Xinhui Li

My name is Xinhui Li (黎欣惠). I am a fifth-year Ph.D. candidate in Electrical and Computer Engineering at Georgia Institute of Technology and Center for Translational Research in Neuroimaging and Data Science (TReNDS), advised by Prof. Vince D. Calhoun and Dr. Rogers F. Silva.

My research sits at the intersection of machine learning and computational neuroscience, developing methods for large-scale neuroimaging analysis to better understand the brain and its disorders.

• Multimodal Representation Learning: I develop multimodal fusion methods that learn latent representations from high-dimensional, multimodal neuroimaging data, with a focus on identifying phenotypic and neuropsychiatric biomarkers.
  • Related work: subspace independent vector analysis, multimodal independent vector analysis, Torch-MISA
• AI for Mental Health: I develop generative models and agentic systems for characterizing and assessing neuropsychiatric disorders, with the goal of improving diagnosis and treatment of mental illness.
  • Related work: functional network connectivity interpolation, AI psychiatrist assistant
• Open and Reproducible Science: As a first-generation scholar, I’m committed to building an open, collaborative research culture. I systematically study variability in neuroimaging preprocessing pipelines and contribute to open-source tools and standards to improve the reproducibility and reliability of neuroimaging analysis.
  • Related work: interpipeline agreement, NMIND consortium
• Neuroimaging Preprocessing: I develop an open-source software toolbox C-PAC for MRI preprocessing and analysis. I also develop deep learning models for brain extraction and segmentation across species, including non-human primate MRI.
  • Related work: DeepBet, DeepSeg

I love art in many forms. I often spend my free time at museums and theaters. I collaborate with talented artists to create artwork related to my research, such as Butterfly Effect and Schizosymphony. Please feel free to reach out if you want to discuss research questions or collaboration opportunities.