1. Introduction
Currently, the use of artificial intelligence (AI) has reached widespread popularity at an unprecedented rate [
1]. AI algorithms have emerged as potential tools in diverse areas of healthcare, including chronic disease management and clinical decision-making [
2,
3,
4]. AI is playing a prominent role in dementia research due to advancements in computing power, novel algorithms, and the availability of big data generated from medical health records and wearable devices [
5,
6,
7]. Recent studies in this domain have primarily focused on developing effective tools to accurately and quickly diagnose dementia, investigate the progression of dementia symptoms, and improve care and support for individuals affected by dementia [
8,
9].
The high number of continuously published articles and the rapid progress of AI research in the field of dementia has created significant challenges when it comes to staying up-to-date. Consequently, it becomes crucial to grasp the applications, significance, trends, and research hotspots within AI research in dementia. Systematic reviews, meta-analyses, and bibliometric analyses are employed to assist authors, clinicians, and policymakers in staying updated with emerging scientific outcomes [
10,
11,
12]. Among these methods, bibliometric analysis is particularly valuable for the quantitative assessment of scientific literature, enabling the identification of key themes and emerging trends within specific research topics [
13,
14,
15]. Bibliometric analyses always offer valuable insights by scrutinizing citations, co-citations, geographical distribution, and word frequency of literature.
The primary objective of this study was to provide a comprehensive overview of AI research in the field of dementia while also identifying future research directions that can benefit the general population, healthcare policymakers, and researchers. To address these goals, this study formulated the following research questions:
RQ1: What are the fundamental characteristics of the published articles? How many articles focusing on AI applications for dementia have been published to date?
RQ2: Who are the most productive authors/co-authors in these areas, and what are their countries of origin?
RQ3: Which journal has published the highest number of articles in this field? Which organizations have made significant contributions to this area of research?
RQ4: What are the most commonly used keywords associated with these publications?
By addressing these research questions, a comprehensive understanding of the current state of AI research in dementia can be obtained, enabling insights into the key contributors, research trends, and impactful literature in this domain.
4. Discussion
This bibliometric analysis aimed to present a comprehensive overview of research on AI-related research in the field of dementia. The analyses encompassed a total of 1094 publications spanning 27 years, from 1997 to the partial year of 2023. The key findings of this study are as follows:
(a) The research publications related to AI in dementia have exhibited a notable growth trend over time;
(b) Developed countries have emerged as the primary as primary drivers of AI research in dementia care. However, due to the increasing aging population, countries worldwide are actively engaged in dementia research;
(c) The majority of highly cited researchers in this area are affiliated with prestigious universities;
(d) Collaboration among universities in the USA demonstrated the highest level, followed by South Korean universities.
Additionally, this study identified the most prevalent research categories, popular keywords, and critical terms within the AI-related research on dementia. These findings underscore the importance of further exploration in this field to enhance AI-based research for managing dementia. By providing a comprehensive summary of the AI research trends in dementia, this work aims to serve as a guiding resource for future research endeavors, fostering advancements in this particular field of study.
The advent of AI research has presented exciting possibilities for the development of efficient and accessible tools that can aid in the early prediction of dementia [
21]. Previous evidence has demonstrated the potential of AI to assist physicians in conducting specific tests and investigations for more effective management of dementia [
22,
23,
24]. Our study findings show that the majority of research has focused on various aspects of dementia, including disease classification, diagnosis, prediction, segmentation, and early detection. Several studies have also highlighted the association between healthy aging and certain types of cognitive decline, such as processing speed, fluid reasoning, and episodic memory [
25,
26,
27]; however, it is worth noting that around 5–15% of individuals eventually develop dementia [
28]. Globally, over 50 million people are currently living with dementia, and the number is projected to rise due to the increasing aging population [
29]. Since there are no specific treatment options for dementia, the development of automated tools for the earliest possible detection is crucial in maximizing the impact of existing and traditional treatment approaches to delay pathological cognitive aging [
30].
The area of AI in dementia research has garnered global attention as dementia has emerged as a significant public health concern [
31]. While high-income countries have been the primary drivers of AI research in dementia, low- and middle-income countries are also increasingly focusing on this area. However, the output of research from low- and middle-income countries remains comparatively lower than that of high-income countries due to limited resources and technological advancements. In recent times, several developed countries have implemented AI healthcare policies that provide guidance on the development and regulation of AI in healthcare, such as the UK Code of Conduct for Data-Driven Health and Care Technology [
32]. It is crucial to encourage researchers from low- and middle-income countries to engage in AI research by offering funding, research opportunities, and access to tools. AI-based automated tools hold promise in improving health outcomes, particularly in low-income countries with limited healthcare resources [
33,
34,
35].
AI research in dementia is predominantly favored by prominent healthcare journals, as indicated by the output and citation counts. The majority of these studies tend to be specific to the field of neurological disorders. Nevertheless, AI-related research also attracts attention from computer science and multidisciplinary journals. It is important to note that open access journals and those with high impact factors tend to receive a greater number of citations [
36,
37]. This can be attributed to the perceived quality and reliability of research published in high-impact-factor journals, as well as the increased visibility of articles in open access journals. Over the past few decades, the occurrence and duration of citation bursts related to AI research on dementia have displayed variation. Certain keywords have consistently remained significant over an extended period, indicating the enduring focus of dementia research. As indicated by the keywords provided by the authors, convolutional neural networks, support vector machines, and random forest models have made a substantial impact on dementia research. These techniques have been widely used in dementia studies for classification, diagnosis, early prediction, and detection [
38,
39,
40]. The field of AI research in dementia has witnessed significant paradigm shifts over the years and continues to evolve. Dementia research exerts strong impacts on healthcare, and our study findings suggest that the impact of AI in dementia research is likely to persist in the coming years.
Over the past few decades, the occurrence and duration of citation bursts related to AI research on dementia have displayed variations. Certain keywords have consistently remained significant over an extended period, indicating the enduring focus of dementia research. As indicated by the keywords provided by the authors, convolutional neural networks, support vector machines, and random forest models have made a substantial impact on dementia research. These techniques have been widely utilized in dementia studies for classification, diagnosis, early prediction, and detection [
38,
39,
40]. The field of AI research in dementia has witnessed significant paradigm shifts over the years and continues to evolve. These research domains exert a strong influence on healthcare, and our study findings suggest that the impact of AI in dementia research is likely to persist in the coming years.
To the best of our knowledge, this study represents the first bibliometric analysis of AI-based research on dementia. However, it is important to acknowledge the limitations of this study. First, there might be a language bias as we focused solely on publications in English. Second, despite our efforts to capture all articles on AI-based research in dementia over the past 27 years, certain sources like gray literature and reports from databases other than WoS may have been omitted. The number of citations reported in this study is based on WoS records and may differ from those in Google Scholar. Additionally, more recent articles are likely to accumulate additional citations in the future, potentially resulting in the exclusion of high-quality studies from our analysis. Lastly, the data used for analysis was solely extracted from WoS, excluding databases such as Scopus (Elsevier) and PubMed. Although WoS includes high-quality journals, there is a possibility that some articles may have been missed.