Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

BRAINS: A Retrieval-Augmented System for Alzheimer's Detection and Monitoring

Rajan Das Gupta, Md Kishor Morol, Nafiz Fahad, Md Tanzib Hosain, Sumaya Binte Zilani Choya, Md Jakir Hossen

TL;DR

The paper tackles early detection and staging of Alzheimer's disease in diverse clinical settings by proposing BRAINS, a retrieval-augmented system that combines a Diagnostic Module with a Case Retrieval Module. Retrieved historical cases are fused with the current input through a Case Fusion Layer and processed by an Inference LLM, enabled by FAISS-based top-$K$ retrieval and cross-attention mechanisms. Fine-tuning with domain-specific data and LoRA adaptation yields substantial accuracy gains (up to $77.30\%$) over baseline LLMs, while analyses reveal trade-offs in multi-clue predictions and the importance of contextual case information. BRAINS demonstrates a scalable, explainable approach to brain-health modelling with potential applicability to broader neurodiagnostic tasks in both high-resource and underserved settings.

Abstract

As the global burden of Alzheimer's disease (AD) continues to grow, early and accurate detection has become increasingly critical, especially in regions with limited access to advanced diagnostic tools. We propose BRAINS (Biomedical Retrieval-Augmented Intelligence for Neurodegeneration Screening) to address this challenge. This novel system harnesses the powerful reasoning capabilities of Large Language Models (LLMs) for Alzheimer's detection and monitoring. BRAINS features a dual-module architecture: a cognitive diagnostic module and a case-retrieval module. The Diagnostic Module utilizes LLMs fine-tuned on cognitive and neuroimaging datasets -- including MMSE, CDR scores, and brain volume metrics -- to perform structured assessments of Alzheimer's risk. Meanwhile, the Case Retrieval Module encodes patient profiles into latent representations and retrieves similar cases from a curated knowledge base. These auxiliary cases are fused with the input profile via a Case Fusion Layer to enhance contextual understanding. The combined representation is then processed with clinical prompts for inference. Evaluations on real-world datasets demonstrate BRAINS effectiveness in classifying disease severity and identifying early signs of cognitive decline. This system not only shows strong potential as an assistive tool for scalable, explainable, and early-stage Alzheimer's disease detection, but also offers hope for future applications in the field.

BRAINS: A Retrieval-Augmented System for Alzheimer's Detection and Monitoring

TL;DR

The paper tackles early detection and staging of Alzheimer's disease in diverse clinical settings by proposing BRAINS, a retrieval-augmented system that combines a Diagnostic Module with a Case Retrieval Module. Retrieved historical cases are fused with the current input through a Case Fusion Layer and processed by an Inference LLM, enabled by FAISS-based top- retrieval and cross-attention mechanisms. Fine-tuning with domain-specific data and LoRA adaptation yields substantial accuracy gains (up to ) over baseline LLMs, while analyses reveal trade-offs in multi-clue predictions and the importance of contextual case information. BRAINS demonstrates a scalable, explainable approach to brain-health modelling with potential applicability to broader neurodiagnostic tasks in both high-resource and underserved settings.

Abstract

As the global burden of Alzheimer's disease (AD) continues to grow, early and accurate detection has become increasingly critical, especially in regions with limited access to advanced diagnostic tools. We propose BRAINS (Biomedical Retrieval-Augmented Intelligence for Neurodegeneration Screening) to address this challenge. This novel system harnesses the powerful reasoning capabilities of Large Language Models (LLMs) for Alzheimer's detection and monitoring. BRAINS features a dual-module architecture: a cognitive diagnostic module and a case-retrieval module. The Diagnostic Module utilizes LLMs fine-tuned on cognitive and neuroimaging datasets -- including MMSE, CDR scores, and brain volume metrics -- to perform structured assessments of Alzheimer's risk. Meanwhile, the Case Retrieval Module encodes patient profiles into latent representations and retrieves similar cases from a curated knowledge base. These auxiliary cases are fused with the input profile via a Case Fusion Layer to enhance contextual understanding. The combined representation is then processed with clinical prompts for inference. Evaluations on real-world datasets demonstrate BRAINS effectiveness in classifying disease severity and identifying early signs of cognitive decline. This system not only shows strong potential as an assistive tool for scalable, explainable, and early-stage Alzheimer's disease detection, but also offers hope for future applications in the field.

Paper Structure

This paper contains 12 sections, 2 equations, 1 figure, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Data Preparing
  4. Alzheimer's Disease Dataset
  5. Model Architecture
  6. Case Retrieval Module
  7. Diagnostic Module
  8. Experiments
  9. Pre-training Setting
  10. Fine-tuning Setting
  11. Results
  12. Conclusion

Figures (1)

  • Figure 1: BRAINS architecture for Alzheimer’s diagnosis. The input case is encoded and used to retrieve similar neurocognitive records. Retrieved cases are fused with the input via the Case Fusion Layer, replacing the <RAGHere> token in the prompt. The fused representation is then passed to the LLM for inference and explanation.