Related Experiment Video
Updated: Aug 12, 2026
09:47
DeepOmicsAE: Representing Signaling Modules in Alzheimer's Disease with Deep Learning Analysis of Proteomics, Metabolomics, and Clinical Data
Published on: December 15, 2023
TriFusion-ADFormer: a deep learning framework for early Alzheimer's disease detection using MRI and cognitive metrics
S Sabari Vasan1, P Jayalakshmi1
1School of Computer Science Engineering and Information Systems, Vellore Institute of Technology, Vellore, India.
Frontiers in Artificial Intelligence
|August 11, 2026
Summary
This study introduces TriFusion-ADFormer, a novel deep learning model that enhances Alzheimer's disease diagnosis by integrating MRI scans, clinical notes, and cognitive scores. The framework achieves 86% accuracy in classifying Alzheimer's disease, mild cognitive impairment, and normal cognition.
Area of Science:
- Neuroscience
- Artificial Intelligence
- Medical Imaging
Background:
- Alzheimer's disease (AD) is a progressive neurodegenerative disorder impacting cognition and neuronal function.
- Early and accurate diagnosis is crucial for effective intervention and patient management.
- Integrating multimodal data for reliable AD classification presents a significant challenge.
Purpose of the Study:
- To develop and evaluate TriFusion-ADFormer, a multimodal deep learning framework for classifying Alzheimer's disease (AD), mild cognitive impairment (MCI), and cognitively normal (CN) subjects.
- To assess the efficacy of fusing structural MRI, MRI-derived clinical text summaries, and cognitive assessment features for improved diagnostic accuracy.
Main Methods:
- The TriFusion-ADFormer framework was developed using a multimodal deep learning approach.
- It extracts features from structural MRI (volumetric measurements) and MRI-derived clinical text summaries.
- Cognitive assessment features (MMSE, GDS, Global CDR, FAQ, NPI-Q) were also incorporated and fused for classification.
Main Results:
- The framework achieved an overall classification accuracy of 86.0% for multiclass AD classification.
- It obtained a Macro AUC of 0.93 and an F1-score of 86.0%.
- MRI-based clinical summaries showed consistency with typical AD-associated structural abnormalities like brain atrophy, supporting framework interpretability.
Conclusions:
- Combining multimodal data (structural MRI, semantic clinical summaries, cognitive scores) significantly enhances Alzheimer's diagnosis accuracy and interpretability.
- The transformer-based framework demonstrates the potential of integrating complementary imaging, semantic, and cognitive features for improved classification of AD, MCI, and CN.