Functional MRI

1
Functional MRI

• Jerry Allison, Ph. D.

Medical College of Georgia
2
BOLD Imaging Technique

• Blood Oxygen Level Dependent contrast can be used
  to map brain function

3
Right Hand Motor Task
4
Outline

• fMRI BOLD Contrast
• An fMRI Exam
• Pulse Sequences for fMRI
• fMRI Acquisition Parameters
• fMRI Artifacts

5
fMRI BOLD Contrast

• Neuronal events (cerebral activation)
• O2 consumption increases (5)
• Cerebral blood flow increases (50)
• Oxygen extraction fraction decreases
• Oxygenation increases in venous blood

6
fMRI BOLD Contrast

• Concentration of paramagnetic deoxyhemoglobin
  decreases
• Intravoxel dephasing decreases
• T2 increases
• T2 weighted image intensity increases (
  few )

7
Latency 5 - 8 seconds may elapse between
neuronal activation and T2 changes
8
Idealized fMRI Study
9
Voxel Time Course
10
fMRI Signal from each voxel is characterized by

• ? Phase
• A Amplitude
• ? Frequency
• T2

11
T2 has two components

• 1/T2 ?????B 1/T2
• T2 nmr spin-spin dephasing
• ?????B magnetic field inhomogeneity
• The BOLD phenomenon changes ?????B

12
Magnetic Field Inhomogeneity results from

• Inhomogeneity of the magnets B0 field
• Variation in magnetic susceptibility of patients
  tissues

13
Magnetic Field Inhomogeneity occurs

• Near the boundaries of tissues having disparate
  susceptibility
• tissue/ air
• tissue/ bone
• The sphenoid sinus causes magnetic susceptibility
  artifacts in EPI images
• Inferior frontal cortex
• Inferior lateral temporal cortex

14
Magnetic Field Inhomogeneity also results from

• The BOLD effect
• Differences in magnetic susceptibility around a
  paramagnetic deoxyhemoglobin molecule
• Changes in magnetic susceptibility around a small
  blood vessel (capillary, venule, small vein) that
  has an increased concentration/fraction of
  oxygenated hemoglobin

15
T2 weighted images are used in fMRI to
demonstrate changes in magnetic susceptibility
associated with the BOLD effect
16
Signal in a T2 weighted BOLD image is affected by

• Blood volume (CBV)
• Blood flow (CBF)
• Arterial hemoglobin concentration
• Venous hemoglobin concentration
• Oxygen extraction rate
• Hematocrit

17
Signal in a T2 weighted BOLD image is thought to
be inversely proportional to the number of
deoxyhemoglobin molecules in the voxel
18
T2 Contrast is available via

• Conventional gradient echo techniques with one RF
  transmission per phase encoded line of k-space
• FLASH
• GRASS
• FISP

19
T2 Contrast is available via

• Single shot EPI techniques with a complete survey
  of k-space (and subsequent image reconstruction)
  for each RF transmission
• EPI - SE techniques
• EPI - GE techniques
• Multi-shot techniques (more than one RF
  transmission per image more than one echo per RF
  transmission)

20
EPI ImagesEPI - GE Technique(T2 weighted)
21
To Map Brain Function

• Acquire T2 weighted images during a brain task
• Images will have slightly higher intensity in
  active brain regions
• Acquire T2 weighted images during a control
  state with the brain task suspended
• Statistically subtract control images from task
  images to map areas of brain activation
  associated with the task

22
An fMRI Exam involves

• Brain Activation Paradigm
• Task presentation systems
• Patient response monitoring
• Image acquisition
• Synchronization
• Data processing

23
A brain activation paradigm

• Control task
• Activation task a task designed to produce brain
  activation (e.g. motor, sensory, language,
  memory)
• Try to avoid
• Habituation
• Learning
• Inattention

24
Event Related fMRI

• fMRI need not be constrained by task on/ task off
  block designs
• By measuring BOLD response to brief stimuli
  (typically presented at irregular intervals), it
  is possible to characterize the hemodynamic
  response function

25
(No Transcript)
26
Noun Verb Task
27
Right Hand Motor Task
28
Left Hand Motor Task
29
Auditory Task
30
Task presentation systems

• Audio system
• w/ Input for external sound sources such as
  computer audio, VCR, stereo, microphone
• Attenuate gradient noise while enabling
  communication
• Non-pneumatic audio offers improved quality
• Electrical stimulation
• A dark room and dark magnet bore

31
Task presentation systems

• Visual presentation
• Slide projector
• LCD panel overhead projector rear screen
  projection
• Large screen LCD MRI projection systems
• Electronic goggles
• MRI compatible corrective lenses

32
fMRI Projection System
33
fMRI I/O Devices
34
Response monitoring

• In order to document whether the patient is doing
  the task
• Key pad
• Joystick
• Track ball

35
MRI Control Room
36
Synchronization

• Task presentation and patient response monitoring
  should be in synchrony with the acquisition of
  T2 weighted MRI images
• Ideally, all of this apparatus should be
  controlled by one host.

37
Signal conditioning

• Wires and tubes that pass in/out of the exam room
  must pass through RF filters or RF waveguides

38
Penetration Panel
39
Image acquisition

• T2 weighted images are acquired while the
  patient alternates between periods having an
  activating task and periods of a control state.

40
Motor task

• Begin imaging
• Control state
• Rest for 30 seconds
• Activating task
• Finger tapping w/ Rt. hand for 30 seconds
• Alternate these activities for 6 minutes

41
Motor task

• Acquire 120 T2 weighted MRI image sets
• Each set
• Transverse oblique
• 32 Slices
• Primary motor cortex to cerebellum

42
Motor task

• Thus for each brain voxel, we have temporal data
  (the voxel time course) having 120 data points (a
  sample every three seconds)
• There can be hundreds of thousands of brain voxels

43
Motor task

• 3840 total EPI images
• Some scanners limit the number of images in one
  study (512, 2048, etc.)

44
Sagittal Localizer
45
We use one of two pulse sequences

• ep2d_fid_66b1190_62.ekc
• ep2d_fid_60b2080_62_64.ekc

46
ep2d_fid_66b1190_62.ekc

• Gradient refocused single shot EPI technique
• Uses optional hybrid gradient overdrive
  amplifiers
• Matrix 128x128

47
ep2d_fid_66b1190_62.ekc

• TE 66 msec
• TR 3 sec
• We acquire 22 slices every 5 seconds

48
ep2d_fid_66b1190_62.ekc

• 22 slices
• 2.0 mm
• skip 1.0
• ascending
• Orientation Transverse oblique
• T --gt C, - 15 o

49
ep2d_fid_66b1190_62.ekc

• Phase encode A ? P (for symmetry)
• FOV 230x230 mm
• Voxels 1.8 x 1.8 x 3.0 mm

50
ep2d_fid_66b1190_62.ekc
51
ep2d_fid_60b2080_62_64.ekc

• Enables up to 64 slices in one file
• We do 32 slices every 3 seconds
• 64 x 64 matrix
• All slices are written into 1 512 x 512 image
  (mosaic mode)
• Reduces I/O time necessary to write data to disk

52
ep2d_fid_60b2080_62_64.ekc
53
GRASS fMRI Technique

• TR 70 msec
• TE 40 msec
• Flip angle 40 o
• Matrix 128 x 256
• Slice thickness 6 mm
• Acquisition 9.4 sec per image

54
Computing options

• MRI scanner software
• AFNI
• STIMULATE
• SPM
• Other approaches
• Brain Voyager
• MEDx

55
Data processing may include the following

• Correction for differences in time of acquisition
  across slices
• Examine the data
• Motion detection
• Plot center-of-mass time course for each slice
• Stimulate
• View Cine loop of T2 weighted images
• As native signal intensity
• As a difference image (a. la. DSA)
• fMRI activation in periphery of brain is an
  indication of motion

56
Data processing may include the following

• Motion correction
• Discard images showing obvious motion
• Image re-registration to correct for rigid body
  motion
• 2D (3 parameters)
• 3D (6 parameters)

57
Data processing may include the following

• Baseline flattening
• 0th order
• 1st order
• 2nd order
• fMRI image calculation
• Clustering of activated voxels

58
Data processing may include the following

• Image fusion
• With T2 weighted images
• With T1 weighted images
• Multi Planar Reconstruction of fused images
• Volume rendering of fused images

59
Motion Peripheral Brain Activation
60
3D Image Registration
61
Reference Curve
62
Cross Correlation
63
fMRI on T2 weighted EPI images
64
fMRI on T1 weighted images
65
fMRI on Orthogonal T1 weighted images
66
EPI (single shot) Sequence Problems

• Nyquist ghosts N/2 ghosts caused by odd and even
  echo asymmetries
• Can apply a correctionby measuring the
  asymmetry with a phase reference FID with the
  phase encode gradient switched off just prior to
  each single shot image

67
GHOSTS!
68
Artifacts in fMRI images

• Subject motion
• Bulk
• Physiologic
• Cardiac
• Respiratory
• Acquisition time (long is bad like mammography)

69
Artifacts in fMRI images

• ROI (Are there magnetic susceptibility artifacts
  in the ROI)
• Geometric distortion
• Poor shimming results in additional geometric
  distortion (particularly in gradient recalled EPI
  images)

70
Susceptibility Artifact in Orthogonal T2 images
71
Phantom Susceptibilityin a T2 Weighted Sequence
72
Physiologic Motion

• Bulk motion
• Settling

73
Physiologic Motion

• Respiration
• Susceptibility changes caused by movement of
  chest during respiration (phase changes of 2o -
  6o at 40 msec)
• Slow varying
• Affects inferior images
• Proportional to B0
• Proportional to TE

74
Physiologic Motion

• Cardiac pulsations
• Non-rigid body movement of brain parenchyma
  caused by cardiac pulsations (Brain changes
  shape)
• Image dependent brain deformation in EPI
• Image dependent motion artifact in FLASH
• Can produce artifacts as large as a few of BOLD
  contrast
• Occurs near CSF (Brainstem, Cerebellum)

75
Motion of as little as 0.1 pixels can seriously
degrade fMRI images because the MRI signal
variation from pixel to pixel is larger than the
BOLD effect that is being measured.
76
Motion prevention

• An Ounce of Prevention is Worth A Pound of Cure
• Padding
• Expandable foam (Alpha Cradle)
• Vacuum bags
• Hammock
• Bite bar
• Contour masks

77
Motion prevention

• Real time
• Pressure sensors
• Infrared systems

78
fMRI Head Fixation