Bibliometric research investigates authorship, publication patterns, citations, and content by applying quantitative measures to a selected body of literature [26].

To address this analysis, python libraries such as Pandas [25] for data manipulation and analysis, matplotlib [27] and seaborn [28] for visualization were employed.

To evaluate the evolution and growth of literature in this field, a graph mapping the number of publications each year and the cumulative number of publications over time was generated (Fig. 2).

Publication Trends Over Time [Each bar in the graph represents the number of articles or studies published in each year. Each point in the graph represents the total number of publications that have been produced up to that year, adding up all the publications from previous years]

Research on the 'cerebellum in emotions' has grown significantly over time. Despite some minor fluctuations, interest in this field has consolidated and increased exponentially. Since 1997, publications have shown a steady increase, exceeding 100 per year as of 2022. This growth may be due to technological or methodological advances and new applications, with a marked acceleration phase coinciding with the discovery of cerebellar cognitive affective syndrome.[29].

To gain a deeper understanding of the evolution of research dissemination and to identify the dominant types of publications during specific periods, stacked charts were generated (Fig. 3), depicting the number of publications per year by publication type (e.g., Journals, Congresses, Books, etc.) and domain (e.g., Neuroscience, Pharmacology, etc.), as shown in Table S1.

Annual Publications by Type [Each type/domain of publication is represented by a different colour]

Initial publications primarily focused on psychiatry, followed by neurology, biology, medicine, and neuroscience. These areas later expanded to include psychology, medical technologies, computational methods, pharmacology, and therapy, reflecting a progressive diversification. Scientific papers predominate over books and conference presentations, highlighted academic journals as the main means of dissemination.

As we mentioned in the previous figure and from a review of the titles and abstract, we note that the finding of the relationship between cerebellar damage and emotional and cognitive deficits, along with studies conducted in humans, prompted a successive wave of investigations that were consolidated by methodological refinements and technological advances.

Identifying the journals with the largest number of publications helps us understand where the most significant advances in the area are being shared and where to look for information. A bar chart (Fig. 4) represents the leading journals with their quantitative impact indicators and the type/domain of the publications on them. As an extension of this figure, a table (Table S2) presenting the impact factors and quartiles of each journal is included in the supplementary material.

Leading Journals [The X-axis shows the number of publications and the Y-axis the name of the journal, the colour still codes the type/domain of papers]

Most publications are concentrated in journals focused on psychiatry, followed by neuroscience and neurology. The leading journals in this area are The Cerebellum [30], dedicated to cerebellar research and related disorders, and the Journal of Affective Disorders [31], which covers affective disorders, including depression, mood spectrum, emotions, anxiety, and stress.

Citations are an important indicator of the influence and impact of research. The most cited publications were selected and represented through a simple bar plot (Fig. 5). For more detailed information on the names of the journals in which each paper was published and the authors, see Table S3.

Leading cited publications [The Y-axis lists each research study along with its publication year, while the X-axis represents the number of citations received. Each bar is coloured according to the type of publications, making it easy to identify the most prolific and interesting areas]

Two publications have a significantly greater impact with citations of 5,679 and 3,740, which is much higher than the others, starting to decrease from 1,873 citations onward. The first of them, published in 2004, although it is not the first to address the role of the cerebellum in the experience of pain, it is one of those that provide the clearest evidence about its activation in the context of empathy towards the pain of others. The second from 1998 on the cognitive-affective cerebellar syndrome was a key point in the recognition of the role of the cerebellum in emotions and marked a change in the traditional view of the cerebellum. Also, most of the highly cited publications are human imaging studies, with great relevance and impact, as they provide key information for the diagnosis, treatment, and understanding of diseases. This explains their high number of citations, given that findings in human imaging are often highly applicable both in research and clinical contexts.

Keyword co-occurrence analysis examines relationships and patterns between terms in a body of literature, visualizing their interconnectedness and revealing thematic structure. It allows identifying major themes, mapping knowledge networks, discovering promising or understudied areas, and focusing research on unexpected or less explored connections.

For the co-occurrence analysis (Fig. 6), we considered the titles of scientific publications and used VOSviewer [32], a software tool specifically designed for the construction and visualization of bibliometric networks.

Keyword co-occurrence network [Each node represents a keyword; its size indicates its importance or frequency, while the lines show co-occurrences in the same context. Colors group semantic clusters linked to specific topics. The most central words in the network are near the center, indicating greater relevance, while peripheral words have less significant connections or deal with more specific topics]

The word “cerebellum” is the main core of the corpus, and “emotion” one of the central nodes, which is not surprising given our search approach. Other relevant words (themes) are “functional connectivity”, “depression”, and “disorder”. There are subthemes connecting the cerebellum with specific conditions such as “major depression”, “schizophrenia” and “anxiety disorders”. One particular cluster is related to “cerebellar cognitive affective syndrome”, which links to other nodes such as “patient”, “ataxia” and “case study”. Other minor semantic clusters connect the cerebellum with cognitive and social functions, while another one links “behavior” with nodes such as “rat” and “mouse”, highlighting the study of the cerebellum in motor and emotional functions through animal models.

Knowing which species have been investigated the most in the field and which have been investigated the least not only identifies trends and applicability of the studies, but also measures the diversity of approaches and methodologies, identifies possible biases in the research, and reflects economic or political influencing factors. A pie chart (Fig. 7) illustrates the distribution of species in the publications. Although computational modeling is not technically a species, it can be included in the analysis, since it represents a source of valuable in-silico data.

Distribution of species [Each segment of the pie chart represents the relative share of each species (including “modeling”) in the total]

Human studies dominate research, reflecting a focus on medicine, psychiatry, and clinical applications. Rats and mice are the second and third most studied species, serving as essential models to explore biological processes and mechanisms underlying emotions, given their similarities to the human brain and practical advantages. Other species are minimally represented, emphasizing the translational focus on human applications.

Computational modeling, though promising, remains underdeveloped, comprising only 0.6% of studies. Existing models, such as the Emotional Cerebellar Model Articulation Controller (CMAC) and its variants, simulate cerebellar functions for complex system control, but further advancements are needed to match the depth of biological research.

A deeper understanding of how research has been oriented over time combines both what is being investigated (topics) and on which species is conducted. First, the thematic category is identified; second, the thematic evolution by species is displayed, by a heat map (Fig. 8). Spikes in specific topics can indicate scientific breakthroughs, technological changes, or discoveries that have driven that topic forward.

Thematic evolution by topic by species [Years on the horizontal axis, and topics and species on the vertical axis. Colours indicate the number of publications for each topic-species combination in a given year, and the colour gradient corresponds to the number of publications]

To offer an even more detailed view, a bubble chart shows the 5 most dominant topics in each area (Fig. 9).

Leading topics [The X-axis displays the macro research areas, while the Y-axis represents specific topics within those areas. The size of each bubble indicates the number of publications related to that topic]

Predominant topics include “neurological disorders”, “affective and mood disorders”, and “emotional processing, arousal, and valence” in humans. The least explored areas so far are “neurodevelopmental disorders”, “substance use disorder”, and “neuromodulation and neurostimulation”. In detail, in the area of “​​affective and mood disorders”, human studies predominate, with a significant increase starting in 2010 and a peak starting in 2022. To a lesser extent, there were publications on this topic in mice and rats, whose studies began in 2007 but without continuity. In the area of ​​ “anxiety and obsessive–compulsive disorders”, publications in humans also predominate, with interest emerging since 2009 and peaking in 2023. Regarding “​​emotional processing, arousal, and valence”, studies in humans have been constant since 2001, showing an increase since 2013. There are also publications in mice and rats, which are smaller and less frequent. In the area related to “fear, anxiety, and reward learning processes”, research is more evenly distributed between humans, mice, and rats, with a slight predominance in humans. In the area we call “general summary”, the focus on humans stands out, with a peak in 2022, followed by those writings where more than one species are mentioned (General). “Neurodevelopmental disorders” are a relatively under-researched topic, and focus especially on humans. Most studies focus on “neurological disorders”, which have seen steady growth since 2006, with a notable peak in 2020. Publications involving rodents in this area are quite limited. Additionally, fields such as “neuromodulation and neurostimulation”, as well as “trauma- and stress-related disorders”, also show a clear preference for human studies. Finally, research on “psychotic disorders and schizophrenia”, and “substance use disorders” has been conducted exclusively on humans.

The analysis of the techniques employed in the investigation of the cerebellum and its role in emotions provides a comprehensive and simplified view of a complex picture. This helps to highlight trends, identify the most researched areas and reveal possible gaps in the use of methodologies, which is essential to guide future research.

For this analysis, we used a Sankey diagram (Fig. 10), which effectively organizes and displays the connections between categories.

Techniques of research [The thickness of the lines indicates the frequency of use of each technique in the studies analyzed. Main Categories: Brain Stimulation Techniques: Includes methods such as rTMS (Repetitive Transcranial Magnetic Stimulation), ECT (Electroconvulsive Therapy), a-tDCS (Anodal Transcranial Direct Current Stimulation), DCS (Direct Current Stimulation), tACS (Transcranial Alternating Current Stimulation), Offline tDCS (Offline Transcranial Direct Current Stimulation), which are used to stimulate or modulate brain activity. Neurophysiological Techniques: include EEG (Electroencephalography), fTCD (Functional Transcranial Doppler), LEPs (Laser-Evoked Potentials), ERP (Event-Related Potentials) methods, which measure brain activity. Neuroimaging Techniques: Comprises advanced methods such as fMRI (Functional Magnetic Resonance Imaging), MRI (Magnetic Resonance Imaging), PET (Positron Emission Tomography), DTI (Diffusion Tensor Imaging), SPECT (Single Photon Emission Computed Tomography), (CT) Scans (Computed Tomography Scans), PWI (Perfusion-Weighted Imaging), T1-weighted Images, DWI (Diffusion-Weighted Imaging), H-MRS (Proton Magnetic Resonance Spectroscopy), In vivo MRS (In vivo Magnetic Resonance Spectroscopy), used to study structures, functions and biochemical processes in the brain. Other techniques include experimental and analytical tools in both animals and humans]

Neuroimaging techniques stand out as essential tools in research because of their noninvasive capacity to analyze brain activity and structure, with emphasis on the interaction of the cerebellum and emotional regions. Although stimulation and neurophysiological methods complement this exploration by modulating and evaluating cerebellar activity in real time, their use is secondary. On the other hand, animal models and preliminary techniques are of limited use, suggesting an approach mostly applied to humans, consistent with the complexity of emotions.

In a bibliometric analysis, identifying the authors with the highest number of publications in a specific area of interest is essential. It helps pinpoint the most influential and active researchers, who often drive advancements and set trends within the field.

Additionally, these authors are frequently part of key collaboration networks, making it easier to recognize scientific communities and their working dynamics. It also reveals which areas or topics an author publishes in most, allowing for insights into their expertise and research focus.

A stacked bar chart (Fig. 11) shows the ranking of authors, assigning equal weight to all to reflect collective contributions without specific hierarchies.

Leading authors [Each bar represents an author and is divided into colouring segments indicating the number of publications according to his/her type/domain of publication. The X-axis shows the number of publications, while the Y-axis shows the names of the authors]

It is known that, at times, the position of authors in a publication may reflect the roles they played in the project (study leader, key support in analysis, experiments or knowledge, supervisor, project manager, and funding). This aspect, detailed in Figures S1 and S2, helps to clarify the contributions and leadership of each author, offering a perspective on their academic impact.

The most prolific and influential author on the subject is Schmahmann, with a greater number of publications in Neurology, which is not surprising since he is known for his pioneering work [33] on the role of the cerebellum in cognition and emotion, and the description of the cerebellar cognitive-affective syndrome (also called Schmahmann syndrome). The five authors with the highest number of publications are Schmahmann, with more than 30 publications; Wang, with 25 publications; Schutter and Timmann, with more than 20 publications each; and Zhang, with 16 publications. This indicates a high level of research activity. In general, most authors have between 5 and 15 publications. It is observed that researchers publish in journals of different related topics, which suggests a trend toward interdisciplinarity.

To understand how knowledge is produced in this field, patterns of collaborations and authorships can be described, specifically how researchers collaborate on publications, focusing on the number of authors involved and the frequency of their contributions across multiple works (Figs. 12 and 13). Figure 12 shows how the total number of publications varies according to the number of authors involved in each publication. On the other hand, Fig. 13 illustrates how many authors are involved in different numbers of collaborations, focusing on the frequency of their collaborative activity.

Distribution of Publications by Author Count [Distribution of publications based on the number of authors, with stacked bars differentiating between type/domain of publications]

Collaboration Patterns by Number of Authors [Scatter plot of the number of collaborations per author. It displays individual collaboration levels]

Most of the publications have between two and eight authors, with a peak of four. Only two publications stand out for their high number of authors (50 and 110) in the field of clinical research. These high-author counts are attributed to multi-center studies, diversity of specializations, large clinical trials, and international collaborations. The prevalence of publications with fewer authors suggests that limited collaborations may enhance specialization in specific research topics.

The majority of authors have few collaborations, as collaborations increase, the number of authors decreases exponentially, forming a “long tail”.This tail represents a small group of highly prolific authors, some with over 100 collaborations. This suggests a collaborative structure where a core group of authors frequently contributes to multiple projects, while most participate occasionally. Additionally, these collaborations involve various combinations of authors rather than fixed teams.

Analyzing the sources of funding sheds light on the resources available to researchers and the priorities of funding agencies, evaluates the relationship between investment and scientific impact, facilitates strategic collaborations, and helps in planning for future investments in research and development. However, it is important to note that not all research articles report funding due to various reasons, such as the use of their own resources, standard institutional infrastructure, or public data that do not require explicit funding. Also, not all journals have strict policies on funding disclosure, mainly in older papers. In other cases, funding may be indirect, coming from previous projects, institutional resources, or general grants not specifically assigned to the work in question.

Table 2 shows the public and private organizations that have funded research on the cerebellum and its role in emotions, based on the information reported in the scientific publications.

Financial support from diverse funding bodies including government agencies, ministries, national research councils, associations, foundations, and independent public and private organizations, are supporting this research field. Particularly in the U.S., Europe, Australia, the U.K., and China. However, limited funding in Latin America, Africa, and parts of Asia highlights disparities in resources and infrastructure, restricting participation in this field.

Author affiliation plays a key role in scientific production, reflecting the resources and infrastructure of institutions rather than the researchers' nationality. Understanding affiliations help identify leading countries and institutions, reveal regional inequalities, guide global collaboration, and inform funding priorities. A world map (Fig. 14) combines geographic and scientific data to illustrate these dynamics created using GeoPandas [34].

Geographic Distribution of Authors' Affiliations [A colour gradation identifies leading and less-developed regions in terms of involvement in this scientific research]

Author's affiliations show significant regional disparities. The U.S. and mainland China lead in publications, followed by some European countries, the U.K., and Canada, supported by strong infrastructure, technology, and funding (observed in Table 2). Countries like Australia, Japan, South Korea, Brazil, and Mexico rank at an intermediate-low level, while regions/economies such as Taiwan (China), Hong Kong (China), Israel, Singapore, India, Argentina, Chile, South Africa, Russia, Saudi Arabia, and Iran show very low output. Africa, more Latin American countries, and Central Asia are largely underrepresented due to limited infrastructure, funding (observed in Table 2), and prioritization of research, further exacerbated by the "brain drain" and difficulties in gaining global recognition.

This section focuses on the temporal analysis of academic collaboration networks and co-authorship dynamics. This analysis exploits two Python libraries designed for the creation, manipulation, and study of complex network structures, NetworkX [35] for exploratory analysis and the calculation of metrics, and Graph-tool [36] for large-scale network visualization.

A network graph is useful to understand a complex system and how its components interact with each other [37]. In a collaboration network, the nodes represent individual entities (authors), while the connections indicate collaborative relationships between them in scientific publications. The Scalable Force-Directed Placement algorithm available in the NetworkX library [35] provides a clear and scalable visualization of the graph structure over different periods. Figure 15 illustrates how the network has evolved, based on four selected periods, allowing us to assess how collaborations develop and change throughout these intervals.

Evolution of the Collaboration Network in Different Periods of Time [Evolution of the network in four periods. Nodes are colored and scaled based on their degree centrality, which reflects the number of connections each node has. Highly connected nodes are blue, peripheral ones with fewer connections are red, and intermediate nodes are purple. The layout minimizes overlap, clustering more connected nodes together. Duplicate edges are removed to avoid clutter, and "bunches of grapes" represent clusters or communities of closely connected nodes]

In the first period 1961–2000 (top left image) the network is small, consisting of compact, isolated groups with limited connections. Key nodes (in purple blue) serve as links between groups, indicating preliminary alliances.

During the period 1961–2012 (top right image), a significant increase in nodes and connections is observed, leading to a dispersed network with no dominant nodes or groups. Clusters operate autonomously, with strong internal connections but limited interactions between clusters, except for a few bridging groups.

Between 1961–2018 (bottom left image), the network grows further, with higher connection density and reduced isolation of peripheral clusters. Emerging hubs act as central nodes, fostering inter-group collaboration and reflecting a growing number of collaborators.

The complete network, spanning from 1961–2024 (bottom right image) achieves high complexity, with numerous nodes and dense clusters. Subgroups are clearly visible, indicating frequent collaborations within areas of expertise or specialized teams. The absence of a dominating node highlights a balanced, distributed structure. Network.

The metrics used in graph theory [37,38,39,40] not only enhance visualizations by providing concrete data but are also essential for characterizing and analyzing the collaboration network, offering a deeper understanding of its structure, key nodes, and temporal dynamics.

We strongly encourage readers to review [37,38,39,40] for theoretical foundation, as well as the Supplementary Table S4 which presents a detailed analysis and explanations of the network's main metrics, assessing its structure and behavior over time. This will provide a better understanding of the analysis.

The metrics confirm quantitively the previous observations in Fig. 15. There is significant growth in the number of publications, researchers and collaborations, reflecting increased scientific activity and interaction among participants. In addition, the network is becoming more cohesive, with larger groups of people working together and an increase in the formation of small, well-connected teams, which act as “hubs” for collaboration in specific areas.

When analyzing the most important connections within the network, it is observed that, although researchers are more distributed and less concentrated in one place, some of them become key intermediaries to connect different areas of knowledge. There is also a general improvement in connectivity among participants, which means that it is easier for anyone within the network to access information or collaborate with others.

Diversity in connections increases, indicating that researchers are collaborating with more people and on more varied topics. At the same time, small groups that used to be very isolated are becoming more integrated, forming a more interconnected network. However, while the network is more resilient to the loss of random connections, it is still vulnerable if certain key individuals or teams that act as important bridges are lost.

Overall, the network not only grows in size, but also becomes more complex, diverse and integrated. This reflects more efficient scientific collaboration, with broader interactions and a structure that fosters both specialization and connection between different areas of knowledge.