CAREER : Human-in-the-Loop Visual Analytics for Biomedical Spatial Profiling

Understanding how our bodies fight disease is essential to advancing medicine and increasing the quality of life. Modern technologies, known as spatial profiling, allow us to scan human tissue to derive genetic and physical information about biological processes. But the resulting data is so large and complex that even experts cannot fully explore it with the naked eye. This project will yield novel human-guided artificial intelligence (AI) approaches to steer and interpret artificial intelligence through interactive visual interfaces. The technology will assist researchers in analyzing biological phenomena, including the causes of diseases (notably cancer), immune responses, and the effects of therapies at the cellular level. Since these phenomena rely on the spatial and temporal organization of cells within the tissue, we will develop an interactive biomedical atlas with embedded analytics to visually analyze these structures and communicate scientific findings to other researchers, practitioners, and the general public. A joint educational plan pairs computer science and biology students to learn and build the research tools of the future. Through these efforts, we aim to revolutionize our understanding of human disease and increase life expectancy for all Americans. Lessons learned in human-AI collaboration and spatial data analysis can expand to other disciplines, including earth and space science. This project develops novel human-in-the-loop visual analytics approaches for spatial-profiling data. By combining imaging with spatially referenced data, the project can map complex multimodal data with greater precision. However, the immense scale and complexity of the multi-modal data, paired with the explorative nature of pre-clinical research, make fully automated analysis unreliable. To bridge this gap, this project proposes interactive visualizations that let experts steer and interpret AI-driven insights, organized into four themes: (1) designing scalable algorithms for joint image and omics representation, (2) integrating machine learning into interactive workflows for expert-led spatial and temporal analysis, (3) creating narrative tools to disseminate scientific findings, and (4) unifying these tools into a single framework validated by real-world biomedical data. The outcomes of this research will drive precision medicine, increasing survival rates and improving public health. Advances in human-centered computing will enhance human-AI collaboration for analyzing imaging and omics data, with applications reaching beyond biomedicine into material, geographic, and astronomical sciences. Our educational initiative bridges human-centered computing and biology, launching new bioinformatics curricula to develop a multidisciplinary workforce at the undergraduate and graduate levels. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria. NSF Award ID: 2543389 | Program: 01002930DB NSF RESEARCH & RELATED ACTIVIT,01003031DB NSF RESEARCH & RELATED ACTIVIT,01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Robert Krueger | Institution: New York University, NEW YORK, NY | Award Amount: $335,278 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2543389 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2543389.html