About

Robin de Graaf, PhD, is a Professor of Radiology and Biomedical Imaging at Yale School of Medicine. His main research focus is the study of cerebral energy metabolism and its relationship to functional activation in human and animal brains. He utilizes NMR spectroscopy, including proton, carbon-13, oxygen-17, and phosphorus-31, as a non-invasive in vivo tool to investigate metabolic processes and fluxes. His work also involves technological and methodological improvements to NMR spectroscopy, such as water suppression, spatial localization, spectral editing, quantification, and shimming. Dr. de Graaf's current research emphasizes addressing challenges and leveraging opportunities of magnetic resonance at very high magnetic fields. This includes developing methods for magnetic field uniformity through dynamic shimming and novel electrical coil arrays, extending 13C NMR techniques for better coverage, sensitivity, and specificity, and advancing software tools for metabolic imaging. His research contributes significantly to biomedical engineering, energy metabolism, and magnetic resonance spectroscopy, with applications in understanding brain function and metabolic disorders.

Research topics

• Computer Science
• Nuclear magnetic resonance
• Biochemistry
• Chemistry
• Materials science
• Physics
• Natural Language Processing
• Biology
• Engineering
• Computational biology
• Medical physics
• Radiology
• Biological system
• Nanotechnology
• Data science
• Epistemology
• Management science
• Medicine
• Psychology
• Linguistics
• Philosophy

Selected publications

• In vivo <scp> ² H </scp> ‐ <scp>MR</scp> spectroscopy and imaging of hepatic metabolic formation of trimethylamine‐N‐oxide

  Magnetic Resonance in Medicine · 2025-04-14 · 2 citations

  articleOpen access

  Abstract Purpose Despite growing evidence of the link between elevated levels of trimethylamine‐N‐oxide (TMAO) and multiple diseases, there is no method with which to spatially monitor its hepatic formation from the interstitially produced trimethylamine (TMA). This study aimed to develop a deuterium metabolic spectroscopy (DMS) and imaging (DMI) approach to detect the TMA‐to‐TMAO metabolism in vivo. Methods The metabolism of 2 H 9 ‐TMA (TMA‐ d 9 ) to 2 H 9 ‐TMAO (TMAO‐ d 9 ) in cells overexpressing the hepatic enzyme flavin‐dependent monooxygenase 3 (FMO3) was monitored in vitro with 2 H‐NMR. Using an ultrahigh‐field (15.2T) MRI scanner, the hepatic metabolism of the orally administered TMA‐ d 9 to TMAO‐ d 9 was studied in mice with DMS and DMI. Results The spectrally resolved 2 H‐NMR peaks of intracellularly produced TMAO‐ d 9 (3.1 ppm) from that of supplemental TMA‐ d 9 (2.7 ppm) could be detected only in cells that overexpressed FMO3. In vivo, DMS and DMI experiments performed after oral administration of TMA‐ d 9 revealed the conversion to high TMAO‐ d 9 levels in the liver of females, which express high levels of FMO3. In contrast, there was no indication of TMAO‐ d 9 production in the liver of males, in agreement with reports of the role of testosterone in downregulating the expression of FMO3. Conclusion This work shows the ability to use 2 H‐MR‐based methodologies to spatially monitor the TMA‐to‐TMAO metabolic pathway in vivo, and thus should be explored further to investigate the role of TMAO in diverse pathologies.

  PublisherOA PDFDOI

• Deuterium MRS for In Vivo Measurement of Lipogenesis in the Liver

  NMR in Biomedicine · 2025-02-24 · 2 citations

  articleOpen access

  ABSTRACT Hepatic de novo lipogenesis (DNL) plays a key role in the pathogenesis of several metabolic diseases that affect the liver. In humans, the detection of deuterium ( 2 H) in triglycerides from very low density lipoprotein collected from blood after administration of deuterated water (D 2 O) is commonly used as an indirect estimate of hepatic DNL. Here, we tested in rats (1) the feasibility to detect 2 H‐labeling directly in liver lipids in vivo by using noninvasive 2 H MRS and (2) to what extent these results correlated with the gold standard measurement of DNL in excised liver tissue. To increase hepatic DNL, half of the animals ( n = 4) underwent a 7‐week dietary intervention in which fructose was provided in drinking water. Deuterium MRS data were acquired from a single voxel placed in the liver. In vivo 2 H MRS data showed 2 H‐labeling in the combined peak of methyl and methylene resonances after 1 week of administrati NBM_70014 on of 5% D 2 O as drinking water. DNL was calculated using 1 H and 2 H NMR data acquired from extracted lipids of excised liver tissue. The 2 H lipid level measured in vivo correlated with the ex vivo estimates of hepatic DNL ( r = 0.81, p = 0.016). These results demonstrate the feasibility of direct detection of deuterium labeling in liver lipids using localized 2 H MRS in vivo and indicate the potential of this approach to measure hepatic DNL. These initial observations provide a basis for the method to be translated and to develop noninvasive, quantitative measurements of hepatic DNL in humans.

  PublisherOA PDFDOI

• Parallel detection of <scp>MRI</scp> and <scp> ¹ H MRSI </scp> for multi‐contrast anatomical and metabolic imaging

  Magnetic Resonance in Medicine · 2025-03-13

  articleOpen access1st authorCorresponding

  Abstract Purpose MRI and MRSI provide unique and complementary information on anatomy, structure, function, and metabolism. The default strategy for a combined MRI and MRSI study is a sequential acquisition of both modalities, leading to long scan times. As MRI and MRSI primarily detect water and metabolites, respectively, the small frequency difference between resonances can be exploited with frequency‐selective RF pulses to achieve interleaved or parallel detection of MRI and MRSI, without an increase in total scan time. Methods Here, we describe the pulse sequence modifications necessary to allow acquisition of T 1 and T 2 ‐weighted MRI and B 0 / B 1 mapping in parallel with MRSI. In general, the MRSI module, including water suppression, can be used unmodified. MRI methods are executed in 3D using 3‐ to 4‐ms frequency‐selective Gaussian RF pulses with acceleration along the third dimension through repetitive small‐angle nutation or multi‐spin‐echo acquisitions. Results Phantom experiments demonstrated artifact‐free 3D MRIs. MRSIs in the absence or presence of MRI elements were identical in sensitivity and spectral resolution (line width) and showed consistent water suppression. Parallel MRI‐MRSI was applied to the brains of tumor‐bearing rats in vivo. High‐contrast, high‐sensitivity metabolic MRSI data at 8 μL nominal resolution was acquired in parallel with 3D T 1 ‐weighted, T 2 ‐weighted, and B 0 / B 1 ‐weighted MRIs for an overall scan duration of 30 min. Conclusion Multi‐contrast MRIs and MRSI can be acquired in parallel by utilizing the small frequency difference between water and metabolites. This opens the possibility for shorter overall scans times, or the acquisition of higher‐resolution or additional contrast MRIs.

  PublisherOA PDFDOI

• Oral Intake of Deuterated Choline at Clinical Dose for Metabolic Imaging of Brain Tumors

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-04-27 · 1 citations

  preprintOpen access

  Abstract Accurate characterization and imaging of brain tumors are essential for effective treatment planning and monitoring. While MRI is widely used because of its high sensitivity for detecting lesions, the range of available types of MR image contrast does not offer high specificity for tumors. Deuterium metabolic imaging (DMI), which combines 2 H magnetic resonance spectroscopic imaging (MRSI) with administration of deuterium-labeled substrates, is a relatively new imaging approach that could provide unique, complementary information to anatomical MRI. Preclinical studies have demonstrated the feasibility of DMI with intravenous (IV) administration of deuterated choline ( 2 H 9 -Cho) for tumor characterization; however, they were performed at doses that exceeded severalfold the daily recommended Cho intake. Here, we investigated the feasibility of oral (PO) administration of 2 H 9 -Cho with a dose set at the recommended upper limit for daily use in humans. DMI was performed in rats with orthotopic glioblastoma tumors following a single, high-dose IV bolus (1 × 285 mg/kg) or low-dose PO administration over three consecutive days (3 × 50 mg/kg). Despite a lower cumulative dose, PO administration resulted in comparable total deuterated Cho ( 2 H 9 -tCho) concentrations in the tumor, and tumor-to-brain image contrast relative to IV administration. Additionally, 2 H and 2D 1 H- 14 N HSQC NMR analyses on excised tumor tissue revealed differences in metabolite contributions to the in vivo 2 H 9 -tCho peak. PO administration led to increased contributions from Cho-derived molecules that were products of tumor metabolism, than during IV infusion of 2 H 9 -Cho. These findings suggest that repeated low-dose PO 2 H 9 - Cho administration can generate high, image contrast between tumor and normal brain, that is predominantly generated by tumor metabolism instead of merely Cho uptake. These results can advance the clinical translation of tCho-DMI as a noninvasive imaging tool for brain tumor characterization by demonstrating the feasibility of an oral intake approach using a clinically relevant dose. Given that Cho is already a widely used and well-tolerated nutritional supplement, oral Cho administration offers a practical, noninvasive alternative to IV infusion that could be conducted alongside regular MRI.

  PublisherOA PDFDOI

• Reproducibility of the Determination of ¹³C‐Labeling of Glutamate and Glutamine in the Human Brain Using selPOCE‐MRS at 7 T Upon [U‐¹³C]–Labeled Glucose Infusion

  NMR in Biomedicine · 2025-04-09 · 1 citations

  articleOpen access

  ABSTRACT Glutamate (Glu) is the major excitatory neurotransmitter in the central nervous system. The measurement of Glu/glutamine (Gln) neurotransmitters in the brain provides valuable insights into the dynamic aspects of neuroenergetics and neurotransmitter cycles and can be accomplished through the detection of 13 C‐labeling of Glu and Gln during the administration of 13 C‐labeled glucose. Our goal is to evaluate the reproducibility of selective proton‐observed, carbon‐edited (selPOCE) MRS at 7 T for the detection of 13 C‐labeled Glu and Gln in the human brain. Data of three healthy participants, who were scanned twice at 7 T while undergoing [U‐ 13 C]‐glucose infusion for 120 min, were used to detect 13 C‐labeled Glu and Gln in the brain, using selPOCE‐STEAM‐MRS. There was a rapid increase of plasma glucose 13 C fractional enrichment (FE) during the first 20 min of infusion, followed by a steady state of plasma glucose 13 C FE until the end of the [U‐ 13 C]‐glucose infusion. The time courses of 13 C‐labeling of Glu and Gln were similar for test/retest. The test/retest variability was 15.8% for 13 C‐Glu and 33.3% for 13 C‐Gln. Knowing the variability of these readings using selPOCE‐STEAM‐MRS can inform the application to future studies on disease‐specific alterations in Glu/Gln cycling.

  PublisherOA PDFDOI

• Multi-echo bSSFP for human cardiac DMI on a clinical 3T MRI scanner

  Journal of Cardiovascular Magnetic Resonance · 2025-01-01

  articleOpen access
  PublisherDOI

• Parallel Detection of Multicontrast MRI and Deuterium Metabolic Imaging for Time-efficient Characterization of Neurologic Diseases

  Radiology · 2025-04-01 · 10 citations

  articleOpen accessSenior author

  Deuterium metabolic imaging was interleaved with fluid-attenuated inversion recovery, T1-weighted, T2-weighted, and susceptibility-weighted MRI to provide a time-efficient clinical imaging protocol that allowed for parallel detection of metabolic and multicontrast anatomical images.

  PublisherOA PDFDOI

• Oral intake of deuterated choline at clinical dose for metabolic imaging of brain tumors

  npj Imaging · 2025-10-24 · 2 citations

  articleOpen access

  Deuterium metabolic imaging (DMI) is a new imaging approach that provides unique, complementary information to anatomical MRI of brain tumors. Preclinical DMI studies have demonstrated excellent image contrast following intravenous infusion of deuterated choline (2H9-Cho) at a severalfold higher dose than recommended for humans. We investigated DMI performance in rat glioblastoma models after oral administration of a 2H9-Cho dose recommended for humans. DMI, following the three daily oral low doses, resulted in 2H9-Cho concentrations in the tumor and tumor-to-normal-brain image contrast comparable to a single, high intravenous dose. Further, ²H and 2D ¹H-14N HSQC NMR on excised tumor tissue revealed that oral administration led to increased contributions from Cho-derived molecules that were products of tumor metabolism compared to intravenous infusion of 2H9-Cho. These results can advance clinical translation of Cho-DMI as a noninvasive imaging tool for brain tumor characterization by demonstrating the feasibility of an oral intake approach using a clinical dose.

  PublisherOA PDFDOI

• Metabolism of Choline and Deuterated Choline Detected by ¹H–¹⁴N 2D Heteronuclear Single-Quantum Coherence (HSQC) NMR

  Analytical Chemistry · 2025-03-20 · 4 citations

  article1st authorCorresponding

  Choline is an essential nutrient that plays a critical role in tumor growth. Choline levels can be detected by proton (1H) MR spectroscopy (MRS) in vivo, whereas active (dynamic) choline metabolism can be studied with deuterated choline and 2H MRS. The detected 1H and 2H choline signals represent the sum of choline, phosphocholine (PC) and glycerophosphocholine (GPC), preventing a detailed characterization of choline metabolism. Here we have developed a two-dimensional (2D) NMR method that allows the simultaneous detection of all protonated and deuterated choline-related compounds in excised tissue. The methodology relies on the high 1H detection sensitivity and chemical shift dispersion to distinguish between choline types, the sensitivity of the 14N chemical shift toward deuteration of nearby methyl groups and the presence of a 1H–14N scalar coupling between 14N and CH2 groups. With optimized sequence parameters, the utility of the method is demonstrated on extracts from cultured cancer cells, blood plasma and rat brain and brain tumor tissues.

  PublisherDOI

• Deuterium Metabolic Imaging Denoising Using a Linear Tangent Space Alignment (LTSA) Model and Performance Analysis

  2025-04-14

  article

  Deuterium metabolic imaging (DMI) is an emerging technique to map metabolism of human body non-invasively. However, conventional DMI acquisition methods are limited by the signal-to-noise ratio (SNR) due to the small gyromagnetic ratio of deuterium and low concentration of deuteriumlabeled metabolites. This work presents a manifold learningbased method that leverages the intrinsic low-dimensional manifold structure of the underlying DMI signals for denoising via a linear tangent space alignment (LTSA) model. Cramer-Rao Lower Bound (CRLB) analysis was performed to characterize the reduction in noise and in the variance of the metabolite concentration estimation achieved by the proposed method, which was validated using numerical simulation studies.

  PublisherDOI

Recent grants

• NIH Grant R21CA118503

  NIH · $327k · 2008

• Robust and highly selective proton MRSI on a clinical 3 T system using a second order gradient insert, for application in schizophrenia

  NIH · $461k · 2023–2026

• NIH Grant R33CA118503

  NIH · $614k · 2010

• Robust 3D MR spectroscopic imaging through multi-coil magnetic field shaping

  NIH · $3.3M · 2012–2022

• NIH Grant R01EB002097

  NIH · $491k · 2006

Frequent coauthors

• Douglas L. Rothman

  Resonance Research (United States)

  184 shared

• Kevin L. Behar

  Yale University

  142 shared

• Henk M. De Feyter

  Yale University

  131 shared

• Graeme F. Mason

  Yale University

  74 shared

• Terence W. Nixon

  Resonance Research (United States)

  73 shared

• Peter B. Brown

  Rhode Island College

  63 shared

• Scott McIntyre

  Resonance Research (United States)

  54 shared

• Christoph Juchem

  46 shared