The role of fNIRS in assessing motor function: A bibliometric review on extremity applications

Abstract

Understanding the mechanisms that control motor movements in the brain is crucial to understand neurological disorders and brain plasticity. Functional near-infrared spectroscopy (fNIRS) is an increasingly popular noninvasive method for investigating these processes. This study examined the application of fNIRS in the field of motor function and limbs and showed that after a limited development period between 2006 and 2010, it has grown rapidly since 2010. The number of publications and citations peaked in 2021 to 2022, with a significant increase in citations in 2023 to 2024. Co-citation analysis revealed 26,950 citations in 982 publications, indicating that these studies made a significant contribution to the scientific literature. Over the last 18 years, 2962 authors from 55 countries have published 982 articles on fNIRS, with an average annual growth of 31.75% in publications and 42.55% in citations since 2010. Furthermore, the analysis of combined fNIRS-electroencephalography studies showed that methodological advances are the main focus of research in the field. Further studies that combine the power of combined neuroimaging methods to monitor both electrical activity (electroencephalography) and hemodynamic changes (fNIRS) are expected. In addition to motor and limb research and brain function studies, studies on these approaches in treating diseases such as Parkinson disease and chronic stroke are also drawing attention, showing the growing importance of fNIRS in motor and limb research, and that it will be an important data tool for future studies.

1. Introduction

In neuroscience, understanding how motor movements are controlled by the brain is crucial for understanding neurological diseases and brain plasticity. Functional near-infrared spectroscopy (fNIRS), a noninvasive method for understanding these processes, has become increasingly popular. This study examined the potential clinical and research applications of fNIRS in the study of motor movements. fNIRS is an easy-to-use and portable brain imaging technology that allows the study of normal brain function and changes that occur in disease, both in the laboratory and real-world settings.^[1–3] Extremity (hand, arm, and leg) studies using fNIRS provide important information for understanding motor skills and developing rehabilitation processes by examining brain activity during motor movements or motor imagery. fNIRS-based brain-computer interface (BCI) systems allow users to communicate their intentions directly to a computer or device by measuring hemodynamic responses in the brain.^[4] fNIRS-based BCI systems can be used as a tool for communication and restoration of motor function through neuroprostheses for people with motor disorders, such as amyotrophic lateral sclerosis and spinal cord injury with limited mobility.^[5] In addition, when fNIRS is used in conjunction with other imaging techniques (e.g., functional magnetic resonance imaging or electroencephalography), it is possible to gain a more complete understanding of the motor areas.^[6,7]

Bibliometric analysis offers the possibility of mapping detailed research, identifying knowledge structures, and providing new insights using scientific and aesthetic applications to recommend academic literature. This method offers a wider range of categories for evaluating the contributions of researchers in the field.^[8,9] Bibliometric studies have been conducted on fNIRS, as in many other disciplines.

Ye et al reported the use of fNIRS for prefrontal korteks-related cognitive functions and various neurological/psychiatric disorders between 2011 and 2022.^[10] Alsharif et al conducted a bibliometric analysis of the use of fNIRS and functional magnetic resonance imaging (fMRI) in neuromarketing.^[11] Li et al conducted a bibliometric analysis of the top 100 most-cited articles on fNIRS research and Yan et al conducted a bibliometric analysis of fNIRS studies from 2000 to 2019.^[12,13] Although previous bibliometric studies have examined fNIRS from more diverse perspectives (e.g., disease-related cognitive functions^[10] or citation-based impact analyses,^[12] these studies have not specifically focused on the applications of fNIRS related to motor function and extremities). The current study addresses this gap by adopting a thematic and scientific measurement perspective, using document co-citation analysis (DCA), keyword analysis, and clustering techniques to explore the conceptual structures and emerging trends in this particular subfield.

Our work provides a more focused and application-specific analysis by offering insights into using fNIRS to comprehend motor function and limb movements, thus rendering it more specialized and pertinent for neuroscience and rehabilitation researchers.

The findings of this study will help to identify general research trends, secondary research topics, and prospective research agendas for future studies. Drawing on the above explanations and the subsequent application of concepts in the literature, we provide a bibliometric overview of fNIRS in extremity studies through descriptive and network analyses.

2. Methodology

Bibliometrics is the quantitative analysis of publications produced by individuals or institutions within a specified timeframe and within a particular subject area, along with the relationships between these publications.^[14] The method provides insight by analyzing the characteristics of research publications, thereby understanding the questions researchers are attempting to answer and the methods they have developed for this purpose.^[15]

Using bibliometric methods, this study analyzed and visualized the relationship between fNIRS and extremity and motor activities in several ways. The primary characteristics of the growth trend in research categories were examined in terms of the interaction between principal topics, co-document citations, and keywords as well as the identification of the most influential and active authors, journals, and countries/regions. The CiteSpace software, a Java application developed by Chaomei Chen from Drexel University, was selected to apply bibliometric methods and perform science mapping. The version of this software that uses advanced visualization and text mining methods for bibliometric data analysis is CiteSpace Advanced 6.3. R3 (update June 27, 2024, available at https://citespace.podia.com/).

A science map comprises a set of nodes that represent analytical elements such as references, keywords, authors, journals, and countries, as well as links that establish connections between these nodes. Cluster analysis was performed in CiteSpace using the log-likelihood ratio to create thematic clusters. Cluster analysis is a widely used statistical method that involves grouping observations considered similar to the same clusters.^[16,17] The modularity Q index (Q) and mean silhouette score (S) were used to evaluate the structural robustness of the network generated by the cluster analysis. Q is used to determine the extent to which the network can be divided into smaller components. S was used to evaluate the quality and homogeneity of the clusters formed within the network.^[16] In addition, a temporal metric, citation burst, was used as a significant indicator of the research impact. Thus, authors, keywords, countries, and document analytical elements that showed significant changes in the literature over time were identified.^[18] Another important metric is the betweenness centrality value, which evaluates the capacity of nodes to establish effective connections with other nodes within the network. In visual networks, nodes with purple circles indicate high intermediary centrality, which is often identified as a hotspot or turning point in a domain.^[19] For descriptive statistics, Excel and SAS (On Demand for Academics-SAS Studio: Last access November 4, 2024) were used.

2.1. Data source and data extraction

The data used in this study were procured from the Web of Science Core Collection on August 13, 2024. No temporal constraints were imposed to observe the comprehensive evolution and transformation of the publications in question. The Web of Science database was selected for its comprehensive coverage of high-quality, peer-reviewed publications across a range of academic disciplines. It is one of the largest and most comprehensive repositories in academic literature, spanning the period from 1900 to the present day.^[20] Keywords and search criteria were determined to identify publications related to fNIRS research on motor control and extremities. The search field was designated as the term “Topic,” which encompasses the title, abstract, author keywords, and keywords plus searches. The search codes were determined to be as follows: (Topic: fNIRS or “functional near-infrared spectrum” or “near-infrared spectrum”) AND (Topic: extremity or motor). The search was limited to articles, reviews, and proceedings written in English. Following the conclusion of the selection process, pertinent data were retained in the “Custom Selection (29)” format, utilizing the “Tab-delimited” configuration for subsequent analysis. The recorded data were converted to Excel format and subjected to a verification process to identify and address any omissions or duplications. Furthermore, to identify and analyze electroencephalography (EEG)-related studies among the selected studies, a new search was performed by adding the AND operator to the existing selection criteria (Topic: Electroencephalography or EEG). A total of 1294 articles were retrieved through the initial database search. These formed the basis of Dataset 1 for further screening. Two authors independently assessed titles and abstracts and systematically screened these results to determine relevance to fNIRS studies on motor function and limb activity. Articles that did not meet the scope (out-of-scope studies) were excluded according to prespecified exclusion criteria (e.g., fNIRS studies that did not include any limb-related motor activity, or those focused solely on unrelated brain domains such as language, emotion, or vision). During this process, if the 2 primary reviewers disagreed or were uncertain about the relevance of a particular study, a third senior author was consulted to resolve the discrepancy.

After this screening step, 982 articles were retained for full-text examination and data extraction. Subsequently, the data were transformed into a format compatible with CiteSpace, stored, and analyzed. Descriptive statistics were performed using Excel, Statistical Package for the Social Sciences, and SAS on an Excel file. A flowchart of the data extraction and processing is shown in Fig. 1.

Figure 1.

Flowchart of the data screening process.

3. Result and discussion

3.1. Publications distributions and growth trends

In approximately 18 years, 2962 authors from 55 different countries published 982 documents on fNIRS studies related to motor function and extremity research. A review of fNIRS studies related to motor function and extremity research in the Web of Science database reveals that the study conducted by Plichta et al, “Event-related visual versus blocked motor task: Detection of specific cortical activation patterns with functional near-infrared spectroscopy,” represents the inaugural investigation in this domain. Figure 2 shows the distribution of fNIRS studies related to motor function and extremity research over time.

Figure 2.

Distribution of publications and citations over the years.

fNIRS studies related to motor function and extremity research have experienced a relatively limited scope and stable trajectory between 2006 and 2010. However, since 2010, there has been a notable increase in the number of publications and citations in the field. In particular, the number of annual publications has increased rapidly since 2015, with exponential growth observed after 2017. The number of publications and citations has increased considerably, particularly since 2019, reaching its maximum in 2021 and 2022. Although many publications persisted in 2023 and 2024, a notable citation surge was also observed. This trend indicates that fNIRS technology has begun to be employed more extensively in motor function and extremity studies and that research in this area has gained momentum. In general, this field, which exhibited relatively static growth in the early periods, has experienced notable growth since the 2010s, driven by technological developments and growing interest from the academic community. Consequently, it has become an important research area in the present day.

When the distribution of the number of publications by year was examined, it was understood that the data were compatible with an exponential or polynomial model. Two different models were used to estimate the trend in the number of publications in the coming years. The R² value, which is a statistical measure that evaluates the fit of the models to the data, was calculated as 0.97 for the exponential model and 0.98 for the polynomial model. These results suggest that the polynomial model provides a better fit to the data (higher R² value). However, the exponential model offers a more balanced approach and is generally preferred for long-term growth projections. In both models, the predicted values indicated an increase in the number of publications and concomitant growth in the field (see Fig. 3).

Figure 3.

Exponential and polynomial growth trend and predictions of publication numbers by year.

There was a significant and strong positive monotonic relationship between the number of publications and citations (Spearman correlation coefficient ρ = 0.98; P = .000). As the number of publications increased, the number of citations also increased consistently. The expected average annual growth rate of fNIRS publications related to motor function and extremity research until 2024 is approximately 31.75%. The expected average annual growth rate of citations for publications of fNIRS studies related to motor function and extremity research until 2024 was calculated to be approximately 42.55%. A comparison of the annual average growth rates of publications and citations reveals that the increase in citation numbers outpaces that in publication numbers. This suggests cross-disciplinary interest in this field.

3.2. Document co-citation analysis

This section analyzes and discusses document co-citations to uncover key documents related to fNIRS studies in the field of motor function and extremity research. Document co-citation is defined as the frequency with which 2 articles from earlier literature are cited together in later literature and help to decipher the intellectual structure of the literature.^[21] The 982 publications related to fNIRS studies in motor function and extremity research had 26,950 references, with an average of 27.44 per publication.

To conduct DCA analysis, a scientific mapping method was employed to create a network visualization with CiteSpace, utilizing 26,950 references from 982 publications. The selection criterion was determined as a g-index to ensure the integrity of the network and the correct identification of important documents, contributing to more meaningful and effective scientific mapping. This approach was used to identify highly cited documents and show a wider citation distribution. In the analysis, the g-index selection criterion scale factor was set as k = 100. The selection criteria and scale factor values were chosen after several trials, during which we aimed to optimize the structural metrics of the network.

Table 1 lists the 20 most-cited documents from the study period. These resources have not only established a foundation for further research on the mechanism of fNIRS studies in motor function and extremity research but have also provided a theoretical basis for such studies. In their study, Noman Naseer and Keum-Shik Hong^[22] examined fNIRS-BCI, explaining the use of this technology in motor imagery and cognitive tasks. They presented examples of signal processing, feature extraction, and classification methods. Emphasizing the potential of fNIRS-BCI systems in neurorehabilitation and communication, the study provides both theoretical and practical contributions to the field.

Table 1.

The most cited references in fNIRS studies on motor function and extremity with centrality values.

Count	Centrality	Year	Cited references
155	0.00	2015	fNIRS-based brain-computer interfaces: a review[22]
107	0.00	2014	A review on continuous wave functional near-infrared spectroscopy and imaging instrumentation and methodology[23]
94	0.00	2012	A brief review on the history of human functional near-infrared spectroscopy (fNIRS) development and fields of application[24]
76	0.01	2013	Classification of functional near-infrared spectroscopy signals corresponding to the right- and left-wrist motor imagery for development of a brain–computer interface[25]
76	0.01	2015	Classification of prefrontal and motor cortex signals for three-class fNIRS–BCI[26]
70	0.00	2020	The present and future use of functional near-infrared spectroscopy (fNIRS) for cognitive neuroscience[27]
67	0.03	2012	Enhanced performance by a hybrid NIRS–EEG brain computer interface[28]
64	0.00	2011	Assessment of the cerebral cortex during motor task behaviors in adults: A systematic review of functional near infrared spectroscopy (fNIRS) studies[29]
60	0.01	2017	Functional near-infrared spectroscopy in movement science: a systematic review on cortical activity in postural and walking tasks[30]
55	0.01	2018	The NIRS Brain AnalyzIR Toolbox[31]
54	0.00	2009	NIRS-SPM: statistical parametric mapping for near-infrared spectroscopy[32]
51	0.10	2014	Online binary decision decoding using functional near-infrared spectroscopy for the development of brain–computer interface[33]
49	0.02	2014	Decoding of 4 movement directions using hybrid NIRS-EEG brain-computer interface[34]
48	0.00	2018	Applications of functional near-infrared spectroscopy (fNIRS) neuroimaging in exercise–cognition science: A Systematic, methodology-focused review[35]
46	0.02	2017	Hybrid EEG–fNIRS-based eight-command decoding for BCI: Application to quadcopter control[36]
46	0.00	2019	Current status and issues regarding preprocessing of fNIRS neuroimaging data: An investigation of diverse signal filtering methods within a general linear model framework[37]
46	0.00	2016	Hybrid EEG-fNIRS asynchronous brain-computer interface for multiple motor tasks[38]
45	0.03	2014	Motion artifacts in functional near-infrared spectroscopy: A comparison of motion correction techniques applied to real cognitive data[39]
43	0.02	2014	Twenty years of functional near-infrared spectroscopy: introduction for the special issue[1]
42	0.02	2011	fNIRS study of walking and walking while talking in young and old individuals[40]

EEG = electroencephalography.

The network obtained for DCA was composed of 2461 nodes and 11,944 links (see Fig. 4).

Figure 4.

Document co-citation analysis (DCA) network visualization.

In the DCA network visualization shown in Fig. 4, the node size indicates the number of citations they have received. Simultaneously, the thickness of the links is representative of the number of collaborations they have engaged in. Furthermore, the colors of the nodes and links transition from cold to warmer tones, symbolizing a temporal shift. The use of cold colors (such as blue and green) denotes citations that have occurred in the past. In contrast, warm colors (yellow, orange, and dark red) signify more current citations. The increased density of nodes and connections in the network, indicated by warmer colors, suggests a notable surge in academic interest in this field in recent years, accompanied by a noteworthy acceleration in the pace of research conducted. The density of the red nodes indicates that these nodes have received a high number of citations, signifying frequent citation bursts. This demonstrates that specific documents have attained a significant position within the existing literature and have exerted considerable influence within their respective fields. Strong connections, particularly around the red dots, highlight the central role of these articles in the network and their influence on other studies. In general, this network structure clearly shows how scientific developments are gaining momentum and which studies are driving this development, revealing both the temporal changes and the influence of important studies in the field.^[33] “Online binary decision decoding using functional near-infrared spectroscopy for the development of brain–computer interfaces.” has the highest centrality value (0.10). This shows that the article plays a key role in the network and has a significant impact on literature. It can be said that it is accepted as a reference point, especially in research on brain-computer interfaces and fNIRS usage and is a basic source for new studies.

Cluster analysis divides publications within a cluster into separate research subfields where they often share similar ideas.^[41] It is important to note that the mere similarity between the findings of these publications does not imply that they are necessarily consistent and compatible with each other. Publications are grouped into the same cluster based on subject similarity, but this does not guarantee that the viewpoints expressed in each publication are entirely aligned. Cluster analysis of the network in Fig. 5 was performed with a modularity Q-index of 0.777 and a weighted average silhouette of 0.8957. It generated a network that could be divided into separate modules, each homogeneous (see Fig. 5A).

Figure 5.

(A) Map of DCA clusters. Clusters represent the topics of the cited reference. (B) Cluster dependency map for 2024. DCA = document co-citation analysis.

The cluster numbers were arranged according to cluster size, starting with the largest cluster, 0. The largest cluster, Cluster #0 (Parkinson disease), consisted of 263 nodes with a S of 0.812. On average, the references comprising this cluster were published in 2015. The most frequently referenced document in this cluster is Herold et al, a study titled “Functional near-infrared spectroscopy in movement science: a systematic review on cortical activity in postural and walking tasks.”^[30] This study systematically summarizes how fNIRS has been used to study brain activity during motor tasks, providing a comprehensive map of research in this area that will help researchers understand the current findings and provide a solid foundation for future studies. The other largest cluster, Cluster #1 (brain area), comprised 262 nodes with a S of 0.801. On average, the references comprising this cluster were published in 2016. The major citing article of the cluster is the feature extraction and classification methods for hybrid fNIRS-EEG brain-computer interfaces.^[42] This study has made significant contributions to data processing by combining fNIRS and EEG technologies by presenting innovative methods for improving the accuracy of hybrid BCI systems and analyzing neural signals more effectively. The first 5 clusters of DCA analyses of fNIRS studies related to motor function and extremity research reflect innovative research in neuroscience and motor control. The strong relationship between the largest cluster on the 2024 dependency map, #0 (Parkinson disease), and the motor rehabilitation cluster is understandable (see Fig. 5B). In Parkinson disease, which is characterized by motor dysfunction, the use of fNIRS in motor rehabilitation methods to alleviate these effects and improve the quality of life of patients is increasing. This indicates that fNIRS will play a crucial role in developing personalized treatment approaches in the future. Furthermore, the strong association between cluster #10 (painful pressure) and the cluster of brain regions highlights the increasing use of fNIRS studies to understand the brain’s reflection of pain perception and the importance of fNIRS in the development of new treatment strategies for pain management.

A citation burst can be interpreted as intense interest from the research community in the underlying work. If a cluster contains a large number of publications with strong citation bursts, the cluster as a whole is considered to cover an active research area, an emerging trend.^[15]

Studies in the fNIRS literature on autism spectrum disorder, signal processing, Parkinson disease, and motor rehabilitation are the current reference points (current references^[25,49–52] see Table 2). Another network metric, Sigma, was obtained by simultaneously considering the betweenness centrality and citation bursts. The Sigma value indicates the novelty and influence of a node in the network of interest.^[17,56] The top-ranked document by sigma is Naseer et al, “Online binary decision decoding using functional near-infrared spectroscopy for the development of brain–computer interface” in Cluster #1, with a sigma of 1.88. The second is the Coyle et al, “Brain-computer interface using a simplified functional near-infrared spectroscopy system” in Cluster #1, with a sigma of 1.64. Cluster #1 (brain area) was a cluster with innovative work.

Table 2.

Top 20 references with the strongest citation bursts.

References	Year	Strength	Begin	End	Sigma	Cluster
Temporal classification of multichannel near-infrared spectroscopy signals of motor imagery for developing a brain–computer interface[43]	2007	19.58	2007	2015	1.47	#6 lyapunov spectrum
A temporal comparison of BOLD, ASL, and NIRS hemodynamic responses to motor stimuli in adult humans[44]	2006	12.99	2007	2014	1.09	#18 neurofeedback training
Brain–computer interface using a simplified functional near-infrared spectroscopy system[45]	2007	16.50	2008	2015	1.64	#1 brain area
NIRS-SPM: Statistical parametric mapping for near-infrared spectroscopy[32]	2009	19.89	2010	2017	1.01	#8 craniocerebral correlation
HomER: a review of time-series analysis methods for near-infrared spectroscopy of the brain[46]	2009	13.97	2010	2017	1.22	#8 craniocerebral correlation
Wireless miniaturized in vivo near infrared imaging[47]	2008	9.84	2010	2015	1.15	#6 lyapunov spectrum
Assessment of the cerebral cortex during motor task behaviors in adults: A systematic review of functional near infrared spectroscopy (fNIRS) studies[29]	2011	15.54	2012	2019	1.00	#12 functional nirs study
fNIRS-based online deception decoding[48]	2012	10.57	2012	2018	1.10	#1 brain area
fNIRS study of walking and walking while talking in young and old individuals[40]	2011	11.14	2013	2019	1.23	#0 parkinsons disease
Kalman estimator- and general linear model-based on-line brain activation mapping by near-infrared spectroscopy[49]	2010	10.38	2013	2018	1.05	#1 brain area
A brief review on the history of human functional near-infrared spectroscopy (fNIRS) development and fields of application[24]	2012	14.24	2014	2020	1.01	#21 functional connectivity analysis
Noise reduction in functional near-infrared spectroscopy signals by independent component analysis[50]	2013	10.44	2014	2018	1.07	#1 brain area
Enhanced performance by a hybrid NIRS–EEG brain computer interface[28]	2012	9.72	2014	2020	1.34	#9 fake feedback
State-space models of impulse hemodynamic responses over motor, somatosensory, and visual cortices[51]	2014	10.47	2015	2018	1.04	#1 brain area
fNIRS-based brain-computer interfaces: a review [22]	2015	16.82	2017	2020	1.06	#1 brain area
The present and future use of functional near-infrared spectroscopy (fNIRS) for cognitive neuroscience[27]	2020	18.48	2022	2024	1.08	#4 signal processing
Best practices for fNIRS publications[52]	2021	12.68	2022	2024	1.11	#5 autism spectrum disorder
Temporal derivative distribution repair (TDDR): A motion correction method for fNIRS[53]	2019	11.10	2022	2024	1.00	#4 signal processing
A consensus guide to using functional near-infrared spectroscopy in posture and gait research[54]	2020	10.02	2022	2024	1.01	#0 parkinsons disease
Time course of sensorimotor cortex reorganization during upper extremity task accompanying motor recovery early after stroke: An fNIRS study[55]	2019	9.71	2022	2024	1.00	#3 motor rehabilitation

EEG = electroencephalography.

3.3. Research hotspots keyword co-occurrence analysis

The logic behind the keyword analysis is that the content of an article is adequately represented by the author’s keywords. The authors’ keywords represent the themes of the research articles. This section uses a combination of techniques, such as occurrence analysis, keyword burst detection analysis, cluster analysis, and timeline visualization, to identify the research focuses, boundaries, and trends. Keywords were obtained from 982 publications and constituted a significant part of this research. The use of high-frequency keywords to highlight research hotspots in a discipline can effectively identify research hotspots and other important topics. A total of 2082 keywords were identified in 982 documents. When the keywords are examined, it is understood that there is no standard in this field and that the same concepts are used in different spellings. For example, words with the same meaning, such as functional near-infrared spectroscopy and functional near-infrared spectroscopy, were classified separately. This complicates the literature analysis. Therefore, the merge command was used in CiteSpace to manually merge identical keywords, thus partially standardizing and conducting a more consistent analysis. Table 3 presents the top 25 keywords and their respective centrality values.

Table 3.

The 25 most frequently occurring keywords.

Counted	Centrality	Year	Keywords
344	0.01	2009	Functional near-infrared spectroscopy
290	0.02	2007	Near infrared spectroscopy
183	0.02	2007	Activation
163	0.04	2008	Motor imagery
149	0.03	2007	Performance
138	0.03	2007	Prefrontal cortex
124	0.07	2006	Cortex
120	0.06	2008	Brain computer interface
115	0.05	2007	FMRI
105	0.01	2012	Classification
103	0.05	2009	FNIRS
99	0.05	2006	Brain
95	0.10	2009	Motor cortex
88	0.05	2007	Motor
77	0.05	2008	Signals
75	0.07	2007	Brain activation
74	0.03	2012	Gait
68	0.02	2014	Recovery
68	0.05	2012	Stroke
62	0.05	2013	Connectivity
61	0.03	2015	Walking
61	0.06	2011	System
59	0.06	2012	Cortical activation
59	0.03	2006	NIRS
57	0.04	2010	Functional connectivity

The most frequent keywords formed the basic concepts of research on motor functions and extremity studies using fNIRS (Table 3). It was the keyword with the highest motor cortex centrality value (0.10). The highest centrality value (0.10) for the motor cortex keyword indicates that this region is at the center of motor control studies, especially fNIRS and motor rehabilitation, and is one of the most influential concepts in the literature.

The strongest citation burst detection technology and algorithm provided in the CiteSpace software yielded 27 strong citation bursts in fNIRS studies on motor functions and extremity studies over the last 18 years. The 27 citation bursts are listed in Table 4.

Table 4.

Top 20 keywords with the strongest citation bursts.

Keywords	Year	Strength	Begin	End	2006–2024
Optical topography	2006	3.11	2006	2012	▃▃▃▃▃▃▃▂▂▂▂▂▂▂▂▂▂▂▂
MRI	2007	3.08	2007	2015	▂▃▃▃▃▃▃▃▃▃▂▂▂▂▂▂▂▂▂
Human brain	2010	4.41	2010	2013	▂▂▂▂▃▃▃▃▂▂▂▂▂▂▂▂▂▂▂
System	2011	3.94	2011	2015	▂▂▂▂▂▃▃▃▃▃▂▂▂▂▂▂▂▂▂
Activation	2007	3.26	2011	2013	▂▂▂▂▂▃▃▃▂▂▂▂▂▂▂▂▂▂▂
Communication	2011	3.21	2011	2016	▂▂▂▂▂▃▃▃▃▃▃▂▂▂▂▂▂▂▂
FMRI	2007	6.03	2012	2014	▂▂▂▂▂▂▃▃▃▂▂▂▂▂▂▂▂▂▂
Signals	2008	7.18	2014	2018	▂▂▂▂▂▂▂▂▃▃▃▃▃▂▂▂▂▂▂
Classification	2012	3.53	2014	2018	▂▂▂▂▂▂▂▂▃▃▃▃▃▂▂▂▂▂▂
Hemodynamic responses	2015	6.64	2015	2019	▂▂▂▂▂▂▂▂▂▃▃▃▃▃▂▂▂▂▂
NIRS	2006	4.35	2015	2017	▂▂▂▂▂▂▂▂▂▃▃▃▂▂▂▂▂▂▂
Auditory cortex	2016	3.66	2016	2017	▂▂▂▂▂▂▂▂▂▂▃▃▂▂▂▂▂▂▂
FNIRS	2009	3.66	2016	2017	▂▂▂▂▂▂▂▂▂▂▃▃▂▂▂▂▂▂▂
Hemodynamic response	2006	3.21	2016	2018	▂▂▂▂▂▂▂▂▂▂▃▃▃▂▂▂▂▂▂
Support vector machine	2017	3.58	2017	2020	▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▂▂▂▂
Children	2010	3.3	2019	2020	▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂▂▂
Cognitive function	2017	3.55	2021	2022	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂
Responses	2009	3.54	2021	2022	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂
Speed	2021	3.26	2021	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃
Hand	2015	3.15	2021	2022	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂
Impairment	2021	3.04	2021	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃
Recovery	2014	8.86	2022	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃
Connectivity	2013	6.7	2022	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃
Rehabilitation	2016	5.21	2022	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃
Reorganization	2007	3.77	2022	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃
Motor function	2022	3.54	2022	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃
Upper limb	2022	3.41	2022	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃

The red grid indicates the duration of citation burst. The dark blue line indicates the period in which the keyword has been active since its first appearance. The keywords that have recently occurred in citation bursts are impairment, recovery, connectivity, rehabilitation, reorganization, motor function, and upper limb, which are closely related to motor dysfunction and rehabilitation studies. These concepts play a critical role in the recovery of motor functions and brain restructuring processes. Upper limb studies focus on recovery and rehabilitation processes following impairment, while connectivity and reorganization emphasize the healing and reorganization of neural connections in the brain.

In this field, we organized the studies in a hierarchical manner using a clustering network. Figure 6 presents a keyword clustering visualization created by visually matching the keyword relationship and clustering labels determined using the log-likelihood ratio.

Figure 6.

(A) The keyword clustering visualization. (B) Timeline view of keyword clusters map.

Different cooccurrence keywords are represented by nodes and different major branches or topics are collectively represented by clusters. The network consists of 3107 nodes and 16,458 connections (Fig. 6A). Co-occurrence keywords were divided into 11 clusters. Clusters with fewer than 10 cluster elements were not included in the image. Cluster analysis was performed with a modularity Q index of 0.4275 and weighted average silhouette of 0.7138. A timeline view was created to observe the evolution of keywords over time (Fig. 6B). The timeline view offers an overview of cluster evolution over time, allowing the user to ascertain whether any development has continued over the years.^[57]

Clusters 0, 1, 3, 5, 6, and 7 were active study areas, and these clusters were hot topics. Dual-task walking, BCI, chronic stroke, unilateral cerebral palsy, functional connectivity analysis, and reverse oxygenation response are groundbreaking research topics in neurotechnology and neurorehabilitation. These topics, which are relevant to conditions such as aging, Parkinson disease, motor deficits, and balance problems, provide valuable insights into brain functionality and plasticity as well as the development of new treatment and rehabilitation methods.

3.4. Most influential authors and co-authorship networks

Authors play an important role in reflecting on research capacity and evaluating the development of an academic field. Therefore, it is important to identify not only authors with good publication skills but also core authors who contribute more to promoting the field’s development.^[58] Co-authorship analysis investigates and quantifies R&D collaboration among authors, institutions, or countries.^[59] A total of 2962 authors contributed to the creation of literature in this field, with an average of 3 authors per document. The most active writer in this field is Hong Keum Shik, who has had a significant impact on 42 publications. The second most influential author in this field was Naseer Noman. Nasser Noman coauthored publication with Hong Keum Shik, “fNIRS-based brain-computer interfaces: a review,” provided a comprehensive review of the development and applications of fNIRS-based BCI.^[22] Another important study by^[25] is titled “Classification of functional near-infrared spectroscopy signals corresponding to the right-and left-wrist motor imagery for development of a brain–computer interface.” This study focused on the classification of fNIRS signals corresponding to right and left wrist motor imagery movements. This study investigates the potential of this classification process for developing BCI.

A total of 2962 authors were from 55 different countries. The USA, which has collaborated on 248 publications and is in a key position to connect other countries, has the highest centrality value of 0.41. The People’s Republic of China has a centrality value of 0.30 with collaboration among 264 publications. South Korea has a centrality value of 0.16 with 109 publications. In proportion to the number of publications, this is lower than in Germany (0.22 centrality value with 82 publications) and the UK (0.21 centrality value with 59 publications), indicating that cooperation is less than in other countries. The top 3 journals in which authors publish their work in this field are Frontiers in Human Neuroscience (66), Frontiers in Neuroscience (39), and Neuroimaging (28).

3.5. fNIRS-EEG combined approaches in motor and extremity research

For motor and limb studies, fNIRS-EEG combinations are often preferred. fNIRS measures brain oxygenation levels, while EEG records neuronal activity; this combination provides a multidimensional and sensitive examination of brain processes during motor movements. This provides a more detailed understanding of the complex brain processes involved in movement planning and execution.

An in-depth analysis of the academic literature pertaining to these fields was conducted to gain a deeper understanding of the specifics of fNIRS and EEG. The initial phase of this analysis involved examining the keywords used in 153 studies. Cluster analysis was conducted using keywords. The resulting clusters were found to be reasonably coherent with a modularity Q value of 0.4817 and a silhouette value of 0.8143. The density of the clusters was visualized using heat maps. The heat maps in Fig. 7, also known as 2D histograms, were used to detect high-density areas. It is clear that there is a high density in the “Methodology-focused review” cluster. It is indicated that methodological aspects are a prominent focus of fNIRS and EEG studies, or that a significant proportion of studies in this field are dedicated to methodological advancements.

Figure 7.

Heatmap of keyword cluster analysis.

“Motor planning,” “bci monitor,” “combined-EEG-fNIRS study,” “brain-computer inference,” and “social brain function” are active study clusters in the fNIRS-EEG combined approaches in motor and extremity research. An important direction for future studies is the combination of EEG and fNIRS, which combines the power of neuroimaging methods. This combined approach to motor and limb research, which monitors both electrical activity (EEG) and hemodynamic changes (fNIRS), provides a more comprehensive dataset on motor movements.

4. Conclusion

Bibliometric analysis of fNIRS studies focusing on motor function and limb movements offers a significant advantage over previous large-scale studies by providing a more specific perspective. While bibliometric studies of fNIRS have examined broad clinical and cognitive applications, our study provides an in-depth analysis in more specific areas, such as motor disorders, motor control, and limb movements. This approach provides more direct and applicable insights into how fNIRS can be used for neurorehabilitation and individuals with motor function loss. The results of this study provide a detailed explanation of how fNIRS technology advances the understanding of motor function and its impact on limb movement in fields such as neurorehabilitation and sports science. The results demonstrate not only the place of fNIRS in the scientific literature but also its increasing importance in motor movement applications. The use of fNIRS as a potential rehabilitation tool for people with motor impairments is particularly important for severe conditions, such as amyotrophic lateral sclerosis, amputees, and spinal cord injuries. In this context, fNIRS-BCI systems stand out as promising methods for restoring motor control and enabling communication via neuroprostheses.

However, it is important that this study provides suggestions for future research. In particular, further studies on the use of fNIRS in conjunction with other imaging techniques (e.g., fMRI and EEG) may provide a more comprehensive understanding of the motor control and rehabilitation processes. At the same time, it is anticipated that methodological developments and technological advances in the field may contribute to a more widespread use of fNIRS and a more precise investigation of motor function.

fNIRS, leveraging its mobility and increasing channel counts, promises realistic tests and solutions for the relationship between cognitive function and motor movement. Its compatibility with standard tests for assessing cognitive function, electrophysiological data, and various imaging tools, such as fMRI, suggests that multi-data structures can be utilized in this field. Additionally, the ability to obtain results in a person’s natural environment using fNIRS is expected to provide significant predictive insights into cognitive function and its relationship with motor function.

To understand brain plasticity, fNIRS will be a crucial tool for elucidating cortical reorganization associated with limb loss and neurodegenerative disorders. In particular, in studies involving limited patient groups, the mobility and problem-solving capacities of fNIRS are highly valuable. The findings of these studies are also considered critical for the design of new rehabilitation system processes and tools.

Improvements in fNIRS preprocessing and processing are essential for accurately capturing hemodynamic responses. It is also possible to use fNIRS as a powerful imaging tool in time-frequency domain definitions and spectral analyses. In motor movement studies, fNIRS offers a strong feature set for AI-based classification and predictive modeling structures.

In conclusion, this bibliometric analysis provides important insights for future studies by detailing the role of fNIRS in motor skills and limb movement research. In addition to opening new avenues of research for scientists, the results of this study also encourage the use of fNIRS technology in applied areas such as neurorehabilitation and sports science.

Author contributions

Conceptualization: Esra Suzen, Fatma Yardibi, Ozlenen Ozkan.

Data curation: Fatma Yardibi.

Formal analysis: Esra Suzen, Omer Halil Colak, Sukru Ozen.

Investigation: Fatma Yardibi, Omer Ozkan, Ozlenen Ozkan, Omer Halil Colak.

Methodology: Esra Suzen, Fatma Yardibi, Ozlenen Ozkan, Omer Halil Colak.

Resources: Omer Ozkan.

Software: Esra Suzen, Omer Halil Colak, Sukru Ozen.

Supervision: Fatma Yardibi, Omer Ozkan, Ozlenen Ozkan.

Validation: Fatma Yardibi, Omer Ozkan.

Visualization: Esra Suzen, Sukru Ozen.

Writing – original draft: Esra Suzen, Fatma Yardibi, Omer Halil Colak.

Writing – review & editing: Esra Suzen, Fatma Yardibi, Omer Ozkan, Ozlenen Ozkan, Omer Halil Colak, Sukru Ozen.