ABSTRACT
Introduction
Writing can provide information on both the written product and the production process, but writing‐related indicators in mild cognitive impairment (MCI) and early Alzheimer's disease (AD) have not been systematically synthesized within a unified framework. This study sought to map writing‐related research in MCI and early AD across two complementary dimensions, namely motor process and content/orthographic indicators, and identify evidence gaps that limit clinical translation.
Methods
Following JBI methodological guidance and PRISMA Extension for Scoping Reviews reporting, we searched PubMed, Web of Science, CiNii, and Igaku Chuo Zasshi (Ichushi‐Web) from inception to October 1, 2025. Eligible studies examined writing characteristics in MCI and/or AD, including early‐stage and broader AD samples relevant to early cognitive decline. Extracted indicators were coded as motor process, content/orthographic, or integrated (both domains analyzed within one framework).
Results
From 3355 records, 24 studies were included. The study context groups were MCI (early) (n = 11), MCI (mixed) (n = 1), AD (early) (n = 6), and AD (cross‐stage) (n = 6). Eight studies primarily used motor process indicators, 14 studies used content/orthographic indicators, and two studies used integrated indicators. Digital measurement studies more frequently emphasized motor process outcomes, whereas paper‐based studies more commonly emphasized content/orthographic outcomes. By study context, motor process indicators were relatively more frequent in MCI (early), whereas content/orthographic indicators predominated in AD (early).
Conclusion
Writing research in MCI and early AD shows context‐dependent outcome selection with limited integration of motor and content/orthographic dimensions. Progress toward clinical use will likely require standardized tasks, shared operational definitions, core indicator sets spanning both domains, and externally validated integrated frameworks to support clinical application.
Keywords: Alzheimer disease, mild cognitive impairment, scoping review, writing
This scoping review identified 24 studies examining writing in mild cognitive impairment and early Alzheimer's disease. Digital studies more often emphasized motor process indicators, whereas paper‐based studies more commonly examined content/orthographic indicators; only two studies integrated both domains, underscoring the need for standardized, multidimensional assessment frameworks.

1. Introduction
Reproducible approaches for detecting early cognitive change and characterizing dementia‐related risk are urgently needed for clinical and research purposes. The global prevalence of dementia continues to rise, increasing the need for appropriate risk assessment during the earliest phases of cognitive decline [1]. Mild cognitive impairment (MCI), widely considered a prodromal stage of dementia, displays substantial interindividual variability concerning both clinical outcomes and progression rates [2, 3]. Consequently, it remains difficult to predict future progression to dementia based on clinical symptoms alone. Although international diagnostic frameworks presume that MCI reflects AD pathology, the diagnosis of MCI relies heavily on specialized assessments and adjunctive examinations [4, 5, 6]. In addition, the definitions and stage boundaries of MCI have not been consistent across studies, and such heterogeneity in study populations complicates the comparison and integration of findings [2].
Neuropsychological assessment has been widely used in clinical practice and research to evaluate cognitive status and dementia‐related risk in individuals with MCI. Brief cognitive screening instruments, particularly the Mini‐Mental State Examination (MMSE), are widely used for initial evaluation and follow‐up [7, 8]. Although these instruments facilitate efficient assessment of global cognitive status, their sensitivity might be insufficient in the earliest stages of decline, during which impairment is subtle, and total scores alone might fail to capture subtle changes [9]. Moreover, the MMSE has limited ability to illuminate the qualitative features of errors and the processes underlying task performance [7, 8]. Therefore, complementary measures that extend beyond total score‐based cognitive assessment are urgently needed to more comprehensively detect early cognitive changes.
Writing is a common daily activity and a potentially useful adjunctive marker for detecting early changes in MCI and early AD because a single task can provide information on both the written product and the production process. Previous studies broadly focused on motor process indicators or content and orthographic indicators [10, 11]. In addition, some studies examined the neural basis of writing impairment, suggesting that early changes in writing might be associated with alterations in brain function [12]. However, studies integrating the written product and the production process within a single analytical framework remain limited. Research on writing in the early stages of MCI remains fragmented because of variation in disease‐stage definitions and indicator frameworks [2, 4]. Furthermore, variation in task design, media, recording methods, and indicators has made comparison across studies more difficult. Therefore, a systematic synthesis of writing research in MCI and early AD is needed to clarify research trends and identify knowledge gaps from the dual perspectives of process and product.
Thus, this scoping review aimed to organize writing studies in MCI and early AD by diagnostic label and reported stage range and provide a structured overview of the evidence. This review first organized writing tasks and the media and recording methods used, and then writing indicators were classified into motor process indicators, content/orthographic indicators, and integrated indicators. In this review, writing included both production process features (motor and temporal characteristics) and product features (content and orthographic characteristics) derived from paper‐based or digital tasks. By presenting the distribution and characteristics of different writing‐based indicators, this review aimed to visualize research imbalances and knowledge gaps and provide a foundation for future efforts toward the development, standardization, and validation of integrated writing‐based indicators.
2. Methods
2.1. Design
The review was conducted in accordance with the JBI methodological guidance for scoping reviews [13] and reported following the PRISMA Extension for Scoping Reviews (PRISMA‐ScR) [14]. The protocol was registered with the University Hospital Medical Information Network (UMIN000059071).
2.2. Information Sources and Search Strategy
The final search was conducted on October 1, 2025, using PubMed, Web of Science, CiNii, and Igaku Chuo Zasshi (Ichushi‐Web). Searches covered records published from database inception to the final search date. The main search concepts were “MCI” and “writing.” The full search strategy for all databases is provided in Appendix S1. We limited the PubMed search to English‐language records. No language restrictions were applied to Web of Science Core Collection, CiNii Research, or Ichushi‐Web searches.
2.3. Eligibility Criteria
Eligibility was defined according to the PCC framework. The participant domain included individuals with MCI and/or AD, including early‐stage and broader AD classifications relevant to writing characteristics in MCI or early AD. Where applicable, studies also included cognitively healthy older adults as comparison groups. The concept domain comprised writing assessment based on product measures, linguistic or syntactic features, motor process measures, or prediction scores. The context included descriptive profiling, between‐group comparison, diagnostic support, screening, and detection of early changes. We excluded studies on children, studies of developmental dyslexia, studies of writing disorders attributable to primary language disorders (e.g., aphasic agraphia), studies without a writing task or writing analysis, and nonhuman studies.
Because this review aimed to map the landscape of writing‐related indicators in early cognitive decline, eligibility was not restricted by study design. We included peer‐reviewed original articles as well as conference abstracts and similar empirical reports, including single‐case reports, provided that they met the predefined PCC criteria and contained sufficient writing‐specific findings for data extraction and indicator coding. When both an abstract and a full article described the same study, the full article was retained. Reports lacking sufficient accessible information for data extraction and indicator coding were excluded.
2.4. Study Selection and Data Extraction
Rayyan (https://www.rayyan.ai/) was used for reference management and screening. Two reviewers (A.S. and T.K.) independently screened titles and abstracts according to the predefined eligibility criteria. Records that could not be clearly excluded at this stage were assessed in full text. The same two reviewers independently conducted the full‐text review. Disagreements were resolved through discussion, and a third author (S.T.) was consulted when necessary.
Two reviewers (A.S. and T.K.) independently extracted data on participant groups, study design, analytical approach, diagnostic and severity measures, writing tasks, administration medium, and writing indicators. Data were managed using Microsoft Excel (Microsoft, Redmond, WA, USA). In this review, writing was operationally defined as including both production process and product features derived from paper‐based or digital tasks.
2.5. Coding
Extracted writing indicators were coded into three predefined categories: motor process indicators (e.g., initiation latency, total duration, writing speed, pauses/pen‐up time, pen pressure); content/orthographic indicators (e.g., number of characters, information content, error types, syntactic measures); and integrated indicators (recorded from studies that analyzed motor process and content/orthographic indicators within the same analytical framework).
Studies were classified as integrated only when integration was explicit, such as that through simultaneous inclusion in the same multivariable model, the use of a composite score, or classification based on integrated feature sets. Studies that measured both domains but analyzed them separately were classified according to the primary outcome reported by the authors. Because the description of MCI and AD severity varied across studies, we did not reclassify diagnoses using a common severity scale. Instead, studies were coded descriptively along two axes: diagnostic label (MCI‐ or AD‐labeled) and stage range (early‐stage restricted, mixed, or stage‐spanning). MCI‐labeled refers to studies explicitly describing the target group as MCI. AD‐labeled describes studies explicitly describing the target group as AD, including mild or early AD.
Early‐stage restricted was assigned when the main analytic sample was limited to Clinical Dementia Rating (CDR) 0.5–1.0 based on the reported mean, range, or inclusion criteria, or alternatively, when the authors explicitly described the sample as very mild, mild, or early and more severe cases were not included. Mixed refers to studies that combined MCI and mild AD within the same analysis or elected not to restrict cases to early stage. Stage‐spanning refers to studies covering a broader severity continuum, such as subjective cognitive impairment to AD or the full AD spectrum. Based on these definitions, studies were grouped into four reviewer‐defined descriptive categories: MCI (early), MCI (mixed), AD (early), and AD (cross‐stage). Even when severity ranges overlapped, distinctions between MCI and AD were based on the diagnostic labels used in the original studies, and no reclassification was performed.
2.6. Data Synthesis
The included studies were synthesized descriptively. We summarized the distribution and patterns of writing tasks, administration media, and writing indicators across the four study context groups. We also examined whether and how motor process and content/orthographic indicators were integrated within the same study to identify gaps in the current literature on integrated writing‐based frameworks.
2.7. Use of AI‐Assisted Technology in Manuscript Preparation
During the preparation of the manuscript, the authors used ChatGPT (OpenAI) for language refinement and expression adjustment. All AI‐assisted output was critically reviewed and revised by the authors, who take full responsibility for the final content.
3. Results
3.1. Study Selection
In total, 3355 records were identified and imported into Rayyan. After removing 541 duplicates, 2814 records were screened for titles and abstracts. Of these, 2711 records were excluded, and 103 records proceeded to the next stage (full‐text retrieval and eligibility confirmation). In this study, 45 (conference abstracts, n = 34; full articles, n = 11) were excluded because the abstract and/or full text could not be obtained or sufficient information for detailed data extraction was unavailable. The remaining 58 records underwent secondary screening (full‐text review). Following full‐text review, 34 records were excluded, and 24 studies were included in the final synthesis. The study selection process is presented in Figure 1.
FIGURE 1.

PRISMA‐ScR flow diagram of the study selection. Records identified from PubMed, Web of Science, CiNii Research, and Igaku Chuo Zasshi (Ichushi‐Web) were deduplicated, screened at the title/abstract level, and assessed for full‐text eligibility. Reasons for full‐text exclusion are shown in the figure.
3.2. Distribution of Study Designs and Analytical Approaches
The included studies comprised 11 MCI (early) studies, 1 MCI (mixed) study, 6 ad (early) studies, and 6 ad (cross‐stage) studies. Most studies were cross‐sectional observational studies, whereas two were longitudinal. Regarding analytical approaches, seven studies used machine‐learning methods for modeling and/or classification. Detailed study‐level characteristics, including study design, analytical approach, and corresponding references, are provided in Appendix S1.
3.3. Writing Tasks, Administration Media, and Indicator Types
Writing tasks varied across studies and included copying, dictation, spontaneous writing (e.g., picture description), and sentence production, with substantial variability in task specification and reporting granularity. The administration media comprised paper and pencil (n = 13), digital measurement using tablets and/or digital pens (n = 9), typing logs (n = 1), and text corpora (n = 1). When examined by administration medium, studies using digital measurement and/or input logs more frequently reported motor process indicators, including initiation latency, total writing time, writing speed, pauses (pen‐up time), and pen pressure, as primary outcomes; representative examples include digital handwriting studies focusing on temporal and kinematic features [10, 15, 16, 17, 18, 19, 20, 21]. In contrast, paper‐based studies more often treated content/orthographic indicators, including output quantity (e.g., number of characters, sentence length, information content), error types (e.g., no response, formal errors, substitutions/omissions), orthographic errors (e.g., okurigana errors), and linguistic/syntactic measures, as primary outcomes, as illustrated by studies examining written‐content, orthographic/error‐based, and sentence‐level output features [11, 12, 22, 23, 24, 25, 26, 27, 28, 29, 30]. Overall, the distribution of indicator types across all included studies was as follows: motor process indicators (n = 8), content/orthographic indicators (n = 14), and integrated indicators (n = 2). Nine studies were conducted in Japan or included Japan as a study setting, and most of these studies used paper‐and‐pencil tasks (n = 8) and reported content/orthographic indicators, such as kanji form errors, no responses, and okurigana errors, as primary outcomes [11, 12, 27, 28, 29, 30]. Detailed study‐level characteristics are provided in Appendix S1.
3.4. Indicator Types as Primary Outcomes by Study Context
The distribution of primary outcome indicator types by study context is presented in Table 1. In MCI [early], motor process indicators (n = 6), content/orthographic indicators (n = 4), and integrated indicators (n = 1) were reported as primary outcomes. In AD (early), all primary outcomes were content/orthographic indicators (n = 6). In AD (cross‐stage), the primary outcomes included motor process indicators (n = 2), content/orthographic indicators (n = 3), and integrated indicators (n = 1). The single study categorized as MCI [mixed] reported content/orthographic indicators as the primary outcome. Only two studies explicitly used integrated indicators, including one each in MCI (early) and AD (cross‐stage). Detailed study‐level information for each included study is provided in Appendix S1.
TABLE 1.
Primary outcome indicator types across study context groups.
| Primary outcome indicator type | MCI (early) | MCI (mixed) | AD (early) | AD (cross‐stage) |
|---|---|---|---|---|
| Motor process indicators | 6 | 0 | 0 | 2 |
| Content/orthographic indicators | 4 | 1 | 6 | 3 |
| Integrated indicators | 1 | 0 | 0 | 1 |
Note: Integrated indicators were defined as studies in which motor process and content/orthographic indicators were analyzed within the same analytical framework (e.g., simultaneous inclusion in the same multivariable model, a composite score, or classification using integrated feature sets). Studies that measured both domains but did not integrate them analytically were classified according to the primary outcome as either motor process‐based or content/orthographic‐based.
4. Discussion
This scoping review mapped writing research from MCI to early AD across two complementary dimensions, namely motor process indicators and content/orthographic indicators, and provides a structured overview of current trends and knowledge gaps. Studies restricted to early‐stage MCI more often adopted motor process indicators as primary outcomes, whereas studies of early AD more consistently emphasized content/orthographic indicators. In AD studies spanning broader disease stages, all three indicator types—motor, content/orthographic, and integrated—were represented. Although some studies explored classification models, the current evidence remains insufficient to directly support clinical implementation because of limited sample sizes, lack of external validation, and heterogeneity in tasks and indicator definitions. Together, these findings suggest that primary outcome selection varies by study context and measurement medium and that studies integrating motor and content/orthographic dimensions within a single analytical framework remain scarce.
As its primary contributions, this review provided an integrated synthesis of writing research in MCI to early AD using a two‐axis framework and visualized the context‐dependent patterns in primary outcome selection. This matters because feasibility and rapidity remain central constraints for routine clinical assessment, and writing tasks are often proposed as low‐burden candidates for early risk characterization. Previous writing studies accumulated along two largely parallel lines, namely those emphasizing motor process features and those emphasizing content/orthographic features, but the field has not sufficiently clarified how these domains might be combined into a framework that supports clinically meaningful interpretation and potential intervention targets. By organizing studies according to diagnostic labels and reported severity ranges and mapping indicators across these two dimensions, this review identified the limited development of integrated frameworks as a major knowledge gap.
The observed differences in primary outcome selection might reflect study aims, hypothesized mechanisms, and information captured by different measurement media. First, the relative predominance of motor process indicators in studies of early‐stage MCI is consistent with designs that use digital pens or similar devices to capture subtle differences in temporal and kinematic data (e.g., pen pressure, writing speed, pen‐up time) [10, 15, 16, 17, 19, 20]. At the same time, motor data alone might be insufficient to determine whether observed abnormalities primarily reflect linguistic/orthographic breakdown or execution‐related process disruption, which limits clinical interpretation and intervention target identification. Second, the consistent emphasis on content/orthographic indicators in early AD is compatible with designs aiming to characterize linguistic and orthographic disruption through detailed error typologies and analyses of orthographic or syntactic features [12, 24, 26, 30]. This issue could be particularly important in Japan, where differences between kanji and kana can strongly influence error patterns, making explicit task design and error classification essential for between‐study and cross‐cultural comparison [11, 12, 27, 28, 29, 30]. Third, the coexistence of motor, content/orthographic, and integrated indicators in AD studies spanning broader disease stages is consistent with the inclusion of wider severity ranges and the use of multiple indicator types within the same research context [21, 25, 27, 28, 29]. Collectively, the combination of study context and measurement medium appears to be associated with primary outcome selection, which might partly explain why integrated studies remain limited.
From a clinical perspective, the present review has underscored the importance of understanding writing changes in MCI to early AD from both motor process and content/orthographic perspectives. Writing is a complex skill that requires the coordination of multiple processes, including fine motor control; retrieval and maintenance of orthographic representations; and lexical, semantic, and syntactic processing. Thus, frameworks that attempt to explain writing change from only one dimension are likely to be incomplete. Meta‐analytic neuroimaging evidence has likewise suggested that writing depends on distributed networks involving both central and peripheral motor processes [31, 32]. However, only two integrated studies were identified in this review, and the current evidence base is not yet sufficient to support direct clinical adoption. A pragmatic next step would be to establish a minimal task set, a core indicator set based on shared operational criteria, and reproducible reporting standards. In practical terms, future studies should report motor process measures (e.g., initiation latency, total duration, speed, pen‐up rate, pen pressure) alongside content/orthographic measures (e.g., error types, number of characters, information content) while explicitly defining the processes each indicator is intended to represent. In addition, because machine learning approaches in small samples are vulnerable to overfitting, future studies should place greater emphasis on external validation and model interpretability.
Several limitations of this study should be noted. First, by design, this was a scoping review, and formal risk‐of‐bias assessment or quantitative synthesis of effect sizes was not included. Therefore, the strength of evidence across individual studies could not be directly compared. Second, diagnostic frameworks for MCI differed across studies, including Petersen‐type MCI [33] and frameworks closer to MCI attributable to AD [4], which limits comparability and generalizability. Third, differences in language and writing systems, task type (e.g., copying, dictation, spontaneous writing), scoring procedures, and outcome reporting constrained direct comparisons across studies. Fourth, because conference abstracts were included to broadly capture early‐stage research, methodological and clinical details varied across reports. In addition, the four study context groups were reviewer‐defined descriptive categories intended for evidence mapping rather than disease restaging, and some misalignment with study‐level clinical nuance is possible. Future validation studies should aim to standardize diagnostic criteria, tasks, indicators, and reporting formats using prospective designs and external validation.
In conclusion, writing research in MCI and early AD has developed along two largely separate lines: motor process indicators and content/orthographic indicators. The present review suggests that outcome selection varies by study context and measurement medium, whereas analytically integrated approaches remain rare. Progress toward clinical application will require standardized tasks, shared operational definitions, core indicator sets spanning both domains, and externally validated integrated frameworks.
Author Contributions
T.K. conceived and designed the study. T.K. and A.S. conducted the study screening and data extraction. T.K. performed the analysis and drafted the manuscript. A.S. contributed to data interpretation and critically revised the manuscript. S.T. contributed to the study design, resolved disagreements during study selection, interpreted the findings, and critically revised the manuscript. All authors read and approved the final manuscript.
Funding
This work was supported by JSPS KAKENHI Grant Number JP25K23322.
Ethics Statement
The authors have nothing to report. This study was a scoping review based on published and publicly available literature.
Consent
The authors have nothing to report.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Appendix S1: Study‐level characteristics of the included studies (n = 24), including writing tasks, measures, and corresponding main text reference numbers where applicable, sorted by study‐context group.
Acknowledgments
The authors thank Crimson Interactive Japan Co. Ltd. for English language editing.
Data Availability Statement
The data supporting the findings of this study are available within the article and its Supporting Information. Additional extracted data are available from the corresponding author upon reasonable request.
References
- 1. World Health Organization , “Dementia. Fact Sheet,” 2025, accessed November 1, 2025, https://www.who.int/news‐room/fact‐sheets/detail/dementia.
- 2. Petersen R. C., “Mild Cognitive Impairment as a Diagnostic Entity,” Journal of Internal Medicine 256 (2004): 183–194. [DOI] [PubMed] [Google Scholar]
- 3. Mitchell A. J. and Shiri‐Feshki M., “Rate of Progression of Mild Cognitive Impairment to Dementia: Meta‐Analysis of 41 Robust Inception Cohort Studies,” Acta Psychiatrica Scandinavica 119 (2009): 252–265. [DOI] [PubMed] [Google Scholar]
- 4. Albert M. S., DeKosky S. T., Dickson D., et al., “The Diagnosis of Mild Cognitive Impairment due to Alzheimer's Disease: Recommendations From the National Institute on Aging–Alzheimer's Association Workgroups on Diagnostic Guidelines for Alzheimer's Disease,” Alzheimer's & Dementia 7 (2011): 270–279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. McKhann G. M., Knopman D. S., Chertkow H., et al., “The Diagnosis of Dementia due to Alzheimer's Disease: Recommendations From the National Institute on Aging–Alzheimer's Association Workgroups on Diagnostic Guidelines for Alzheimer's Disease,” Alzheimer's & Dementia 7 (2011): 263–269. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Sperling R. A., Aisen P. S., Beckett L. A., et al., “Toward Defining the Preclinical Stages of Alzheimer's Disease: Recommendations From the National Institute on Aging–Alzheimer's Association Workgroups on Diagnostic Guidelines for Alzheimer's Disease,” Alzheimer's & Dementia 7 (2011): 280–292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Folstein M. F., Folstein S. E., and McHugh P. R., ““Mini‐Mental State”. A Practical Method for Grading the Cognitive State of Patients for the Clinician,” Journal of Psychiatric Research 12 (1975): 189–198. [DOI] [PubMed] [Google Scholar]
- 8. Tombaugh T. N. and McIntyre N. J., “The Mini‐Mental State Examination: A Comprehensive Review,” Journal of the American Geriatrics Society 40 (1992): 922–935. [DOI] [PubMed] [Google Scholar]
- 9. Mitchell A. J., “A Meta‐Analysis of the Accuracy of the Mini‐Mental State Examination in the Detection of Dementia and Mild Cognitive Impairment,” Journal of Psychiatric Research 43 (2009): 411–431. [DOI] [PubMed] [Google Scholar]
- 10. Werner P., Rosenblum S., Bar‐On G., Heinik J., and Korczyn A., “Handwriting Process Variables Discriminating Mild Alzheimer's Disease and Mild Cognitive Impairment,” Journals of Gerontology. Series B, Psychological Sciences and Social Sciences 61 (2006): P228–P236. [DOI] [PubMed] [Google Scholar]
- 11. Akanuma K., Meguro K., Meguro M., Chubaci R. Y. S., Caramelli P., and Nitrini R., “Kanji and Kana Agraphia in Mild Cognitive Impairment and Dementia: A Trans‐Cultural Comparison of Elderly Japanese Subjects Living in Japan and Brazil,” Dementia & Neuropsychologia 4 (2010): 300–305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Hayashi A., Nomura H., Mochizuki R., et al., “Neural Substrates for Writing Impairments in Japanese Patients With Mild Alzheimer's Disease: A SPECT Study,” Neuropsychologia 49 (2011): 1962–1968. [DOI] [PubMed] [Google Scholar]
- 13. Peters M. D. J., Marnie C., Tricco A. C., et al., “Updated Methodological Guidance for the Conduct of Scoping Reviews,” JBI Evidence Synthesis 18 (2020): 2119–2126. [DOI] [PubMed] [Google Scholar]
- 14. Tricco A. C., Lillie E., Zarin W., et al., “PRISMA Extension for Scoping Reviews (PRISMA‐ScR): Checklist and Explanation,” Annals of Internal Medicine 169 (2018): 467–473. [DOI] [PubMed] [Google Scholar]
- 15. Toffoli S., Abbate C., Lunardini F., et al., “Handwriting in Mild Cognitive Impairment: Reliability Assessment and Machine Learning‐Based Screening,” JMIR Aging 8 (2025): e73074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Li S., Li K., Liu J., et al., “A New Method for Community‐Based Intelligent Screening of Early Alzheimer's Disease Populations Based on Digital Biomarkers of the Writing Process,” Frontiers in Computational Neuroscience 19 (2025): 1564932. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. An S. K., Jang H., Kim H. J., Na D. L., and Yoon J. H., “Linguistic, Visuospatial, and Kinematic Writing Characteristics in Cognitively Impaired Patients With Beta‐Amyloid Deposition,” Frontiers in Aging Neuroscience 15 (2023): 1217746. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Meulemans C., Leijten M., Van Waes L., et al., “Cognitive Writing Process Characteristics in Alzheimer's Disease,” Frontiers in Psychology 13 (2022): 872280. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. El‐Yacoubi M. A., Garcia‐Salicetti S., Kahindo C., et al., “From Aging to Early‐Stage Alzheimer's: Uncovering Handwriting Multimodal Behaviors by Semi‐Supervised Learning and Sequential Representation Learning,” Pattern Recognition 86 (2019): 112–133. [Google Scholar]
- 20. Kahindo C., El‐Yacoubi M. A., Garcia‐Salicetti S., et al., “Semi‐Global Parameterization of Online Handwriting Features for Characterizing Early‐Stage Alzheimer and Mild Cognitive Impairment,” IRBM 39 (2018): 421–429. [Google Scholar]
- 21. Cilia N. D., De Stefano C., Fontanella F., and Siniscalchi S. M., “How Word Semantics and Phonology Affect Handwriting of Alzheimer's Patients: A Machine Learning Based Analysis,” Computers in Biology and Medicine 169 (2024): 107891. [DOI] [PubMed] [Google Scholar]
- 22. Sajjadi S. A., Khan Z., El‐Khoury A., et al., “Association of Language Markers With Future Cognitive Impairment and Presence of Limbic Predominant Age‐Related TDP‐43 Encephalopathy,” Alzheimer's & Dementia: The Journal of the Alzheimer's Association 21 (2025): e70450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Sand Aronsson F., Kuhlmann M., Jelic V., and Östberg P., “Is Cognitive Impairment Associated With Reduced Syntactic Complexity in Writing? Evidence From Automated Text Analysis,” Aphasiology 35 (2021): 900–913. [Google Scholar]
- 24. Rodríguez‐Ferreiro J., Martínez C., Pérez‐Carbajal A. J., and Cuetos F., “Neural Correlates of Spelling Difficulties in Alzheimer's Disease,” Neuropsychologia 65 (2014): 12–17. [DOI] [PubMed] [Google Scholar]
- 25. González‐Nosti M., Cuetos F., and Martínez C., “Oral and Written Naming in Alzheimer's Disease: A Longitudinal Study,” Current Alzheimer Research 15 (2018): 1142–1150. [DOI] [PubMed] [Google Scholar]
- 26. Forbes‐McKay K. E. and Venneri A., “Detecting Subtle Spontaneous Language Decline in Early Alzheimer's Disease With a Picture Description Task,” Neurological Sciences 26 (2005): 243–254. [DOI] [PubMed] [Google Scholar]
- 27. Sakano N., Notoya M., Tanabe S., et al., “A Characteristic and the Assessment of the Writing Impairments in Subjects With Alzheimer's Disease,” Journal of the Tsuruma Health Science Society, Kanazawa University 37, no. 2 (2013): 35–43. [Google Scholar]
- 28. Sato A., Imamura T., Kudo Y., et al., “Dysgraphia in Alzheimer's Disease: Relation Between Each Type of Error in a Dictation Task and the Scores of Attention Tasks,” Japanese Journal of Neuropsychology 28, no. 4 (2012): 266–273. [Google Scholar]
- 29. Saito F., Hayashi A., Mori E., et al., “Selective Agraphia for Kanji in Alzheimer's Disease: A Case Study,” Higher Brain Function Research 28, no. 4 (2008): 392–403. [Google Scholar]
- 30. Akanuma K., Meguro K., Hashimoto R., et al., “Spontaneous Writing in Very Mild Alzheimer's Disease: Analyses of MMSE by Older Residents in a Community,” Higher Brain Function Research 24, no. 3 (2004): 360–367. [Google Scholar]
- 31. Purcell J. J., Turkeltaub P. E., Eden G. F., and Rapp B., “Examining the Central and Peripheral Processes of Written Word Production Through Meta‐Analysis,” Frontiers in Psychology 2 (2011): 239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Planton S., Jucla M., Roux F. E., and Démonet J. F., “The “Handwriting Brain”: A Meta‐Analysis of Neuroimaging Studies of Motor Versus Orthographic Processes,” Cortex 49 (2013): 2772–2787. [DOI] [PubMed] [Google Scholar]
- 33. Petersen R. C., Smith G. E., Waring S. C., et al., “Mild Cognitive Impairment: Clinical Characterization and Outcome,” Archives of Neurology 56 (1999): 303–308. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix S1: Study‐level characteristics of the included studies (n = 24), including writing tasks, measures, and corresponding main text reference numbers where applicable, sorted by study‐context group.
Data Availability Statement
The data supporting the findings of this study are available within the article and its Supporting Information. Additional extracted data are available from the corresponding author upon reasonable request.
