Early Detection of Neurodegenerative Disease via AI-Decoded Microglial Activation from 3D Brain Imaging

Jachin Thilak, Sai Ganbote, Sinchana Joshi, Tejshree Ratnakar

Abstract: Neurodegenerative diseases like Alzheimer?s and Parkinson?s often progress silently for years before diagnosis, reducing treatment efficacy. This study proposes an AI-powered diagnostic tool using a 3D Vision Transformer (ViT3D) to detect early microglial activation from PET and fMRI imaging. The model was trained on synthetic PET datasets containing simulated inflammation hotspots and achieved a validation ROC-AUC of 0.99, outperforming conventional methods. This work highlights the potential of attention-based deep learning to identify early neuroinflammatory changes, offering a non-invasive pathway for preclinical screening and intervention in neurodegenerative conditions.

Keywords: Microglial activation; Neuroinflammation; 3D Vision Transformer (ViT3D); PET/fMRI neuroimaging; Preclinical detection of neurodegeneration

How to Cite?: Jachin Thilak, Sai Ganbote, Sinchana Joshi, Tejshree Ratnakar, "Early Detection of Neurodegenerative Disease via AI-Decoded Microglial Activation from 3D Brain Imaging", Volume 14 Issue 10, October 2025, International Journal of Science and Research (IJSR), Pages: 1371-1375, https://www.ijsr.net/getabstract.php?paperid=SR251022062156, DOI: https://dx.doi.org/10.21275/SR251022062156