TRD4: Platforms for Multimodal and Multi-scale Imaging Data

Principal Investigators: Andreia Faria, PhD, Jeremias Sulam, PhD, Michael Miller, PhD
Co-investigators: Tilak Ratnanather, Laurent Younes

Over the past two decades, brain MRI has seen major advances in acquisition, processing, and analysis. Yet, critical challenges remain—particularly in integrating information across macro-, meso-, and micro- scales and developing predictive models that require more advanced tools. Rapidly evolving data science technologies, including deep learning, offer new opportunities to address these gaps, but demand not just specialized tools, but unified frameworks for data integration and exploration.

Recognizing this need, TRD4 has developed tools to integrate and analyze multimodal, multiscale brain MRI using data science and deep learning. These include frameworks for image analysis, lesion detection, and predictive modeling, enabling efficient pipelines for translational research. TRD4 has also supported collaborative projects, services, and the broader neuroimaging community by deploying key technologies with a strong emphasis on accessibility, reproducibility, and scalability. We believe this is an ideal moment to invest in hybrid approaches that combine conventional engineering with data-driven methods to improve precision, accuracy, and efficiency in neuroimaging research.

Contents:

• Statistical models of anatomical structures based on learned imaging priors for image reconstructions
• Technologies to integrate multi-scale, multimodal, and time-domain data
• Expansion of multimodal MRI analysis with integrated AI tools and data objects for clinical and biomarker research