Magnetic resonance imaging in mood disorders: a bibliometric analysis from 1999 to 2020

Globally, mood disorders are highly prevalent, and are associated with increased morbidity and mortalities. Magnetic resonance imaging is widely used in the study of mood disorders. However, bibliometric analyses of the state of this field are lacking. A literature search in the web of science core collection (WoSCC) for the period between 1945 and 2020 returned 3073 results. Data extracted from these publications include, publication year, journal names, countries of origin, institutions, author names and research areas. The bibliometric method, CiteSpace V and key words analysis were used to visualize the collaboration network and identify research trends, respectively. Since it was first reported in 1999, the use of magnetic resonance imaging in studies on mood disorders has been increasing. Biological psychiatry is the core journal that has extensively published on this topic, while the UNIV PITTSBURGH, USA, has the highest published papers on this topic. Keyword analysis indicated that studies on depression, bipolar disorders, and schizophrenia, with a focus on specific brain regions, including amygdala, prefrontal cortex and anterior cingulate cortex are key research topics. Brain structure and network, sex differences, and treatment-associated brain changes are key topics of future research.


Introduction
Mood disorders are a class of psychiatric conditions that alter moods, energy and motivation. This group of disorders include depression, bipolar disorder, premenstrual syndrome, premenstrual dysphoric disorder, and the autism spectrum of disorders among others. Globally, mood disorders affect > 300 million people with devastating consequences [1,2]. Previous studies have reported the clinical prevalence of mood disorders [3,4], with children, adolescents, elderly people and patients with neurological conditions being prone to depressive disorders [5,6]. The pathophysiology of mood disorders is complex and not properly elucidated [7,8]. Moreover, it is influenced by genetic and environmental factors that may disrupt neurotransmitter homeostasis [7,8]. Therefore, mood disorders may result from negative feelings that leave a lasting influence on mental processes associated with perception, cognition and motor systems [9,10]. In affective disorders, brain morphological alterations largely occur on the fronto-limbic cortex, hippocampus and amygdala. These structures modulate an individual's emotional and cognitive function, and mood disorder manifestation [11]. Neuroimaging has revealed that brain networks that control emotional behaviors can also influence mood disorder pathophysiology [12]. Advances Mingzhou Gao and Hui Sun have contributed equally to this work and should be considered co-first authors.
* Dongmei Gao gcy_112@163.com * Mingqi Qiao qmingqi@163.com 1 3 in in vivo neuroimaging techniques have, therefore, greatly improved our understanding of mood disorder pathophysiology [13]. There are various clinical imaging techniques, such as PET, DTI [14], ASL [15], and 1H-MRS [16]. However, magnetic resonance imaging (MRI), one of the neuroimaging techniques that are widely used in neuroscience to visualize neural activities, and to identify mental disease biomarkers [17] has unique advantages. In particular, MRI is now widely used to evaluate mood disorders [18,19], such as depression [20] and premenstrual dysphoric disorder [21]. Advances in this technology will enable quantification of biochemical components of a functioning brain that are not detectable [22]. However, despite progress in the use of MRI to study mood disorders, accompanying bibliometric analysis is lacking.
Bibliometrics is a statistical analysis and quantitative tool for identifying research hotspots and frontiers in a particular field within a short time by extracting quantitative information on study distributions by country/region, institution, author, and journal [23]. Now, bibliometric methods in Citespace have been widely used in medical fields, such as acupuncture [24] and regenerative medicine [25]. The Cit-eSpace software was invented by Professor Chaomei Chen in early 2004. It is based on scientometrics and knowledge visualization, and is effective in mining scientific literature for critical information. It is used to establish co-occurrence network maps of authors, keywords, institutions, countries, and subject categories as well as co-citation networks of cited authors, cited references, and cited journals. For any discipline of interest, CiteSpace can greatly inform on basic knowledge, identify seminal studies in the area, uncover current research themes, and contextualize research advances [26].
Therefore, this study aimed at comprehensively analyzing the current status and developing trends of related references derived from Web of Science database from 1999 to 2020 in global research on mood disorders by magnetic resonance imaging through a bibliometric and visual analysis.

Data collection and screening
Online literature search was performed in web of science core collection (WoSCC). To minimize bias from daily database updates, the search was done on the same day (April 29th, 2020) using the following search terms: ( Only published articles were selected, and publication language was unrestricted ( Fig. 1).
Through descriptive analysis, we characterized the selected studies by publication year, country, journal, and authors and then constructed knowledge maps using the CiteSpace software (version 5.3.R4). Finally, citation growth rate or keywords were evaluated using burst in CiteSpace V.

General information of publications
Due to its quality curation of research studies, WoS is a reliable source of studies for bibliometric analysis of scientific literature [27,28]. Our first round of literature search found 54,346 and 331,162 reports on mood disorders and MRI, respectively. Among them, a total of 3073 original publications involving MRI in mood disorder studies were selected for further analysis. There was an increase in the number of publications on this topic between 1999 (28) and 2019 (236). Notably, in the last 4 months of 2020, 66 original studies were published (Fig. 2a). Citation scores of these studies for the 1999 (16) and 2019 (14,915) years exhibited a significant increase. An increase in citation scores by 4039 resulted in a total score of 142,188, with an h-index of 166 and on average, 46.27 references per item (Fig. 2b). These results indicate that MRI is increasingly becoming popular as a mood disorder research tool.

Funding analysis
Adequate research funding is crucial for scientific advancement and development of future researchers [29]. Research grant applications is critical in advancing one's academic career [30]. The 3073 published studies included in this study were supported by a total of 2976 grants. The CiteSpace journal co-citation analysis identified 1029 highly cited journals and 10,270 network lines. Biological psychiatry was the most cited journal (Fig. 3). Moreover, Biological psychiatry, Neuroimage, American journal of psychiatry, Archives of General Psychiatry, and Journal of neuroscience published many groundbreaking studies that influenced research direction (Tables 2and3). Biological psychiatry is society of biological psychiatry's official journal that is geared towards enhancing research excellence in areas studying the nature, causes, mechanisms, and management of thought, emotion, or behavioral disorders. Biological psychiatry is highly regarded in psychiatric neuroscience and was ranked 2nd among psychiatric journals in 2019 by google scholar, with an impact factor of 7.27.

Scientific Research Collaboration Analysis
Research collaboration is crucial in interdisciplinary research and helps overcome scientific challenges, leading to major research breakthroughs. Research collaboration network refers to the appearance of authors from different research institutions, or countries (regions) in the same paper.

Author analysis
CiteSpace analysis uncovered 919 nodes (corresponding to 919 prolific authors), and 1642 cooperative lines, indicating intense collaboration between authors (Fig. 4). Node size is directly proportional to the number of publications, while connecting line thickness is directly proportional to the level of collaboration. Phillips ML was the most productive author, with 59 publications, followed by Soares JC (54)   Co-author networks are used to analyze joint research in any field of interest. We found a total of 1029 highly cited authors and 1468 network lines, with Drevets WC being top, with 761 co-authorships (Fig. 5, Table 4).

Country analysis
CiteSpace analysis identified 57 countries and regions, with 70 collaboration links (Fig. 6), indicating high crosscountry (region) research links. Top countries by publication numbers were USA, Germany, England, China, Canada, Italy, Netherlands, Australia, Brazil and Switzerland. Interestingly, while Germany is the 2nd (481 publications), no German institutions are in the top 10 list of institutions (Fig. 6, Table 5).  Table 5).

Research area analysis
The studies included covered a total of 50 research fields, including immunology (171), followed by cell biology (65), and biochemistry-molecular biology (60). The top 20 research fields involving MRI use in mood disorders are  shown in Table 6. Neurosciences, neurology and psychiatry were the main research areas (Table 6).

Cited reference analysis
Analysis of cited reference co-citation identified 324 nodes and 555 links (Fig. 8). The top 10 cited publications on    (Table 7).

Research hotspots and frontiers
Within any given time period, knowledge maps of keywords co-occurrence reflect hot topics, while trending keywords indicate frontier topics. We used CiteSpace V to evaluate keywords in titles, abstracts, as well as article keyword sections, and to build keyword co-occurrence networks.

Key diseases of mood disorders
Individuals experience various emotions, such as sadness, irritation, happiness, and elation. However, people with mood disorders often experience strong emotions, for unusually long periods of time to the extent that emotions negatively impact their daily routines. Depression and bipolar disorders are the main mood disorders. Our analysis revealed 658 and 570 studies on depression and bipolar disorder, respectively. Depression is a significant clinical and public health concern [31] and its pathophysiology has not been clearly elucidated. Environmental as well as cellular and molecular factors trigger functional, and/or structural brain alterations [32] that may contribute to these disorders. Cognitive behavioral therapy is the best-studied psychotherapeutic approach, although its mechanisms of action are unknown [33]. Bipolar disorder is characterized by recurrent episodes of elevated moods and depression, along with changes in activity levels. Bipolar (affective) disorder, (previously manic depressive illness) is extremely difficult to manage [34]. Additionally, major depression and schizophrenia are major research topics. Schizophrenia is a heterogeneous psychiatric disorder with wide-ranging clinical and biological manifestations [35].

Key brain regions associated with mood disorders
Neuroimaging is increasingly becoming important in the study of the pathophysiology of mood disorders, with most studies focusing on the amygdala, prefrontal cortex and anterior cingulate cortex. The involvement of amygdala in major depression pathogenesis has been reported in previous studies [36]. While amygdala volume alterations have been implicated in mood disorders [37], Tomas et al. suggested that amygdala volume changes may reflect underlying illness or represent artifacts from medication or comorbidities [38]. The prefrontal cortex (PFC), which is involved in thinking and behavioral regulation has also been implicated in depression [39,40]. Based on anatomical connectivity and functional specialization, the prefrontal cortex is divided into two subregions: ventromedial prefrontal cortex (vmPFC) and dorsolateral sectors (dlPFC) [41]. The human anterior cingulate cortex (ACC) is involved in response to pain affecting oneself, or others. However, its mechanisms of activity are poorly understood [42]. The mean gray matter volume of the "subgenual" ACC (sgACC) cortex is abnormally low in individuals with major depressive disorder (MDD) and bipolar disorder [43]. Psychopathic traits have been negatively correlated with amygdala-ventral anterior cingulate cortex connectivity for angry versus neutral faces [44]. The anterior cingulate cortex may influence initiation, motivation, and goal-directed behaviors [45].

Research frontiers
Keyword bursts provide a means of forecasting research direction. CiteSpace V analysis of burst keywords revealed 209 burst terms and 40 keywords by the end of 2020, and predicted the research hotspots shown in Fig. 10.  . 9 Keyword co-occurrence map of MRI use in mood disorder studies from 1999 to 2020. Node size represents keyword co-occurrence frequency. Node size corresponds to keyword frequency 1 3

Discovering abnormal brain structure and network
In vivo neuroimaging of neural networks that putatively regulate normal emotions have shown that they are implicated in the pathogenesis of mood disorders [12]. Therefore, functional and structural brain pathology may contribute to mood disorders [46,47]. Reduced gray matter volume has been reported in the posterior cingulate cortex and superior temporal gyrus in bipolar disorder [48], while white matter abnormalities have been associated with major depression [49]. Severe depression cases exhibit ventricular enlargement, sulcal widening, and reduced frontal lobe, hippocampus and caudate nucleus volumes [50]. Advances in brain imaging of mood disorders have been mainly due to studying regional brain abnormalities, and more recently, studying whole brain connectome. Various MRI strategies, including 'functional' MRI, which detects changes in blood flow to a specific area of the brain, blood-oxygen-level-dependent imaging (BOLD) fMRI, which examines changes in oxygenated or deoxygenated blood, and resting state fMRI have significantly improved our knowledge of neuronal activity. Mood disorder patients exhibit morphological abnormalities or morphometry in various visceromotor network structures [51]. Young MDD reported significant reduction in DRNcingulate cortex connectivity relative to healthy controls and a correlation between DRN amygdala/hippocampus complex connectivity and depressive symptom severity [42]. Wayne C [46] proposed a neural model in which dysfunctions of MPFC and the circuits connecting it to other cortical and limbic structures account for mood disorder associated neural mechanisms.

Sex differences
Biological differences between sexes affect emotions and social behavior [52]. Lifetime depression and anxiety disorder rates are twice higher in women than men [53,54]. Gender impacts mood and anxiety disorders and may, therefore, offer some insights into the mechanistic basis of affective disturbances in men and women [55][56][57]. Besides, women's moods are significantly affected by changes in menstrual cycle. Depression, anxiety and irritability are the three most studied symptoms of PMDD in women of childbearing age [58]. During the perinatal period, mood disturbance are commonly reported by women [59]. Depression is a common complication of pregnancy and the postpartum period [60]. Moreover, menopausal transition is correlated with mood changes ranging from distress to minor depression to major depressive disorders in women [61].The role of reproductive steroids in mood regulation among women may become the most potential research direction [62]. Given that sex differences affect various aspects of the brain, including its structure, function, and stress responses, studies should aim at establishing gender-associated differences that have an effect on mood disorders [63,64].

Effect of treatment measures on brain activity
Effective therapies for emotional disorders are urgently needed. Although various pharmacological and non-pharmacological therapeutic approaches, including electroconvulsive therapy, cognitive behavioral therapy, ketogenic diets [65], circadian treatment [66], bright-light therapy [67] and drug naïve (ketamine, lithium salts) [68,69] are available, their efficacies are widely varied between individuals.

Evidence-based medicine with meta-analysis
Meta-analysis may help discern important findings across disparate studies, thereby, providing actionable information to clinicians [70,71]. For instance, multiple antidepressants are available for MDD, and their prescription should be informed by the best available data. A meta-analysis by Andrea Cipriani [72] on the relative efficacy and tolerability of 21 antidepressants used to treat major depressive disorders found that all antidepressants were more efficacious than placebo in adults. A meta-analysis study also showed that the relationship between creativity and mood disorders  1  fMRI  862  11  Emotion  372  2  Depression  658  12  Disorder  365  3  Bipolar disorder  570  13  Functional connectivity  342  4  Amygdala  530  14  Metaanalysis  323  5  Mood disorder  517  15  Activation  317  6  Prefrontal cortex  496  16  Abnormality  312  7  Brain  455  17  Major depressive disorder  301  8  Major depression  440  18  Mood  263  9  Schizophrenia  419  19  Anterior cingulate cortex  233  10  MRI  386  20 Magnetic resonance imaging 221 varied according to the research methodology used [73]. Low doses of partial dopamine agonist therapies represent a relatively safe and effective alternative therapies [74]. Additionally, a meta-analysis showed that lithium, which is commonly used to manage mood disorders, may elevate the risk of low urine concentrations, hypothyroidism, hyperparathyroidism, and weight gain [75].

Conclusions
We performed a bibliometric analysis of publications involving the use of MRI in mood disorder studies over the last 2 decades. Within this period, the number of such publications rose rapidly. The US, England, and Canada have high publication rates and centrality in this area. Additionally, research collaborations on this topic are very active. Our findings reveal Soares JC as the most prolific author, while the US emerged as the global leader in research investment. Mounting evidence shows that MRI is an effective tool for elucidating neuronal activity. The US department of human health services is the leading funding agency on this topic. We also found that depression, bipolar disorder and schizophrenia are the major focus areas when evaluating major brain regions, including the amygdala, prefrontal cortex and anterior cingulate cortex. Our findings indicate that in future, studies will likely focus on brain structure and network, sex differences, brain changes upon treatment and meta-analyses.