The trend in the number of publications and citations from 2004 to 2024 is depicted in Fig. 1 Overall, both the annual publication and citation counts show an increasing trend. The trend in publication counts exhibits some fluctuations, with each period spanning 5–6 years showing incremental growth. There are slight declines in 2011, 2016, and 2022, while in 2010, 2015, and 2021, the publication count increased significantly compared to the previous year, reaching peak values in those years as indicated by the peaks in the graph, with a peak of 136 articles/year reached in 2023. The citation frequency has increased year by year from 2004 to 2022, with significant increases observed in 2018–2019 and 2020–2021. The peak occurred in 2022 (7098 citations/year), with 2023 showing a similar level to 2021 (Fig. 1A).

Trends in the published articles and citation counts of circadian rhythms in depression from 2004 to 2024

(A) The annual publication quantity and citation frequency of circadian rhythms in depression from 2004 to 2024

(B) The annual publication quantity, cumulative publication quantity, and their polynomial fitting curves for researches of circadian rhythms in depression from 2004 to 2024

Additionally, a polynomial fitting was performed on the cumulative annual publication count as shown in Fig. 1B, with the fitting formula: y = 2.6979 × 2 + 20.421x − 9.1684, and the goodness of fit (R²) was calculated to be 0.9994.

To analyze the geographical origin of relevant publications, visualize the spatial distribution of research outcomes in the field, and understand global collaboration in this area, Supplementary Table 1 provides insight. In the field of circadian rhythms in depression research, the leading countries are the United States and the United Kingdom, both of which hold prominent positions in terms of publication volume and citation frequency internationally. For a more intuitive understanding, we conducted a detailed analysis of the top 30 countries by publication volume, as shown in Supplementary Fig. 1.

The top ten authors ranked by publication volume and co-citation frequency were presented in Supplementary Table 2. The most prolific authors contributing to publications in this field include Eus J. W. Van Someren from the Netherlands Institute for Neuroscience, P. A. Geoffroy from Hôpital Bichat Claude-Bernard in France, Nakao Iwata from Fujita Health University in Japan, and Tsuyoshi Kitajima. All four authors have published 12 or more relevant articles in this field. Among the authors ranked high in co-citation frequency, Aj Lewy from Oregon Health and Science University in the United States (with 280 citations) and F. Benedetti from the University of Trieste in Italy (with 273 citations) receive significant attention. We also employed VOSviewer to visualize the collaboration network among these authors, as depicted in Supplementary Fig. 2.

In Supplementary Table 3, we reported the top ten institutions ranked by the number of publications and total citations. It is evident from this table that institutions based in the United States dominate both in terms of publication volume and citation frequency. The top two institutions in terms of publication volume are the University of California, San Diego (38 papers) and the University of Pittsburgh (36 papers), both located in the United States, followed by the University of Toronto (30 papers) in Canada. In terms of total citations, the top four institutions are all based in the United States: The University of California, Los Angeles (3290 citations), Stanford University (2971 citations), the University of California, San Diego (2360 citations), and the University of Pittsburgh (2291 citations).

Visual analysis of the collaboration network among these institutions in Supplementary Fig. 3A reveals several clusters. Supplementary Fig. 3B further incorporates the temporal dimension into the spatial analysis presented in Supplementary Fig. 3A by highlighting the sequence of collaboration through colors.

Supplementary Table 4 presents the top ten journals in terms of publication and citation frequency in the field of circadian rhythms in depression. The journal with the highest output is Chronobiology International (77 articles), followed by Journal of Affective Disorders (50 articles) and Psychoneuroendocrinology (39 articles). Supplementary Fig. 4A visualizes the relationships among journals that have published literature related to circadian rhythms in depression. Besides, Supplementary Fig. 4B categorizes journals related to researches of circadian rhythms in depression into three groups based on co-citation frequency. In the double-overlay of Supplementary Fig. 4C, citing journals are connected to cited journals through lines of varying thickness.

Circadian rhythm appears most frequently (634 occurrences), nearly twice as much as the second-ranked keyword. Following closely are depression (339 occurrences), sleep (185 occurrences), melatonin (144 occurrences), and bipolar disorder (131 occurrences), all appearing over 100 times, indicating significant research hotspots in this field. The top 10 keywords in terms of occurrence frequency and total link strength were presented in Supplementary Table 5.

Figure 2 illustrates the co-occurrence relationships and strengths among keywords, with closely related keywords grouped into clusters. In Fig. 2A, the keyword circadian rhythm is at the center, with clinical research-related nouns such as behavior, serotonin, clock genes, and mouse forming the red cluster. The light blue cluster in the upper left mainly consists of nouns related to diseases and pathogenesis, such as melatonin, ramelteon, inflammation, Parkinson’s disease, and Alzheimer’s disease. The green cluster is associated with daily life and sleep issues: sleep, mood, insomnia, activity, and quality of life. Additionally, keywords related to adolescent life rhythms, such as bipolar disorder, chronobiology, chronotype, adolescence, children, and ADHD, form the blue cluster on the right. Keywords primarily associated with treating psychological problems or rhythm disorders, such as bright light therapy, chronotherapy, and mental health, form the purple cluster, while those related to mental stress form the yellow cluster: cortisol, stress, diurnal rhythm, and schizophrenia. Figure 2B and C provide insights into the contribution strengths of different keywords over time. Popular keywords in recent years include adolescence, mental health, chronotype, inflammation, sleep quality, aging, and pain.

Keywords co-occurrence network mapping of circadian rhythms in depression from 2004 to 2024

(A) The keyword map of circadian rhythms in depression visually displays the connections among studied keywords. Nodes, distinguished by various colors, represent different keyword clusters. Node size reflects co-occurrence frequency, and connections between nodes depict relationships among keywords

(B) The figure depicts keywords’ recent contributions to researches of circadian rhythms in depression relative to their overall output from 2004 to 2024, with a red bias indicating increased influence and a blue bias suggesting decreased activity in the field. The color scale reflects the ratio of keywords over the past five years, highlighting institutions with notable impacts or reduced involvement in this study

(C) Keywords heat map on researches of circadian rhythms in depression

(D) The diagram illustrates the 25 primary keywords characterized by pronounced bursts of citations, denoted by red spikes on the timeline. These spikes signify sudden surges in citation counts, signaling pivotal moments of emerging crucial questions or solutions within the field

Additionally, we list the top 25 keywords with the strongest burst citations in Fig. 2D. The burst periods of the top-ranked keywords, such as diurnal rhythm (burst strength 8.07), secretion (burst strength 7.84), and seasonal affective disorder (burst strength 9.97), all fall between 2004 and 2013. It’s worth noting the keywords below the graph: validity (burst strength 5.31), physical activity (burst strength 6.63), anxiety (burst strength 5.43), sleep quality (burst strength 7.82), and mental health (burst strength 9.3), which remain at the forefront of research to this day.

Table 1 lists the basic information of the top fifteen highly cited articles. The most cited article is a review titled “Practice parameters for the indications for polysomnography and related procedures: An update for 2005” by Kushida, CA et al. [26], published in Sleep in 2005 (1121 citations). In this article, the authors updated the practice parameters for polysomnography and related procedures, covering diagnostic categories related to sleep-disordered breathing, hypersomnia, parasomnias, periodic limb movement disorder, insomnia, depression, and circadian rhythm sleep disorders. This review laid the foundation for subsequent research in this field. The second most cited article is “The two-process model of sleep regulation: a reappraisal” [27] (755 citations), published in 2016 by Borbély, AA et al. in the Journal of Sleep Research. Despite being the most recent among the top fifteen highly cited articles, it demonstrates significant influence in the field. The authors proposed a reappraisal of the two-process model, highlighting its continued relevance in organizing thoughts on sleep regulation along two axes and its role in integrating the disciplines of sleep and rhythm research. Additionally, they provided evidence from animal experiments suggesting the involvement of the suprachiasmatic nucleus in regulating circadian rhythms in humans, thereby offering a theoretical foundation and research direction for the field of researches of circadian rhythms in depression. The third article, “Sleep and depression” [28] (747 citations), is a review published in the Journal of Clinical Psychiatry in 2005 by Tsuno, N et al. The authors conducted a comprehensive review of literature published in English or French between 1964 and 2005 on the relationship between sleep disorders and depression. Their analysis indicated a significant association between changes in sleep and depression, suggesting the need for further development of comprehensive depression models and new treatment approaches.

In CiteSpace, we analyzed the temporal relationships between different studies. As shown in Fig. 3A, the publication time of highly cited articles and their potential relationships are illustrated. Articles marked with purple circles signify their potential impact in driving the development of the field or serving as key turning points in academic research. This helps researchers in the field better understand the trends in its development.

Highly cited references co-citation network mapping of circadian rhythms in depression from 2004 to 2024

(A) Application of citespace, literature relation network diagram. The different colors of the nodes indicate additional years, from blue to red, the later. The node size indicates the frequency of the reference

(B) The keyword heatmap displays topics, with smaller numbers indicating larger clusters, and #0 representing the largest cluster. Node size reflects co-citation frequency, and the links between nodes signify co-citation relationships

(C) The diagram illustrates the 25 primary references characterized by pronounced bursts of citations, denoted by red spikes on the timeline. These spikes signify sudden surges in citation counts, signaling pivotal moments of emerging crucial questions or solutions within the field

In Fig. 3B, we further explore the potential relationships between the main contents of these studies, dividing them into 16 clusters based on their relevance. The most extensive theme is #0 human, which owes its emergence to early themes such as #1 bipolar disorder, #2 actigraphy, #4 circadian disruption, and #6 chronotype. Subsequently, from #0 human, there are explorations into related themes such as #3 melatonin, #7 genetics, #11 structure-activity relationships, and #13 chronotypes.

In Fig. 3C, the burst situation of citations is illustrated. The earliest burst occurred in 2006 with Nievergelt, CM et al.’s article “Suggestive evidence for association of the circadian genes PERIOD3 and ARNTL with bipolar disorder” published in the American Journal of Medical Genetics Part B-Neuropsychiatric Genetics [29] (burst strength 8.44). The longest burst duration is associated with Soria, V et al.‘s 2010 publication in Neuropsychopharmacology titled “Differential Association of Circadian Genes with Mood Disorders: CRY1 and NPAS2 are Associated with Unipolar Major Depression and CLOCK and VIP with Bipolar Disorder” [30] (burst strength 10.38), and their 2018 publication in Lancet Psychiatry titled “Association of disrupted circadian rhythmicity with mood disorders, subjective wellbeing, and cognitive function: a cross-sectional study of 91105 participants from the UK Biobank” authored by Lyall, LM et al. [31] (burst strength 10.73). These two articles had bursts lasting for six years, and notably, the latter still receives significant attention to this day. Additionally, the four articles at the bottom of the graph are still in their burst periods.