Unsupervised learning of acquisition variability in structural connectomes via hybrid latent space modeling

1. 1Apply dimensionality reduction techniques to structural connectomes
2. 2Use hybrid latent space modeling to separate acquisition effects from biological variation
3. 3Configure deep learning models to represent high-dimensional connectomes in a low-dimensional space
4. 4Test the models on dMRI data from different sites and scanners
5. 5Run unsupervised learning algorithms to identify patterns in the data

Data scientists and neuroimaging researchers on a team can benefit from this technique to improve the accuracy of their structural connectome analysis, and software engineers can help implement the deep learning models

💡 Hybrid latent space modeling can explicitly separate acquisition-related effects from biological variation in structural connectomes

💡 Separate acquisition variability from biological variation in structural connectomes using hybrid latent space modeling!