Machine Learning for Brain Mapping
Deep learning prediction of progression of atrophy on brain images. The goal of this study is to develop a predictive deep learning biomarker of future early brain atrophy rate in Alzheimer's Disease-related neurogeneration using multimodal imaging.
![Unet](/sites/g/files/dgvnsk7901/files/media/images/unet_fig.png)
Machine learning for multi-parametric mapping
Magnetic resonance vascular fingerprinting quantitatively measures microvascular blood oxygen saturation (SO2), cerebral blood volume (CBV), and vessel radii (R). Matching simulated signals to in-vivo data is computationally expensive, therefore, we leverage deep learning to alleviate the burden. We train a neural network with simulated dictionaries to simultaneously estimate multiple vascular parameters.
![DL_mrvf](/sites/g/files/dgvnsk7901/files/media/images/machinelearning_mrvf.png)