Medical Imaging

Lab 3:
Magnetic
Resonance Imaging
(MRI)
Le Nhat Tan MEng.
Department of Biomedical Engineering
Faculty of Applied Science
HCMUT - VNUHCM
[email protected]
Content

1. Review: Principle & Technical Design, Image

Acquisition.
2. MRI Image Storage (DICOM & NifTI)
3. Practical Session
1. MRI Image Generator
2. fMRI Image Visualization
2
 1. Review
1.1. What is Magnetic Resonance?

Magnetic Properties of Magnetic Resonance

- Protons (especially in hydrogen nuclei) have a magnetic moment When a proton is exposed to a specific frequency of radio waves
because they are charged particles and behave like tiny magnets. (radiofrequency pulse), it absorbs energy and moves to a
higher energy state.
- Most of the time, protons are randomly oriented, but when After the radiofrequency pulse is turned off, the proton relaxes
subjected to a magnetic field, they tend to align along the field. back to its original, lower-energy state, releasing the absorbed
energy as radio waves.
1. Review
1.2. Principle & Technical Design – Magnetic Field
Semiconductor Magnet
• Superconducting magnet made from materials like
niobium-titanium (NbTi), which become
superconductive at very low temperatures
(typically using liquid helium for cooling).

• Field strengths of 1.5T, 3T, or higher (7T in some

advanced research applications). Higher field
strengths provide better image resolution but
may also increase patient discomfort and system
complexity.

• Homogeneity: The magnetic field should be as

uniform as possible throughout the scanner’s
volume.
 1. Review
1.2. Principle & Technical Design – RF Pulses

RF coils can be used as transmitters and receivers

• Transmit Coil: The RF coil is responsible for
transmitting the radiofrequency pulse into the patient
• Receive Coil: receiving the emitting signal from
releasing energy process of relaxing atoms. They can
be: surface coils, phased array coils, volume coils.

[1] Tas, J. M. C., Modelling and analyzing the z-gradient coil in an MRI scanner 30 Apr 2005
1. Review
1.2. Principle & Technical Design – Gradient coil
• To determine the location of the signals
within the body, the MRI system uses
gradient magnetic fields, which allow for
spatial encoding of the signal.
• These gradient fields vary the magnetic field
strength across the body, allowing the
system to pinpoint the exact location from
which each signal originated.
• Gradient coils generate magnetic fields
that vary linearly in space along a specific
direction. These gradients allow MRI
systems to spatially encode the signals,
essentially "tagging" each proton with its
specific location in the body
 1. Review
1.3. MRI Sequences Types – T1 Weighted
Focusing on the longitudinal relaxation (T1 relaxation), refers to the process where protons return to
their original alignment with the magnetic field (recovery of longitudinal magnetization).
Different tissues in the body recover at different rates, resulting in varying signal intensities on T1-
weighted images.

T1-weighted imaging is typically used for anatomical imaging,

as it provides clear and detailed views of the body's structure.
• Fat appears bright: Tissues like fat recover quickly (short T1),
so they appear bright on T1-weighted images.
• Water appears dark: Tissues with high water content, like
cerebrospinal fluid (CSF), have a longer T1 relaxation time and
appear darker.
• Gray matter and white matter in the brain have intermediate
T1 relaxation times and thus appear as varying shades of gray.
• Lesions or tumors may appear bright or dark depending on
their T1 properties and contrast agent usage.
 1. Review
1.3. MRI Sequences Types – T2 Weighted
Focusing on the transverse relaxation time (T2), refers to the time it takes for protons to lose phase
coherence in the transverse plane after being disturbed by an RF pulse.
T2-weighted imaging highlights this difference in relaxation times, providing contrast between various
tissues in the body

T1-weighted imaging is essential imaging technique for detecting

certain types of pathologies, especially in the brain and spinal cord.
• Water appears bright: Tissues with high water content, such as
cerebrospinal fluid (CSF), edema, and certain types of lesions (such
as tumors), have long T2 times and appear bright on T2-weighted
images.
• Fat appears dark: Fat has a shorter T2 relaxation time and loses
coherence more quickly, resulting in a weaker signal and a darker
appearance on T2-weighted images.
• Muscles and other tissues have intermediate signal intensities,
appearing as shades of gray.
• Lesions and inflammation can appear bright due to the high water
content in abnormal tissues.
 1. Review
1.3. MRI Sequences Types – FLAIR
FLAIR (Fluid Attenuated Inversion Recovery) is a specialized MRI sequence that is primarily used to
suppress the signal from cerebrospinal fluid (CSF), which allows for better visualization of brain
abnormalities, especially in conditions like multiple sclerosis (MS), brain tumors, and other neurological
disorders.

FLAIR is a type of inversion recovery (IR) pulse sequence to

selectively nullify the signal from certain tissues. Specifically, it
suppresses the signal from free-flowing water, such as CSF.
=> This makes the surrounding brain tissue more visible and
enhances the contrast of lesions or abnormalities, such as those in
the white matter.
• Bright lesions in the white matter, which could indicate multiple
sclerosis (MS) plaques, edema, or tumors.
• Ventricles are dark (since CSF is suppressed), providing better contrast
between the surrounding brain tissue and any lesions near or adjacent to
the ventricles.
• Cortical areas and gray matter appear dark to medium gray, depending
on the tissue type
2. Medical Image Storage
2.1. DICOM
DICOM (Digital Imaging and Communications in Medicine) is a standard for
handling, storing, printing, and transmitting medical imaging data.

Pixel Representation: The pixel data in

DICOM files is represented in a gray-
scale or RGB format (for color images
like ultrasound or X-rays). The data can
be encoded in different bit depths (e.g.,
8-bit, 16-bit) depending on the type of
imaging modality and the required
resolution.
 2. Medical Image Storage
2.2. NifTI
NIfTI (Neuroimaging Informatics Technology Initiative) is a file format used to
store and share neuroimaging data.

• A NIfTI file contains both image data (in the form of a multi-
dimensional array) and metadata (information about the data
such as voxel size, dimensions, and origin).

• NIfTI is designed to handle multi-dimensional data, especially in

the context of brain imaging, where data is often stored in 3D (for
static scans) or 4D (for time-series data, such as fMRI).

• NIfTI is widely used in neuroimaging for storing structural MRI

scans, functional MRI (fMRI) data, and diffusion tensor
imaging (DTI) data.

Ruizhi Han et al, in IEEE Access, vol. 8, pp. 214646-214657, 2020

 3. Practical Session
3.2. fMRI Image Visualization - Data

Kaggle is a data science competition platform and

online community for data scientists and machine
learning practitioners under Google LLC.

Data from the Coursera Course: Neurohacking In R taught by

Dr. Elizabeth Sweeney , Rice Academy Postdoctoral Fellow,
Ciprian M. Crainiceanu, Professor and John Muschelli III ,
Assistant Scientist
3. Practical Session
3.2. fMRI Image Visualization

1.Download Dataset
2.Load data from DICOM/NifTI file
3. Visualization
4. Analysis
3. Report Requirements

1. Theory in T1, T2, FLAIR Image Re-

construction; DICOM and NifTI image
storage.
2. Discussion on Image Quality of DICOM,
NifTI, and JPG image file.