NeuroNest: Community based Cognitive Health Monitoring & Early Dementia Detection System

Abstract

Dementia represents a major public health concern worldwide, particularly as the global population continues to age. The condition is characterized by a gradual deterioration in cognitive functions including memory retention, reasoning ability, decision making, and language processing. As the disease progresses, individuals experience increasing difficulty performing routine daily activities, which significantly affects their independence and overall quality of life. Early identification of cognitive decline is essential because medical and behavioral interventions are most effective when applied during the initial stages of impairment. However, conventional diagnostic procedures generally require clinical visits and trained specialists, which limits accessibility and frequency of assessment. These challenges highlight the need for scalable digital solutions capable of supporting remote cognitive health monitoring. The NeuroNest platform is developed to address these limitations by providing an intelligent web-based environment for cognitive evaluation, behavioral analysis, and continuous monitoring. To address the limitations of conventional dementia screening methods, this research proposes NeuroNest, an intelligent web-based cognitive assessment and monitoring platform. NeuroNest is designed to evaluate multiple cognitive domains through a set of interactive brain-training games that measure memory, attention, language ability, and reaction speed. By transforming cognitive testing into interactive digital activities, the system encourages consistent user engagement while simultaneously collecting valuable performance data for analysis. The proposed platform aims to provide a scalable, accessible, and user-friendly solution for cognitive health monitoring. By enabling remote participation and automated analysis of cognitive performance data, NeuroNest has the potential to improve early dementia screening and facilitate continuous cognitive monitoring. The integration of interactive digital assessments with secure data management and analytical reporting makes NeuroNest a promising approach for supporting both clinical research and real-world healthcare applications in the field of cognitive health.

Introduction

NeuroNest is an AI-enabled, web-based cognitive health platform designed to support the early detection and longitudinal monitoring of cognitive decline and dementia. It addresses limitations of conventional screening methods such as MMSE and MoCA, which generally require trained clinicians and in-person assessments.

The system combines interactive cognitive games, facial-behavior analysis, and machine-learning algorithms to generate a more comprehensive assessment of cognitive health. Users complete tasks measuring reaction time, memory, attention, language comprehension, and recall, while a webcam—used with permission—captures facial behavior through Face Mesh technology.

Key Components

1. Cognitive Assessment
   • Reaction-time tests measure processing speed and attention.
   • Pattern-memory tests evaluate recall accuracy and memory speed.
   • Language tasks assess comprehension and verbal processing.
   • Metrics such as accuracy, reaction time, errors, and completion time are recorded.
2. Facial Behavior Analysis
   • A Face Mesh model tracks approximately 468 three-dimensional facial landmarks.
   • Features such as eye-blink frequency, gaze direction, mouth movement, facial activity, and head orientation are extracted.
   • These features provide additional behavioral information alongside cognitive-test performance.
3. Machine Learning
   • Cognitive and facial features are normalized, processed, and combined into a unified feature vector.
   • Models such as Random Forest, SVM, Logistic Regression, and Decision Tree can be used for classification.
   • The system categorizes users into risk groups such as:
     • Normal/Low Cognitive Risk
     • Mild Cognitive Impairment/Moderate Risk
     • High Dementia Risk
4. Longitudinal Monitoring
   • Session information, scores, reaction times, and other metrics are stored securely.
   • Repeated assessments allow the system to identify trends, such as progressively slower reaction times or declining memory accuracy.
   • Dashboards visualize these changes, while automatically generated PDF reports can support clinician review.

Results

The prototype successfully demonstrated the core assessment and data-management functions. Example results included a ~450 ms reaction time in a reaction-time task and 80% accuracy when recalling four of five items in a memory task. Data were correctly stored and displayed through patient dashboards, with trends across multiple sessions visualized through graphs. PDF reports were also generated and verified against dashboard data.

The Face Mesh component reportedly achieved 20–25 frames per second on standard laptop hardware, indicating that real-time facial landmark tracking is technically feasible for the proposed platform.

Overall Conclusion

NeuroNest demonstrates the potential of combining digital cognitive testing, computer vision, machine learning, and longitudinal data analysis into an accessible remote screening platform. Its main contribution is the integration of cognitive-game performance with facial behavioral features rather than relying solely on conventional cognitive scores.

However, the system should be regarded as a screening and decision-support tool rather than a diagnostic system. The reported machine-learning risk categories require validation using sufficiently large, clinically diagnosed datasets and independent testing before they can be used to make medical decisions. Factors such as age, education, culture, language, lighting, facial differences, privacy, and webcam quality should also be considered.

Conclusion

This paper presented NeuroNest, a web-based cognitive assessment system designed to aid early dementia detection and monitoring. NeuroNest integrates interactive cognitive games with a full-stack web architecture to provide accessible screening tools outside of clinical settings. Patients and seniors can perform memory, attention, and language tasks via engaging online games, while doctors and family members can monitor the user’s performance remotely through dashboards and reports. In testing, NeuroNest successfully recorded reaction times, accuracy scores, and completion times for each game, storing them in a structured database. The system’s user interface enabled smooth interaction and intuitive navigation. The collected data were accurately reflected in the generated reports, allowing easy interpretation of cognitive trends. These findings align with literature suggesting that digital game-based tools can achieve screening accuracy similar to traditional tests. NeuroNest thus demonstrates the potential of web technologies to support continuous cognitive health monitoring and early intervention.

References

Journal Paper: [1] N. Piyaamornpan et al., “Web-Based Application for Cognitive and Functional Assessments in Dementia Screening: Mixed Methods, User-Centered Development Approach,” JMIR Hum. Factors, vol. 13, 2026, e85454. [2] F. G. Scaramuzzi et al., “Digital Screening for Early Identification of Cognitive Impairment: A Narrative Review,” WIREs Cogn. Sci., vol. 17, 2025, e1500. [3] Y Y. Chen et al., “Video Games and Gamification for Assessing Mild Cognitive Impairment: Scoping Review,” JMIR Ment. Health, vol. 12, 2025, e40521. Proceeding paper: [4] E. Stroulia et al., “Machine Learning Analysis of Engagement Behaviors in Older Adults with Dementia Playing Mobile Games: Exploratory Study,” JMIR Serious Games, vol. 13, 2025, e54797. [5] B. Liu, X. Li, H. Cai, “The effects of video games on cognitive function in older adults with mild cognitive impairment: a meta-analysis,” Front. Aging Neurosci., vol. 17, 2026, Art. 1756970. [6] A. Cejudo et al., “AI and Wearables for Early Detection of Cognitive Impairment and Dementia: Systematic Review,” J. Med. Internet Res., vol. 28, 2026, e86262.

Copyright

Copyright © 2026 Mr. Ajay Kushwaha, Mr. Mayur Dete, Mr. Shreyas Chavan, Prof. Sneha Sankhe. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.