Mapping Research Trends from 20 Years of Publications in Rhythmic Auditory Stimulation

This study aims to create an all-around insight into the evolutions, status, and global trends of rhythmic auditory stimulation (RAS) research via enhanced bibliometric methods for the 2001–2020 time period. Articles concerning RAS were extracted from the Web of Science database. CiteSpace, Bibliometrix, VOSviewer, and Graphpad Prism were employed to analyze publication patterns and research trends. A total of 586 publications related to RAS between 2001 and 2020 were retrieved from the Web of Science database. The researcher Goswami U. made the greatest contribution to this field. The University of Toronto was the institution that published the most articles. Motor dysfunction, sensory perception, and cognition are the three major domains of RAS research. Neural tracking, working memory, and neural basis may be the latest research frontiers. This study reveals the publication patterns and topic trends of RAS based on the records published between 2001 and 2020. The insights obtained provided useful references for the future research and applications of RAS.


Introduction
Rhythmic auditory stimulation (RAS) is a therapeutic technique that uses predictable time cues to improve different outcomes related to movement [1,2]. RAS is characterized by acoustic stimuli that synchronize motor response through rhythmic cueing, and the synchronous connection is produced by neuronal substrates with effects on movements, therefore the synchronization is created between the subject's movements and external auditory stimuli through an internal timing process [3][4][5]. Consequently, the subject will perform the time-coordinated movements and the rhythmic sequence programmed by the RAS model [4,6].
A large number of clinical research and systematic reviews have examined the effectiveness of RAS in the rehabilitation of Parkinson's disease (PD) [5,7], stroke [8], language disorders [9], and other sensorimotor dysfunctions. Combined strategies (e.g., integrating auditory cues and visual stimuli) have also been investigated to optimize the use of the technique.
Bibliometric analysis is used for analyzing literature characteristics, research impact, and trends over time [10]. In comparison to other reviewing methods, bibliometrics provide a more quantitative way of measuring research impact, so are seen as relatively objective. Bibliometric analysis has been widely applied to public health, rehabilitation, and medical fields [11][12][13]. A growing number of practitioners have focused on conducting RAS research over the past two decades. However, a report on RAS using bibliometric analysis has not been found yet. Research relating to RAS has been published in 234 journals. PLoS One published the most RAS-relevant studies, with 32 articles (Table 1). The Frontiers in Neuroscience ranked second with 20 articles. The Experimental Brain Research, Frontiers in Human Neuroscience, Human Movement Science, Journal of Neuroscience, and Neuroimaging tied for the third, publishing 16 articles on RAS research, respectively. The Frontiers in Neurology (n = 15), European Journal of Neuroscience (n = 14), and Gait & Posture (n = 14) ranked from eighth to tenth. Amongst the ten journals, there were six neuroscience journals, one multidisciplinary journal, one psychological journal, one neuroimaging journal, and a rehabilitation journal. Authors from 40 countries have contributed to research on RAS. As shown in Figure  2, the United States was the most productive country in RAS research and ranked the first in publication outputs, accounting for around 30% (n = 176) of the total. The United Kingdom (n = 60), Germany (n = 54), Canada (n = 53), Italy (n = 33), France (n = 30), The Netherlands (n = 25), Australia (n = 24), South Korea (n = 23), and China (n = 22) ranked from second to tenth and were all far behind on article totals. Figure 3 shows the collaborative relationships of the Top 20 most productive countries within RAS research. It is obvious  Authors from 40 countries have contributed to research on RAS. As shown in Figure 2, the United States was the most productive country in RAS research and ranked the first in publication outputs, accounting for around 30% (n = 176) of the total. The United Kingdom (n = 60), Germany (n = 54), Canada (n = 53), Italy (n = 33), France (n = 30), The Netherlands (n = 25), Australia (n = 24), South Korea (n = 23), and China (n = 22) ranked from second to tenth and were all far behind on article totals. Figure 3 shows the collaborative relationships of the Top 20 most productive countries within RAS research. It is obvious that the United States plays an irreplaceable leading role, which has built relatively stable collaborative relationships with at least ten countries. What should also be noticed is that the UK and Germany also demonstrated the high centrality to collaborative research in RAS. The collaborations displayed a significantly regional characteristic.
Aligned with the contribution of countries, Table 2 shows the 10 most contributive institutions in the RAS research field. The University of Toronto was the most productive institution, which had 42 publications related to RAS. The Colorado State University and Radboud University Nijmegen tied for second with 26 publications that the United States plays an irreplaceable leading role, which has built relatively stable collaborative relationships with at least ten countries. What should also be noticed is that the UK and Germany also demonstrated the high centrality to collaborative research in RAS. The collaborations displayed a significantly regional characteristic.  Aligned with the contribution of countries, Table 2 shows the 10 most contributive institutions in the RAS research field. The University of Toronto was the most productive institution, which had 42 publications related to RAS. The Colorado State University and that the United States plays an irreplaceable leading role, which has built relatively stable collaborative relationships with at least ten countries. What should also be noticed is that the UK and Germany also demonstrated the high centrality to collaborative research in RAS. The collaborations displayed a significantly regional characteristic.  Aligned with the contribution of countries, Table 2 shows the 10 most contributive institutions in the RAS research field. The University of Toronto was the most productive institution, which had 42 publications related to RAS. The Colorado State University and

Research Topics and Hotspots
The keywords co-occurrence analysis produced a network map to reflect high-frequency keywords and topics within RAS research. As Figure 4 shows, there are 3 clusters for 60 high-frequency keywords within the field of RAS research. The keywords in cluster 1 (colored by red) are mainly related to movement dysfunctions, cluster 2 (colored by blue) covers keywords related to perception, and cluster 3 (colored by gold) tends to be more related to cognition. This represents that previous research is mainly based on these three directions.
The frequency of citations could reflect the research topics within the field. The top ten cited articles in the field of RAS were shown in Table 4. Among the ten articles, there were two stroke studies and two PD studies, and the other six were related to perception and cognitive process. The article that received the highest citation, "Rhythm and beat perception in motor areas of the brain", was published in the Journal of Cognitive Neuroscience in 2007 and received a total of 501 citations. Most of the articles (i.e., 7/10) were published in journals in the field of neuroscience, and the other three appeared in medical and multidisciplinary journals. The top ten cited RAS articles in the past five years (2016-2020) were shown in Table 5. The ten articles tended to focus more on the neural mechanisms of RAS. The most cited article was "Motor origin of temporal predictions in auditory attention", which was published in Proceedings of the National Academy of Sciences of the United States of America in 2017. Among these ten articles, four were published in neuroscience journals, four in multidisciplinary journals, one in a biological journal, and one in a psychological journal.
The keywords co-occurrence analysis produced a network map to reflect high-frequency keywords and topics within RAS research. As Figure 4 shows, there are 3 clusters for 60 high-frequency keywords within the field of RAS research. The keywords in cluster 1 (colored by red) are mainly related to movement dysfunctions, cluster 2 (colored by blue) covers keywords related to perception, and cluster 3 (colored by gold) tends to be more related to cognition. This represents that previous research is mainly based on these three directions. The frequency of citations could reflect the research topics within the field. The top ten cited articles in the field of RAS were shown in Table 4. Among the ten articles, there were two stroke studies and two PD studies, and the other six were related to perception and cognitive process. The article that received the highest citation, "Rhythm and beat perception in motor areas of the brain", was published in the Journal of Cognitive Neuroscience in 2007 and received a total of 501 citations. Most of the articles (i.e., 7/10) were published in journals in the field of neuroscience, and the other three appeared in medical and multidisciplinary journals. The top ten cited RAS articles in the past five years (2016-2020) were shown in Table 5. The ten articles tended to focus more on the neural mechanisms of RAS. The most cited article was "Motor origin of temporal predictions in auditory attention", which was published in Proceedings of the National Academy of Sciences of the United States of America in 2017. Among these ten articles, four were published in neuroscience journals, four in multidisciplinary journals, one in a biological journal, and one in a psychological journal.

1.
Rhythm and beat perception in motor areas of the brain Grahn     Burst detection was applied to investigate the research hotspots further. Burst keywords can be identified as indicators of emerging trends [42]. Figure 5 presents keywords with the 25 strongest citation bursts between 2011 and 2020. The most recent burst keywords were working memory, tracking, and neural basis.
Clinically, RAS has been widely applied as a rehabilitation strategy. According to the analysis of high-frequency keywords, PD was the most focused disease within RAS research ( Figure 6). A total of 97 publications focused on PD. Stroke and speech impairment ranked second and third with 50 and 47 publications, while the other six were well below: dyslexia (18 articles), dementia (18 articles), multiple sclerosis (10 articles), apraxia (6 articles), cognitive impairments (6), cerebral palsy (5 articles), and depression (5 articles). All the diseases were neurological and psychiatric disorders. Clinically, RAS has been widely applied as a rehabilitation strategy. According to the analysis of high-frequency keywords, PD was the most focused disease within RAS research ( Figure 6). A total of 97 publications focused on PD. Stroke and speech impairment ranked second and third with 50 and 47 publications, while the other six were well below: dyslexia (18 articles), dementia (18 articles), multiple sclerosis (10 articles), apraxia (6 articles), cognitive impairments (6), cerebral palsy (5 articles), and depression (5 articles). All the diseases were neurological and psychiatric disorders.   Clinically, RAS has been widely applied as a rehabilitation strategy. According to the analysis of high-frequency keywords, PD was the most focused disease within RAS research ( Figure 6). A total of 97 publications focused on PD. Stroke and speech impairment ranked second and third with 50 and 47 publications, while the other six were well below: dyslexia (18 articles), dementia (18 articles), multiple sclerosis (10 articles), apraxia (6 articles), cognitive impairments (6), cerebral palsy (5 articles), and depression (5 articles). All the diseases were neurological and psychiatric disorders.

Discussion
The bibliometric analysis aimed to investigate the developmental trends and hotspots of research on 586 RAS articles during 2001-2020 by using VOSviewer, Bibliometrix, and Citespace. This work summarized research trends, topics, sources, and contributions for RAS.
An exponential growth has been observed in the output of RAS research. If the overall growth rate that was observed also applies to the future, we can expect the publication volume in the RAS research field to double about every five years. This strong growth of research output may be partially attributed to the increased prevalence of neurologic and psychiatric diseases (those were major types of disease applying RAS treatment). Given this, researchers and practitioners worked on exploiting RAS to facilitate rehabilitation. Intervention methods and analysis strategies are being developed [16,43,44]. The surge of RAS research in 2014 might be influenced by the release of the Global Burden of Disease (GBD) 2010 study at the end of 2012 [45]. The GBD 2010 found 114% and 68% increases in the mortality of psychiatric and neurological disorders from 1990 [46][47][48][49].
Among the top-performing journals, more than half were neuroscience journals. It seems that RAS researchers preferred to submit studies to the neuroscience area. An explanation is that for promoting research in the relevant field, some of the neuroscience journals have set auditory cognition and/or brain stimulation research sections or special columns. Most RAS studies were published in multidisciplinary journals also reflected a trend. This could be attributed to how the multidisciplinary journals tended to have a larger number of articles published than subject-specific journals. In addition, RAS has always been considered a transdisciplinary subject [50]. Publishing in multidisciplinary journals may expand the influence of studies.
Of the top productive countries, the U.S. ranked first in the number of RAS publications. Five American affiliations came in the top 10 institutions in the research of RAS, symbolizing that the U.S. possessed the most elite institutions around the world. However, there is only one American author in the top ten authors, indicating the contributions of US researchers were more even. This may be because the concept of integrating RAS into rehabilitation was initially proposed by an American researcher [51] thus the U.S. went deeper in this field as compared with the rest of the world. It is also worth mentioning that the boost of RAS clinical research should be attributed to a Canadian researcher, Dr. Michael Thaut, who has revolutionized the use of auditory stimulation in rehabilitation around the world.
Keyword analyses can be used for detecting research topics, analyzing research hotspots, as well as monitoring the research frontier transitions [52]. Through the analysis of the high-frequency keywords in the RAS research of the past 20 years, the studies were closely related to the three themes: motor dysfunction, cognition, and perception. This hypothesis was also supported by burst detection.

Motor Dysfunctions
Facilitating walking rehabilitation is a major use of RAS [7,32]. RAS demonstrated significant effects on improving muscle coordination and balance performance to patients with PD and stroke [53,54]. Early researchers focused more on the patterns of auditory cues [4,6]. More recently, researchers began to investigate the integrated effects of RAS and other techniques, such as exercise and robot-assisted therapeutic devices [55]. The comprehensive efficacy of RAS is constantly being verified in clinical trials for the rehabilitation of various movement disorders/dysfunctions [7,[56][57][58].

Cognition
Early studies regarding the beneficial effects of auditory stimuli on cognition (e.g., learning and memory) had been inconsistent [59]. However, more recent neurobiological research revealed the significant roles of RAS for cognitive performance of Alzheimer's disease, developmental language disorders, and also PD [9, 23,45,60]. To boost cognitive processing, a temporal sampling framework for specific disorders can be integrated into auditory stimuli [9]. Additionally, bass sounds were examined to be better at allocating attention than high-frequency sounds, providing incomplete evidence for the convention of using bass instruments to carry the rhythmic foundations of music [61].

Perception
The broad application of electrophysiological technologies (e.g., electroencephalogram) allowed researchers to explore the associations between RAS and sensory perception. Most early studies attempted to explain how neural sensory, and motor contributed to behavioral beat synchronization [62,63]. Over time, neural tracking, the mechanism of neural signals synchronizing to rhythmic sensory inputs [64], has been found to excel in rhythmic coordination of perception and action [61]. Clinically, RAS has been examined to be effective on improving perceptual performance, especially speech and visual perception [45,46].
We tried to provide information for future research and clinical uses based on quantitative bibliometrics, but a few limitations should be noticed. One is the findings of the study may be limited by the scope of the analytic tools. We only retrieved data from the WoS database, some vital information might have been missed. Another one is that the bibliometric data might not be adequate to assess the actual contribution of the research/researchers. Even though the quantitative metrics (e.g., number of publications) can reflect the popularity, the results should be explained carefully.

Conclusions
This study may help practitioners and researchers discover the publication patterns and emerging trends of RAS and present reference values for future research and applications. The most influential author, institutions, journals, and countries were Goswami U., University of Toronto, PLoS ONE, and the U.S., respectively. Motor dysfunction, sensory perception, and cognition are the three major clinical domains of RAS research. Neural tracking, working memory, and neural basis may be the latest research frontiers.  Institutional Review Board Statement: Ethics approval is not required to the study as it does not analyze data collected from human or animal subjects.