Saturday 08 March 2025
Scientists have made a significant breakthrough in developing a new approach for predicting Alzheimer’s disease and determining brain age. By combining data from medical images and surface meshes of the brain, researchers were able to create a more accurate model for identifying individuals at risk of developing the condition.
The study used a combination of machine learning techniques, including convolutional neural networks (CNNs) and graph neural networks (GNNs), to analyze data from three public datasets. The first dataset consisted of 875 images from individuals with normal cognitive function and those with Alzheimer’s disease. The second dataset included 82 images from individuals with mild cognitive impairment, while the third dataset contained 207 images from individuals with Alzheimer’s disease.
The researchers used a CNN to extract features from the medical images, such as shape and texture, which were then combined with features extracted from the surface meshes of the brain using a GNN. The resulting feature embeddings were then passed through a multi-layer perceptron (MLP) to predict brain age and diagnose Alzheimer’s disease.
The results showed that the fusion model outperformed standalone models in terms of accuracy and precision. For brain age regression, the mean absolute error was reduced by 1.6% compared to the best-performing standalone model. In addition, the area under the receiver operating characteristic (ROC) curve for Alzheimer’s disease classification increased from 0.79 to 0.86.
The study highlights the potential of combining different data representations and machine learning techniques to improve predictive accuracy. By leveraging the strengths of both medical images and surface meshes, researchers can develop more accurate models for identifying individuals at risk of developing Alzheimer’s disease.
In addition to its clinical applications, this research has implications for our understanding of brain aging and the progression of neurodegenerative diseases. The study suggests that changes in brain shape and structure may be an early indicator of cognitive decline, providing a new avenue for researchers to explore.
The findings of this study are also significant because they demonstrate the potential of fusion models to improve predictive accuracy in other fields. For example, medical imaging datasets often contain different types of data, such as MRI and CT scans, which can be combined using similar techniques to improve diagnostic accuracy.
Overall, this research represents an important step forward in developing more accurate models for predicting Alzheimer’s disease and determining brain age. By combining the strengths of machine learning and neuroimaging, researchers are one step closer to identifying individuals at risk of developing this devastating condition.
Cite this article: “Breakthrough in Predicting Alzheimers Disease Using Combined Medical Images and Surface Meshes of the Brain”, The Science Archive, 2025.
Alzheimer’S Disease, Machine Learning, Medical Imaging, Brain Age, Neuroimaging, Convolutional Neural Networks, Graph Neural Networks, Multi-Layer Perceptron, Fusion Model, Predictive Accuracy
