Effects of robot-assisted task-oriented training on upper limb function and activities of daily living in patients with stroke: A systematic review

Abstract

Objective

This study was undertaken to systematically examine the effects of robot-assisted, task-oriented training on upper limb function and activities of daily living in patients with stroke.

Methods

A systematic search was conducted across PubMed, EMBASE, SCOPUS, CINAHL, and Ovid LWW databases in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines, including literature published from 2014 to 2025. Eligible studies were randomized controlled trials that compared an experimental group receiving robot-assisted task-oriented training with a control group undergoing conventional treatment or alternative interventions and evaluated upper limb function and activities of daily living outcomes. The Physiotherapy Evidence Database scale was used to assess the methodological quality of the literature. This systematic review was registered with the Open Science Framework (DOI: 10.17605/OSF.IO/4DT6G).

Results

Ten studies were included in the analysis. Robot-assisted interventions consistently improved upper extremity motor function, particularly when implemented as an adjunct to conventional therapy or integrated with advanced technologies such as functional electrical stimulation. However, activities of daily living–related improvements varied across studies, and functional improvements were confirmed in only three studies. The effectiveness of robotic intervention depended on the intervention modality, stroke phase, and technological integration.

Conclusions

Robot-assisted task-oriented training can effectively improve upper limb function in patients with stroke, and in some cases, combining it with conventional therapy may produce synergistic effects. Nevertheless, evidence regarding improvements in activities of daily living and long-term maintenance effects remains limited. Further high-quality randomized trials focusing on optimizing specific clinical dimensions and facilitating the transfer of motor recovery to activities of daily living are required.

Keywords

Activities of daily living robot rehabilitation stroke task-oriented training upper limb function

Introduction

Stroke is a condition in which brain cells are damaged due to blockage of cerebral blood vessels or bleeding.¹ Because damaged brain cells do not regenerate, the resulting physical disabilities are long-term and permanent.² Paralysis due to stroke impairs upper extremity function, and only 5% of patients achieve full functional use of the paralyzed arm.³ In addition, reduced upper limb function negatively impacts independent activities of daily living (ADLs) in affected patients, causing emotional distress, including depression, which further lowers quality of life.^4,5

Virtual reality–based exercise therapy,^6,7 mirror therapy,⁸ mental practice,⁹ upper limb muscle strengthening training,¹⁰ constraint-induced movement therapy,¹¹ task-oriented training (TOT),¹² and robot-assisted exercise therapy^13,14 are being implemented to restore upper limb function in patients with stroke. In particular, TOT offers the advantage of requiring repeated performance of functional tasks, thereby complementing repetitive training based on simple physical movements. TOT is an effective intervention for motor learning because it enhances exercise concentration and provides knowledge of exercise outcomes.^15,16 In addition, TOT improves motor control through active neuroplasticity and helps resolve patient-related functional limitations by enhancing task performance.^15,17

Robot-assisted therapy (ROT) has recently emerged as a widely implemented intervention for achieving functional recovery in patients with stroke, as it can standardize training intensity, precision, and repetition, thereby increasing intervention consistency and reducing the physical burden on therapists.¹⁸ In particular, robot-based TOT is considered more effective than general task performance training for improving upper limb function in patients with stroke, as it enables precise control and biofeedback.¹⁹

Calabrò et al.²⁰ conducted a multicenter training comparing the intervention effects of robot-assisted task training using exoskeleton and end-effector devices and reported significant improvements in upper limb function on the paretic side in patients with chronic stroke, including those with subacute stroke. Similarly, Cho and Song²¹ reported that robot-assisted reach training effectively improved upper limb function and kinematic movements in patients with chronic stroke. However, a large-scale comparative study by Rodgers et al.,²² which evaluated robot-assisted training, enhanced upper limb therapy (EULT), and usual care, reported that robot-assisted task training did not demonstrate greater effect on upper limb function than general rehabilitation training. Accordingly, robot-assisted TOT is limited by low consistencies of its reported intervention effects, which may be attributed to factors such as the timing of training after stroke onset, the type of robot used, and training intensity and frequency. Therefore, we conducted this systematic literature review to evaluate the effects of robot-based TOT on upper limb function and ADLs in patients with stroke, focusing on randomized controlled trials (RCTs) published from 2014 to 2025. Additionally, we aimed to propose clinical implications and future research directions based on the latest evidence.

Methods

The systematic review protocol was based on the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2020²³ and was registered with the Open Science Framework (DOI: 10.17605/OSF.IO/4DT6G).

Search strategy and data resources

A systematic literatures search was conducted in the EMBASE, Ovid LWW, Scopus, PubMed, and CINAHL electronic databases from May 2025 to June 2025. Only articles published in English between 2014 and 2025 were considered. To maximize search sensitivity, a broad initial search strategy was used with the primary concepts of population (“stroke” OR “cerebrovascular accident”) AND intervention (“task-oriented training” OR “task-oriented exercise”). The specific criteria for “robot-assisted” interventions and “randomized controlled trials” were applied during the manual screening phase.

Inclusion and exclusion criteria

Eligibility criteria were established according to the Patient Intervention Comparison Outcome (PICO) framework as follows:

Patients. Patients with stroke;

Interventions. Robot-assisted task-oriented exercise;

Comparisons: Comparison with conventional therapy or other interventions;

Outcomes: Upper limb function and ADLs.

Reviews, meta-analyses, letters, and conference proceedings were excluded. In addition, studies that did not include stroke survivors, lacked details on interventions or results, were case reports, or did not have available full texts were also excluded from the review.

Screening, selection, and extraction process

PRISMA 2020 guidelines were followed for study selection. Two researchers independently screened titles and abstracts and evaluated full-text articles. Inter-rater agreement for study selection was calculated (Cohen’s Kappa = 0.84). Any disagreements were resolved by consensus with a third independent researcher. Following study selection, the two reviewers independently extracted data, including study characteristics, intervention protocols, and outcome measures. Duplicate and irrelevant records were removed during the identification stage. Publications were first identified through title and abstract screening, followed by full-text review to identify studies consistent with the objectives of the present review.

Assessment of quality

The methodological quality of the selected studies was assessed using the Physiotherapy Evidence Database (PEDro) scale.²⁴ This scale is appropriate for evaluating clinical interventions and comprises 10 criteria: specified eligibility criteria; random allocation; concealed allocation; baseline group similarity; blinding of participants, therapists, and assessors; a dropout rate of less than 15%; intention-to-treat analysis; between-group statistical comparisons; point estimates; and variability data. Two reviewers independently assessed the risk of bias and methodological quality using the PEDro scale. Disagreements were resolved by consensus or by consultation with a third independent reviewer. Due to considerable heterogeneity among the included studies, particularly in terms of types of robotic devices, TOT protocols, and outcome measures, meta-analysis was not feasible. Therefore, a narrative synthesis was conducted to systematically summarize the findings.

Results

Study selection

A total of 1112 papers were identified as potential candidates during the database search. After excluding 529 duplicate papers and 503 papers based on title and abstract screening, 80 articles remained. Of these, 33 articles without available full texts were excluded, leaving 47 articles for review. Following full-text evaluation, 37 studies were excluded. Finally, 10 studies were included in this review. The PRISMA flow diagram is presented in Figure 1.

Figure 1.

PRISMA flowchart of study selection. PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Quality assessment and risk of bias

The PEDro scale was used to evaluate the quality of the selected studies. Of the 10 articles, 8 were of good quality (scores, 6–8 points), whereas the remaining 2 were of fair quality (scores, 4–5 points). A qualitative assessment of individual PEDro criteria showed that domains such as baseline comparability and between-group statistical comparisons were well satisfied; however, criteria related to the blinding of participants and therapists (criteria 5 and 6) were not met in the included trials. The detailed PEDro quality assessment results are presented in Table 1.^25–34

Table 1.

Assessment of methodological qualities as determined using the PEDro scale.

Study (year)	1	2	3	4	5	6	7	8	9	10	11	Total score
Hsieh et al.²⁵	Y	Y	Y	Y	Y	N	N	N (31%)	Y	Y	Y	7
Li et al.²⁹	Y	Y	N	Y	N	N	Y	Y (9%)	N	Y	Y	6
Timmermans et al.²⁷	Y	Y	Y	Y	N	N	N	Y (0%)	Y	Y	Y	7
Shi et al.³³	Y	Y	N	Y	N	N	Y	Y (0%)	Y	Y	Y	7
Frisoli et al.³¹	Y	Y	N	Y	N	N	N	N (15%)	N	Y	Y	4
Ambrosini et al.²⁸	Y	Y	Y	Y	N	N	Y	Y (11%)	N	Y	Y	7
Tomic et al.³⁰	Y	Y	N	Y	N	N	Y	Y (0%)	Y	Y	Y	7
Klamroth-Marganska et al.³²	Y	Y	Y	Y	N	N	Y	Y (5%)	Y	Y	Y	8
Keeling et al.³⁴	Y	N	N	Y	N	N	N	Y (0%)	Y	N	Y	4
Perini et al.²⁶	Y	Y	N	Y	N	N	Y	Y (0%)	Y	Y	Y	7

PEDro: Physiotherapy Evidence Database.

Study characteristics

Ten studies involving patients with stroke who underwent TOT with robotic devices were analyzed. Intervention characteristics, stroke phases, outcomes, and results are summarized in Table 2.^25–34

Table 2.

Characteristics of the studies included in the systematic review.

Study (year)	Age (year, mean ± SD)	Sample size (sex (M/F))	Time elapsed since stroke (mean ± SD)	Intervention	Outcomes
Hsieh et al.²⁵	E: 49.28 ± 10.90C: 52.87 ± 10.40	E: 16 (11/5)C: 15 (7/8)	E: 2.56 ± 1.69 (month)C: 2.21 ± 1.11 (month)	E: bilateral priming combined with a task-oriented approachC: task-oriented approach	FMA-UE, BBT, grip strength, FIM
Li et al.²⁹	E: 63.6 ± 11.4C: 63.7 ± 9.4	E: 20 (16/4)C: 20 (17/3)	E: 67.35 ± 47.31 (day)C: 70.55 ± 49.37 (day)	E: conventional upper limb rehabilitation therapy (40 min) + TOT assisted by a force feedback hand rehabilitation robot (20 min)C: conventional upper limb rehabilitation therapy (40 min) + TOT (20 min)	FMA-hand, ARAT, grip strength (N), BI
Timmermans et al.²⁷	E: 61.8 ± 6.8C: 56.8 ± 6.4	E: 11 (8/3)C: 11 (8/3)	E: 2.8 ± 2.9 (year)C: 3.7 ± 3.0 (year)	E: technology-supported task-oriented arm trainingC: arm-hand training program	FMA, ARAT
Shi et al.³³	E: 55.3 ± 16.8C: 57.5 ± 13.7	E: 17 (11/6)C: 17 (12/5)	E: 4.8 ± 1.8 (year)C: 4.3 ± 2.2 (year)	E: TOT by manipulating small objects with the robotic handC: hand-functional exercises without objects using the robot	FMA-UE, ARAT, BBT, grip strength, MAS
Frisoli et al.³¹	E: 62 ± 12C: 70 ± 11	E: 11 (7/4)C: 11 (8/3)	E: 30 ± 20 (month)C: 37 ± 24 (month)	E: robotic-aided therapyC: manual physical therapy	FMA
Ambrosini et al.²⁸	E: 60.9 ± 13.7C: 67.8 ± 12.2	E: 36 (25/11)C: 36 (25/11)	E: 63.4 ± 58.8 (day)C: 62.1 ± 76.9 (day)	E: advance conventional therapy (ACT) 60 min + 30 min with RETRAINER (robot)C: advance conventional therapy (ACT) 90 min	ARAT, BBT
Tomic et al.³⁰	E: 56.5 ± 7.4C: 58.3 ± 5.2	E: 13 (12/1)C: 13 (9/4)	E: 35.3 ± 9.7 (day)C: 37.3 ± 7.7 (day)	E: conventional rehabilitation (30 min) + arm assist (AA) robotic training (30 min)C: conventional rehabilitation (30 min) + occupational therapy (30 min)	FMA-UE, BI
Klamroth-Marganska et al.³²	E: 55 ± 13C: 58 ± 14	E: 38 (21/17)C: 35 (25/10)	E: 52 ± 44 (month)C: 40 ± 45 (month)	E: robotic therapy with ARMinC: common neurorehabilitation treatment	FMA-UE, grip strength
Keeling et al.³⁴	E: 54.3 ± 15.6C: 57.7 ± 15.1	E: 9 (8/1)C: 10 (9/1)	E: 42.7 ± 25.5 (day)C: 36.8 ± 15.5 (day)	E: robot therapy + standard of careC: standard of care	FMA-UE, ARAT, FIM
Perini et al.²⁶	E: 58.7 ± 20.6C: 61.4 ± 9	E: 9 (5/4)C: 9 (4/5)	E: 35.3 ± 44.5 (month)C: 42 ± 44.7 (month)	E: TO activity session with MeCFES + planar robotC: conventional upper limb rehabilitation therapy	FMA-UE, BBT

ARAT: Action Research Arm Test; BBT: Box and Block Test; BI: Barthel index; C: control group; E: experimental group; F: female; FIM: Functional Independence Measure; FMA-UE: Fugl-Meyer Assessment for Upper Extremity; M: male; TOT: task-oriented training.

All the 10 studies were RCTs, and the mean participant age was 59.2 years, reflecting predominantly middle-aged and older adults. Various robotic devices were used across the studies. Five studies were conducted during the subacute phase, five during the chronic phase, and none during the acute phase. Outcome measures included assessments of upper extremity function, including the Fugl-Meyer Assessment for Upper Extremity (FMA-UE) test, the Action Research Arm Test (ARAT), the Box and Block Test (BBT), and grip strength. Additional outcomes included ADL measures, such as the Barthel Index (BI) and the Functional Independence Measure (FIM).

The most commonly used assessment tool was the FMA-UE test, which was employed in 9 of the 10 studies. Five studies used the ARAT, four used the BBT, and four assessed grip strength. ADLs were evaluated in four studies, with two using the BI and two using the FIM.

Robot application type and clinical dimensions

The findings of the included RCTs were categorized and evaluated using three primary clinical dimensions: (a) intervention modality (standalone robot-assisted task-oriented therapy vs conventional rehabilitation with adjunct robot-assisted TOT); (b) stroke phase (subacute vs chronic); and (c) technological integration, including approaches incorporating functional electrical stimulation (FES) and electromyography (EMG)-driven control.

Standalone robot-assisted therapy versus adjunct therapy

Seven of the included studies applied robot‑assisted TOT as a standalone intervention and compared it with conventional therapy. In these studies, within‑group improvements were consistently observed in upper limb motor impairment, primarily measured using the FMA‑UE; however, several studies reported no significant between‑group differences compared with conventional therapy.^25–27 Conversely, the three studies used robot‑assisted training as an adjunct to conventional therapy and compared the combined approach with conventional therapy alone. All three reported significant between‑group differences, particularly in functional capacity measures such as the ARAT and BBT.^28–30

Efficacy based on stroke phase (subacute versus chronic)

The included trials were evenly distributed, with five studies targeting the subacute phase and five targeting the chronic phase. Robot-assisted TOT was associated with reductions in motor impairment, as measured by the FMA-UE, across both phases. Studies involving patients with chronic stroke demonstrated significant within-group improvements in upper limb functional capacity, including improvements in ARAT.^27,31–33 Similarly, studies conducted during the subacute phase reported significant within-group improvements.^{25,26,29,30,34} However, between-group superiority over equivalent intensive conventional therapy varied across trials.

Integration of advanced technologies (functional electrical stimulation and electromyopraphy-driven systems)

Regarding technological characteristics, two studies combined robotic therapy with FES, and both reported significant within-group and between-group functional improvements.^26,28 In another study, a personalized EMG-driven soft robotic hand was used to manipulate real-world objects.³³ These technologically integrated approaches demonstrated significant between-group superiority in motor function and functional capacity compared with conventional therapy or passive robotic assistance in control groups.

Discussion

This systematic review analyzed 10 experimental studies evaluating upper limb function and ADLs in patients with stroke who underwent robot-assisted TOT between 2014 and 2025. The aim of the review was to investigate the effectiveness of this training and inform the development of intervention programs. The studies were categorized as robot-based therapy alone, conventional therapy combined with robot therapy, robot therapy combined with FES, and studies that included robotic interventions in the control group. The effects of these different robot-assisted approaches were then compared. Overall, robot-assisted interventions had a positive impact on upper limb function, and some studies reported improvements in ADLs.

Compared with existing high-level evidence, such as the RATULS trial, the Cochrane review by Mehrholz et al.,³⁵ and other systematic reviews,^36–38 our findings are consistent in demonstrating improvements in impairment. However, these prior studies often evaluated the general effects of robot-assisted therapy by including a broad range of training modalities, such as simple or passive repetitive movement exercises. In contrast, the present review focuses exclusively on robot-assisted TOT, which emphasizes active motor learning and goal-directed functional task practice. Furthermore, unlike earlier reviews that included older literature, this review extends prior evidence by including trials published between 2014 and 2025. The findings were further categorized into specific clinical dimensions, including standalone versus adjunct interventions, stroke phase, and integration of advanced technologies. Notably, recent RCTs represent a paradigm shift toward intention-driven motor learning. For example, the integration of FES with robotics²⁸ and the use of EMG-driven soft robotic devices to manipulate real-world objects³³ require active patient engagement in sensorimotor integration. These targeted approaches suggest a trajectory that overcomes the functional limitations identified in earlier large-scale reviews.

Effect of standalone robot-assisted task-oriented therapy

In the seven studies evaluating standalone robot-assisted task-oriented therapy, measures of motor impairment consistently demonstrated significant improvements (Table 2). Robotic systems can deliver interventions that are more intensive than conventional physical therapy.¹⁴ Dimyan and Cohen³⁹ suggested that interventions such as robot-assisted training are essential for enhancing motor rehabilitation based on a neuroplasticity rationale, which suggests that precise movement training and repetitive task practice using robots contribute to motor learning reorganization.

However, some studies reported no significant differences between robot-assisted therapy and control groups. For example, Timmermans et al.²⁷ and Hsieh et al.²⁵ found no significant between-group differences in specific impairment or functional capacity measures, such as the FMA-UE and the BBT. In these studies, substantial recovery was also observed in patients receiving conventional therapy, which narrowed the comparative margin for robotic intervention. Interestingly, a meta-analysis by Chien et al.³⁶ reported no significant difference between robot-assisted therapy and conventional therapy in patients with subacute stroke, suggesting that outcomes may depend on patient characteristics and intervention design.

Furthermore, robot-assisted therapy resulted in nonsignificant ADL improvements in trials conducted by Hsieh et al,²⁵ and Keeling et al.³⁴ reported no significant between-group differences in FIM scores. The discrepancy between observed impairment-level recovery and functional independence may be because of the nature of the robotic devices. Although robotic devices excel at providing high-dose repetitive training, they often lack the capacity to address the unpredictable, multitasking demands required for real-world ADLs.⁴⁰ Given that these findings on daily functioning are based on only a limited number of trials, further studies are warranted to draw definitive conclusions regarding ADL outcomes. Ultimately, these conflicting results suggest that standalone robotic therapy may not consistently guarantee superior outcomes across clinical settings. Therefore, future studies should focus on identifying the specific patient characteristics and optimal intervention parameters required to maximize the efficacy of robotic therapy. Additionally, to enhance the transfer of motor gains to independence in ADLs, emerging evidence suggests that robotic therapy should be combined with training in unconstrained, real-world environments.^41,42

Synergistic effects of conventional therapy with additional robot therapy

In the three studies that implemented robotic training as an adjunct to conventional therapy, the combined approach consistently yielded superior motor recovery compared with conventional therapy. Particularly, Li et al.²⁹ and Tomic et al.³⁰ reported significant between-group advantages in upper limb functional assessments, including the FMA-UE and the ARAT (Table 2). These findings suggest that robot-assisted training can complement conventional therapy and accelerate functional recovery. Furthermore, a meta-analysis by Jin et al.⁴³ reported that robot-assisted TOT significantly improved FMA-UE scores (standardized mean difference = 1.01; 95% confidence interval: 0.57–1.45), consistent with our findings. However, another study reported that robot-assisted therapy is less cost-effective compared with conventional therapy.⁴⁴

However, ADL outcomes were mixed despite clear improvements in motor impairment with adjunct robot therapy. Although Li et al. reported significant between-group superiority, as determined by Barthel Indices,²⁹ Tomic et al. observed no significant difference.³⁰ These findings indicate that improvements in arm movement do not necessarily translate into independence in daily functioning and suggest that longer or more comprehensive interventions may be required. These results are broadly consistent with a meta-analysis by Boardsworth et al.,⁴⁵ which reported small but significant improvements in upper limb function but limited effects on ADLs.

Influence of stroke phase on therapeutic outcomes

Robot‑assisted TOT improved motor impairment in the subacute or chronic phase. In the chronic phase, patients achieved substantial functional recovery, suggesting that high-intensity, repetitive robotic training can successfully overcome the plateau in motor recovery by promoting neuroplasticity.⁴⁶ In contrast, trials conducted in the subacute phase reported variable results compared with equally intensive conventional therapy. This variability may reflect the confounding effects of spontaneous neurological recovery characteristic of the subacute phase. During this early phase, patients with stroke receiving conventional therapy may also experience significant spontaneous gains, thereby reducing the relative advantage of robotic interventions.⁴⁷ Therefore, although robotic therapy is highly effective across all stages, its relative superiority over conventional care is more evident in patients with chronic stroke.

Impact of integrating advanced neurorehabilitation technologies

The integration of advanced technologies, such as FES and EMG-driven systems, was associated with clear improvements in motor function and functional capacity. Interventions used by Ambrosini et al.,²⁸ Perini et al.,²⁶ and Shi et al.³³ actively engaged residual voluntary effort in patients with stroke rather than providing passive mobilization.⁴⁸ By synchronizing robotic assistance with patients’ active motor commands, these approaches promote stronger coupling between cortical descending signals and peripheral sensory feedback. This sensorimotor integration is a key mechanism underlying effective motor learning and may explain why robots equipped with active, neural-triggered interfaces produce more robust functional improvements and greater cerebral cortex activation than standard robotic assistance.⁴⁹

Differences in interpretation based on assessment tools

The FMA was the most commonly used tool in this review. Although there were consistent reports of improvements in upper limb function, the extent of improvement in ADL measures (FIM and BI) was unclear. These findings are consistent with those of Mehrholz et al.³⁵ in their Cochrane review, which demonstrated improvements in upper limb function but no significant differences in ADLs. Schepers et al.⁵⁰ reported that both the FIM and the BI are effective for assessing changes in ADLs. However, the limited extent of these changes reported suggests that improvements in daily living functions may not be achievable by enhancements in movement alone. These findings indicate that more comprehensive intervention strategies are needed to translate functional improvement directly into ADL performances, and that training strategies based on functional and short-term movement improvements should be implemented in parallel.

The participants in the included studies were randomly assigned to experimental and control groups. The mean PEDro score across the 10 studies was 6.4, indicating high methodological quality (PEDro scale, 6–10). This scale evaluates internal validity and methodological quality of studies, and the mean score of the included studies exceeded the PEDro database average of 5.⁵¹ Therefore, despite the inherent difficulty of blinding participants and therapists due to the nature of the interventions, the internal validity of this review is considered acceptable.

Future research directions

Robot-assisted therapy may be effective for upper limb recovery in patients with stroke. However, studies to date have not optimized the conditions required to maximize its effects. Furthermore, although many studies have focused on the subacute and chronic stages of stroke, limited evidence is available on the application of robotic therapy in the acute stage. Therefore, further investigation is required to understand the effects of robotic interventions during early rehabilitation. In addition, future studies should address the relative sensitivity of evaluation tools and the heterogeneity among patient groups.

Study limitations

The limitations of this systematic review are as follows:

Sample sizes were relatively small, increasing the risk of false-negative results. In addition, as identified in the quality assessment, the inability to blind participants and therapists to robotic interventions introduces a structural risk of performance bias.

The included studies varied in terms of robot device type, intervention intensity, training frequency and duration, and stroke phase (subacute vs chronic). This heterogeneity may limit comprehensive interpretation.

Upper limb function measures, such as the FMA and ARAT, are appropriate for precise functional assessment. However, ADL measures, such as the BI or FIM, are better suited for evaluating gross functional levels and may be inadequate for capturing fine motor changes.

Few studies assessed ADL outcomes, limiting the ability of our review to determine whether functional recovery is associated with quality of life.

Of the 10 studies, 7 reported outcomes immediately postintervention; therefore, evidence on long-term follow-up and maintenance effects is limited.

The literature search was restricted to English-language publications, which may have introduced language bias by excluding relevant trials published in other languages.

Conclusions

This systematic literature review was conducted to examine the effects of robot-assisted TOT on upper extremity function and ADLs in patients with stroke. The findings indicate that robot-assisted interventions consistently improve upper extremity function; however, they have limited effects on indicators of ADL. Furthermore, clinical outcomes were influenced by intervention modality, such as standalone versus adjunct therapy, stroke phase, and the integration of advanced technologies. However, evidence regarding the efficacy of this technique during the acute phase and long-term maintenance remains limited. Therefore, future high-quality randomized trials should focus on optimizing these specific clinical dimensions and developing strategies to facilitate the transfer of motor recovery to meaningful functional independence.

Footnotes

Acknowledgments

Generative AI tools were used exclusively for language editing and proofreading during the preparation of this manuscript, and the authors take full responsibility for the final content.

Author contributions

Yong-Ho Cho, Myoung-Ho Lee, and Yong-Jun Cha contributed to the study conception and design. Myoung-Ho Lee performed data collection and the statistical analysis. Yong-Ho Cho and Yong-Jun Cha critically revised the manuscript. Yong-Jun Cha supervised the project and approved the final version. All authors read and approved the final manuscript.

Data availability statement

All data analyzed during this study are included in this article. No new primary data were created in this systematic review.

Declaration of conflicting interests

The authors declare no potential conflicts of interest with respect to the research, authorship, or publication of this article.

Ethics approval and consent to participate

The need for informed patient consent was waived due to the retrospective nature of the study.

Clinical trial registration number: Not applicable.

Open Science Framework registration DOI: 10.17605/OSF.IO/4DT6G.

Human Ethics and Consent to Participate declarations: Not required.

Informed Consent: Not required.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Yong-Jun Cha

References

Potter

TBH

Tannous

Vahidy

FS.

A contemporary review of epidemiology, risk factors, etiology, and outcomes of premature stroke. Curr Atheroscler Rep 2022; 24: 939–948.

Feigin

Krishnamurthi

Parmar

; GBD 2013 Stroke Panel Experts Groupet al. Update on the global burden of ischemic and hemorrhagic stroke in 1990-2013: the GBD 2013 study. Neuroepidemiology 2015; 45: 161–176.

Bonifer

Anderson

Arciniegas

DB.

Constraint-induced movement therapy after stroke: efficacy for patients with minimal upper-extremity motor ability. Arch Phys Med Rehabil 2005; 86: 1867–1873.

Pollock

Farmer

Brady

, et al. Interventions for improving upper limb function after stroke. Cochrane Database Syst Rev 2014; 2014: CD010820.

Almeida

OP.

Stroke, depression, and self-harm in later life. Curr Opin Psychiatry 2023; 36: 371–375.

Chen

Effectiveness of using virtual reality-supported exercise therapy for upper extremity motor rehabilitation in patients with stroke: systematic review and meta-analysis of randomized controlled trials. J Med Internet Res 2022; 24: e24111.

Chen

, et al. Effectiveness, safety and patients' perceptions of an immersive virtual reality-based exercise system for poststroke upper limb motor rehabilitation: a proof-of-concept and feasibility randomized controlled trial. Digit Health 2023; 9: 20552076231203599.

Wen

, et al. Therapeutic role of additional mirror therapy on the recovery of upper extremity motor function after stroke: a single-blind, randomized controlled trial. Neural Plast 2022; 2022: 8966920.

Barclay

Stevenson

Poluha

, et al. Mental practice for treating upper extremity deficits in individuals with hemiparesis after stroke. Cochrane Database Syst Rev 2020; 5: Cd005950.

10.

Kerimov

Coskun Benlidayi

Ozdemir

, et al. The effects of upper extremity isokinetic strengthening in post-stroke hemiplegia: a randomized controlled trial. J Stroke Cerebrovasc Dis 2021; 30: 105729.

11.

Saygili

Guclu-Gunduz

Eldemir

, et al. Effects of modified-constraint induced movement therapy based telerehabilitation on upper extremity motor functions in stroke patients. Brain Behav 2024; 14: e3569.

12.

Almhdawi

Mathiowetz

White

, et al. Efficacy of occupational therapy task-oriented approach in upper extremity post-stroke rehabilitation. Occup Ther Int 2016; 23: 444–456.

13.

Murakami

Honaga

Kono

, et al. New artificial intelligence-integrated electromyography-driven robot hand for upper extremity rehabilitation of patients with stroke: a randomized, controlled trial. Neurorehabil Neural Repair 2023; 37: 298–306.

14.

Guarino

Richards

, et al. Robot-assisted therapy for long-term upper-limb impairment after stroke. N Engl J Med 2010; 362: 1772–1783.

15.

Halford

Jakubiszak

Krug

, et al. What is task-oriented training? a scoping review. SJOT 2024; 4: 1–23.

16.

Arman

Tarakci

, et al. Effects of video games–based task-oriented activity training (Xbox 360 Kinect) on activity performance and participation in patients with juvenile idiopathic arthritis: a randomized clinical trial. Am J Phys Med Rehabil 2019; 98: 174–181.

17.

Hussain

Fatima

Ahmad

, et al. Effects of task oriented rehabilitation of upper extremity after stroke: a systematic review. J Pak Med Assoc 2022; 72: 1406–1415.

18.

Mazzoleni

Sale

Tiboni

, et al. Upper limb robot-assisted therapy in chronic and subacute stroke patients: a kinematic analysis. Am J Phys Med Rehabil 2013; 92: e26–e37.

19.

Gerardin

Bontemps

Babuin

, et al. Bimanual motor skill learning with robotics in chronic stroke: comparison between minimally impaired and moderately impaired patients, and healthy individuals. J Neuroeng Rehabil 2022; 19: 28.

20.

Calabrò

Morone

Naro

, et al. Robot-assisted training for upper limb in stroke (ROBOTAS): an observational, multicenter study to identify determinants of efficacy. J Clin Med 2021; 10: 5245.

21.

Cho

Song

WK.

Robot-assisted reach training with an active assistant protocol for long-term upper extremity impairment poststroke: a randomized controlled trial. Arch Phys Med Rehabil 2019; 100: 213–219.

22.

Rodgers

Bosomworth

Krebs

, et al. Robot assisted training for the upper limb after stroke (RATULS): a multicentre randomised controlled trial. Lancet 2019; 394: 51–62.

23.

Page

McKenzie

Bossuyt

, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Syst Rev 2021; 10: 89.

24.

Maher

Sherrington

Herbert

, et al. Reliability of the PEDro scale for rating quality of randomized controlled trials. Phys Ther 2003; 83: 713–721.

25.

Hsieh

Wang

, et al. Bilateral robotic priming before task-oriented approach in subacute stroke rehabilitation: a pilot randomized controlled trial. Clin Rehabil 2017; 31: 225–233.

26.

Perini

Bertoni

Thorsen

, et al. Sequentially applied myoelectrically controlled FES in a task-oriented approach and robotic therapy for the recovery of upper limb in post-stroke patients: A randomized controlled pilot study. Technol Health Care 2021; 29: 419–429.

27.

Timmermans

Lemmens

Monfrance

, et al. Effects of task-oriented robot training on arm function, activity, and quality of life in chronic stroke patients: a randomized controlled trial. J Neuroeng Rehabil 2014; 11: 45.

28.

Ambrosini

Gasperini

Zajc

, et al. A robotic system with EMG-triggered functional eletrical stimulation for restoring arm functions in stroke survivors. Neurorehabil Neural Repair 2021; 35: 334–345.

29.

Lian

Chen

, et al. Effect of task-oriented training assisted by force feedback hand rehabilitation robot on finger grasping function in stroke patients with hemiplegia: a randomised controlled trial. J Neuroeng Rehabil 2024; 21: 77.

30.

Tomić

Savić

Vidaković

, et al. ArmAssist robotic system versus matched conventional therapy for poststroke upper limb rehabilitation: a randomized clinical trial. Biomed Res Int 2017; 2017: 7659893.

31.

Frisoli

Barsotti

Sotgiu

, et al. A randomized clinical control study on the efficacy of three-dimensional upper limb robotic exoskeleton training in chronic stroke. J Neuroeng Rehabil 2022; 19: 14.

32.

Klamroth-Marganska

Blanco

Campen

, et al. Three-dimensional, task-specific robot therapy of the arm after stroke: a multicentre, parallel-group randomised trial. Lancet Neurol 2014; 13: 159–166.

33.

Shi

, et al. Task-oriented training by a personalized electromyography-driven soft robotic hand in chronic stroke: a randomized controlled trial. Neurorehabil Neural Repair 2024; 38: 595–606.

34.

Keeling

Piitz

Semrau

, et al. Robot enhanced stroke therapy optimizes rehabilitation (RESTORE): a pilot study. J Neuroeng Rehabil 2021; 18: 10.

35.

Mehrholz

Pohl

Platz

, et al. Electromechanical and robot‐assisted arm training for improving activities of daily living, arm function, and arm muscle strength after stroke. Cochrane Database Syst Rev 2018; 9: CD006876.

36.

Chien

Chong

Tse

, et al. Robot‐assisted therapy for upper‐limb rehabilitation in subacute stroke patients: a systematic review and meta‐analysis. Brain Behavior 2020; 10: e01742.

37.

Kwakkel

Kollen

Krebs

HI.

Effects of robot-assisted therapy on upper limb recovery after stroke: a systematic review. Neurorehabil Neural Repair 2008; 22: 111–121.

38.

Veerbeek

Langbroek-Amersfoort

Van Wegen

, et al. Effects of robot-assisted therapy for the upper limb after stroke. Neurorehabil Neural Repair 2017; 31: 107–121.

39.

Dimyan

Cohen

LG.

Neuroplasticity in the context of motor rehabilitation after stroke. Nat Rev Neurol 2011; 7: 76–85.

40.

Basteris

Nijenhuis

Stienen

, et al. Training modalities in robot-mediated upper limb rehabilitation in stroke: a framework for classification based on a systematic review. J Neuroeng Rehabil 2014; 11: 111.

41.

Maier

Ballester

Verschure

Principles of neurorehabilitation after stroke based on motor learning and brain plasticity mechanisms. Front Syst Neurosci 2019; 13: 74.

42.

Takebayashi

Takahashi

Amano

, et al. Robot-assisted training as self-training for upper-limb hemiplegia in chronic stroke: a randomized controlled trial. Stroke 2022; 53: 2182–2191.

43.

Jin

Chen

Effectiveness of robot-assisted task-oriented training intervention for upper limb and daily living skills in stroke patients: a meta-analysis. PLOS One 2025; 20: e0316633.

44.

Wagner

Peduzzi

, et al. An economic analysis of robot-assisted therapy for long-term upper-limb impairment after stroke. Stroke 2011; 42: 2630–2632.

45.

Boardsworth

Rashid

Olsen

, et al. Upper limb robotic rehabilitation following stroke: a systematic review and meta-analysis investigating efficacy and the influence of device features and program parameters. J NeuroEng Rehabil 2025; 22: 164.

46.

Gassert

Dietz

Rehabilitation robots for the treatment of sensorimotor deficits: a neurophysiological perspective. J Neuroeng Rehabil 2018; 15: 46.

47.

Cramer

SC.

Repairing the human brain after stroke: I. Mechanisms of spontaneous recovery. Ann Neurol 2008; 63: 272–287.

48.

Micera

Caleo

Chisari

, et al. Advanced neurotechnologies for the restoration of motor function. Neuron 2020; 105: 604–620.

49.

Yang

Choi

JB.

Effect of 3-dimensional robotic therapy combined with electromyography-triggered neuromuscular electrical stimulation on upper limb function and cerebral cortex activation in stroke patients: a randomized controlled trial. Bioengineering (Basel) 2023; 11: 12.

50.

Schepers

Ketelaar

Visser-Meily

, et al. Responsiveness of functional health status measures frequently used in stroke research. Disabil Rehabil 2006; 28: 1035–1040.

51.

Yamato

Arora

Stevens

, et al. Quality, language, subdiscipline and promotion were associated with article accesses on Physiotherapy Evidence Database (PEDro). Physiotherapy 2018; 104: 122–128.