MRI AT HIGH MAGNETIC FIELDS Kâmil Uğurbil University of Minnesota
CENTER for MAGNETIC RESONANCE RESEARCH (CMRR)
Blood Vessel Distribution in Rat Brain brain slice (ink injection) Venous structure: T 2 *-weighted GE EPI (150 x 150 x 1000 µm 3 ) Ogawa S, Lee T-M, Kay AR, Tank DW.. Proc Natl Acad Sci USA 1990;87:9868-9872.
MAGNETIC FIELD AROUND BLOOD VESSELS 0% deoxy-hb 10% deoxy-hb 30% (Arteries)
BLOOD OXYGEN LEVEL DEPENDENT CONTRAST (BOL One VOXEL in the IMAGE DEOXY-Hemoglobin (Paramagnetic)
INCREASED NEURONAL ACTIVITY INCREASE in REGIONAL BLOOD FLOW (& Volume) LOWER DEOXYHEMOGLOBIN CONTENT per unit volume in the BRAIN if Cerebral Oxygen Consumption (CMRO 2 ) does not increase commensurately
MAGNETIC FIELD AROUND BLOOD VESSELS Increased Neuronal Activity: deoxy-hb 30% deoxy-hb 10%
From Ogawa et al 1992 PNAS paper From Ogawa et al 1992 PNAS Paper. From Kwong et al 1992 PNAS paper Title Ogawa et al PNAS 1992: Figure 2 images superimposed
Opaque anatomical Image In Grey scale SILENT WORD GENERATION
STATE-of-the-ART in 1990 PET Difference Images Raichle, M. Trends in Neurosciences, 2009. 32(2): p. 118-26.
Scales of the Brain
Ocular Dominance Columns OPTICAL IMAGING, MONKEY V2 V1 Ts o et al., Science 1990
Ogawa S, Menon RS, Tank DW, Kim SG, Merkle H, Ellermann JM, Ugurbil K. Biophys J 1993;64(3):803-812. B 0 B 0 B 0 2 13
EXTRAVASCULAR BOLD R 2 * (1/s) Gradient Echo 14 10.5T RATIO of Capillary vs. Venule/Vein 60 & 100 µ 12 7T contribution 10 to FUNCTIONAL 5 µ SIGNAL 8 CHANGES in GE BOLD 1.5T R* 2 (1/s) 6 4 2 Vessel diameter ν vw = κ B 0 f dhb κ = constant f dhb = fraction of dhb in blood Deoxyhemoglobin Change ΔS S O e TE R 2 * ( R 2 * / ν vw ) (κ B 0 TE) Δf dhb 0 0 20 40 60 80 100 120 1 Frequency Frequency ν Difference, on the vessel Hz wall, (Hz) vw B o Functional Signal Change Slope CONSTANT Green lines mark expected values for 30% deoxyhb in all vessels K. Uludağ, B. Müller-Bierl, K. Uğurbil Neuroimage (2009) 48(1): p. 150-65.
EXTRAVASCULAR BOLD (GRADIENT ECHO) RATIO of Functional Signal contribution from the two different vessel types RATIO RATIO ΔS Capillary/Venule of SLOPES 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1.5T 7T 10.5T 5 µ vs. 60 or 100 µ Vessel diameter 20 40 60 80 100 120 1 B o f dexyhb Frequency ν on the vessel wall, (Hz) vw B o ΔS S O e TE R 2 * ( R 2 * / ν vw ) (κ B 0 TE) Δf dhb Functional Signal Change Slope Deoxyhemoglobin Change CONSTANT Green lines mark expected values for 30% deoxyhb in all vessels K. Uludağ, B. Müller-Bierl, K. Uğurbil Neuroimage (2009) 48(1): p. 150-65.
EXTRAVASCULAR BOLD (GRADIENT ECHO) RATIO of Functional Signal contribution from the two different vessel types RATIO RATIO ΔS Capillary/Venule of SLOPES 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1.5T 7T 10.5T 5 µ vs. 60 or 100 µ Vessel diameter 20 40 60 80 100 120 1 B o f dexyhb Frequency ν on the vessel wall, (Hz) vw B o Green lines mark expected values for 30% deoxyhb in all vessels K. Uludağ, B. Müller-Bierl, K. Uğurbil Neuroimage (2009) 48(1): p. 150-65.
EXTRAVASCULAR BOLD GRADIENT Echo (GE) SPIN Echo (SE) R 2 * (1/s) 14 12 10 8 6 4 2 0 1.5T 7T 9.4T 16.4T 0 20 40 60 80 100 120 1 Vessel diameter 60&100 µ 5 µ R 2 2 (1/s) 8 6 4 2 1.5T 7T 9.4T 16.4T Vessel diameter 60 µ 0 0 20 40 60 80 100 120 1 Frequency Difference, on the vessel Hz wall, (Hz) Frequency Frequency on the Difference, vessel wall, Hz (Hz) ν vw ν vw 5 µ 100 µ 2% blood volume Green lines mark expected values for 30% deoxyhb K.Uludağ, B. Müller-Bier, K. Uğurbil Neuroimage (2009) 48(1): p. 150-65.
MR detected Mapping Signals and Physiologic Changes induced by Neuronal activity ΔS/S 4 SE fmri, TE set to tissue T 2 total fmri signal (micro-vasculature) total fmri signal (macro-vasculature) ΔCBV=0 total fmri signal [%] 3 2 1 d) ΔCBV 16% 0 0 2 4 6 8 10 12 14 16 field strength (T) K. Uludağ, B. Müller-Bierl, K. Uğurbil Neuroimage (2009) 48(1): p. 150-65.
Ocular Dominance Columns OPTICAL IMAGING, MONKEY HIGH Field SPIN ECHO fmri (when the resolution is High enough) V2 V1 Ts o et al., Science 1990
Ocular Dominance Columns OPTICAL IMAGING, MONKEY S B 0 TE X S B 0 TE LOW FIELD HIGH FIELD GE fmri x >1 for B 0 < ~ 10T x 1 as B 0 >> 10T Ts o et al., Science 1990
High Res. GE fmri 1.5 T courtesy of Mark Haacke JMRI 9: 475-482 (1999). 0.4x0.4x2 mm 3 High Res. GE fmri 7T 0.5x0.5x2 mm 3
Ocular Dominance (ODC) and Orientation maps in Human V1 (7 Tesla) ODC REPRODUCIBILITY 2 days 3 days Left Right Yacoub, Shmuel, et al. Neuroimage (2007) 37(4): 1161-77
Ocular Dominance (ODC) and Orientation maps in Human V1 (7 Tesla) ODC Orientation Left Phase Right 1 mm 1 mm Yacoub, Shmuel, et al. Neuroimage (2007) 37(4): 1161-77 Yacoub, Harel, Uğurbil PNAS (2008) 105(30): 10607-12
Orientation Domains in the Primary Visual Cortex Monkey Optical Imaging Human fmri (SE, 7 Tesla) ~4 mm ~4 mm Yacoub, Harel, Uğurbil PNAS 2008
Challenge: To Acquire High Field Human Data 4 TESLA Barfuss et.al. NMR Biomed:3(1)1990 From SIEMENS CORPORATE RESEARCH
INITIAL RESULTS from SIEMENS at 4 TESLA Object dimension L ~ λ for RF Barfuss et.al. NMR Biomed:3(1)1990 (DATA from SIEMENS)
4T vs. 7T : B1 NON-UNIFORMITY & SNR T. Vaughan et al MRM 46, 24-30 (2001) 4T 7T 1.5 1.4 2.0 1.4 1.4 Mean SNR (boxes) from 6 identical comparison studies in approximately fully relaxed Gradient recalled echo images
Transmit B 1 Maps with a Volume Coil at 7T 0.06 B 1 (µt/v) 0.04 0.02 0.0
Simulation of B 1 + (300 MHz; 7 Tesla) Deionized Water L ~ λ Yang et al. MRM 47 (5), 982-989 (2002) CMRR / PennState collaboration
Simulation of B 1 + 20 mm NaCl Yang et al. MRM 47 (5), 982-989 (2002) CMRR / PennState collaboration
Simulation of B 1 + 50 mm NaCl Yang et al. MRM 47 (5), 982-989 (2002) CMRR / PennState collaboration
EXCERPT from the ABSTRACT: and in the human head. Temporal progression of the RF field indicates that the standing wave and associated dielectric res- onance occurring in a pure water phantom near 300 MHz is greatly dampened in the human head due to the strong decay of the electromagnetic wave. The characteristic image intensity distribution in the human head is the result of spatial phase distribution and amplitude modulation by the interference of the RF traveling waves determined by a given sample- coil config- uration. The numerical calculation method is validated with
7T Body: Head: λ = 12 cm L ~ 40 20 cm L/λ~1.7 ~ 3.3
7 T B 1 Field Contours in a TEM BODY COIL UNLOADED L/λ > ~ 1 Body ~ 3λ x 6λ LOADED Traveling wave regime Vaughan, J.T., et al. MRM, 2009. 61(1): p. 244-8
7 T B 1 + Field Contours in a TEM BODY COIL Body coil Transmit Only Local array Receive only d, cm 60 LOADED d 36 23 10 Magnet isocenter Body Coil Tx Local Coil Rx Vaughan, J.T., et al. MRM, 2009. 61(1): 244-8
E field B 1 field Simulations by Jinfeng Tian,CMRR, U Minnesota Vaughan, J.T., et al. MRM, 2009. 61(1): p. 244-8
Relative B 1 Phase (experimental) Current carrying element ground plane Van de Moortele et al MRM 2005; 54: 1503-1518
Relative B 1 Phase M k + B 1,k Van de Moortele et al MRM 2005; 54(6): 1503-1518
Relative Transmit B 1 AMPLITUDE 300 MHz [a.u] M M B + + 1,k B 1,k k k Van de Moortele et al Magn Reson Med 2005; 54(6): 1503-1518
Relative Transmit B 1 AMPLITUDE 300 MHz [a.u] M k + B 1,k Van de Moortele et al Magn Reson Med 2005; 54(6): 1503-1518
Relative RECEIVE B 1 AMPLITUDE 300 MHz N j B 1, j N j B 1, j 64 MHz Van de Moortele et al Magn Reson Med 2005; 54(6): 1503-1518
Transmit B 1 Maps with a Volume Coil at 7T 0.06 B 1 (µt/v) 0.04 0.02 0.0
3 Tesla: Actual Flip Angle Map
Multichannel TRANSMIT & RECEIVE Adriany Vaughan, et al, J.T., MRM et al. 2005;53(2):434-445 MRM, 2009. 61(1): p. 244-8 Van Snyder, Moortele C.J., et et al. al MRM, 2005; 2009. 54(6): 61(3): 1503-1518 p. 517-24 Metzger, G.J., et al. MRM, 2008. 59(2): p. 396-409 16x T/R Switches up to 32 Low Noise Preamps. Receive Digital Receiver 16 Independent RF Channels 16 x 1 kw amp 16 Channel RF Safety Monitoring
Multiband RF Pulse Performance: ptx vs. Circularly Polarized L Curves for RF Energy vs. B1 Homogeneity (RMSE) MB=2, 1 spoke 0.6 RF Energy 0.4 0.2 Same fidelity Same energy ~70% Decrease 0.0 0.2 0.4 0.6 0.8 Root Mean Square Error (RMSE) in Transmit B1 X. Wu, S. Schmitter, E. J. Auerbach, S.Moeller, K. Uğurbil, and P-F Van de Moortele ISMRM 2013, #74; MRM 2013; Magn Reson Med 70:630 638 (2013)
Parallel Transmit (ptx) with Multiband Pulses, 7 T: Improving B 1 + Homogeneity for MB2 Wu et al. Magn Reson Med 70:630 638 (2013) Experimental Flip angle map 16 channel GRE sag view Circularly Pol. ptx, 1 spoke
1x16 L curves of peak 10 g SAR vs. excitation error (RMSE) Xu et al 2014 ISMRM 1x8 1x8 1x16 2x8 2x8 2x8, axial 2x8, coronal
Parallel Imaging Performance as a Function of Field Strength Geometry Factor in the Center; FOV 25 cm Geometry Factor 2.8 2.2 1.6 1 1.0 10.5 10.57.5 6 7.5 5 4.5 4 3 1.5 1 B o, Tesla 2 R B o, Tesla 1.5 1 10 7 4 SNR 4.5 PI = SNR 2 full g R 3 4 R 5 Wiesinger F, Van de Moortele PF, Adriany G, De Zanche N, Ugurbil K, Pruessmann KP. Magn Reson Med 2004;52(5):953-964
Higher Fields Favors Parallel Imaging Performance SNR PI = SNR full g R B o, Tesla R (Reduction Factor) Wiesinger et al, NMR Biomed 2006 8/10/15 JT Vaughan-CMRR-UMN 49
7T Cardiac Cine: Parallel Imaging with a local 32-Channel Receive Array R=2 R=5 Snyder, DelaBarre &Vaughan
Proc Natl Acad Sci U S A, 74, 5551-5553 (1977) PCr P i ATPγ + ADP P i ATPγ
HUMAN VISUAL CORTEX 7 Tesla P i PCr ATP Syn. Creatine Kinase P i γ-atp (A-B) SNR Chemical Shift Resolution * * (B) Measurement of Pi + ADP ATP γ-atp H. Lei, et al, PNAS 100, 14409-14 (2003) (A)
Functional MRI Mapping of Hemifield Visual Stimulation [4-13 C] Glu [4-13 C] Glu Slice 1 64 min Slice 2 48 min 32 min 16 min W. Chen, et al. MRM 2001; 45; 349-355 0 min
J. Budde S. Gunamony R. Pohmann MPI, Tübingen 9.4 T vs. 3T 9.4T, 21 slices, projection thickness 14.3mm 3T, 14 slices, projection thickness 14mm
9.4T, 21 slices, projection thickness 14.3mm 3T, 14 slices, projection thickness 14mm 9.4 T vs. 3T MPI, Tübingen
TOF angiography at 7 Tesla at 0.4 mm isotropic resolution S. Schmitter, CMRR
7 Tesla T2 Weighted ANATOMIC IMAGING in The Brain Three-dimensional Segmentation of the Internal Structures of the Human HIPPOCAMPUS Henry, T.R., M. Chupin, S. Lehericy, J. Strupp, M.A. Sikora, Z.- Y. Sha, K. Ugurbi, and P.- F. Van de Moortele: Radiology, 2011; 261(1): p. 199-209. Image acquisition: University of Minnesota, CMRR; Image Segmentation: CNRS UPR 640 - LENA, University Pierre and Marie Curie, Paris, France;Center for NeuroImaging Research, University Pierre and Marie Curie - Paris, France; Neuroradiology, Pitie-Salpetriere Hospital, Paris, France
4 Tesla 7 Tesla
7 Tesla 60
CENTER for MAGNETIC RESONANCE RESEARCH (CMRR)