Related Experiment Video
Updated: Dec 11, 2025
09:47
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
1.5K
Predicting rate of cognitive decline at baseline using a deep neural network with multidata analysis
Sema Candemir1, Xuan V Nguyen1, Luciano M Prevedello1
1The Ohio State University College of Medicine, Laboratory for Augmented Intelligence in Imaging, Department of Radiology, Columbus, Ohio, United States.
Journal of Medical Imaging (Bellingham, Wash.)
|August 25, 2020
Summary
This study uses machine learning to predict cognitive decline rates in mild cognitive impairment patients using initial clinical and MRI data. The model successfully identified patients with different rates of cognitive deterioration.
Area of Science:
- Neuroscience
- Artificial Intelligence
- Medical Imaging
Background:
- Mild cognitive impairment (MCI) is a precursor to Alzheimer's disease.
- Predicting the rate of cognitive decline in MCI patients is crucial for timely intervention.
- Current methods often focus on disease conversion or classification, not decline rate prediction.
Purpose of the Study:
- To develop and validate a machine-learning system for predicting cognitive decline rates in MCI patients.
- To assess the utility of baseline clinical and magnetic resonance imaging (MRI) data for this prediction.
- To differentiate between slowly and rapidly deteriorating MCI patient groups.
Main Methods:
- A supervised hybrid neural network model was developed.
- A 3D convolutional neural network analyzed MRI volumes.
- Nonimaging clinical data were integrated at the fully connected layer.
- The Alzheimer's Disease Neuroimaging Initiative dataset was used for experiments.
Main Results:
- A correlation between initial visit data and cognitive decline was confirmed.
- The system achieved an area under the receiver operator curve of 0.70 for predicting cognitive decline classes.
- The model effectively utilized baseline MRI, MMSE, and demographic data.
Conclusions:
- This is the first study to predict cognitive decline rates ('slowly deteriorating/stable' vs. 'rapidly deteriorating') using only baseline clinical and demographic data.
- The findings suggest that early prediction of cognitive decline trajectory is feasible with routinely collected data.
- This approach offers a novel perspective compared to studies focusing on MCI-to-Alzheimer's conversion.
