About

Dr. R. Todd Constable is the Co-Director of the Yale MR Research Center and holds the Elizabeth Mears and House Jameson Professor of Radiology and Biomedical Imaging as well as Professor of Neurosurgery. His research in the Magnetic Resonance Imaging Group focuses on technical advances in imaging, including imaging sequences applied to the body and brain, methodology development in molecular imaging, cardiac imaging, and functional brain imaging. The group's mission is to develop novel MR imaging methods with both clinical and basic science applications, addressing fundamental questions about brain function, tissue damage, and recovery. Dr. Constable's laboratory is engaged in examining the relationship between the increment in the functional MR signal measured during tasks and the influence of baseline brain activity on this increment. His projects include understanding negative blood oxygenation level dependent (BOLD) signal changes, the relationship between surface EEG signals and fMRI, and the localization of electrical discharges in epilepsy patients. His work also involves developing new MR imaging strategies for faster and more efficient parallel imaging, with a focus on understanding brain connectivity and function through advanced imaging techniques.

Research topics

• Medicine
• Psychology
• Neuroscience
• Internal medicine
• Psychiatry
• Physical therapy
• Gerontology
• Endocrinology
• Clinical psychology

Selected publications

• Subject-Specific Low-Field MRI Synthesis via a Neural Operator

  ArXiv.org · 2026-03-26

  articleOpen accessSenior author

  Low-field (LF) magnetic resonance imaging (MRI) improves accessibility and reduces costs but generally has lower signal-to-noise ratios and degraded contrast compared to high field (HF) MRI, limiting its clinical utility. Simulating LF MRI from HF MRI enables virtual evaluation of novel imaging devices and development of LF algorithms. Existing low field simulators rely on noise injection and smoothing, which fail to capture the contrast degradation seen in LF acquisitions. To this end, we introduce an end-to-end LF-MRI synthesis framework that learns HF to LF image degradation directly from a small number of paired HF-LF MRIs. Specifically, we introduce a novel HF to LF coordinate-image decoupled neural operator (H2LO) to model the underlying degradation process, and tailor it to capture high-frequency noise textures and image structure. Experimental results in T1w and T2w MRI demonstrate that H2LO produces more faithful simulated low-field images than existing parameterized noise synthesis models and popular image-to-image translation models. Furthermore, it improves performance in downstream image enhancement tasks, showcasing its potential to enhance LF MRI diagnostic capabilities.

  PublisherOA PDF

• 48. Targeting Functional Networks With Brain-Based Psychometric Scales

  Biological Psychiatry · 2026-04-25

  articleSenior author
  PublisherDOI

• Subject-Specific Low-Field MRI Synthesis via a Neural Operator

  arXiv (Cornell University) · 2026-03-26

  preprintOpen accessSenior author

  Low-field (LF) magnetic resonance imaging (MRI) improves accessibility and reduces costs but generally has lower signal-to-noise ratios and degraded contrast compared to high field (HF) MRI, limiting its clinical utility. Simulating LF MRI from HF MRI enables virtual evaluation of novel imaging devices and development of LF algorithms. Existing low field simulators rely on noise injection and smoothing, which fail to capture the contrast degradation seen in LF acquisitions. To this end, we introduce an end-to-end LF-MRI synthesis framework that learns HF to LF image degradation directly from a small number of paired HF-LF MRIs. Specifically, we introduce a novel HF to LF coordinate-image decoupled neural operator (H2LO) to model the underlying degradation process, and tailor it to capture high-frequency noise textures and image structure. Experimental results in T1w and T2w MRI demonstrate that H2LO produces more faithful simulated low-field images than existing parameterized noise synthesis models and popular image-to-image translation models. Furthermore, it improves performance in downstream image enhancement tasks, showcasing its potential to enhance LF MRI diagnostic capabilities.

  PublisherDOI

• Localized Gradients in an Open Field-Cycling Low-Field MRI: First Images

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  articleSenior author

  Motivation: RF spatial encoding has previously been used to acquire images in our open low-field scanner1. To develop breast imaging with this device and achieve higher spatial resolution and lower scan time, localized spatial encoding cup-shaped gradients are used for spatial encoding. Goal(s): To demonstrate imaging feasibility in this non-uniform B0 field cycling magnet using localized gradients. Approach: Non-linear left-right and superior-inferior gradients are optimized on a breast imaging FOV, and a single-shot turbo spin echo pulse sequence is used. Results: The first gradient-encoded images in an open rampable nonuniform electromagnet were demonstrated. Impact: MRI has high breast cancer detection sensitivity, but due to its high cost, breast MRI is rarely conducted. By generating images with low-input-current breast-cup-shaped gradients, we show that affordable low-field MRI is a viable option for breast cancer detection.

  PublisherDOI

• Sequential multi-slice imaging strategy using field-cycling

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  articleSenior author

  Motivation: To reduce scan time without compromising signal-to-noise ratio through multi-slice imaging within each polarization cycle of a novel inhomogeneous B0 field cycling magnet. Goal(s): To demonstrate feasibility of multi-slice imaging acquisition using field cycling for slice selection. Approach: Spin polarization is performed at maximum field strength (0.4T). During the imaging phase, multiple slice modules can be defined during which the field is sequentially modulated to achieve slice selection. Results: Proof of concept for 3-slice multi-slice imaging with field cycling is presented, demonstrating feasibility. Impact: Image quality and scan time are prevalent concerns with low-field systems. Multi-slice imaging is a powerful tool to reduce scan time without compromising SNR or necessitating additional hardware.

  PublisherDOI

• Prospective Multicenter Analysis of Neuroimaging and Behavioral Outcomes in Infants with Craniosynostosis: Initial Results and Proof of Concept Study

  Plastic & Reconstructive Surgery Global Open · 2025-09-26

  articleOpen access
  PublisherDOI

• Supervised brain node and network construction under voxel-level functional imaging

  Imaging Neuroscience · 2025-01-01

  articleOpen access

  Recent advancements in understanding the brain's functional organization related to behavior have been pivotal, particularly in the development of predictive models based on brain connectivity. A major analytical strategy in this domain involves a two-step process by first constructing a connectivity matrix from predefined brain regions, and then linking these connections to behaviors or clinical outcomes. Although some advances considered subject-specific functionally homogeneous nodes without relying on predefined regions of interest (ROIs), all these approaches with unsupervised node partitions predict outcomes inefficiently with independently established connectivity. In this paper, we introduce the Supervised Brain Parcellation (SBP), a brain node parcellation scheme informed by the downstream predictive task. With voxel-level functional time courses generated under resting-state or cognitive tasks as input, our approach clusters voxels into nodes in a manner that maximizes the correlation between inter-node connections and the behavioral outcome, while also accommodating intra-node homogeneity. We rigorously evaluate the SBP approach using resting-state and task-based fMRI data from both the Adolescent Brain Cognitive Development (ABCD) study and the Human Connectome Project (HCP). Our analyses show that SBP significantly improves out-of-sample connectome-based predictive performance compared to conventional step-wise methods under various brain atlases. This advancement holds promise for enhancing our understanding of brain functional architectures with behavior and establishing more informative network neuromarkers for clinical applications.

  PublisherOA PDFDOI

• Multi-echo RF spatial phase encoding for gradient-free imaging in a nonuniform B0-field at low-field

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2024-08-14

  articleSenior author

  The cost and complexity of MR scanners can be significantly reduced by eliminating conventional linear gradient coils, and instead using RF coils for spatial encoding. Here we have developed a novel multi-echo RF phase encoding pulse sequence that exploits the Bloch-Siegert shift for nonlinear spatial encoding in a nonuniform, field-cycling, low-field MR system. Phantoms of varying sizes were successfully imaged using this pulse sequence, demonstrating that this technique can be used to perform gradient-free imaging in an inhomogeneous B0-field at low-field.

  PublisherDOI

• Bayesian subtyping for multi-state brain functional connectome with application on preadolescent brain cognition

  Biostatistics · 2024-11-01 · 1 citations

  articleOpen access

  Converging evidence indicates that the heterogeneity of cognitive profiles may arise through detectable alternations in brain functional connectivity. Despite an unprecedented opportunity to uncover neurobiological subtypes through clustering or subtyping analyses on multi-state functional connectivity, few existing approaches are applicable to accommodate the network topology and unique biological architecture. To address this issue, we propose an innovative Bayesian nonparametric network-variate clustering analysis to uncover subgroups of individuals with homogeneous brain functional network patterns under multiple cognitive states. In light of the existing neuroscience literature, we assume there are unknown state-specific modular structures within functional connectivity. Concurrently, we identify informative network features essential for defining subtypes. To further facilitate practical use, we develop a computationally efficient variational inference algorithm to approximate posterior inference with satisfactory estimation accuracy. Extensive simulations show the superiority of our method. We apply the method to the Adolescent Brain Cognitive Development (ABCD) study, and identify neurodevelopmental subtypes and brain sub-network phenotypes under each state to signal neurobiological heterogeneity, suggesting promising directions for further exploration and investigation in neuroscience.

  PublisherOA PDFDOI

• Supervised brain node and network construction under voxel-level functional imaging

  arXiv (Cornell University) · 2024-07-30

  preprintOpen access

  Recent advancements in understanding the brain's functional organization related to behavior have been pivotal, particularly in the development of predictive models based on brain connectivity. Traditional methods in this domain often involve a two-step process by first constructing a connectivity matrix from predefined brain regions, and then linking these connections to behaviors or clinical outcomes. However, these approaches with unsupervised node partitions predict outcomes inefficiently with independently established connectivity. In this paper, we introduce the Supervised Brain Parcellation (SBP), a brain node parcellation scheme informed by the downstream predictive task. With voxel-level functional time courses generated under resting-state or cognitive tasks as input, our approach clusters voxels into nodes in a manner that maximizes the correlation between inter-node connections and the behavioral outcome, while also accommodating intra-node homogeneity. We rigorously evaluate the SBP approach using resting-state and task-based fMRI data from both the Adolescent Brain Cognitive Development (ABCD) study and the Human Connectome Project (HCP). Our analyses show that SBP significantly improves out-of-sample connectome-based predictive performance compared to conventional step-wise methods under various brain atlases. This advancement holds promise for enhancing our understanding of brain functional architectures with behavior and establishing more informative network neuromarkers for clinical applications.

  PublisherOA PDFDOI

Frequent coauthors

• Cheryl Lacadie

  Resonance Research (United States)

  19 shared

• Rajita Sinha

  9 shared

• Dustin Scheinost

  Yale University

  8 shared

• Dongju Seo

  Yale University

  8 shared

• Janice J. Hwang

  Yale University

  8 shared

• Renata Belfort‐DeAguiar

  Yale University

  7 shared

• Meena M. Makary

  Athinoula A. Martinos Center for Biomedical Imaging

  6 shared

• Daniel S. Barron

  Brigham and Women's Hospital

  6 shared

Labs

• Magnetic Resonance Imaging GroupPI