why high-field mri?humanbrainmapping.org/files/2016/ed/course materials... · 2016. 6. 14. ·...
TRANSCRIPT
1
Voxel volume 2 x 2 x 2 mm3 = 8
2 x 2 x 2 mm3 1 x 1 x 1 mm3 8 x 8 x 8 mm3 4 x 4 x 4 mm3
Voxel volume 1 x 1 x 1 mm3 = 1
Why high-field MRI?
• Signal-to-noise ratio (SNR)
• Contrast (anatomical & functional)
Benefits of high-field MRI
2
Outline
• High-resolution vascular imaging
• High-resolution fMRI applications
• Anatomical contrast @ 7T MRI
• Clinical applications of @ 7T MRI
Veins are dark due to elevated concentration of deoxyhemoglobin (BOLD)
A T1-weighted 3D gradient echo acquisition.
TR/TE = 32.4/5.5 ms
readout bandwidth 20 kHz
FOV: 20 x 20 x 5mm
Matrix: 256 x 256 x 32
Image resolution: 0.078 x 0.078 x 0.156 mm
San time: acquisition time 4:25 min/scan,
total scan time 35:20 min (n=8).
axial
Cat, 9.4T
coronal
High Resolution Vascular Imaging
Bolan, et al. NeuroImage 2006
axial axial
Excitation slab
Fresh blood that flows into the slice is fully relaxed, and therefore bright
Cat, 9.4T
coronal
In a T1-weighted sequence, long T1 spins are saturated because they don't have time to recover
High Resolution Vascular Imaging (Time-of-flight / inflow effects)
Bolan, et al. NeuroImage 2006
3
axial sagittal Cat, 9.4T
coronal
In a T1-weighted sequence, long T1 spins are saturated because they don't have time to recover
High Resolution Vascular Imaging (Time-of-flight / inflow effects)
Veins are dark due to elevated concentration of deoxyhemoglobin (BOLD)
Arteries are bright due to unsaturated fresh blood fast inflow into slab
Bolan, et al. NeuroImage 2006
Arteries
Vessel Classification
High Resolution Vascular Imaging (Time-of-flight / inflow effects)
In a T1-weighted sequence, long T1 spins are saturated because they don't have time to recover
Cat, 9.4T
Veins are dark due to elevated concentration of deoxyhemoglobin (BOLD)
Arteries are bright due to unsaturated fresh blood fast inflow into slab
Veins
axial
Bolan, et al. NeuroImage 2006
Maximum Intensity Projections (MIP)
MIP of the 3D subtraction (Pre/Post MION) Cat, 9.4T
In-vivo High Resolution Vascular Imaging
Bolan et al., NeuroImage 2006
MIP (subtraction of Pre/Post MION)
4
3D Vessel Reconstruction
Cat, 9.4T
Vessel Classification
Veins Arteries
Bolan et al., NeuroImage 2006
In-vivo High Resolution Vascular Imaging
Weber et al., Cereb. Cortex 2008
Scanning electron micrographs of a vascular corrosion cast from monkey visual cortex (superior temporal gyrus)
Ex-vivo
High Resolution Vascular Imaging (SEM )
Cat, 9.4T Multi slice BOLD fMRI: Voxel size: 0.25 x 0.25 x 0.5 mm
Harel et al, Front Neuroenergetics 2010
BOLD signal and the underlying vascular system
5
Functional ImageAnatomical Image
Ogawa, Tank, Menon, Ellermann, Kim, Merkle, Ugurbil. PNAS.1992
7 Tesla 4 Tesla
Advantages of high-field for fMRI
Yacoub, 2003
Yacoub et al., 2003
SE-BOLD 7T; 1 x 1 x 2 mm
Spatial Specificity of fMRI Signals
Yacoub, 2003
6
Brodmann's "areas" Surface
White Matter
Brodmann, 1909
• Cell size • type • density
Cortical Lamina
gray matter / “tissue”
I
II / III
IV
V
VI
Korbinian Brodmann (1868-1918)
SE-BOLD GE-BOLD
High Resolution BOLD fMRI
0.7
0.1
0.5
0.1
Cat Magnet: 9.4T In-plane Resolution: 150 x 150 µm2
1mm
Harel et al. 2006
Surface
White Matter
Harel et al, NeuroImage 2006
MRI Histology (Nissl)
Harel et al. 2006
7
Low magnification photographs
High magnification Histology (Nissl)
I II & III IV V VI
Harel et al, NeuroImage 2006
Harel et al. 2006 Cat, 9.4T
GE-BOLD
Layer Specificity of BOLD fMRI (Cat, 9.4T, ∆S)
Harel et al, NeuroImage 2006
coronal
axial
In-vivo High Resolution Vascular Imaging
MIP (subtraction of Pre/Post MION)
Cat, 9.4T Bolan et al, NeuroImage 2006
8
Harel et al. 2006 Cat, 9.4T
SE-BOLD GE-BOLD
Harel et al, NeuroImage 2006
Layer Specificity of BOLD fMRI (Cat, 9.4T, ∆S)
Functional ImageAnatomical Image
Functional-units (cortical columns) are believed to be the basic computational unit in the brain
Ogawa, Tank, Menon, Ellermann, Kim, Merkle, Ugurbil. PNAS.1992
Surface
White matter
Neurons with similar response properties tend to cluster together in ‘columns’ extending through the entire cortex
LGN
9
Optical Imaging - Monkey
Recording Chamber
Ts’o et al., 1990
LR LRLR
Optical Imaging - Monkey
Monocular stimulation
Ocular Dominance Columns – 2DG
(Sokoloff et al., 1976) radioactively labeled glucose
Dechent & Frahm, 2000 Menon et al., 1997 Cheng et al., 2001 Goodyear & Menon, 2001
Horton & Hedley-Whyte, 1984; Horton, 2006
Anatomical (post-mortem) ODC Spatial Organization
Cheng et al., 2001
Functional (fMRI)
10
SE-BOLD 7 T
1 mm
Monocular stimulation
Yacoub et al., 2007
Image resolution: 0.5 x 0.5 x 3 mm
ODC Spatial Organization
Cheng et al., 2001
3 cm EPI Image
Electrophysiological Recording
Hubel & Wiesel, 1968
Cell Discharge
“Ice cube” model
orientations ODC
Optical Imaging - Cat
Bonhoeffer & Grinvald, 1991; 1993; Blasdel 1992
“Pinwheels” “Ice cube” model
Hubel & Wiesel, 1968 orientations ODC
11
0°
Phase Map - fMRI
1 mm
2π
Pha
se
2π
0°
Does the fMRI-Phase Map Reveal Orientation Specific Activation
Zones?
Bonhoeffer & Grinvald, 1991; 1993
Orientation Map – Optical Imaging Phase Map - fMRI
1 mm
Pha
se
1 mm
2π
0°
Yacoub, Harel & Ugurbil, PNAS 2008
12
Pha
se
fMRI - human
Optical Imaging - cat
2π
0°
1 mm
Yacoub, Harel & Ugurbil, PNAS 2008
ODC Phase Map
1 mm 1 mm
Scalebar = 0.5 mm
Ipsi- C
ontra-
Phase 0° 2π
Phase Map ODC Map Phase Map + ODC borders
Yacoub, Harel & Ugurbil, PNAS 2008
13
Yacoub, Harel & Ugurbil, PNAS 2008 Obermayer & Blasdel, JNS 1993
~4 mm ~4 mm
Zimmermann et al., 2011
Anatomical Imaging at Ultra-high Field MRI
14
1.5 T (clinical) 7 T
Noam Harel, University of Minnesota / CMRR
Benefits of high-field (7 Tesla) MRI
7T, T2W, 0.4 x 0.4 x 2 mm3
Benefits of High-field MRI
Susceptibility-Weighted Imaging (SWI)
Duyn J et al. PNAS 2007;104:11796-11801
Phase contrast in the primary visual cortex
15
T2W T1W SWI
Anatomical contrast @ 7 Tesla
STN: subthalamic nucleus SN: substantia nigra RN: red nucleus
STN
SN RN
Abosch, et al., Neurosurgery 2010
STN = Subthalamic Nucleus SN = Substantia Nigra
Susceptibility-Weighted Imaging @ 7T
STN
SN
Magnet: 7T Resolution: 0.4 x 0.4 x 0.8 mm
SWI @ 7T (In-vivo)
Schaltenbrand and Wahren Atlas (Ex-vivo)
Lamina pallidi medialis
GPi GPe
GPi
GPe
GP = Globus pallidus
Susceptibility-Weighted Imaging @ 7T
Abosch, et al., Neurosurgery 2010
16
Detections of Brain Structures with 7 Tesla MRI
Patient-Specific Anatomical Model
SWI
T2W
Thalamus (Tha) Subthalamic nucleus (STN) Substantia nigra (SN) Red nucleus (RN) Globus Pallidus GPi GPe
Put GPe
IC
VIM
Ca
SN
ZI STN Successful DBS surgery is
critically dependent on precise placement of DBS electrodes into target structures
McIntyre et al., Brain Stimulation 2014
Limitation of Current Procedure
2D - Stereotactic atlas Microelectrode recording (MER)
Patient awake during surgery
Success of DBS surgery is critically dependent on the precise placement of the electrodes into the target structures
17
7T MRI Patient-Specific Model Shape and Size Variability of STN
PD 031 PD 032 PD 033
PD 034 PD 036 PD 037
STN
SN
RedN
Location, shape, size & orientation
one-size-fits-all does not apply!
PD 031
PD 032
PD 033
PD 034 PD 036
PD 037
STN
SN
RedN
7T MRI Patient-Specific Model Shape and Size Variability of STN
Duchin, Sapiro, Vitek, Harel
RedN SN
STN DBS
DBS lead
18
Summary
• High-resolution vascular imaging
• High-resolution fMRI applications
• Anatomical contrast @ 7T MRI
• Clinical applications of @ 7T MRI
U of Minnesota / CMRR Essa Yacoub Patrick Bolan Steen Moeller
Christophe Lenglet Yuval Duchin Remi Patriat Kamil Ugurbil
Neurosurgery Aviva Abosch Jon McIver
Michael Park
Duke University Guillermo Sapiro
Neurology Jerrold Vitek Scott Cooper
Work supported in part by the National Institutes of Health (R01NS085188, P41RR08079, P30 NS057091, R01 NS081118) the W.M. Keck Foundation, and MIND institute.