Related Experiment Video
Updated: Aug 25, 2025
09:47
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
1.2K
Generalizable deep learning model for early Alzheimer's disease detection from structural MRIs.
Sheng Liu1, Arjun V Masurkar2,3, Henry Rusinek4,5
1Center for Data Science, NYU, 60 Fifth Avenue, 5th Floor, New York, NY, 10011, USA.
Scientific Reports
|October 17, 2022
Summary
This study introduces a 3D deep learning model for early Alzheimer's disease detection using MRI scans. The AI model accurately identifies disease stages and predicts progression faster than traditional methods.
Area of Science:
- Neuroimaging
- Artificial Intelligence
- Medical Diagnostics
Background:
- Early diagnosis of Alzheimer's disease (AD) is crucial for effective patient management and clinical trial success.
- Current diagnostic methods can be time-consuming and may not capture the full complexity of early-stage neurodegeneration.
- Structural magnetic resonance imaging (MRI) offers valuable insights into brain changes associated with AD.
Purpose of the Study:
- To develop and validate a novel 3D deep convolutional neural network (CNN) for accurate early detection of Alzheimer's disease.
- To compare the performance and efficiency of the deep learning model against a traditional region-based analysis model.
- To investigate the model's capability in predicting disease progression in individuals with mild cognitive impairment (MCI).
Main Methods:
- Development of a 3D deep CNN utilizing structural MRI data.
- Creation of a comparative reference model based on volumetric and thickness measurements of known AD-affected brain regions.
- Validation of both models on independent internal (Alzheimer's Disease Neuroimaging Initiative - ADNI) and external (National Alzheimer's Coordinating Center - NACC) cohorts.
Main Results:
- The 3D deep CNN achieved an area-under-the-curve (AUC) of 85.12% in distinguishing cognitively normal individuals from those with MCI or mild AD dementia.
- The model demonstrated an AUC of 62.45% for the more challenging task of detecting MCI.
- The deep learning approach was significantly faster than the traditional model and showed potential in forecasting disease progression.
Conclusions:
- Deep neural networks can automatically identify predictive imaging biomarkers for Alzheimer's disease from structural MRIs.
- The developed 3D deep CNN offers a highly accurate and efficient tool for early Alzheimer's disease detection.
- The model's ability to predict progression highlights its potential clinical utility in managing neurodegenerative diseases.
