Related Experiment Video
Updated: May 30, 2025
08:43
Application of Granger Causality Analysis of the Directed Functional Connection in Alzheimer's Disease and Mild Cognitive Impairment
Published on: August 7, 2017
7.8K
Evaluating conversion from mild cognitive impairment to Alzheimer's disease with structural MRI: a machine learning
Daniela Vecchio1, Federica Piras1, Federica Natalizi1,2,3
1Neuropsychiatry Laboratory, Department of Clinical Neuroscience and Neurorehabilitation, IRCCS Santa Lucia Foundation, Rome 00179, Italy.
Brain Communications
|January 31, 2025
Summary
This study used MRI scans and machine learning to predict Alzheimer's disease progression in patients with mild cognitive impairment. Key brain region volumes, particularly the right entorhinal cortex, can help identify individuals at higher risk.
Area of Science:
- Neuroimaging
- Computational neuroscience
- Biomarkers
Background:
- Alzheimer's disease (AD) is a progressive neurodegenerative disorder with no cure.
- Current AD diagnostic biomarkers are often invasive, costly, or time-consuming.
- Early diagnosis is critical for patient outcomes and managing cognitive decline.
Purpose of the Study:
- To develop and validate MRI-based computational methods for early Alzheimer's disease diagnosis.
- To identify quantitative MRI-based rules predicting conversion from amnestic mild cognitive impairment (aMCI) to AD.
- To apply machine learning algorithms for predicting aMCI to AD conversion.
Main Methods:
- Collected T1-weighted brain MRI images from 104 aMCI patients.
- Extracted 146 volumetric grey matter measures (regions of interest - ROIs).
- Utilized Random Forest for feature selection, followed by Support Vector Machine and Decision Tree classification.
Main Results:
- Achieved 86% average accuracy in differentiating aMCI converters (aMCI-c) from non-converters (aMCI-s) using SVM and DT.
- Identified right entorhinal cortex (EC-r) volume below 1286 mm³ as a key predictor for aMCI-c.
- Highlighted left lateral occipital (LOC-l) and other temporal lobe regions as significant for classification.
Conclusions:
- Morphometry of EC-r and LOC-l significantly predicts aMCI to AD conversion.
- Established preliminary volumetric thresholds for early AD identification using non-invasive MRI.
- Findings suggest MRI can serve as a prognostic tool for dementia onset, aiding clinical practice.
