📰🧠🔬 ALZHEIMER’S DISEASE STUDY HIGHLIGHTS – Alzheimer’s disease (AD) diagnosis model based on fusion of heterogeneous brain imaging and genomic data by Zhang Z, et al.

This study investigated whether combining brain imaging, genetic, clinical and cognitive data using machine-learning models could improve the early screening of AD.

Key findings:
• Multimodal data produced better predictive performance than genetic data alone.
• The gender-corrected multimodal model achieved approximately 94% accuracy in cross-validation.
• Cognitive and neuropsychological assessment scores contributed most strongly to the model’s predictions.
• Volumes of AD-related brain regions, including the hippocampus, entorhinal cortex and fusiform gyrus, were among the important imaging features.
• Several genetic markers, including IFI27, SERPINA3 and APOE4, also contributed to the model.
• “Ensemble” machine-learning models demonstrated the strongest overall performance.

Key takeaways:
These findings highlight the potential of multimodal data and machine learning to support earlier Alzheimer’s screening. However, further research using larger, balanced and clinically representative datasets is needed before this approach can be applied in clinical practice.

📄 Access the full article at: Zhang Z, Zhang R, Yang W, lv K, Wu M and Xu L (2026). AD diagnosis model based on fusion of heterogeneous brain imaging and genomic data. Front. Neurosci. 20:1719390. https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2026.1719390/full. This is an open-access article under a CC BY 4.0 license.