Cognitive impairment screening in aging China: a narrative review of the Hong Kong Brief Cognitive Test and its future with artificial intelligence

Introduction

The world is undergoing a profound demographic shift, and nowhere is this more acute than in China (1). The number of individuals aged 65 and over reached 220 million in 2024, constituting 15.6% of the total population and signaling the nation’s entry into a deeply aged society (2). This trend brings a commensurate surge in age-related health conditions, most notably dementia (3-6). With an estimated 16.99 million dementia patients, China accounts for nearly 30% of the global total, representing a serious public health challenge (7). This imposes severe economic and emotional burdens, with the average annual cost per patient estimated at $19,000 and significant psychological distress reported by family caregivers (8,9).

Given that dementia is a progressive neurodegenerative condition with no known cure, the clinical and public health consensus is unequivocal: early detection and intervention are the only effective strategies to mitigate its devastating impact (10,11). Recognizing cognitive decline at its earlier stages, such as Mild Cognitive Impairment (MCI), allows for timely management, support, and planning (12,13). In response, China has elevated cognitive health to a national priority, progressively emphasizing early detection in national health plans. This culminated in strategic frameworks like the “National Action Plan for Addressing Dementia in the Elderly (2024–2030)” (14,15), which mandates cognitive screening for elderly community residents. The success of such large-scale initiatives hinges on the availability of a screening tool that is not only accurate but also rapid, easy to administer, and culturally appropriate.

For decades, paper-and-pencil tests have been the cornerstone of cognitive screening. However, the most widely used instruments, developed primarily in Western contexts, have significant limitations that create a knowledge and implementation gap for community-based screening in China (16,17).

The Mini-Mental State Examination (MMSE), while historically significant, suffers from several critical flaws. It is well-documented for its lack of sensitivity in detecting MCI, often yielding “normal” scores in individuals with genuine deficits (18-21). Furthermore, its performance is heavily influenced by education level, which can lead to the misclassification of individuals with lower educational attainment as impaired (22,23). While Chinese versions have been developed, studies show they possess only moderate reliability (test-retest reliability estimated at 0.78), and their utility is further complicated by commercial copyright claims (18,19).

The Montreal Cognitive Assessment (MoCA) was developed to address the MMSE’s insensitivity to MCI (24). However, its standard version is undermined by a strong Western cultural and linguistic bias (24,25). Tasks such as naming a lion, rhinoceros, or camel are often inappropriate for elderly Chinese individuals, testing cultural exposure rather than cognitive function (25). This has led to a proliferation of different Chinese versions [e.g., the Hong Kong MoCA (HK-MoCA), MoCA-Basic Chinese], each with its own cultural adaptations (21,25-27). While necessary, these adaptations have resulted in a lack of standardization, with different versions yielding different optimal cut-off scores. For instance, the validated HK-MoCA uses a cut-off of 21/22 for MCI detection, whereas other studies in mainland China have proposed different thresholds, creating clinical ambiguity (26,28). Even other brief tools recommended for community use, like the Ascertain Dementia 8 questionnaire and the Brief Community Screening Instrument for Dementia, have demonstrated insufficient sensitivity for early detection (29,30). This clear gap highlights the urgent need for a single, nationally standardized tool that is culturally tailored and sensitive to the earliest signs of decline.

Despite the availability of Western tools (MMSE, MoCA) and brief Chinese adaptations, there remains a lack of a single, nationally standardized, culturally tailored tool sensitive to early cognitive decline—this gap undermines China’s large-scale dementia screening efforts. This review aims to address this gap by synthesizing the evidence for the Hong Kong Brief Cognitive Test (HKBC) as a superior alternative and, in a novel contribution, proposing a detailed, evidence-based roadmap for its future as an artificial intelligence (AI)-powered tool to meet China’s large-scale screening needs. Through this work, we hope to promote the effective utilization of this highly promising, locally-developed tool. This endeavor intends to provide scientific reference and practical support for addressing the dementia challenges of China’s aging population and for the effective implementation of the National Dementia Action Plan [2024–2030]. We present this article in accordance with the Narrative Review reporting checklist (available at https://jmai.amegroups.com/article/view/10.21037/jmai-2025-207/rc).

The HKBC: a culturally-engineered solution

To address the documented failings of its predecessors, a team of experts led by Professor Chiu at the Chinese University of Hong Kong developed the HKBC in 2018 (19). As the first cognitive screening tool pioneered by Chinese scholars with independent intellectual property, the HKBC was engineered as a direct response to the aforementioned challenges.

The HKBC is a 30-point cognitive screening test with an average administration time of approximately 7 minutes, designed for older adults, particularly those with low educational levels. It assesses nine key cognitive domains through culturally relevant tasks tailored for Chinese populations. For example, orientation (5 points) is evaluated by asking for the current month, day of the week, season, region, and specific location (e.g., hospital or street name). Memory assessment includes immediate recall (1 point) and delayed recall (8 points), using a practical list-learning task with familiar items like “mantou” (Chinese steamed bread). Language and semantic memory (4 points) are tested through naming culturally familiar objects (e.g., a button or bicycle tire) and describing their functions, alongside a 1-minute verbal fluency test for animals (4 points). General knowledge (1 point) is assessed by asking for the name of the current national president, reflecting awareness of contemporary affairs. Executive function (2 points) is evaluated using the Luria “fist-edge-palm” sequence, while visuospatial abilities (3 points) are tested through the Clock Drawing Test (CDT), which includes clock-drawing, clock-setting (indicating 8:20), and clock-reading (reading 10:10). Recent memory (2 points) is assessed by recalling a recent news event. Each task is carefully designed to minimize the influence of education or literacy, with adaptations like providing pre-drawn clock faces for those unable to write numbers. This culturally sensitive design ensures the HKBC accurately measures cognitive ability rather than educational attainment, making it an effective and equitable screening tool for cognitive impairment in older Chinese adults (19).

Table 1 provides a systematic comparison of MMSE, MoCA, and HKBC, distilling data from numerous validation studies to highlight their relative strengths and weaknesses for large-scale application in China. This comparative analysis reveals that while each tool has merits, the HKBC’s combination of cultural appropriateness, efficiency, balanced psychometric properties, and—most importantly—the recent establishment of national norms makes it the most suitable candidate for China’s national dementia screening program. The lack of standardization for the MMSE and MoCA is not a minor technicality; it is a fundamental barrier. A national public health strategy requires a single, unified metric to compare prevalence rates across provinces, monitor the effectiveness of interventions, and pool data for large-scale research. The HKBC’s national standardization is therefore not merely an advantage; it is a prerequisite for its use as the backbone of the National Action Plan.

Methods

A literature search was conducted in two major databases: PubMed (primarily for English-language publications) and Wanfang Data (for Chinese-language publications). The search covered publications from the inception of the HKBC in January 2018 to September 2025. This narrative and critical review considered studies of any design (e.g., validation, cross-sectional, cohort, review) published in either English or Chinese, which focused on the development, validation, or application of the HKBC. Ultimately, nine articles were selected for inclusion to inform this review (19,33,35-41). The search strategy is summarized in Table 2.

Findings from the literature

Validation, standardization, and application of HKBC

Validation studies have consistently shown the HKBC’s superior diagnostic accuracy. The initial study in a Cantonese-speaking population in Hong Kong found that a cut-off of 21/22 yielded excellent results for detecting major neurocognitive disorder (sensitivity =0.88, specificity =0.84) and mild neurocognitive disorder (sensitivity =0.89, specificity =0.81) (19). A subsequent validation in mainland China confirmed its robust performance in a Mandarin-speaking cohort, with optimal cut-offs demonstrating high accuracy for distinguishing amnesic MCI (aMCI) from healthy controls (sensitivity =0.79, specificity =0.85) and mild Alzheimer’s disease (AD) from aMCI (sensitivity =0.85, specificity =0.96) (33).

A landmark contribution to its national applicability is the recent establishment of national normative data (35). This large-scale study, conducted across six regions in China, confirmed the influence of age and education on HKBC scores and provided standardized T-scores for clinical use. This development is a critical milestone, transforming the HKBC from a promising instrument into a nationally standardized tool that allows for more precise, demographically-adjusted interpretation of cognitive performance across the country.

Beyond validation studies, the HKBC is increasingly being adopted as a primary cognitive outcome measure in diverse clinical and epidemiological research across mainland China. It has been used to assess older adults’ cognitive function in studies investigating risk factors like physical inactivity and depression in Beijing and the relationship between excessive daytime sleepiness and cognition in Shangrao (36,39). Its growing prominence is further evidenced by its inclusion in academic discourse on digital test development (40). Significantly, its role as a trusted benchmark is demonstrated in recent digital test development, where it has served both as a criterion validity measure for an electronic version of the Hopkins Verbal Learning Test-Revised (38) and as the core reference standard for developing a new machine learning-based digital screening tool, chosen precisely because it is considered to have the “highest validity and reliability in identifying the earliest stages of subtle cognitive decline” (40). This expanding body of evidence is echoed in reviews; for instance, a major scoping review not only listed the HKBC alongside the MoCA and MMSE as one of the “ideal screening tools with versatile performance” but also specifically concluded that among these top-tier tools, the HKBC shows the “highest validity and reliability” for detecting the initial signs of cognitive change (37). It also highlights the excellent inter-rater reliability of HKBC, further demonstrating the tool’s consistent high performance (37). Another review listed HKBC alongside audiovisual MoCA and German Quick MCI screen as “Class A recommendations” out of a pool of 30 different versions of screening instruments for the test’s high content validity, internal consistency, and criterion validity (41). These reviews cement HKBC’s status as a leading instrument for cognitive assessment in Chinese older adults.

Discussion

Limitations of the paper-and-pencil HKBC

Despite the HKBC’s clear superiority in content and cultural design, its administration via a conventional paper-and-pencil format imposes fundamental limitations that hinder its potential for mass screening. The challenges are not merely superficial but are rooted in systemic inefficiencies that create a mismatch between an excellent tool and an unscalable method.

The original authors of HKBC had noted the issue of subjective scoring, which can introduce rater variability and requires specialized training (19). The root cause of this problem, however, extends beyond individual judgment. It reflects the systemic inefficiency of relying on a manual testing paradigm for a national-scale problem. To implement the National Action Plan using a paper test, China would need to train, deploy, and continuously quality-control a massive workforce of technicians across a vast and diverse country. This represents a formidable logistical and economic bottleneck for the Ministry of Health, making true scalability unfeasible.

Furthermore, the loss of dynamic, process-related data is a significant limitation of the traditional HKBC. This limitation has profound implications for the test’s clinical utility. A single, static score from a paper test is a crude snapshot in time. Neurodegenerative diseases, however, are a dynamic process that unfolds over years. By failing to capture process-related data—such as the hesitation times, drawing kinematics, and stroke order in the CDT—the paper test loses its ability to provide crucial prognostic information. These digital biomarkers can help predict an individual’s rate of decline, which is critical for patient management, clinical trial stratification, and public health resource planning. The loss of this fine-grained information is a significant missed opportunity for more precise, personalized assessment.

Ultimately, the core limitation is a fundamental mismatch between technology and ambition. A paper-and-pencil test, regardless of its content quality, is an analog tool ill-suited to the digital-era task of screening over 220 million people. Its scalability is inherently constrained. This reframes the central argument: the goal should not simply be to digitize the HKBC for convenience, but to fundamentally re-engineer it using AI to overcome the inherent limitations of the manual testing paradigm itself. This technological leap is necessary to align the tool’s potential with the nation’s public health goals.

The next frontier: an AI-powered, multimodal HKBC

The fusion of traditional neuropsychological testing with modern AI and sensor technology offers a transformative solution, moving assessment from simple computerization to true digital intelligence (42,43). The future of the HKBC lies in its evolution from a static paper form into a dynamic, digitalized, and intelligent screening system based on a multimodal approach.

Existing applications and evidence of AI in cognitive impairment screening

The proposal for an AI-driven HKBC is grounded in a robust and rapidly growing body of scientific evidence demonstrating the power of AI to enhance cognitive assessment. AI and machine learning are increasingly used to optimize diagnostic procedures, predict cognitive decline, and outperform traditional tests by analyzing complex, high-dimensional data that are imperceptible to human evaluators (44). Specific applications provide a direct proof-of-concept for the proposed multimodal approach.

In the domain of visuospatial and motor analysis, studies on the digital CDT (dCDT) have shown that machine learning algorithms can analyze drawing dynamics—such as pen speed, pressure, stroke order, and hesitation times—to classify non-MCI, various MCI subtypes, and AD with accuracies exceeding 83% (45,46). Deep learning has also been successfully applied to automate the scoring of the Pentagon Drawing Test, making it more sensitive for detecting cognitive impairments in Parkinson’s disease (47). This evidence provides strong support for the digital capture and AI-based analysis of the HKBC’s CDT and Luria “fist-edge-palm” tasks.

Similarly, speech and language analysis has emerged as a powerful, non-invasive modality for detecting cognitive decline. AI models can analyze both acoustic and linguistic features of speech to identify patterns associated with dementia. Acoustic approaches analyze prosodic features like speech rate, pause duration, and vocal jitter, with traditional classifiers achieving accuracies of 70–80% (48). Linguistic approaches use natural language processing (NLP) to analyze semantic content, word choice, and syntactic complexity from transcribed speech, with transformer-based models like Bidirectional Encoder Representations from Transformers showing high performance (49). When these modalities are combined, AI systems can achieve impressive results. Recent studies have demonstrated that AI-powered speech analysis can predict the progression from MCI to AD with over 78% accuracy, and multimodal models integrating both acoustic and linguistic features often achieve classification accuracies above 85% (50,51). This body of work provides direct validation for applying AI to analyze responses from the HKBC’s verbal fluency, recall, and naming tasks.

A proposed framework for an ai-powered, multimodal HKBC

Building on this evidence, a digital HKBC could be developed to collect rich, synergistic data from 3 key modalities on a standard tablet computer:

• ⅠSemantics & speech: verbal responses to tasks like memory recall and verbal fluency would be recorded. AI algorithms would perform parallel analyses: NLP models would assess semantic content and complexity, while other models would analyze paralinguistic features (vocal biomarkers) like speech rate, pause duration, and pitch.
• ⅡGesture & posture: the performance of the Luria “fist-edge-palm” test would be captured by the device’s camera. Computer vision models could then analyze the fluidity, accuracy, and speed of the sequence, providing an objective measure of executive function.
• ⅢImage & drawing process: the CDT would be performed on the tablet with a stylus, capturing not only the final image but the entire drawing process, including pen pressure, speed, stroke order, and hesitation times (pen kinematics).

Deep learning models, trained on this multimodal data from clinically diagnosed patients and healthy controls, could then generate an automated, instantaneous, and highly accurate classification of a user’s cognitive status. This would eliminate scorer subjectivity, enhance precision by leveraging digital biomarkers, and enable massive scalability through deployment on common electronic devices.

Potential obstacles and mitigation strategies

While the vision for an AI-powered HKBC is compelling, its realization faces significant challenges that extend beyond the superficial issues of user familiarity. These obstacles are best understood as an interconnected “implementation trilemma” involving technical feasibility, algorithmic bias, and clinical translation, where attempts to solve one problem can often exacerbate another.

Technical feasibility

The development of a robust medical AI tool is a formidable technical undertaking. The primary bottleneck is the need for a massive, high-quality, and well-annotated training dataset (52). This dataset must be multimodal—containing synchronized audio, video, and kinematic data—and each case must be linked to a gold-standard clinical diagnosis (e.g., from neuropsychological testing, cerebrospinal fluid biomarkers, or positron emission tomography imaging). Assembling such a dataset is a costly and logistically complex endeavor. A second major challenge is infrastructure integration. The tool must be able to operate within China’s diverse and often fragmented clinical IT landscape, which includes numerous legacy systems (53). A lack of standards for data sharing and interoperability makes seamless integration difficult (54). Finally, the “black box” nature of many deep learning models presents a validation challenge. Proving the algorithm’s robustness, reliability, and generalizability across different hardware, operating systems, and patient populations requires extensive, multi-site prospective clinical validation, a process that is both time-consuming and expensive (52).

In addition, digitalization risks excluding the most vulnerable elderly: those with low digital familiarity, health-related issues like impaired vision, or motor control difficulties. This is particularly challenging for tasks like drawing on a slippery glass screen, an issue more pronounced for individuals with motor impairments, such as stroke survivors. Mitigation requires user-centered design (e.g., large fonts) and practical solutions like applying paper-like screen protectors, which add friction to the screen to improve stylus control and mimic the feel of paper.

Algorithmic bias

A critical ethical and scientific challenge is algorithmic bias. AI models learn from and can amplify biases present in their training data (55). If the training data is not representative of the target population, the resulting tool will perform inequitably. In the Chinese context, the critical axes of potential bias are not primarily racial, as often discussed in Western literature, but rather education level, urban versus rural residency, and regional/dialectical differences (55). For example, an AI model trained predominantly on speech data from educated, urban, Mandarin-speaking individuals may systematically misclassify older adults from rural areas who speak a local dialect and have less formal education. This would not only be unjust but would also undermine the tool’s clinical utility in the very populations that are often underserved. Mitigating this risk requires a deliberate and resource-intensive national effort to collect a demographically balanced and representative training dataset from across China’s diverse regions.

Clinical translation thresholds

Even a technically perfect and unbiased tool can fail if it cannot cross the threshold into clinical practice. This involves three key hurdles. First, AI diagnostic tools are regulated as medical devices and will require approval from China’s National Medical Products Administration, a process that involves navigating a complex and evolving regulatory framework (56). Second, the tool must integrate seamlessly into existing clinical workflows in primary care clinics, community health centers, and hospitals (57). A tool that is cumbersome or exists outside of the main electronic health record system will not be adopted, a lesson learned from the failure of early clinical decision support systems (54). Third, and perhaps most importantly, the tool must earn the trust of both clinicians and patients (53). Clinicians will have valid concerns about medicolegal accountability, the opacity of the algorithm’s decision-making process, and the potential for technology to supplant clinical judgment (52). Building this trust through transparency, education, and robust evidence of real-world utility is a critical, non-technical barrier to adoption.

Addressing this implementation trilemma requires a coordinated, systems-level strategy that simultaneously tackles data collection, equitable design, and clinical integration. A sequential approach is destined to fail.

A roadmap for responsible clinical implementation

A structured, phased approach is essential for responsible clinical translation (58).

❖Phase 1: pre-implementation. This involves defining the clinical use case, preparing diverse datasets, and conducting retrospective validation using local data to ensure performance and identify dataset shift issues.

❖Phase 2: peri-implementation. This phase focuses on user-centered design, technical integration with Electronic Health Records, and conducting prospective pilot studies in controlled clinical settings.

❖Phase 3: post-implementation. After deployment, a system for continuous performance monitoring is required to detect model drift and evaluate the tool’s clinical impact and fairness across demographic subgroups.

Limitations of this review

This study is a narrative review and, as such, has inherent limitations. The literature search was systematic but not exhaustive, meaning some relevant studies may have been omitted, which could introduce selection bias. Furthermore, as a narrative review, we did not perform a formal quality assessment of the included studies, which likely vary in their methodological rigor. The conclusions drawn are therefore based on our qualitative synthesis and interpretation of the available evidence.

Conclusions

This review makes a novel contribution by proposing a detailed, evidence-based roadmap for HKBC’s evolution into an AI-powered multimodal tool, directly aligning with China’s National Dementia Action Plan [2024–2030]. The HKBC represents a significant advancement in culturally sensitive screening for cognitive impairment in China. However, to meet the immense challenge of an aging population, the next leap must be a technological one. An AI-powered, digitalized HKBC offers a clear path forward. By harnessing multimodal data, such a tool could overcome the limitations of traditional methods, providing a rapid, objective, and highly accurate means of identifying individuals at the earliest stages of cognitive decline.

This innovation is a necessary evolution. Its development and deployment, however, must be guided by an actionable framework that prioritizes clinical utility, user accessibility, and ethical fairness. Addressing the challenges of usability, technical validation, and algorithmic bias is paramount. Only through such a responsible and comprehensive approach can an AI-powered HKBC fulfill its potential to safeguard the cognitive health of all of China’s aging citizens and realize the ambitious goals of the national dementia strategy.

None.

Footnote

Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://jmai.amegroups.com/article/view/10.21037/jmai-2025-207/rc

Peer Review File: Available at https://jmai.amegroups.com/article/view/10.21037/jmai-2025-207/prf

Funding: This study was supported by the Wuhan Municipal Health Commission and Bureau of Science and Technology Innovation of Wuhan Municipality (grant No. WX23A99), National Natural Science Foundation of China (grant No. 71774060), and the Young Top Talent Program in Public Health from Health Commission of Hubei Province (grant No. EWEITONG[2021]74, PI: B-LZ).

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (https://jmai.amegroups.com/article/view/10.21037/jmai-2025-207/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Yuan MD, Liu JF, Zhong BL. Prevalence of prolonged grief disorder and its symptoms among bereaved individuals in China: a systematic review and meta-analysis. Gen Psychiatr 2024;37:e101216. [Crossref] [PubMed]
2. National Bureau of Statistics. Statistical Bulletin of the People’s Republic of China on the 2024 National Economic and Social Development 2025. Available from: https://www.stats.gov.cn/sj/zxfb/202502/t20250228_1958817.html
3. Zhang HG, Fan F, Zhong BL, et al. Relationship between left-behind status and cognitive function in older Chinese adults: a prospective 3-year cohort study. Gen Psychiatr 2023;36:e101054. [Crossref] [PubMed]
4. Wang ZQ, Fei L, Xu YM, et al. Prevalence and correlates of suspected dementia in older adults receiving primary healthcare in Wuhan, China: A multicenter cross-sectional survey. Front Public Health 2022;10:1032118. [Crossref] [PubMed]
5. Ren Z, Nie L, Du Y, et al. Intertwined depressive and cognitive trajectories and the risk of dementia and death in older adults: a competing risk analysis. Gen Psychiatr 2024;37:e101156. [Crossref] [PubMed]
6. Taylor JL. Exercise and the Brain in Cardiovascular Disease: A Narrative Review. Heart and Mind 2023;7:5-12.
7. Zhi N, Ren R, Qi J, et al. The China Alzheimer Report 2025. Gen Psychiatr 2025;38:e102020. [Crossref] [PubMed]
8. Jia J, Wei C, Chen S, et al. The cost of Alzheimer's disease in China and re-estimation of costs worldwide. Alzheimers Dement 2018;14:483-91. [Crossref] [PubMed]
9. Alzheimer’s Disease Branch of Chinese Aging Well Association. 2019 Survey Report on Family Survival Status of Alzheimer’s Disease Patients in China. Beijing: Chinese Aging Well Association; 2020.
10. Huang L, Li Q, Lu Y, et al. Consensus on rapid screening for prodromal Alzheimer's disease in China. Gen Psychiatr 2024;37:e101310. [Crossref] [PubMed]
11. Agarwal R, Tully PJ, Mahajan R. Cognitive Function in Atrial Fibrillation: A Narrative Review of Evidence and Mechanisms. Heart and Mind 2024;8:100-10.
12. Pavel A, Paun R, Matei V, et al. Quality of Life in People With Subjective Cognitive Decline. Alpha Psychiatry 2023;24:60-4. [Crossref] [PubMed]
13. Nie J, Yang Y, Gao Y, et al. Newly self-administered two-step tool for screening cognitive function in an ageing Chinese population: an exploratory cross-sectional study. Gen Psychiatr 2023;36:e100837. [Crossref] [PubMed]
14. National Health Commission of China. Joint Announcement of the National Action Plan for Addressing Dementia in the Elderly (2024-2030) by the National Health Commission and 14 Other National Ministries of China Beijing, China: National Health Commission of China; 2025. Available online: https://www.nhc.gov.cn/lljks/c100158/202501/4b4476877f21467fbfcea71bc90d8313.shtml
15. Wang J, Zeng HY, Gao JX, et al. Enhancing Dementia Awareness and Screening, and Reducing Stigmatizing Attitudes towards Dementia in Urban China: The Role of Opinion Leader Intervention in Community-Dwelling Older Adults. Alpha Psychiatry 2025;26:38857. [Crossref] [PubMed]
16. Kim HG, Kim Y, Edmonds EC, et al. Comparison of Cognitive Impairment Diagnosis Criteria in Clinical Settings: Conventional vs. Neuropsychological. Alpha Psychiatry 2024;25:212-9. [Crossref] [PubMed]
17. Zhong BL, Chiu HF. Ageism, dementia, and culture. Int Psychogeriatr 2023;35:1-2. [Crossref] [PubMed]
18. Guan Q, Sheng F, Yu Z, et al. Cognitive assessments by using MMSE and MoCA: a analysis on the differences. Chinese Journal of Integrative Medicine on Cardio-cerebrovascular Disease 2021;19:1577-80.
19. Chiu HFK, Zhong BL, Leung T, et al. Development and validation of a new cognitive screening test: The Hong Kong Brief Cognitive Test (HKBC) Int J Geriatr Psychiatry 2018;33:994-9. [Crossref] [PubMed]
20. Zhong BL, Chen SL, Tu X, et al. Loneliness and Cognitive Function in Older Adults: Findings From the Chinese Longitudinal Healthy Longevity Survey. J Gerontol B Psychol Sci Soc Sci 2017;72:120-8. [Crossref] [PubMed]
21. Pang T, Xia B, Zhao X, et al. Cost-benefit and discriminant validity of a stepwise dementia case-finding approach in an Asian older adult community. Gen Psychiatr 2023;36:e101049. [Crossref] [PubMed]
22. Siddiqui M, Nyahoda T, Traber C, et al. Screening for Cognitive Impairment in Primary Care: Rationale and Tools. Mo Med 2023;120:431-9. [PubMed]
23. Ng TP, Niti M, Chiam PC, et al. Ethnic and educational differences in cognitive test performance on mini-mental state examination in Asians. Am J Geriatr Psychiatry 2007;15:130-9. [Crossref] [PubMed]
24. Hu JB, Zhou WH, Hu SH, et al. Cross-cultural difference and validation of the Chinese version of Montreal Cognitive Assessment in older adults residing in Eastern China: preliminary findings. Arch Gerontol Geriatr 2013;56:38-43. [Crossref] [PubMed]
25. Chen J, Ye Z, Yuan M, et al. Application and progress of Montreal cognitive assessment in screening mild cognitive impairment. Chinese Journal of Psychiatry 2017;50:386-9.
26. Wong A, Yiu S, Nasreddine Z, et al. Validity and reliability of two alternate versions of the Montreal Cognitive Assessment (Hong Kong version) for screening of Mild Neurocognitive Disorder. PLoS One 2018;13:e0196344. [Crossref] [PubMed]
27. Tan JP, Wang X, Zhang S, et al. Accuracy of the Short-Form Montreal Cognitive Assessment Chinese Versions. Front Aging Neurosci 2021;13:687824. [Crossref] [PubMed]
28. Yeung PY, Wong LL, Chan CC, et al. A validation study of the Hong Kong version of Montreal Cognitive Assessment (HK-MoCA) in Chinese older adults in Hong Kong. Hong Kong Med J 2014;20:504-10. [Crossref] [PubMed]
29. Li X. Application study of the Brief Community Screening Instrument for Dementia on elderly in Hubei community. Wuhan: Huazhong University of Science & Technology; 2012.
30. Li T, Wang HL, Yang YH, et al. [The reliability and validity of Chinese version of AD8]. Zhonghua Nei Ke Za Zhi 2012;51:777-80. [PubMed]
31. Li H, Peng A, Lai W, et al. Impacts of education level on Montreal Cognitive Assessment and saccades in community residents from Western China. Clin Neurophysiol 2024;161:27-39. [Crossref] [PubMed]
32. Jia X, Wang Z, Huang F, et al. A comparison of the Mini-Mental State Examination (MMSE) with the Montreal Cognitive Assessment (MoCA) for mild cognitive impairment screening in Chinese middle-aged and older population: a cross-sectional study. BMC Psychiatry 2021;21:485. [Crossref] [PubMed]
33. Sun W, Wu Q, Chen H, et al. A Validation Study of the Hong Kong Brief Cognitive Test for Screening Patients with Mild Cognitive Impairment and Alzheimer's Disease. J Alzheimers Dis 2022;88:1523-32. [Crossref] [PubMed]
34. Tsai JC, Chen CW, Chu H, et al. Comparing the Sensitivity, Specificity, and Predictive Values of the Montreal Cognitive Assessment and Mini-Mental State Examination When Screening People for Mild Cognitive Impairment and Dementia in Chinese Population. Arch Psychiatr Nurs 2016;30:486-91. [Crossref] [PubMed]
35. Zheng Y, Zhu J, Li N, et al. Norms of Hong Kong Brief Cognitive Test and the Utility for Screening Cognitive Impairment Among Chinese Older Adults. Int J Geriatr Psychiatry 2025;40:e70056. [Crossref] [PubMed]
36. Lin Y, Wang H, Tian Y, et al. [Factors influencing cognitive function among the older adults in Beijing]. Beijing Da Xue Xue Bao Yi Xue Ban 2024;56:456-61. [Crossref] [PubMed]
37. Chun CT, Seward K, Patterson A, et al. Evaluation of Available Cognitive Tools Used to Measure Mild Cognitive Decline: A Scoping Review. Nutrients 2021;13:3974. [Crossref] [PubMed]
38. Jiang L, Xu M, Xia S, et al. Reliability and validity of the electronic version of the Hopkins verbal learning test-revised in middle-aged and elderly Chinese people. Front Aging Neurosci 2023;15:1124731. [Crossref] [PubMed]
39. Wu J, Wu Z, Xie C, et al. A high propensity for excessive daytime sleepiness independent of lifestyle is associated with cognitive performance in community-dwelling older adults. Front Psychiatry 2023;14:1190353. [Crossref] [PubMed]
40. Gu D, Lv X, Shi C, et al. A Stable and Scalable Digital Composite Neurocognitive Test for Early Dementia Screening Based on Machine Learning: Model Development and Validation Study. J Med Internet Res 2023;25:e49147. [Crossref] [PubMed]
41. Wen S, Cheng D, Zhao N, et al. Psychometric properties of screening tools for mild cognitive impairment in older adults based on COSMIN guidelines: a systematic review. BMC Geriatr 2025;25:401. [Crossref] [PubMed]
42. Veneziani I, Marra A, Formica C, et al. Applications of Artificial Intelligence in the Neuropsychological Assessment of Dementia: A Systematic Review. J Pers Med 2024;14:113. [Crossref] [PubMed]
43. Um Din N, Maronnat F, Zolnowski-Kolp V, et al. Diagnosis accuracy of touchscreen-based testings for major neurocognitive disorders: a systematic review and meta-analysis. Age Ageing 2025;54:afaf204. [Crossref] [PubMed]
44. Díaz-Guerra DD, Hernández-Lugo MC, Broche-Pérez Y, et al. AI-assisted neurocognitive assessment protocol for older adults with psychiatric disorders. Front Psychiatry 2024;15:1516065. [Crossref] [PubMed]
45. Ding Z, Lee TL, Chan AS. Digital Cognitive Biomarker for Mild Cognitive Impairments and Dementia: A Systematic Review. J Clin Med 2022;11:4191. [Crossref] [PubMed]
46. Binaco R, Calzaretto N, Epifano J, et al. Machine Learning Analysis of Digital Clock Drawing Test Performance for Differential Classification of Mild Cognitive Impairment Subtypes Versus Alzheimer's Disease. J Int Neuropsychol Soc 2020;26:690-700. [Crossref] [PubMed]
47. Sirilertmekasakul C, Rattanawong W, Gongvatana A, et al. The current state of artificial intelligence-augmented digitized neurocognitive screening test. Front Hum Neurosci 2023;17:1133632. [Crossref] [PubMed]
48. Wang C, Tachimori H, Yamaguchi H, et al. A multimodal deep learning approach for the prediction of cognitive decline and its effectiveness in clinical trials for Alzheimer's disease. Transl Psychiatry 2024;14:105. [Crossref] [PubMed]
49. Shakeri A, Farmanbar M. Natural language processing in Alzheimer's disease research: Systematic review of methods, data, and efficacy. Alzheimers Dement (Amst) 2025;17:e70082. [Crossref] [PubMed]
50. Ding K, Chetty M, Noori Hoshyar A, et al. Speech based detection of Alzheimer’s disease: a survey of AI techniques, datasets and challenges. Artif Intell Rev 2024;57:325.
51. Amini S, Hao B, Yang J, et al. Prediction of Alzheimer's disease progression within 6 years using speech: A novel approach leveraging language models. Alzheimers Dement 2024;20:5262-70. [Crossref] [PubMed]
52. Kotter E, Ranschaert E. Challenges and solutions for introducing artificial intelligence (AI) in daily clinical workflow. Eur Radiol 2021;31:5-7. [Crossref] [PubMed]
53. Poon EG, Lemak CH, Rojas JC, et al. Adoption of artificial intelligence in healthcare: survey of health system priorities, successes, and challenges. J Am Med Inform Assoc 2025;32:1093-100. [Crossref] [PubMed]
54. Adler-Milstein J, Aggarwal N, Ahmed M, et al. Meeting the Moment: Addressing Barriers and Facilitating Clinical Adoption of Artificial Intelligence in Medical Diagnosis. NAM Perspect 2022;2022: [Crossref] [PubMed]
55. Aquino YSJ, Carter SM, Houssami N, et al. Practical, epistemic and normative implications of algorithmic bias in healthcare artificial intelligence: a qualitative study of multidisciplinary expert perspectives. J Med Ethics 2025;51:420-8. [Crossref] [PubMed]
56. Nair M, Svedberg P, Larsson I, et al. A comprehensive overview of barriers and strategies for AI implementation in healthcare: Mixed-method design. PLoS One 2024;19:e0305949. [Crossref] [PubMed]
57. Cabral BP, Braga LAM, Conte Filho CG, et al. Future Use of AI in Diagnostic Medicine: 2-Wave Cross-Sectional Survey Study. J Med Internet Res 2025;27:e53892. [Crossref] [PubMed]
58. Hassan N, Slight R, Morgan G, et al. Road map for clinicians to develop and evaluate AI predictive models to inform clinical decision-making. BMJ Health Care Inform 2023;30:e100784. [Crossref] [PubMed]