Alzheimer's disease (AD) presents an escalating global health crisis, currently affecting over 50 million individuals worldwide, a figure projected to surge as the global population ages (Cano et al., 2021). The progressive cognitive decline characteristic of AD significantly impairs daily functioning and independence. Amidst the emergence of disease-modifying anti-amyloid therapies for early-stage AD, the imperative for timely detection, risk stratification, and prompt intervention has become paramount. Within this context, Mild Cognitive Impairment due to Alzheimer's Disease (MCI due to AD) represents a pivotal, prodromal phase—a crucial window for therapeutic strategies. Individuals diagnosed at this stage face a substantial annual conversion rate to full dementia, estimated at approximately 15% (Xu X. et al., 2024), underscoring the critical need for accurate and accessible screening during this initial clinical period.

Pioneering Digital Diagnostics for Cognitive Decline

A collaborative research effort has recently unveiled a novel digital screening methodology for MCI due to AD, leveraging a touchscreen-based maze hand-interaction kinetic paradigm. This innovative approach integrates digital biomarkers from both visuospatial/executive and episodic memory domains to facilitate early identification. The findings of this original research article were published on June 3, 2026, in the Neurodegeneration section of *Frontiers in Neuroscience*, Volume 20, and can be accessed via DOI: https://doi.org/10.3389/fnins.2026.1788324.

The study's development involved researchers from several key institutions: the School of Medical Technology and Information Engineering, Zhejiang Chinese Medical University, Hangzhou, China; the School of Information Engineering, Hangzhou Medical College, Hangzhou, China; the Zhejiang Engineering Research Center for Brain Cognition and Brain Diseases Digital Medical Instruments, Hangzhou Medical College, Hangzhou, China; the Centre for Cognitive and Brain Sciences, University of Macau, Taipa, Macau SAR, China; and the Department of Neurology, The Second Medical Center and National Clinical Research Center for Geriatric Diseases, Chinese PLA General Hospital, Beijing, China. Thamires Naela Cardoso Magalhães is also credited for contributing to the understanding of these digital biomarkers.

The Maze Paradigm: Unpacking the Methodology

To develop this screening tool, a specially designed maze task was administered to a cohort of 40 patients with a clinical diagnosis of MCI due to AD, alongside a control group of 40 healthy individuals (HC). During this assessment, various behavioral data points were meticulously collected. Subsequently, two distinct categories of digital biomarkers were extracted for analysis. The first category captured visuospatial and executive functions, quantified by metrics such as task completion time (VSETT) and average movement speed (VSES). The second domain, episodic memory, was assessed using measures like episodic memory total time (EMTT) and the number of correct choices (EMCC).

Significant digital biomarkers, distinguishing between the patient and control groups, were identified through comprehensive between-group comparisons. Their collective classification efficacy was subsequently evaluated utilizing binary logistic regression and receiver operating characteristic (ROC) analysis.

Robust Results Validate Screening Efficacy

Initial findings demonstrated promising discriminative performance for the integrated digital biomarker model within the entire study cohort, yielding an Area Under the Curve (AUC) of 0.899 (with a 95% Confidence Interval: 0.831–0.967). To mitigate potential optimism often associated with biomarker selection and model development within a single dataset, the researchers conducted rigorous internal validation. This process involved a full-pipeline repeated stratified five-fold cross-validation, where all 16 candidate digital biomarkers were included, and biomarker selection was iterated within each training fold.

Following this robust internal validation, the model consistently maintained strong discriminative capabilities. It achieved a mean cross-validated AUC of 0.842, accompanied by an empirical 95% interval ranging from 0.779–0.878. Furthermore, the validated model exhibited an accuracy of 0.783, a sensitivity of 0.772, and a specificity of 0.795.

These compelling results strongly indicate that the proposed touchscreen maze-based digital assessment offers a promising and objective avenue for supporting the early screening of MCI due to AD.

Beyond Memory: A Multi-Domain Approach

The early clinical presentation of MCI due to AD is not confined to a single isolated cognitive deficit but rather encompasses impairments across multiple cognitive domains. Episodic memory impairment, characterized by difficulties in learning and recalling newly acquired information (Alegret et al., 2022; Gagliardi et al., 2023; Pelgrim et al., 2021), is widely recognized as a hallmark feature of early AD. This deficit is closely linked to early degenerative changes within the medial temporal lobe, particularly affecting the hippocampus and entorhinal cortex, alongside disruptions in connectivity with posterior nodes of the default mode network (DMN) (Albertina et al., 2024; Chandra et al., 2025; Ekstrom and Hill, 2023).

Additionally, impairments in visuospatial and executive functions are prominent in the initial stages of AD. Executive dysfunction manifests as deficits in planning, working memory, cognitive flexibility, and response inhibition, correlating with abnormalities in the fronto-parietal network (Coleman et al., 2023). Visuospatial decline impacts spatial orientation, visual integration, and navigation (Li et al., 2023), with corresponding abnormalities in the posterior parietal cortex, occipitotemporal junction, and dorsal visual pathway (Beyh et al., 2022; Schoonover et al., 2024). These functions are intrinsically linked to real-world navigation and visually guided behaviors, requiring complex interplay of perception, planning, attentional control, and flexible decision-making.

Fotoğraf: Breakthrough Touchscreen Maze Unlocks Early Detection of Alzheimer's Risk

Critically, the simultaneous presence of deficits across episodic memory, visuospatial processing, and executive control offers superior predictive value for dementia progression compared to impairment in a single cognitive domain (Knopman et al., 2024). Therefore, assessment tools capable of integrating these diverse cognitive domains are particularly valuable for capturing the heterogeneous yet convergent cognitive profile of early AD.

Overcoming Screening Challenges with Digital Innovation

Currently, standard screening for AD often relies on brief neuropsychological scales such as the Mini-Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA). However, these tools are limited by subjective scoring, educational and cultural biases, practice effects, and insufficient sensitivity to detect subtle, nascent cognitive changes (Wang et al., 2024). While advanced biomarkers, including neuroimaging (Leuzy et al., 2025), cerebrospinal fluid analysis (Wang Z. et al., 2025), and blood-based assays (Therriault et al., 2024), offer high diagnostic accuracy, their considerable cost, invasiveness, and restricted accessibility hinder their widespread adoption in large-scale community or primary-care screening programs.

Digital biomarkers, defined as objective and quantifiable behavioral or physiological data captured through digital technologies, represent a compelling alternative to bridge this diagnostic gap (Avram et al., 2020). In AD research, metrics derived from wearable sensors, handwriting dynamics, and eye-tracking have shown potential for early detection by revealing subtle alterations in motor control, kinematics, and visual exploration that often elude traditional paper-and-pencil tests (Lin et al., 2023; Oyama et al., 2019; Shah et al., 2025). These methods are invaluable for their capacity to quantify task performance continuously and objectively, moving beyond reliance solely on total scores or categorical clinical judgments.

However, existing digital approaches vary in their focus and practical utility. Virtual reality (VR) navigation tasks, while offering high ecological validity for assessing spatial memory, encounter obstacles related to cost, technical complexity, and potential simulator sickness (Carelli et al., 2011; Machado et al., 2019). Eye-tracking paradigms, despite their sensitivity to early visual processing deficits, are susceptible to performance degradation due to head posture changes—accounting for approximately 45% of total variance—and are further impacted by calibration and other variables (Wang X. et al., 2025). In contrast, tablet-based serious games can engage multiple cognitive domains in an enjoyable format but may lack the kinematic granularity necessary to dissect specific motor planning or execution deficits. An optimal digital screening approach for early AD, therefore, must strike a balance between ecological validity, cognitive specificity, quantitative precision, and practical feasibility.

Among various behavioral paradigms, maze tasks are particularly well-suited for evaluating the cognitive profile of early AD. They inherently engage visuospatial processing, executive planning, and spatial memory within a single, ecologically relevant task. While traditional mazes, such as the Porteus maze and Morris water maze, face limitations in standardization and practical application (Morris, 1984; Ott et al., 2003; Porteus, 1945), touchscreen-based digital maze tasks offer significant advantages. They maintain ecological validity akin to real-world navigation (Germine et al., 2019; Iliadou et al., 2021), enable the automated extraction of rich, high-resolution kinematic data, and are scalable, cost-effective, and straightforward to implement in clinical settings.

This study's new paradigm, combining a visuospatial-executive module with a landmark-based episodic recognition module, was specifically designed to reflect this multi-domain impairment pattern, aiming to support more sensitive detection of early AD-related cognitive changes.

Latest Updates on this Story

Researchers continue to refine digital biomarker methodologies, with this study representing a significant step forward in the urgent quest for non-invasive early Alzheimer's detection. The breaking news of validated digital tools like this maze paradigm offers hope for expanded access to screening for mild cognitive impairment, providing critical insights for current news and future clinical applications. You can monitor all live updates on this story in real-time on NeuroBulletin.com.

Related Topics

🔹 Alzheimer's Disease 🔹 Mild Cognitive Impairment (MCI) 🔹 Digital Biomarkers 🔹 Early Detection Methods 🔹 Neurodegeneration Research 🔹 Touchscreen Diagnostics 🔹 Cognitive Assessment 🔹 Brain Health Technology

About NeuroBulletin News

NeuroBulletin.com stands as the leading independent resource for cutting-edge research and developments in neuroscience and brain health. Our coverage spans a wide array of topics, from novel diagnostic tools like the one presented here to therapeutic advancements and foundational brain science, delivering meticulously edited news to our global readership. We are dedicated to providing accurate, objective, and premium insights into the complex world of the brain and neurological disorders.

Frequently Asked Questions

What is Mild Cognitive Impairment due to Alzheimer's Disease (MCI due to AD)?

MCI due to AD is an early stage of cognitive decline that is more severe than typical age-related memory loss but not severe enough to interfere with daily life and independence. It is considered a prodromal stage of Alzheimer's disease, meaning individuals are at a significantly higher risk of progressing to full dementia.

How does this touchscreen maze method work for early screening?

This method involves a customized maze task administered on a touchscreen. It collects behavioral data, from which digital biomarkers are extracted to measure visuospatial/executive functions (like task completion time and movement speed) and episodic memory (like total time and correct choices). These biomarkers help differentiate individuals with MCI due to AD from healthy controls.

What are the advantages of using digital biomarkers in AD screening?

Digital biomarkers offer several advantages, including objectivity, quantifiability, and the ability to capture subtle alterations missed by traditional tests. They can provide continuous, high-resolution data, overcome subjective scoring limitations, and are generally more scalable and accessible than costly or invasive advanced diagnostic methods.

How accurate is this new screening approach?

The integrated digital biomarker model demonstrated strong performance, with an initial AUC of 0.899. After rigorous internal validation using five-fold cross-validation, the model maintained good discriminative power, showing a mean cross-validated AUC of 0.842, an accuracy of 0.783, a sensitivity of 0.772, and a specificity of 0.795, indicating its reliability.