Multimodal Data Integration

Brain structure data holds significant importance in neuropsychology research. However, brain MRI scans, the primary source of brain structure data, are often limited or absent from existing neuropsychology datasets. Also, the high cost of MRI scanning restricts the amount of data that can be collected in neuropsychology research.

In this project, we aim to develop deep generative models that can predict brain MRI scans (white/gray matter and CSF segmentations) based on the features of human subjects that are available in the target dataset and are potentially relevant to human brain structures, including demographic features, behavioral data, and cognitive test results.

This project is a collaboration between DDIS and other multidisciplinary scholars working in the context of the UZH DynAge URPP https://www.dynage.uzh.ch/en.html.