Deep Learning Model for Prediction of Progressive Mild Cognitive Impairment to Alzheimer's Disease Using Structural...

Area of Science:

• Neurology
• Medical Imaging
• Artificial Intelligence

Background:

• Alzheimer's disease (AD) is a leading cause of dementia, characterized by progressive memory loss and cognitive decline.
• Early detection of AD and mild cognitive impairment (MCI) is crucial for timely intervention and disease management.
• Current diagnostic methods can be invasive or time-consuming, highlighting the need for efficient, non-invasive techniques.

Purpose of the Study:

• To develop a computer-aided deep learning method for distinguishing AD and MCI from cognitively normal individuals using structural MRI (sMRI).
• To evaluate the efficacy of different deep learning models, including a custom CNN, VGG-16, and ResNet-50, for multiclass classification of neurodegenerative conditions.

Main Methods:

• Utilized a multiclass classification approach with 3D T1-weight brain sMRI images from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database.
• Extracted axial brain slices and processed them using Convolutional Neural Networks (CNNs).
• Employed pre-trained VGG-16 and ResNet-50 models as feature extractors, with a densely connected classifier for final classification.

Main Results:

• The study successfully developed a deep learning model capable of classifying individuals into AD, MCI, or cognitively normal categories based on sMRI data.
• Performance metrics indicated the potential of the proposed method for accurate differentiation between these cognitive states.
• Transfer learning using ImageNet-trained models (VGG-16, ResNet-50) demonstrated effectiveness in feature extraction for neuroimaging analysis.

Conclusions:

• Deep learning models, particularly when leveraging pre-trained networks, show significant promise for the early and accurate detection of Alzheimer's disease and mild cognitive impairment from structural MRI.
• This computer-aided approach offers a non-invasive and potentially scalable tool to aid clinicians in diagnosing neurodegenerative conditions.
• Further research and validation are warranted to integrate this technology into clinical practice for improved patient outcomes.