Two decades of AI-driven motion capture in rehabilitation: Mapping research networks,thematic hotspots,and future trajectories

Abstract

Background

Motion capture technology integration in artificial intelligence (AI)-driven rehabilitation represents a rapidly expanding interdisciplinary field with significant potential for advancing movement analysis and motor recovery. A comprehensive bibliometric mapping of this domain is currently lacking, limiting systematic understanding of its development trajectory and key contributors.

Objective

To provide a WoS-indexed bibliometric analysis of AI applications in motion capture for rehabilitation, identifying research trends, collaboration networks, key contributors, and emerging research frontiers from 2004 to 2023.

Methods

A total of 3,500 relevant publications indexed in the Science Citation Index Expanded (SCI-EXPANDED) and Social Sciences Citation Index (SSCI) databases were retrieved and analyzed. Bibliometric and visualization analyses were performed using VOSviewer (version 1.6.19) and CiteSpace (version 6.2.4) to map collaboration networks, co-citation relationships, and keyword co-occurrence patterns.

Results

Annual publication output demonstrated consistent growth from 2004 to 2023, with cumulative publications exceeding 3,500. The United States (1,014 publications) and China (722 publications) dominated research output, although collaboration patterns differed substantially. The University of Chinese Academy of Sciences led institutional contributions (52 publications). Keyword clustering revealed prominent research themes centered on brain-computer interfaces, machine learning, EEG-based signal processing, and real-time rehabilitation feedback systems. Temporal analysis demonstrated a paradigm shift from fundamental neurophysiological investigations toward computationally-driven and AI-integrated rehabilitation frameworks.

Conclusions

This bibliometric analysis provides a WoS-indexed mapping of AI-driven motion capture research in rehabilitation. The identified research hotspots and collaboration patterns offer a foundational reference for future investigations, despite limitations related to database coverage and language scope. Continued interdisciplinary collaboration and standardized methodological frameworks are essential to accelerate clinical translation.

Keywords

motion capture rehabilitation medicine artificial intelligence digital health bibliometrics research trends gait analysis

Highlights

• First WoS-indexed bibliometric overview of AI-driven motion capture in rehabilitation: This study consolidates and maps a scattered yet rapidly expanding research field.

• Key thematic shifts identified: Research has progressively transitioned from basic neurophysiological studies toward AI-integrated, real-time, and data-driven rehabilitation frameworks.

• Global collaboration patterns mapped: The United States serves as a global research hub, while China demonstrates accelerated output with predominantly regional collaboration patterns.

• Actionable guidance for future research: Identified hotspots in EEG, brain-computer interfaces, and multimodal sensing point toward the most impactful directions for technology-driven rehabilitation.

1. Introduction

The convergence of motion capture technology and rehabilitation practices has emerged as a significant area of interdisciplinary research over the past two decades. This intersection represents a promising frontier in digital healthcare, where technological innovations are poised to transform traditional approaches to movement analysis and motor recovery.^1,2 Motion capture systems enable precise, quantitative detection of human movement, offering unprecedented opportunities for developing personalized rehabilitation strategies informed by real-time data. This capability is particularly critical in addressing motor impairments where conventional rehabilitation approaches are limited by subjectivity and low resolution of movement assessment.³

The rapid advancement of artificial intelligence (AI) and machine learning technologies has further amplified the potential of motion capture in rehabilitation contexts. AI algorithms are now capable of extracting clinically meaningful features from complex movement datasets, enabling automated assessment of motor function, prediction of rehabilitation trajectories, and adaptive feedback in real time.^4,5 Beyond rehabilitation, AI-based predictive modeling has demonstrated utility in adjacent domains—most notably in injury prediction frameworks within sports science, where machine learning has been applied to forecast musculoskeletal injury risk and optimize return-to-play protocols.⁶ Such cross-domain applications further underscore the translational potential of AI-driven movement analysis. Convolutional neural networks, recurrent architectures, and reinforcement learning frameworks have each found application in this domain, and their integration with wearable sensors and markerless motion capture systems is reshaping both research and clinical practice.^7,8

Despite this growth, the research landscape remains fragmented across disciplines including biomedical engineering, neuroscience, physical therapy, and computer science. A comprehensive synthesis of publication patterns, collaboration structures, and thematic evolution is therefore essential not only for researchers seeking to navigate the field but also for clinicians, policymakers, and funding agencies aiming to direct resources toward the most impactful areas.^9,10 Bibliometric analysis provides a validated quantitative framework for such synthesis, enabling systematic mapping of research output, intellectual networks, and emerging frontiers.¹¹

To address this gap, we conducted a systematic bibliometric analysis of publications from 2004 to 2023 focusing on the intersection of motion capture technology and rehabilitation. Using VOSviewer and CiteSpace as complementary scientometric platforms, this study maps collaboration networks, identifies research hotspots, and traces the thematic evolution of the field. The objective is to provide a data-driven foundation for understanding the trajectory of AI-integrated motion capture research in rehabilitation and to guide future interdisciplinary investigations and clinical translation efforts.

2. Materials and methods

2.1. Search strategy and data collection

All relevant publications published between January 1, 2004 and December 31, 2023 were retrieved from the Science Citation Index Expanded (SCI-EXPANDED) and Social Sciences Citation Index (SSCI) databases within the Web of Science (WoS) Core Collection. To minimize bias arising from real-time database updates, all records were downloaded on a single date. The search strategy was defined as: TS = (“motion capture” OR “movement detection”) AND TS = (“rehabilitation” OR “rehab” OR “movement training” OR “motor adaptation”). This deliberately inclusive strategy was designed to capture the broader landscape of motion capture applications in rehabilitation from which AI-driven methodologies have progressively emerged over the two-decade study period. While the search did not require explicit AI, machine learning, or deep learning terminology, subsequent keyword co-occurrence analysis confirms that AI-related terms dominate the recent literature within the retrieved corpus, validating the relevance of the dataset to the study’s focus on AI-driven research. The publication year range was limited to 2004–2023, and only English-language articles and reviews were included.

All retrieved records were exported in “Plain Text File” format with “Full Record and Cited References” selected. Inclusion criteria: (1) original research articles or systematic/scoping reviews; (2) published in English; (3) indexed in WoS SCI-EXPANDED or SSCI; (4) addressing motion capture technology applied in a rehabilitation context. Exclusion criteria: (1) conference abstracts, book chapters, editorials, letters, or retracted publications; (2) studies focused exclusively on healthy populations without rehabilitation relevance; (3) animal studies; (4) duplicates. Manual screening was subsequently performed by two independent reviewers to exclude irrelevant records based on titles and abstracts; discrepancies were resolved by consensus. The detailed search flowchart is presented in Figure 1. Inter-rater reliability was assessed using Cohen’s kappa coefficient, yielding κ = 0.85, indicating high agreement.

Figure 1.

Flow chart for the publications selection.

Animal studies were excluded through a two-stage procedure: a Web of Science document-type filter restricting results to human studies was applied during database export, and any residual animal studies were identified and excluded during the subsequent manual title and abstract screening phase conducted by two independent reviewers.

2.2. Bibliometric and visualization analysis

Bibliometric analyses were performed using VOSviewer (version 1.6.19) and CiteSpace (version 6.2.4). VOSviewer was employed to construct co-authorship networks, bibliographic coupling maps, and keyword co-occurrence networks. For co-authorship networks, a minimum threshold of 5 publications per author and 25 citations was applied. For keyword co-occurrence, only keywords appearing at least 15 times across the corpus were included. Normalization was performed using the association strength method. In the resulting visualizations, node size reflects publication count, line thickness indicates collaboration strength, and node color differentiates clusters.¹² CiteSpace was used to generate temporal knowledge maps, detect citation burst patterns, and perform keyword clustering using the log-likelihood ratio (LLR) algorithm. For CiteSpace, node selection was set to Top 50 per slice with a two-year time slice. The threshold for citation burst detection was set at γ = 0.25. These complementary tools enabled both static mapping of collaborative structures and dynamic analysis of research trend evolution.

Prior to analysis, institutional names and author identifiers were standardized to address discrepancies arising from varying affiliation reporting conventions and transliteration differences. All processed data were imported into both software packages in plain text format.

Cluster labels were assigned through manual author interpretation of the highest-weighted central nodes within each cluster as identified by CiteSpace’s LLR algorithm, rather than automated algorithmic extraction. Each label reflects the dominant topical theme represented by the five highest-frequency keywords in that cluster. The direct automated LLR cluster labels alongside the corresponding author interpretations are provided in Supplementary Table S1 to ensure analytical transparency and reproducibility.

3. Results

3.1. Publication trends and global distribution

The bibliometric analysis identified 3,500 relevant publications spanning 2004 to 2023. Both annual and cumulative publication volumes demonstrated a consistent upward trajectory over this period, with a pronounced acceleration beginning around 2018, coinciding with widespread adoption of deep learning frameworks and the increasing availability of wearable motion capture devices (Figure 2). Cumulative publications exceeded 3,500 by the end of the study period, reflecting exponential growth in academic interest.

Figure 2.

The number of annual and accumulative publications.

Geographically, the United States dominated research output with 1,014 publications, followed by China with 722 publications (Table 1). European nations, particularly England (332 publications) and Germany (286 publications), also demonstrated robust research activity. Notably, publication volume did not consistently predict citation impact. The Netherlands, for example, ranked 14th in total output yet achieved the highest mean citation rate at approximately 30 citations per publication, suggesting that high-impact research is not confined to the most prolific nations. The international collaboration network (Figure 3) identified the United States as the central hub, maintaining dense collaborative linkages with both European and Asian research communities. China, despite its high output, demonstrated a more regionally concentrated collaboration pattern, primarily engaging with East Asian institutions.

Table 1.

The top 15 countries/regions with the most publications.

Rank	Country	Publications	Total citations	Mean citations/publication
1	USA	1014	24126	24
2	China	722	9884	14
3	UK	332	7990	24
4	Germany	286	8106	28
5	Canada	245	5013	20
6	Italy	234	5965	25
7	Australia	204	4011	20
8	South Korea	179	2111	12
9	Japan	173	3216	19
10	Spain	161	2960	18
11	France	146	2551	17
12	Switzerland	131	2227	17
13	India	121	1745	14
14	Netherlands	111	3327	30
15	Taiwan	105	1347	13

Figure 3.

(a): Growth trends of the top 10 countries from 2014 to 2023; (b): The network visualization map of collaborations between countrise/regions.

3.2. Institutional contributions and impact

Among contributing institutions, the University of Chinese Academy of Sciences led in publication volume with 52 articles, followed by Aalborg University with 40 publications (Table 2). Chinese institutions collectively accounted for four of the top 15 contributing organizations, underscoring China’s growing research presence. Harvard University, the University of Tokyo, and Stanford University achieved the highest mean citation rates per publication (52, 42, and 35, respectively), demonstrating that institutional impact is not solely a function of output volume. The institutional collaboration network (Figure 4(a)) revealed tight clustering among North American, Western European, and East Asian institutions, with evidence of established transnational partnerships in both methodology development and clinical application domains.

Table 2.

The top 15 institutions with the most publications.

Rank	Organization	Country	Publications	Total citations	Mean citations/publication
1	University of Chinese Academy of Sciences	China	52	721	14
2	Aalborg University	Denmark	40	1397	35
3	University of Toronto	Canada	34	478	14
4	Swiss Federal Institute of Technology	Switzerland	33	422	13
5	Harvard Medical School	USA	30	468	16
6	Zhejiang University	China	30	523	17
7	Northwestern University	USA	28	748	27
8	Shanghai Jiao Tong University	China	27	482	18
9	University of Melbourne	Australia	27	614	23
10	Beihang University	China	26	786	30
11	Harvard University	USA	26	1361	52
12	Imperial College London	UK	26	333	13
13	McGill University	Canada	25	699	28
14	Stanford University	USA	25	871	35
15	University of Tokyo	Japan	25	1049	42

Figure 4.

(a): Geographic distribution of academic Institutions, (b): Author distribution and collaboration networks.

3.3. Author collaboration networks

Author co-authorship analysis identified 34 researchers with five or more publications in the field. The top 10 contributing authors collectively produced 148 articles and accumulated 4,688 citations (Table 3). High-impact authors such as Dario Farina (19 publications, 1,110 citations) and Imran Khan Niazi (18 publications, 754 citations) served as nucleation points for distinct research clusters. The author collaboration network (Figure 4(b)) revealed that these researchers frequently function as bridges across disciplinary and institutional boundaries, connecting engineering-oriented and clinically-oriented research communities. Collaboration clusters corresponded to specific methodological domains, including EEG-based interfaces, inertial sensor applications, and markerless motion analysis.

Table 3.

The top 15 authors with the most publications.

Rank	Author	Publications	Citations	Mean citations/publication
1	Farina, Dario	19	1110	58
2	Niazi, Imran Khan	18	754	42
3	Dremstrup, Kim	13	1044	80
4	Mrachacz-kersting, Natalie	13	802	62
5	Jiang, Ning	12	1042	87
6	Jochumsen, Mads	12	270	23
7	Xie, Longhan	9	73	8
8	Foreman, K. bo	8	83	10
9	Hager, Charlotte K.	8	50	6
10	Nakamura, Yoshihiko	8	95	12
11	Wang, Wei	8	97	12

3.4. Keyword analysis and thematic evolution

Keyword co-occurrence analysis using VOSviewer identified 176 terms appearing at least 15 times across the corpus (Figure 5(a)). Central nodes included “gait analysis”, “machine learning”, “brain-computer interface”, “deep learning”, and “rehabilitation”, reflecting the field’s broad interdisciplinary scope. The temporal overlay visualization (Figure 5(b)) revealed a clear chronological shift: earlier publications (2004–2014) were dominated by terms associated with basic motion kinematics and neurophysiology, while more recent publications (2018–2023) are characterized by computationally intensive terms including “convolutional neural network”, “transfer learning”, and “real-time feedback”.

Figure 5.

(a): Network visualization of keyword Co-occurrence, (b): Temporal evolution of research Themes, (c): Density visualization of research hotspots.

CiteSpace keyword clustering using the log-likelihood ratio algorithm generated eight thematic modules with a modularity Q-value of 0.42 and mean silhouette S-value of 0.71, both exceeding the recommended thresholds for reliable clustering (Q > 0.3, S > 0.5). Cluster labels were assigned through manual author interpretation of the highest-weighted central nodes within each cluster as identified by CiteSpace’s LLR algorithm, rather than automated algorithmic extraction. Each label reflects the dominant topical theme represented by the five highest-frequency keywords in that cluster. The dominant cluster centered on “motion capture systems”, while other prominent clusters included “EEG signal processing”, “wearable sensor rehabilitation”, “fall risk assessment”, and “AI-assisted gait analysis” (Figure 5(c)). Citation burst analysis identified “OpenPose-based pose estimation”, “inertial measurement unit gait monitoring”, and “virtual reality rehabilitation” as terms with the most recent and sustained citation surges, pointing to these as priority areas in the current research frontier.

4. Discussion

This bibliometric analysis provides a WoS-indexed mapping of AI-integrated motion capture research in rehabilitation from 2004 to 2023, identifying the field’s developmental trajectory, leading contributors, and emerging thematic directions. As the first systematic bibliometric study of this specific domain within the WoS Core Collection, our findings offer a foundational reference for researchers and clinicians seeking to navigate an expanding and interdisciplinary literature.

4.1. Publication trends and geographic patterns

The exponential growth in publications—particularly the acceleration beginning around 2018—parallels major technological milestones including breakthroughs in deep learning, the democratization of wearable sensors, and the expanding availability of large-scale motion datasets.^13,14 The geographic distribution of research output highlights the United States and China as dominant contributors, yet their collaborative behaviors differ markedly. The United States functions as a global knowledge hub with dense international linkages, while China’s research network is more regionally concentrated.^15,16 This pattern mirrors broader observations in global biomedical research, where increasing Chinese output has not yet been uniformly accompanied by proportional expansion of international collaborative reach.^17,18

The finding that high citation impact is not confined to high-output nations is clinically and strategically important. Countries such as the Netherlands and Switzerland achieve disproportionate citation influence, suggesting that research quality and targeted specialization may offer viable pathways to global impact even for smaller research communities.^19,20

4.2. Thematic shifts and research hotspots

The keyword temporal analysis reveals a profound epistemological transition in the field. Early-stage research was primarily characterized by foundational biomechanical investigations—studies of kinematics, joint angles, and gait cycle parameters. The progressive emergence of AI-related terms from approximately 2017 onward reflects a field-wide shift toward data-driven and computationally sophisticated research paradigms.^21,22 Machine learning—particularly deep learning architectures including convolutional neural networks and recurrent networks—has fundamentally expanded the capacity of motion capture systems to extract clinical meaning from high-dimensional movement data.²³

Research hotspots in EEG-based brain-computer interfaces and multimodal sensor fusion indicate a convergence of neuroscience, engineering, and clinical rehabilitation that is generating novel approaches to motor recovery assessment and intervention.^24,25 The prominence of real-time feedback systems in recent citation bursts further indicates the field’s migration from laboratory-based investigation toward deployable clinical and home-based applications.^26,27

4.3. Clinical and technical implications

The integration of markerless motion capture with AI algorithms—exemplified by citation-bursting tools such as OpenPose—is particularly significant for clinical settings where traditional marker-based optical capture is impractical.²⁸ These developments lower the barrier to high-resolution movement assessment in outpatient and community rehabilitation contexts. Similarly, AI-enhanced analysis of inertial measurement unit (IMU) data is enabling continuous ambulatory monitoring that yields ecologically valid movement metrics across daily activities.^29,30 The external responsiveness and metrological validity of wearable sensors in continuous monitoring scenarios represent an important practical consideration; recent work has examined the responsiveness of wearable devices to exercise-induced physiological changes,³¹ underscoring that deployment of these technologies in ecological settings requires rigorous device-level validation prior to clinical adoption.

Virtual reality integrated with motion capture represents another clinically meaningful trajectory. The gamification of rehabilitation exercises through VR environments has been shown to improve patient engagement and adherence, while motion capture provides the objective movement data necessary for adaptive difficulty adjustment and progress monitoring.³² These integrated systems also hold implications for posture-related conditions. For instance, the management of text neck syndrome has been reconceptualized from posture correction toward posture change strategies facilitated by real-time motion feedback,³³ illustrating how motion capture-enabled clinical systems can drive broader shifts in rehabilitation paradigms. The emergence of these integrated systems necessitates the development of validated, standardized assessment protocols to enable cross-study comparison and regulatory approval.

4.4. Collaboration gaps and future directions

The collaboration network analysis identifies important structural gaps, particularly the relative insularity of Chinese research networks and the underrepresentation of clinical rehabilitation institutions within primarily engineering-dominated author clusters. Bridging these gaps through deliberately interdisciplinary and internationally distributed research consortia represents a key priority for accelerating translation from computational innovation to clinical impact.^34,35 Funding agencies and journal editorial boards could play important roles in incentivizing such collaboration.

Future investigations should prioritize prospective validation of AI-integrated motion capture systems in real-world clinical populations, with attention to generalizability across diverse patient demographics and rehabilitation settings. The development of open-access standardized motion capture datasets—analogous to ImageNet for computer vision—would substantially facilitate benchmarking and model development across the field.³⁶ In this regard, the broader field of sensor-based AI in sports and medical science has called for standardized frameworks for assessing both sensor technologies and AI algorithms, emphasizing the need for methodological transparency, reproducibility, and cross-platform comparability.³⁷ These principles are equally applicable to motion capture-based rehabilitation research, where heterogeneous hardware, algorithms, and outcome metrics currently impede cross-study synthesis.

4.5. Limitations

Several limitations should be considered when interpreting these findings. This study represents a WoS-indexed bibliometric mapping rather than comprehensive coverage of the entire literature. The exclusive reliance on WoS databases may underrepresent publications from non-English-speaking regions and conference proceedings that are particularly prevalent in engineering disciplines; databases such as Scopus or IEEE Xplore may offer complementary coverage,^38,39 future studies should replicate this analysis using Scopus and IEEE Xplore to mitigate this limitation. Replication of this analysis with expanded database coverage is encouraged to validate the network structures and thematic clusters identified here. Second, the analysis was restricted to English-language publications, potentially missing significant contributions from Asian and European research communities. Third, bibliometric methods by their nature quantify academic output and citations rather than directly measuring technological innovation or clinical effectiveness, which may not be fully captured by traditional publication metrics.^40,41 Finally, this analysis was not prospectively registered prior to data collection. This decision was made because the study was initiated as an exploratory scoping exercise, prior to the formal codification of bibliometric pre-registration norms. We acknowledge this as a limitation affecting methodological transparency, and future bibliometric analyses in this domain are strongly encouraged to pre-register on platforms such as the Open Science Framework.

5. Conclusion

This systematic bibliometric analysis maps the trajectory of AI-integrated motion capture research in rehabilitation from 2004 to 2023, revealing a field characterized by exponential growth, geographic diversification, and a fundamental thematic transition toward computationally-driven and clinically-translatable research frameworks. The identified research hotspots—encompassing gait analysis, brain-computer interfaces, EEG signal processing, and real-time motion feedback—provide actionable guidance for future investigations. Continued interdisciplinary collaboration, development of standardized evaluation methodologies, and investment in prospective clinical validation will be essential to realize the transformative potential of AI-driven motion capture in rehabilitative care.

Supplemental material

Supplemental material - Two decades of AI-driven motion capture in rehabilitation: Mapping research networks, thematic hotspots, and future trajectories

Supplemental material for Two decades of AI-driven motion capture in rehabilitation: Mapping research networks, thematic hotspots, and future trajectories by Xiaojing Huang, Jing Xu, Lingyan Chen in Digital Health.

Footnotes

ORCID iD

Xiaojing Huang

Author contributions

Xiaojing Huang led study design and data analysis. Xiaojing Huang, Jing Xu and Lingyan Chen oversaw the project, provided critical revision, and approved the final manuscript. All authors reviewed the submission and take responsibility for the accuracy and integrity of the reported data.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Shanghai Municipal Health Commission (grant number: 202340161), the Shanghai Society of Rehabilitation Medicine (grant number: 2022KJCX006), and the Pudong New Area Health Commission for the Special Disease of Rotator Cuff Injury (grant number: PWZzb2022-24).

Declaration of conflicting interests

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

Data Availability Statement

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.*

Supplemental material

Supplemental material for this article is available online.

References

Porciuncula

Roto

, et al. Wearable movement sensors for rehabilitation: a focused review of technological and clinical advances. PM&R 2018; 10(9): S220–S232.

Lum

Burgar

Van der Loos

, et al. Robot-assisted movement training compared with conventional therapy techniques for the rehabilitation of upper-limb motor function after stroke. Arch Phys Med Rehabil 2002; 83(7): 952–959. https://doi.org/10.1053/apmr.2001.33101

Rahman

Ahmed

Uddin

, et al. AI-driven stroke rehabilitation systems and assessment: a systematic review. IEEE Trans Neural Syst Rehabil Eng 2022; 31: 192–207.

Tramontano

Morone

Palomba

, et al. Effectiveness of a sensor-based technology in upper limb motor recovery in post-acute stroke neurorehabilitation: a randomized controlled trial. J Biol Regul Homeost Agents 2020; 34(5 Suppl 3): 165–174.

Bernaldo de Quiros

Leocadio

Suarez-Garcia

, et al. Quantification of movement in stroke patients under free-living conditions using wearable sensors: a systematic review. Sensors 2022; 22(3): 1050. https://doi.org/10.3390/s22031050

Gaddour

Nticha

Mtawaa

, et al. Injury Prediction in Football: How Artificial Intelligence Is Shaping the Present and Transforming the Future in Africa. Sports Health 2026; 18(2): 247–249. https://doi.org/10.1177/19417381251410084

den Hartog

van der Krogt

van der Burg

, et al. Home-based measurements of dystonia in cerebral palsy using smartphone-coupled inertial sensor technology and machine learning: a proof-of-concept study. Sensors (Basel) 2022; 22(12): 4386. https://doi.org/10.3390/s22124386

Senadheera

Hettiarachchi

Haslam

, et al. AI applications in adult stroke recovery and rehabilitation: a scoping review using AI. Sensors (Basel) 2024; 24(20): 6585. https://doi.org/10.3390/s24206585

Langerak

Regterschot

GRH

Selles

, et al. Requirements for home-based upper extremity rehabilitation using wearable motion sensors for stroke patients: a user-centred approach. Disabil Rehabil Assist Technol 2024; 19(4): 1392–1404. https://doi.org/10.1080/17483107.2023.2183993

10.

Scott

Seyres

Philp

, et al. Healthcare applications of single-camera markerless motion capture: a scoping review. PeerJ 2022; 10: e13517. https://doi.org/10.7717/peerj.13517

11.

Chen

Song

. Visualizing a field of research: a methodology of systematic scientometric reviews. PLoS One 2019; 14(10): e0223994. https://doi.org/10.1371/journal.pone.0223994

12.

van Eck

Waltman

. Software survey: VOSviewer, a computer program for bibliometric mapping. Scientometrics 2010; 84: 523–538. https://doi.org/10.1007/s11192-009-0146-3

13.

Wang

. Visualization analysis of big data research based on CiteSpace. Soft Comput 2020; 24(11): 8173–8186. https://doi.org/10.1007/s00500-019-04384-7

14.

Fong

DTP

Chan

. The use of wearable inertial motion sensors in human lower limb biomechanics studies: a systematic review. Sensors 2010; 10(12): 11556–11565. https://doi.org/10.3390/s101211556

15.

Scalise

Bora

Zancanella

, et al. Virtual reality as a therapeutic tool in spinal cord injury rehabilitation: a comprehensive evaluation and systematic review. J Clin Med 2024; 13(18): 5429. https://doi.org/10.3390/jcm13185429

16.

Osgouei

Soulsby

Bello

. Rehabilitation exergames: use of motion sensing and machine learning to quantify exercise performance in healthy volunteers. JMIR Rehabil Assist Technol 2020; 7(2): e17289. https://doi.org/10.2196/17289

17.

Hubaut

Guichard

Greenfield

, et al. Validation of an embedded motion-capture and EMG setup for the analysis of musculoskeletal disorder risks during manhole cover handling. Sensors 2022; 22(2): 436. https://doi.org/10.3390/s22020436

18.

Baldinger

Lippmann

Senner

. Artificial intelligence-based motion capture: current technologies, applications and challenges. Artif Intell Sports Mov Health 2024: 161–176.

19.

Wang

Cao

Zhou

, et al. A comprehensive review on motion trajectory reconstruction for EEG-based brain-computer interface. Front Neurosci 2023; 17: 1086472. https://doi.org/10.3389/fnins.2023.1086472

20.

Orban

Elsamanty

Guo

, et al. A review of brain activity and EEG-based brain-computer interfaces for rehabilitation application. Bioengineering 2022; 9(12): 768. https://doi.org/10.3390/bioengineering9120768

21.

Piszcz

Rojek

Mikołajewski

. Impact of virtual reality on brain-computer interface performance in IoT control—review of current state of knowledge. Appl Sci 2024; 14(22): 10541. https://doi.org/10.3390/app142210541

22.

Thopalli

Mahadevan

Sundaresan

, et al. Advances in computer vision for home-based stroke rehabilitation. Computer Vision. Chapman and Hall/CRC, 2024, pp. 109–127.

23.

Noorbakhsh-Sabet

Zand

Zhang

, et al. Artificial intelligence transforms the future of health care. Am J Med 2019; 132(7): 795–801. https://doi.org/10.1016/j.amjmed.2019.01.017

24.

Menolotto

Komaris

Tedesco

, et al. Motion capture technology in industrial applications: a systematic review. Sensors (Basel) 2020; 20(19): 5687. https://doi.org/10.3390/s20195687

25.

Pirasteh

Shamseini Ghiyasvand

Pouladian

. EEG-based brain-computer interface methods with the aim of rehabilitating advanced-stage ALS patients. Disabil Rehabil Assist Technol 2024; 19: 1–11. https://doi.org/10.1080/17483107.2024.2316312

26.

Aziz

Hussain

Majeed

, et al. Wearable IoT devices in rehabilitation: enabling personalized precision medicine. In: Medical Robotics and AI-Assisted Diagnostics for a High-Tech Healthcare Industry. IGI Global, 2024, pp. 281–308.

27.

Lorenz

Skjaeret-Maroni

. A review of combined functional neuroimaging and motion capture for motor rehabilitation. J Neuroeng Rehabil 2024; 21(1): 3. https://doi.org/10.1186/s12984-023-01294-6

28.

Cao

Hidalgo

Simon

, et al. OpenPose: realtime multi-person 2D pose estimation using part affinity fields. IEEE Trans Pattern Anal Mach Intell 2021; 43: 172–186. https://doi.org/10.1109/TPAMI.2019.2929257

29.

Prasanth

Caban

Keller

, et al. Wearable sensor-based real-time gait detection: a systematic review. Sensors (Basel) 2021; 21: 2727. https://doi.org/10.3390/s21082727

30.

Karikari

Koshechkin

. Review on brain-computer interface technologies in healthcare. Biophys Rev 2023; 15(5): 1351–1358. https://doi.org/10.1007/s12551-023-01138-6

31.

Dhahbi

Saad

. Revolutionizing “text neck syndrome” management: paradigm shifting from “posture correction” to “posture change”. Acta Kinesiologica 2024; 18(3): 37–40.

32.

Willingham

Seifert

Carter

, et al. Leveraging emerging technologies to expand accessibility and improve precision in rehabilitation and exercise for people with disabilities. Int J Environ Res Public Health 2024; 21(1): 79. https://doi.org/10.3390/ijerph21010079

33.

Ardigò

Palermi

Padulo

, et al. External Responsiveness of the SuperOp^TM Device to Assess Recovery After Exercise: A Pilot Study. Front Sports Act Living 2020; 2: 67. https://doi.org/10.3389/fspor.2020.00067

34.

Diaz-Asper

Rogers

Smith

, et al. A framework for language technologies in behavioral research and clinical applications: ethical challenges, implications, and solutions. Am Psychol 2024; 79(1): 79–91. https://doi.org/10.1037/amp0001195

35.

Mennella

Fadhel

Ahmed

, et al. Ethical and regulatory challenges of AI technologies in healthcare: a narrative review. Heliyon 2024.

36.

Wang

Smith

White

, et al. The time is now: a FASTER approach to generate research evidence for technology-based interventions in the field of disability and rehabilitation. Arch Phys Med Rehabil 2021; 102(9): 1848–1859. https://doi.org/10.1016/j.apmr.2021.04.009

37.

Dhahbi

Chamari

. The Algorithmic Athlete: A Call to Standardize Assessment of Sensor Technologies and Artificial Intelligence. Int J Sports Physiol Perform 2026; 21(4): 505–506. https://doi.org/10.1123/ijspp.2025-0547

38.

Beatty

Woo

, et al. A new era in cardiac rehabilitation delivery: research gaps, questions, strategies, and priorities. Circulation 2023; 147(3): 254–266. https://doi.org/10.1161/CIRCULATIONAHA.122.061046

39.

Abiri

Borhani

Sellers

, et al. A comprehensive review of EEG-based brain-computer interface paradigms. J Neural Eng 2019; 16(1): 011001. https://doi.org/10.1088/1741-2552/aaf12e

40.

Dasgupta

Bora

Zancanella

, et al. Machine learning for optical motion capture-driven musculoskeletal modelling from inertial motion capture data. Bioengineering 2023; 10(5): 510. https://doi.org/10.3390/bioengineering10050510

41.

Fang

Bora

Zancanella

, et al. EMG-centered multisensory-based technologies for pattern recognition in rehabilitation: state of the art and challenges. Biosensors 2020; 10(8): 85. https://doi.org/10.3390/bios10080085

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.02 MB