Computer Vision News Computer Vision News 38 From MICCAI 2024 - DGM4MICCAI Workshop SynthBrainGrow: Synthetic Diffusion Brain Aging for Longitudinal MRI Data Generation in Young People Anna Zapaishchykova is a PhD candidate at the Artificial Intelligence in Medicine (AIM) Lab at Harvard MGB, where she specializes in pediatric brain cancer research through the application of AI and medical imaging; Executive Officer at the MICCAI Student Board; and Editorial Board Trainee at Radiology: Artificial Intelligence. Fresh from an online appearance at MICCAI 2024, she speaks to us about her paper on synthetic diffusion brain aging. In this paper, Anna proposes a method to age structural brain MRI by two years using diffusion brain models. Due to a scarcity of data, the method currently focuses on young people, using the popular ABCD dataset (ages 9-16), and has been tested on the long579 dataset (ages 7-9). “Unfortunately, there is a lack of big longitudinal data,” Anna tells us. “We need a lot of data to train this model. Right now, it would only work on teenagers, but we have plans to expand it forward.” Expanding the method beyond teenage brains could open new avenues for exploring how brains develop over time and have implications for a wide range of neurocognitive studies. The brain’s development is influenced by various factors, such as genetics, environment, and stress, making it develop slightly differently from its typical trajectory. SynthBrainGrow can shed light on these changes, enabling scientists to compare how a synthetically aged brain differs from a real, naturally aged brain.
RkJQdWJsaXNoZXIy NTc3NzU=