifePhysical therapyCetin 2024Patients with chronic neck pain n = 41VRVR training × 20 minMotor control exercises × 40 minSensory Organization TestVR demonstrates significant effects on postural control, posture, and kinesiophobia in patients with chronic neck pain.No adverse events.Motor control exercises × 20 minLimits of stability, unilateral stance testsOculus Go device3 sessions × 6 weeks3 sessions × 6 weeksGait speedPosture measures KinesiophobiaTereja [55]Patients with non‐specific chronic neck painVRVR‐based neck exercisesConventional neck exercisesConditioned pain modulationBoth groups demonstrated significant improvements in pain intensity, rotation range of motion, Neck Disability Index, and anxiety. Virtual reality group demonstrated greater improvements in kinesiophobia.No adverse events.Cervical range of motion n = 442 sessions × 4 weeks2 sessions × 4 weeksNeck Disability IndexKinesiophobia Pain Anxiety Symptoms ScaleOrthopedic rehabilitation: Postoperative careMazurek 2023Post‐total hip or knee arthroplastyVR HTC ViveVR for relaxation and mood enhancementStandard care Conventional rehabilitation programHospital Anxiety and Depression ScaleThe findings emphasize the promising role of VR therapy as a beneficial addition to the rehabilitation process for older adults’ post‐hip and knee arthroplasty.Not reported.Perceived Stress Scale n = 6820 min × 2 sessions × 4 weeksGeneralized self‐efficacy scaleConventional rehabilitation programBarthel index Physical function, stress, and pain assessmentsShim [56]Post‐total knee arthroplastyAR UINCARE HomeAR‐based home exercise program 12 weeksHome exercise program with a brochure 12 weeks Gait speed AR improves the functional outcomes, pain, and quality of life of patients after total knee arthroplasty. AR‐based rehabilitation may be useful treatment as an alternative to conventional rehabilitation.No adverse events.Pain, range of motion, muscle strength n = 56Berg Balance Scale Quality of life Western Ontario and McMaster UniversitiesOsteoarthritis IndexShim [57] Post‐rotator cuff repair n = 115 AR UINCARE HomeAR‐based home exercise program 12 weeks Home exercise program with a brochure 12 weeks Simple Shoulder TestAR‐based rehabilitation shows better improvement on shoulder function compared to conventional rehabilitation.No adverse events.Pain, range of motion, muscle strength, grip strengthShoulder function Quality of lifeYoon [58] Post‐total knee arthroplasty n = 36 VRVR balance exercisesContinuous passive motion device Static and dynamic balance ability VR training produced better early balance ability and knee function than the control group.Not reported.20 min × 5 days × 2 weeksContinuous passive motion deviceTimed Up and Go TestExercise therapyExercise therapyKnee functionOrthopedic conditions—otherOzlu [59] Patients with knee osteoarthritis n = 73 VR Oculus QuestVR exergamesUltrasound and transcutaneous electrical nerve stimulationPain and functionalityThe disease‐specific gamification through VR added to the conservative treatment has a positive effect on pain, functionality, and balance.Three patients had mild nausea and one patient had a headache15 min × 5 days × 3 weeks6‐minute walk testUltrasound and transcutaneous electrical nerve stimulationBerg Balance ScaleWankhade [60] Patients with frozen shoulder n = 50 VR Oculus RiftVR exergames Joint mobilizationJoint mobilizationPain and range of motionVR shows more remarkable improvement in reducing pain, improving movement of shoulder.Not reported.15–20 min daily × 2 weeks15–20 min daily × 2 weeksPain and range of motion Shoulder Pain and Disability IndexPediatric rehabilitationBortone [61] Children with neuromotor impairments (cerebral palsy or developmental dyspraxia) n = 8 VR Oculus Rift VK2 and wearable haptic devicesVR rehabilitation Mobeybox and Marble labyrinth games Sensorimotor and cognitive training 1 h × 2 sessions × 4 weeks Conventional rehabilitation Cognitive therapeutic exercise based on neurocognitive rehabilitation theory 1 h × 2 sessions × 4 weeks Nine Hole Peg Test Kinesiological assessment (movement error and smoothness) VR and wearable haptic devices is a viable alternative to conventional therapy for improving upper extremity function in children with neuromotor impairments.No adverse events.Khamees [62] Children with a hand burn n = 72 ARAR‐enhanced hand rehabilitation booklet 1 monthRoutine hand rehabilitation with educational booklet 1 monthJebsen‐Taylor Hand Function TestRehabilitation of children with hand burns, using printed educational booklet with related AR application, improves their hand function.Not reported.Liu [63] Children with autism spectrum disorder n = 24 MR MR aquarium (a real aquarium with a virtual underwater world)MR cognitive training Traditional training course 30 min × 4 weeksTraditional training course 30 min × 4 weeksThe Chinese version of the psycho‐educational profile (third edition) Post‐training questionnaireMR training environment can effectively attract and maintain the attention of children, bring better ease of use, multi‐sensory feedback, and then assist the rehabilitation.Not reported.Malick [64] Children with spastic hemiplegic cerebral palsy n = 30 AR Microsoft Kinect v2 sensorAR game: Balance It 15 min × 3 sessions × 8 weeksAR game: Bubble Pop AR game: Scoop's 15 min × 3 sessions × 8 weeksUpper extremity range of motion Upper extremity muscle strengthThe AR games intervention was beneficial and effective for improving the ROM of majority of the joints and strength of various muscles of upper extremity.No adverse events.Menekseoglu [65] Children with hemiplegic cerebral palsy n = 36 VR Oculus QuestVR upper extremity training Upper extremity exercise 60 min × 2 sessions × 6 weeksUpper extremity exercise 60 min × 2 sessions × 6 weeksAssisting Hand Assessment ABILHAND‐Kids Quality of Upper Extremity Skills Test Modified Ashworth scaleVR group demonstrated significant improvements on upper limb function, quality of life, and active joint range of motion.Three children had headaches; one child had nausea and dizziness. All these adverse effects were short lasting and did not require additional intervention.Shen [66] Children with traumatic brain injury n = 26 VR HTC ViveVR‐based interactive cognitive training for executive functionVR gaming without executive function trainingNIH Toolbox Cognitive Battery Conners’ Continuous Performance test PedsQL Core Scale Child Behavior Checklist Behavior Rating Inventory of Executive Function 2Results supported potential efficacy of the VR‐based interactive cognitive training, particularly for moderate and severe traumatic brain injuries.One child had headache. Characteristics of the included studies. Patients with incomplete spinal cord injury VR Both groups significantly improved Barthel Index. The VR group demonstrated significant improvements on upper extremity motor function than the control group. Gait speed Post‐rotator cuff repair n = 115 Home exercise program with a brochure 12 weeks Post‐total knee arthroplasty n = 36 Static and dynamic balance ability Patients with knee osteoarthritis n = 73 Patients with frozen shoulder n = 50 Children with neuromotor impairments (cerebral palsy or developmental dyspraxia) n = 8 Conventional rehabilitation Cognitive therapeutic exercise based on neurocognitive rehabilitation theory 1 h × 2 sessions × 4 weeks Nine Hole Peg Test Kinesiological assessment (movement error and smoothness) Children with a hand burn n = 72 Children with autism spectrum disorder n = 24 Children with spastic hemiplegic cerebral palsy n = 30 Children with hemiplegic cerebral palsy n = 36 Children with traumatic brain injury n = 26 The 58 included randomized controlled trials were scored 6–8 on the PEDro scale. Less than half of studies (n = 22) performed concealed allocation. All studies specify the eligibility criteria, present comparable baseline characteristics of the participants, and perform statistical analyses appropriately. Most studies did not blind participants or therapists due to the nature of rehabilitation intervention, which requires active involvement of both parties. Thirty‐four studies blinded the assessors. Nine studies did not have adequate follow‐up (key outcome measures were obtained from less than 85% of the participants from initial allocation), and 10 did not use the intention‐to‐treat analysis. Details of the PEDro assessment are reported in Table 2. Table 2The Physiotherapy Evidence Database (PEDro) scale assessment results of the included studies.Eligibility criteriaRandom allocationConcealed allocationBaseline comparabilityBlind subjectsBlind therapistsBlind assessorsAdequate follow‐upIntention‐to‐treatBetween group comparisonsPoint estimates and variabilityTotal scoreAmin [34]Yes10100011116An [32]Yes10110011117Appel [21]Yes11100010117Austin 2020Yes11100011116Bahat [52]Yes11100110117Bahat [53]Yes11100101117Bahat [54]Yes11100111118Bortone [61]Yes10100111117Campo 2022aYes10100011116Campo 2022bYes10100011116Cetin 2024Yes10100101116Choi [35]Yes11100111118Crosbie [36]Yes11100111118Duff [67]Yes10100001116Evron [37]Yes11100001116Goel 2021Yes11100010116Gulcan 2022Yes10100110116Hegazy [38]Yes10110111118Hsu [39]Yes11100110117Huang [40]Yes11100111118Jeon [24]Yes11100010116Jo 2024Yes10100111117Kang [68]Yes11100110117Khamees [62]Yes10100011116Ku [25]Yes11100110117Lee [41]Yes10100111117Lee [20]Yes11100101117Li [69]Yes11100111118Lima [26]Yes11100110117Lim [33]Yes10100011116Liu [63]Yes10100011116Liu 2023Yes10100111117Malick [64]Yes10110011117Mekbib [42]Yes11100101117Menekseoglu [65]Yes10100111117Micarelli [70]Yes10100011116Ogun [43]Yes10110100116Optale [27]Yes10100111117Ozdil [48]Yes10100011116Ozlu [59]Yes10100111117Park [28]Yes10100111117Pelaez [44]Yes10100111117Qiu [29]Yes10100011116Rutkowski 2021Yes11100111118Schuch [31]Yes10100111117Shen [66]Yes10100111117Shim [56]Yes10100101116Shim [57]Yes10100111117Shin [45]Yes11100111118Stammler [46]Yes10100011116Stankiewicz [49]Yes10100011116Tereja [55]Yes10100011116Viziano [50]Yes10100111117Wankhade [60]Yes10100011116Yan [51]Yes10100011116Yoon [58]Yes11100101117Zhu [30]Yes11100011117 The Physiotherapy Evidence Database (PEDro) scale assessment results of the included studies. Geriatric Rehabilitation Ten randomized controlled trials investigated the effects of XR in geriatric rehabilitation. Two [22, 23] studies reported that VR exergaming program as an add‐on intervention to usual care significantly improved balance ability as measured by the Tinetti scale; in addition, one [23] of them demonstrated the significant effects of VR exergaming on functional strength (measured by the Five times sit‐to‐stand test) and mobility (measured by the Timed Up and Go test). Lima et al. [26] reported that for older adults with balance deficits, VR exergaming produced similar benefits on balance‐related outcomes (static and dynamic balance, functional reach, dizziness, and fear of falling) as dose‐matched conventional multimodal circuit exercises. Furthermore, Qiu et al. [29] investigated the application of VR to enhance Tai Chi among 300 community‐dwelling older adults. They found that compared to traditional Tai Chi, VR‐enhanced Tai Chi exhibited a significantly greater reduction of depression as measured by the Geriatric Depression Scale. Three studies used AR‐based exercise programs for community‐dwelling older adults. Ku et al. [25] reported that the AR group demonstrated significantly better improvements in balance (measured by the Berg Balance Scale), mobility (measured by the Timed Up and Go test), and fall risk index (measured by posturography). Similarly, Jeon et al. [24] showed significant improvements in favor of AR on balance, lower extremity strength, mobility, endurance, and exercise self‐efficacy. Besides, Zhu et al. [30] reported that a 4‐week AR memory training program significantly improved working memory among Chinese healthy older adults. Three studies [21, 27, 28] focused on older adults with cognitive and memory impairment. Optale et al. [27] compared VR memory training to dose‐matched music therapy among older adults with memory deficits. They found that the VR memory training group demonstrated significant improvements in verbal memory, long‐term recall, and global cognition; whereas the control group showed progressive decline. Similarly, Park et al. [28] reported that MR‐based cognitive training was superior to dose‐matched computer‐assisted cognitive training in improving visuospatial working memory. Appel et al. [21] applied VR to patients with dementia in an acute care hospital. VR films were displayed to patients for up to 20 min every 1–3 days to promote nature visualization and psychological well‐being. VR was overall acceptable and enjoyable to patients. Results found that VR had a significant effect on reducing aggressiveness, including physical aggression and loud vociferation, while the effects on other behaviors and psychological symptoms, falls, length of stay, and quality of life were comparable to standard care. Neurological Rehabilitation Thirty‐one randomized controlled trials investigated the effects of XR in neurological rehabilitation. Neurological conditions included Parkinson's disease, spinal cord injury, stroke, and vestibular dysfunction. Gulcan et al. [71] used an AR‐augmented treadmill for gait training among patients with Parkinson's disease. The study group received AR gait training combined with conventional therapy, and the control group received conventional therapy only. The training was applied to both groups three sessions per week for a total of 6 weeks. They found that both groups showed similar amounts of improvement in the Berg Balance Scale, the Activity‐Specific Balance Confidence Scale, but the AR group demonstrated more improvements on spatiotemporal gait parameters. Another study by Schuch et al. [31] reported that a 5‐week VR balance training program did not produce significantly different outcomes in functional mobility compared to a structured exercise program for Parkinson's disease. Three studies applied VR to motor rehabilitation for patients with incomplete spinal cord injury. An et al. [32] reported that VR soccer games significantly improved the results of the chair stand test, Timed Up and Go test, 10‐meter walk test, and kick speed test compared to conventional rehabilitation. Goel et al. [72] reported that VR training produced significantly improved sitting balance compared to functional electrical stimulation. For upper extremity function [33], VR demonstrated significant improvements in grip and pinch strength, and muscle strength of elbow and wrist extensors. One study [73] used VR to manage neuropathic pain experienced by patients with spinal cord injury. Participants reported significantly lower pain intensity after viewing nature scenes in VR compared to 2D laptop screen application; an increased level of presence was associated with greater pain reduction. Thirteen studies used VR for stroke rehabilitation. The majority of these studies used VR interventions for the upper extremity and hand. VR‐based mirror therapy demonstrated significantly improved manual dexterity than conventional rehabilitation [39, 74]. Four studies [34, 38, 42, 43] reported that VR significantly promoted motor function of the upper extremity. Conversely, Crosbie et al. [36] and Huang et al. [40] reported no significant difference in upper extremity function between the VR and conventional therapy groups. Ogun et al. [43] found significant between‐group differences in self‐care skills (measured by Performance Assessment of Self‐care Skills) but not functional independence (Functional Independence Measure). For lower extremity function, Kang et al. [68] reported that VR treadmill training improved balance and gait significantly compared to regular treadmill training, and Pelaez‐Velez et al. [44] reported that VR exergames significantly improved trunk control, lower extremity, balance, and mobility. Moreover, two studies [35, 45] found VR‐based spatial neglect training more effective than conventional rehabilitation. For patients with mild post‐stroke cognitive impairment, Liu et al. [63] reported that VR puzzle game therapy demonstrated significant improvements in executive and visual‐spatial function. Six studies used AR for stroke rehabilitation. Evron et al. [37] employed AR mobility exercises with the Selfit system, which projected different scenarios to the floor and captured patients’ movements with motion sensors. They reported that the AR group demonstrated significantly better improvement in the Timed Up and Go test; both therapists and patients demonstrated high levels of satisfaction with the AR system. Similarly, AR postural control training [41] yielded significantly better gait parameters than conventional rehabilitation; AR gait adaptation training [47] demonstrated superior effects on obstacle avoidance, balance ability, and functional mobility compared to regular training. Lee et al. [20] found that an AR‐enhanced home exercise program produced better outcomes of balance and fear of falling compared to a written and pictorial home exercise program. Li et al. [69] reported that adding a game‐based cellphone AR system training to conventional occupational therapy was effective in improving both upper extremity and cognitive function. Furthermore, Stammler et al. [46] used AR‐based active exploration training for patients with unilateral spatial neglect, and they found that the AR training produced significantly better outcomes than standard therapy, which remained stable for 2 months after the completion of treatment. One study used MR for stroke rehabilitation. Duff et al. [67] developed the Adaptive MR Rehabilitation system for upper extremity training, which provided precise kinematic measurements and engaged multimodal feedback for self‐assessment during therapeutic activities. They found that both the MR and control groups demonstrated improved outcomes on upper extremity function scales, while only the MR group showed a consistent improvement in the kinematic parameters of reaching, indicating better quality of movement. For vestibular dysfunction, studies used VR for benign paroxysmal positional vertigo [48, 51] and unilateral vestibular hypofunction [49, 50, 70]. Both Ozdil et al. [48] and Yan et al. [51] reported that VR was effective in mitigating the residual symptoms, including balance impairments and dizziness, after the canalith repositioning procedure. A VR‐based vestibular exercise program, including substitution, habituation, and balance exercises, demonstrated improvements in vestibular‐ocular reflex, balance ability, and dizziness symptoms in both short‐ [70] and long‐term [50]. Stankiewicz et al. [49] reported that VR vestibular rehabilitation was as effective as conventional vestibular rehabilitation, with significantly higher levels of patient satisfaction. Orthopedic Rehabilitation Eleven randomized controlled trials investigated the effects of XR in orthopedic rehabilitation. Among them, five studies used VR for neck pain. Cetin et al. [75] compared VR training to motor control exercises for chronic neck pain. Compared to the control group, the VR group demonstrated more positive effects on exercise compliance, balance control, and kinesiophobia. Similarly, Tereja et al. [55] reported that VR demonstrated significant effects in reducing kinesiophobia; while there were no significant between‐group differences in pain intensity, neck function, and psychological distress (pain catastrophizing, fear‐avoidance beliefs, and anxiety). Two studies compared VR‐based kinetic training to conventional kinetic training for chronic neck pain, at an in‐clinic and a home‐based setting, respectively. For the in‐clinic training study [52], both VR and control demonstrated significant improvements in neck pain intensity, neck function, and range of motion throughout the intervention, with only a few between‐group differences detected. For the home‐based training study [53], VR demonstrated significant advantages in cervical motion velocity, pain intensity, health status, and movement accuracy. Another study applied self‐VR kinetic training to pilots with flight‐associated neck pain [54]. It found that compliance of the participants was poor, and no differences were observed in self‐reported measures related to neck pain and cervical strength. Four studies focused on postoperative rehabilitation. Shim et al. [56, 57] used AR to augment home exercise programs and applied it to patients post‐total knee arthroplasty [56] and rotator cuff repair [57]. In both studies, the AR‐based home exercise program was compared with the conventional home exercise program with a brochure. They found that all physical and function‐related outcomes were significantly improved in both groups for patients post‐total knee arthroplasty, with no significant between‐group differences. For patients post‐rotator cuff repair, AR showed better improvement in shoulder function. Mazurek et al. [76] reported that for older adults post‐total hip or knee arthroplasty, eight VR sessions demonstrated significantly reduced psychological distress and improved functional independence. Furthermore, for older adults post‐total knee arthroplasty, VR balance training produced better early balance ability and knee function compared to the control group [58]. Two studies used VR exergames for conservative management of orthopedic conditions. Ozlu et al. [59] reported that the addition of VR training produced significant improvements in pain reduction, lower extremity function, and balance ability for patients with osteoarthritis. For patients with adhesive capsulitis [60], VR exercises combined with joint mobilization demonstrated significantly better outcomes in pain and range of motion compared to joint mobilization only. Pediatric Rehabilitation Six randomized controlled trials investigated the effects of XR in pediatric rehabilitation. Four studies used XR for upper extremity rehabilitation. Bortone et al. [61] reported for children with neuromotor impairments (cerebral palsy and developmental dyspraxia), VR significantly improved manual dexterity and movement quality of upper extremity. Similarly, for children with hemiplegic cerebral palsy, VR‐mediated gamified rehabilitation significantly improved upper extremity function and quality of life [65]. For children with hand burns [62], the intervention group performed AR‐enhanced exercises with a printed booklet, and the control group received a printed booklet only. Results revealed that the AR group demonstrated significantly better hand function as measured by the Jebsen‐Taylor hand function test. Furthermore, Malick et al. [64] allocated children with hemiplegic cerebral palsy into three groups, each group receiving an AR game. Following 8 weeks of intervention, all the groups demonstrated significant improvements in upper extremity range of motion and muscle strength. Two studies used XR in cognitive training programs. Shen et al. [66] developed an innovative VR interactive cognitive training for executive function rehabilitation and examined its feasibility and efficacy among children with traumatic brain injuries. Participants reported adequate satisfaction with this system regarding pleasure and motivation; low levels of physical exertion and simulator sickness. Results also supported the efficacy of this system, particularly for children with moderate and severe brain injuries. Liu et al. [77] presented an MR aquarium, which enhanced a real aquarium with a virtual underwater environment, to children with autism spectrum disorder. They found that MR training can effectively attract and maintain the attention of children, bring better ease of use, provide multi‐sensory feedback, and assist cognitive rehabilitation. Discussion This systematic review provides a comprehensive synthesis of the current evidence regarding the application of XR technologies, comprising VR, AR, and MR, in rehabilitation medicine. The 58 included randomized controlled trials cover a wide range of patient populations, interventions, and clinical outcomes across geriatric, neurological, orthopedic, and pediatric rehabilitation. More studies used VR than AR and MR. XR can be implemented in all stages of rehabilitation in different settings, including acute care hospitals, inpatient and outpatient rehabilitation, residential care facilities, and patients’ homes. Overall, the findings indicate that XR has the potential to enhance various aspects of rehabilitation outcomes, including motor and cognitive function, pain management, post‐surgical recovery, psychological well‐being, and quality of life. The results of this review highlight several promising applications of XR in rehabilitation. In geriatric populations, XR demonstrated substantial benefits in improving balance, strength, mobility, and cognitive function. These findings align with previous research supporting the positive effects of serious games on physical [78] and cognitive [79] aspects of older adults. XR provides immersive and interactive environments that enhance the effectiveness of game‐based learning and training processes, fostering engagement and motivation for rehabilitation interventions. A study [80] demonstrated that gamified interventions can be used to improve the motivation for rehabilitation and facilitate the transition from hospital to home for older adults. Moreover, XR's capacity to improve psychological outcomes, such as reducing depression through VR‐enhanced Tai Chi [29], underscores its potential to address the mental health needs of geriatric patients. However, it should be recognized that older adults may encounter potential barriers when using XR in rehabilitation, such as visual and hearing deficits, limited digital literacy, psychological barriers, and resistance to change. It is crucial to take these factors into consideration through adaptive design and customization to make XR more inclusive and effective for older adults. Neurological rehabilitation, which accounted for the largest body of evidence in this review, demonstrated that XR interventions can facilitate recovery of motor function in stroke, spinal cord injury, and Parkinson's disease; reduce neuropathic pain in spinal cord injury; mitigate cognitive and perceptual deficits in stroke; and enhance treatment effects of vestibular dysfunction. Current neuroscience research suggests rehabilitation professionals use timely, intensive, and highly repetitive task‐specific practices to harness experience‐dependent neural plasticity and promote functional independence [81]. Compared to conventional rehabilitation approaches that can be tedious and labor‐intensive, XR technologies open the door to immersive and engaging therapeutic experiences that enable highly controlled environments for targeted task practice and skill relearning. Moreover, XR technologies can support therapists by reducing their workload, as they allow for self‐directed rehabilitation sessions [20] that patients can engage in independently or with minimal supervision. Therapists can use XR to monitor progress, track patient performance, and adjust treatment plans based on clinical outcome data, enabling them to focus more on personalized interventions and strategic modifications. Furthermore, A systematic review [82] synthesizes the effects of VR interventions on cortical reorganization in stroke rehabilitation, elucidating the underlying neurophysiological mechanism of VR. Future studies should investigate the effects of AR and MR on neural plasticity to facilitate a comprehensive understanding of their mechanism in neurological rehabilitation. In orthopedic rehabilitation, VR and AR interventions were primarily used to aid postoperative rehabilitation and facilitate conservative management of chronic musculoskeletal conditions including neck pain, knee osteoarthritis, and frozen shoulder. Overall, VR demonstrates positive effects on reducing pain and disability among patients with chronic neck pain, consistent with the findings in a previous systematic review [83]. However, for a group of pilots with flight‐associated neck pain, there were no significant differences between self‐VR kinetic training and control groups in symptom improvement due to high baseline kinematic performance and low adherence [54]. The negative findings of this study indicate the importance of considering specific impairments, devising targeted interventions, and securing patients’ compliance to attain optimal outcomes. An AR‐based digital rehabilitation system (UINCARE Home+) demonstrated positive outcomes in patients post‐total knee arthroplasty [56] and rotator cuff tear [57]. The AR system provides different levels of therapeutic exercises based on the postoperative timeframe, and uses a three‐dimensional depth camera to capture patients’ movements and provides real‐time feedback. Therapists can leverage these technological advantages to improve efficiency in remote therapeutic monitoring. Pediatric rehabilitation also showed positive outcomes with XR, particularly in improving upper extremity function in children with neuromotor impairments and enhancing cognitive function in children with brain injuries or autism spectrum disorders. Immersive and gamified exercises can help children achieve therapeutic goals through interactive tasks that keep them engaged, supporting more sustained participation than traditional therapy alone. The external focus of attention provided by XR can facilitate intrinsic motivation and implicit motor learning [84]. Studies have shown that it was feasible to implement an intensive bimanual training program in VR for children with hemiplegic cerebral palsy [85], and apply the training program in the home setting [86]. While developing XR rehabilitation programs for children, it is important to consider the active ingredients of computer play‐based interventions [87], including intensive practice, task specificity, flexibility to individualization, feedback, social play equalization, neural plasticity changes, problem‐solving, motivation, and external support. A thoughtfully designed XR program also needs to balance the entertainment feature and the therapeutic effectiveness of the technology to maximize its acceptance and utility in the pediatric population. The findings from this review have important implications for clinical practice. XR technologies offer an innovative, patient‐centered approach to rehabilitation that can be tailored to individual characteristics, providing augmented feedback, and promoting active patient engagement. For clinicians, the ability to use XR to simulate various therapeutic scenarios or exercises, while simultaneously tracking patient performance, can enhance the precision and efficiency of rehabilitation interventions. Importantly, safety and tolerability should be carefully considered when implementing XR‐based rehabilitation interventions. Across the included studies, adverse events were typically reported passively rather than systematically assessed using predefined monitoring procedures. These safety considerations may be particularly important for older adults and individuals with neurological disorders, who may be more susceptible to cybersickness, sensory conflict, or balance instability when exposed to immersive environments. Accordingly, future clinical implementation of XR interventions should incorporate structured safety screening, monitoring protocols, and appropriate supervision to ensure safe and effective use in rehabilitation settings. Practical implementation issues should also be considered when integrating XR technologies into clinical practice. Clinicians may require training to operate XR systems, interpret system‐generated feedback, and incorporate XR‐based exercises into individualized rehabilitation programs. Infrastructure requirements, including compatible hardware, technical support, and adequate space, may influence the feasibility of XR deployment in clinical settings. In addition, integrating XR interventions into existing clinical workflows requires careful planning to ensure that these technologies complement routine therapy. Adoption may also be limited in low‐resource or rural settings where access to advanced equipment and technical expertise is restricted. Improving accessibility may depend on the development of affordable XR technologies and implementation models adapted to resource‐limited healthcare environments. Future research should further clarify the role of XR in rehabilitation across different clinical contexts. Key research priorities include identifying which patient populations may benefit most from XR interventions and determining the most relevant outcome measures for evaluating their effectiveness. In addition, there is a need to standardize the reporting of XR intervention characteristics, including hardware, training protocols, and dosage parameters, to improve comparability across studies. Comparative effectiveness studies are also warranted to evaluate differences between VR, AR, and MR modalities. Furthermore, future work should examine the cost‐effectiveness of XR‐based rehabilitation and investigate long‐term outcomes to better understand the sustainability of treatment effects. This systematic review has some limitations. First, the number of included studies and associated sample sizes are limited, especially for studies using AR and MR. Further high‐quality randomized controlled trials are warranted to investigate the specific effects of different XR technologies to inform clinical practice. Second, only articles in English were searched and included in this review. Potentially eligible studies in other languages could be missed. Third, studies rated as “fair” or “poor” on the PEDro scale were excluded a priori to reduce the risk of bias and improve the methodological rigor of the synthesized evidence. However, this decision may introduce selection bias and potentially limit the generalizability of the findings. Moreover, fewer than half of the studies performed concealed allocation, and most studies did not blind participants or therapists, which could introduce potential bias. In addition, some studies had incomplete follow‐up or did not use intention‐to‐treat analysis. These methodological limitations may lead to an overestimation of treatment effects and should be considered when interpreting the findings of this review. Conclusions This systematic review demonstrates the significant potential of XR technologies, including VR, AR, and MR, to enhance rehabilitation outcomes across diverse patient populations and clinical contexts. The evidence suggests that XR may improve motor and cognitive functions, pain management, psychological well‐being, and overall quality of life, with promising applications identified in geriatric, neurological, orthopedic, and pediatric rehabilitation. Overall, XR technologies offer a transformative approach to rehabilitation that emphasizes patient‐centered care and enhances therapeutic engagement, demonstrating the potential to redefine rehabilitation practices and improve outcomes for diverse patient populations. Author Contributions Jie Hao: conceptualization, data curation, formal analysis, investigation, project administration, software, visualization, writing – original draft, writing – review and editing. Yao Yao: data curation, investigation, validation, visualization, writing – original draft. Ka‐Chun Siu: conceptualization, data curation, formal analysis, investigation, methodology, supervision, validation, visualization, writing – original draft, writing – review and editing. Funding The authors have nothing to report. Disclosure All Authors read and approved the final version of the manuscript. Ka‐Chun Siu had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis. Ethics Statement Ethical approval was not required because this study is a systematic review based on previously published data. Conflicts of Interest The authors declare no conflicts of interest. Transparency Statement The lead author, Ka‐Chun Siu, affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained. Search strategy for PubMed (“Virtual Reality”[Mesh] OR “virtual reality” OR “virtual environment” OR “augmented reality” OR “mixed reality” OR “extended reality” OR “head mounted devic*” OR “head mounted displa*” OR “Oculus rift” OR “Meta quest” OR “Google glass*” OR “Hololens” OR “HTC vive”) AND (“Rehabilitation”[Mesh] OR “Physical therapy modalities”[Mesh] OR “Occupational therapy”[Mesh] OR “Activities of daily living”[Mesh] OR “Walking”[Mesh] OR “gait”[Mesh] OR “rehabilitation” OR “physiotherapy” OR “occupational therapy”) Search strategy for Embase (‘virtual reality’/exp/mj OR ‘virtual reality head mounted display’/exp/mj OR ‘virtual reality’:ti,ab OR ‘augmented reality’:ti,ab OR ‘mixed reality’:ti,ab OR ‘virtual environment’:ti,ab OR ‘extended reality’:ti,ab OR ‘google glass*’:ti,ab OR ‘head mounted devic*’:ti,ab OR ‘head mounted displa*’:ti,ab OR ‘oculus rift’:ti,ab OR ‘meta quest’:ti,ab OR ‘hololens’:ti,ab OR ‘HTC vive’:ti,ab) AND (‘rehabilitation’/exp/mj OR ‘rehabilitation’:ti,ab OR ‘physiotherapy’:ti,ab OR ‘occupational therapy’:ti,ab) AND [article]/lim AND [english]/lim AND [humans]/lim AND [clinical study]/lim Search strategy for Scopus (TITLE‐ABS‐KEY (“virtual reality” OR “virtual environment” OR “augmented reality” OR “mixed reality” OR “extended reality” OR “head mounted devic*” OR “head mounted displa*” OR “Oculus rift” OR “Meta quest” OR “Google glass*” OR “Hololens” OR “HTC vive”) AND TITLE‐ABS‐KEY (“rehabilitation” OR “physiotherapy” OR “occupational therapy”)) AND (LIMIT‐TO (DOCTYPE, “ar”)) AND (LIMIT‐TO (LANGUAGE, “English”)) AND (LIMIT‐TO (SRCTYPE, “j”)). Data Availability Statement Data sharing not applicable to this article as no datasets were generated or analysed during the current study. References Associated Data Data Availability Statement Data sharing not applicable to this article as no datasets were generated or analysed during the current study.