Welcome to the Molecular Imaging/Magnetic Resonance Technology Lab (MIMRTL)
Located at the University of Wisconsin–Madison, our lab is dedicated to advancing medical imaging through cutting-edge research and innovative technologies. We focus on several key areas, described in more detail below.
Image Reconstruction and Data Modeling
We develop and apply sophisticated models for measured data to reconstruct higher-quality images. Our approaches incorporate physiological information into dynamic datasets, enabling us to extract new quantitative insights. Key research topics include:
- Iterative Reconstruction Algorithms: Improving image quality and reducing artifacts, especially in challenging scenarios like low-dose imaging.
- Statistical Modeling: Creating accurate models of image data to handle incomplete or noisy measurements.
- Machine Learning-Based Techniques: Leveraging deep learning to enhance reconstruction accuracy and speed.
- Motion-Robust Reconstruction: Developing methods to correct for patient movement, ensuring clearer and more reliable imaging results.
These techniques are beneficial in clinical and research settings where data quality may be compromised by noise, motion, or limited sampling.
Machine Learning and Deep Learning
We harness the power of machine learning and deep learning to expand the capabilities of medical imaging. Our work includes:
- Deep Learning-Based Image Reconstruction: Accelerating image acquisition and improving clarity, enabling faster and more precise diagnostics.
- Multimodal Data Integration: Combining data from different imaging modalities, as well as language and vision models, to enrich our understanding of complex biological processes.
- Predictive Modeling: Using large datasets to develop models that assist in image interpretation, disease classification, and personalized treatment planning.
- Embedding Model Development: Investigating and creating advanced embedding techniques for better representation and analysis of high-dimensional medical imaging data.
By applying these AI-driven strategies, we strive to boost diagnostic accuracy and efficiency across a range of medical applications.
Magnetic Resonance Imaging (MRI)
We explore the frontiers of MRI technology, with a particular focus on designing and optimizing novel pulse sequences and radiofrequency (RF) coils.
- Pulse Sequence Development: Pulse sequences are patterns of RF pulses and magnetic field gradients that manipulate nuclear spins. By customizing these sequences, we can selectively highlight different tissue properties, improve image contrast, and reduce scan times.
- RF Coil Design: RF coils are essential for transmitting and receiving signals in MRI systems. We create specialized coil designs to improve signal sensitivity and overall image quality.
- Electron MRI (EMRI): Our team is developing a low-field EMRI system. EMRI detects resonance transitions of unpaired electrons, providing insights into tissue oxygenation and redox status. This approach is especially valuable for understanding hypoxia in cancer.
Positron Emission Tomography (PET)
Our PET research focuses on novel methodologies and techniques to improve sensitivity, resolution, and diagnostic utility. PET imaging involves the use of radioactive tracers that highlight metabolic processes in the body. By integrating PET with other imaging modalities, we gain a more comprehensive view of disease processes:
- PET/CT and PET/MRI: Combining PET with CT or MRI offers both functional and structural information, enhancing diagnostic accuracy.
- Methodology and Technique Development: We design advanced acquisition protocols and data analysis methods to capture more detailed information about physiological and molecular processes.
Stay Updated
Visit our News page to learn more about our recent work, publications, and ongoing projects. We look forward to sharing our latest findings and collaborations with you.
