MRI: FOUNDATIONS

9.1 Historical Introduction of MRI

What is MRI? (Essence)

Magnetic Resonance Imaging (MRI) is a medical imaging technique that produces high-quality images of the inside of the human body, especially soft tissues.

MRI is based on Nuclear Magnetic Resonance (NMR)—a spectroscopic method originally used by chemists and physicists to study molecules, not patients.

From Chemistry to Medicine (Timeline Logic)

• 1946 – Felix Bloch & Edward Purcell independently discovered magnetic resonance
  → Nobel Prize in 1952

• 1950–1970 – NMR used for chemical and molecular analysis, not imaging

• 1971 – Raymond Damadian
  → Discovered different relaxation times between normal tissue and tumors
  → This was the medical turning point

• 1973

  • CT introduced by Hounsfield
    → Proved hospitals would invest in expensive imaging

  • Paul Lauterbur produced the first MR image using gradients

• 1975 – Richard Ernst
  → Introduced Fourier Transform, phase and frequency encoding
  → This is the foundation of modern MRI

• 1977

  • Damadian demonstrated early MRI scanning

  • Peter Mansfield developed Echo-Planar Imaging (EPI)

• 1980–1986

  • Body MRI achieved

  • Scan time reduced from 5 minutes → 5 seconds

• 1987

  • Real-time cardiac MRI

  • Magnetic Resonance Angiography (MRA) without contrast

• 1992

  • Functional MRI (fMRI) developed
    → Brain function mapping

• 2003

  • Lauterbur & Mansfield awarded Nobel Prize in Medicine

Why MRI? (Compared to X-ray and CT)

X-ray

• Excellent spatial resolution

• Poor contrast resolution

• Best for bone

CT

• Better contrast than X-ray

• Still limited for soft tissue

MRI (Key Advantage)

• Superior soft tissue contrast

• Multiplanar imaging (any plane without moving the patient)

• No ionizing radiation

• Adjustable parameters for specific pathology

9.2 MRI Hardware (Big Picture)

An MRI scanner is built around one central requirement:

Hardware choice depends on:

• Purpose

• Image quality needed

• Cost

• Patient comfort

9.3 Magnet Types

Two Ways to Create Bo (Main Magnetic Field)

1. Permanent Magnets

2. Electromagnets

   • Resistive

   • Superconducting

9.3.1 Permanent Magnets

• Made of ferromagnetic material

• Field strength: up to 0.4 Tesla

• Open design

• Advantages:

  • Comfortable

  • Less claustrophobia

• Disadvantages:

  • Heavy

  • Lower field strength → lower image quality

9.3.2 Electromagnets

A. Resistive Magnets

• Wire coils with large electric current

• Field strength: up to 0.6 Tesla

• Advantages:

  • Cheaper

  • Open design

• Disadvantages:

  • High power consumption

  • Heat generation

  • Cooling required

B. Superconducting Magnets (Most Important)

• Made of niobium-titanium / niobium-tin

• Cooled with liquid helium

• At 4 K (−269°C) → zero resistance

• Field strength: up to 12 Tesla

• Clinical standard: 1.5 Tesla

Advantages:

• High field homogeneity

• High image quality

• Stable magnetic field

Disadvantages:

• Expensive

• Bore (tunnel-like) design

9.4 Shimming (Field Perfection)

MRI demands a perfectly uniform magnetic field
→ Variations must be within parts per million (ppm)

What is Shimming?

Shimming = fine adjustment of the magnetic field to improve homogeneity.

Two Types

1. Passive Shimming

   • Steel pieces added

   • Mechanical correction

2. Active Shimming

   • Electric currents through shim coils

   • Electrical correction

9.5 RF Coils (Signal Makers)

RF coils:

• Transmit RF pulses

• Receive MR signals

They directly affect image quality.

9.5.1 Volume Coils

• Large field of view

• Homogeneous RF field

• Used for:

  • Whole body

  • Head

• Lower SNR for small regions

9.5.2 Surface Coils

• Placed directly over anatomy

• High SNR

• High resolution

• Limited depth penetration

9.5.3 Quadrature Coils

• Two coils at 90°

• √2 signal improvement

• Circular polarization

• Most modern volume coils use this design

9.5.4 Phased Array Coils

• Multiple surface coils combined

• Large coverage + high SNR

• Used in modern MRI

• Best for spine, body imaging

9.6 Other Hardware

Faraday Cage

• Shields MRI from radiofrequency interference

• Prevents artifacts

• Protects outside equipment

9.7 Atomic Structure (Why Hydrogen?)

Atoms

• Made of:

  • Nucleus (protons + neutrons)

  • Orbiting electrons

Hydrogen

• Most abundant in the body

• Present in water and fat

• Has one proton

• Ideal for MRI because:

  1. High abundance

  2. High gyromagnetic ratio

9.8 Magnetization

A spinning charged particle creates a magnetic field
→ Like Earth itself

Hydrogen protons act as tiny bar magnets

When placed in Bo:

• Protons align:

  • Parallel (low energy)

  • Anti-parallel (high energy)

More choose low energy
→ Net magnetization forms

9.9 Larmor Frequency (Very Important)

Equation

ω0=γB0\omega_0 = \gamma B_0ω0​=γB0​

• ω₀ = Larmor frequency

• γ = Gyromagnetic ratio (Hydrogen = 42.57 MHz/T)

• B₀ = Magnetic field strength

Example:

• 1.5T → 63.86 MHz

Net Magnetization Vector

• Aligned along Z-axis

• Called longitudinal magnetization (M₀)

9.10 In-Phase and De-Phase (Foundation of Signal)

• In-phase → spins aligned → strong signal

• Out-of-phase → spins misaligned → signal loss

Final Old-Tutor Advice

Don’t memorize MRI.
Understand the logic:

Magnet → Protons → RF → Signal → Image

9.11 RF PULSE — THE MOMENT EVERYTHING CHANGES

Before RF:

• Protons precess around B₀

• Net magnetization points along Z-axis

• No signal. No image.

Now listen carefully.

What is an RF pulse, really?

An RF pulse is:

• A small magnetic field

• Applied perpendicular to B₀ (X or Y direction)

• Oscillating at the Larmor frequency

No match = no effect.

Why oscillating, not static?

• A static field would cause chaotic motion

• An oscillating field at the exact frequency causes a clean rotation

This clean rotation tips the net magnetization away from Z.

90° RF Pulse

• Rotates net magnetization from Z-axis into XY-plane

• Produces maximum transverse magnetization

• Starts:

  • T2 decay immediately

  • T1 recovery immediately

9.12 EXCITATION — ENERGY IN

Before acquisition

A short pre-scan determines:

• Exact Larmor frequency

• System calibration

This is necessary because:

• Field strength slightly varies

• Precision matters in MRI

What excitation really means

• Protons absorb RF energy

• They move to a higher energy state

• Net magnetization is flipped

Flip Angle (FA):

• Controlled by:

  • RF pulse strength

  • RF pulse duration

• Can be 1° to 180°

• Standard teaching example: 90°

Resonance analogy (important)

Like:

• Opera singer shattering glass

• Only the correct frequency works

Wrong frequency → nothing happens

9.13 RELAXATION — ENERGY OUT

Now we reach the heart of MRI contrast.

Relaxation =
Return to equilibrium

Protons:

• Do not like high energy

• Always seek lowest energy state

Two independent processes happen simultaneously:

9.13.1 T1 RELAXATION (SPIN–LATTICE)

What T1 describes

• Recovery of longitudinal magnetization (MZ)

• Energy transferred to surrounding tissue (lattice)

Immediately after 90° pulse:

• MZ = 0

• All magnetization in XY-plane

With time:

• MZ recovers

• System returns to equilibrium

T1 definition

Mathematically:

Mz=M0(1−e−t/T1)M_z = M_0 (1 - e^{-t/T1})Mz​=M0​(1−e−t/T1)

At:

• t = T1 → 63%

• t = 5T1 → ~99%

Why tissues differ

• Fat: tightly bound → short T1

• Water / CSF: loosely bound → long T1

This difference = contrast

9.13.2 T2 RELAXATION (SPIN–SPIN)

This is where students panic.
Don’t.

Key rule (memorize this sentence):

What happens right after 90° pulse

• Spins are:

  • In XY-plane

  • In-phase

• Strong signal is produced

Immediately:

• Spins begin to lose phase coherence

• Transverse magnetization decays

This loss of coherence = T2 relaxation

Why spins de-phase

• Protons interact with neighboring protons

• Local magnetic fields differ slightly

• Some spins:

  • Speed up

  • Slow down

End result:

• Vectors point in all directions

• Net transverse signal → zero

T2 definition

Equation:

Mxy=M0e−t/T2M_{xy} = M_0 e^{-t/T2}Mxy​=M0​e−t/T2

Important truths

• T2 is:

  • Much faster than T1

  • Tens of milliseconds

• Fat:

  • Dephases quickly → short T2

• Water:

  • Dephases slowly → long T2

9.13.2.2 T2 RELAXATION (IMPERFECTION EFFECT)*

Now reality interferes.

T2* includes:

• True T2 effects

• PLUS:

  • Magnetic field inhomogeneity

  • Susceptibility differences

  • Chemical shift

  • Gradients

Relationship:

T1>T2>T2∗T1 > T2 > T2*T1>T2>T2∗

Or:

1T2∗=1T2+γΔB\frac{1}{T2*} = \frac{1}{T2} + \gamma \Delta BT2∗1​=T21​+γΔB

Spin-Echo vs Gradient-Echo

• Spin Echo (180° pulse):

  • Removes inhomogeneity effects

  • Shows true T2

• Gradient Echo:

  • No refocusing pulse

  • Shows T2*

9.14 ACQUISITION — LISTENING TO THE BODY

During relaxation:

• Protons emit RF signals

• Receive coils detect them

Critical rule

Why?

• If B₀ passes through coil → huge induced current

• RF signal is drowned in noise

This is Faraday’s Law, not superstition.

FID (Free Induction Decay)

• Signal immediately after excitation

• Strong at first

• Rapidly decays due to T2*

9.16 FOURIER TRANSFORM — TRANSLATION

MRI data is first:

• In time domain

Image needs:

• Frequency domain

Fourier Transform:

• Converts time → frequency

• Separates signals by:

  • Frequency

  • Phase

  • Amplitude

9.17–9.19 GRADIENTS — LOCATION, LOCATION, LOCATION

If B₀ were perfectly uniform:

• All protons spin the same

• No spatial information

• No image

Gradients solve this.

Three gradients

Slice Selection

• Gradient ON + RF pulse

• Only spins with matching frequency respond

• Slice thickness depends on:

  • Gradient strength

  • RF bandwidth

Thin slice:

• Better detail

• Lower SNR

Phase Encoding

• Applied briefly

• Creates phase differences

• Repeated many times

• Determines image resolution

Frequency Encoding

• Applied during signal readout

• Creates frequency differences

• Allows localization using Fourier Transform

9.20 K-SPACE — THE TRUTH BEHIND THE IMAGE

K-space:

• Raw data matrix

• NOT an image

• Center:

  • Contrast information

• Edges:

  • Detail (resolution)

When K-space is full:

• Fourier Transform → image appears

FINAL MASTER SUMMARY (THIS IS GOLD)

• RF pulse → excitation

• T1 → energy loss (Z)

• T2 → phase loss (XY)

• Gradients → location

• Fourier → image

• K-space → raw truth

9.21 GRADIENT ECHO (GRE) PULSE SEQUENCE

(Understand this once, and half of MRI sequences stop being scary)

What is a Pulse Sequence?

A pulse sequence is simply:

It shows:

• RF pulses

• Gradient pulses

• Signal acquisition
  all plotted against time

Think of it as a musical score for the scanner.

Why Gradient Echo?

Gradient Echo (GRE):

• Uses gradients, not a 180° RF pulse, to form the echo

• Faster than Spin Echo

• More sensitive to T2* effects

Key Features of GRE (Memorize This Table)

Why Small Flip Angle?

Two consequences (very important):

1. Smaller signal

2. Shorter TR possible

Short TR → short scan time

Step-by-Step GRE Sequence Logic

1. RF pulse + Slice Select Gradient (GS)

   • Excites a slice

   • Introduces phase dispersion

2. Slice Rephasing Gradient

   • Corrects slice-selection dephasing

3. Phase Encoding Gradient (GPh)

   • Changes from TR to TR

   • Each step fills one k-space line

4. Readout Dephasing Gradient

   • Prepares spins for echo

5. Readout Gradient (GRO) + Acquisition

   • Signal sampled

   • Echo peak occurs at TE

This entire cycle fills one line of k-space.

Very Important Exam Line

Repeat this 256 times → k-space full → Fourier Transform → image.

9.22 GRADIENT SPECIFICATIONS (WHY HARDWARE MATTERS)

If MRI were a car:

• Gradients are the engine

Three Key Gradient Specs

Why They Matter

• Faster gradients → shorter TE, TR

• Enables:

  • Fast imaging

  • EPI

  • DWI

  • fMRI

9.23 MRI IMAGE QUALITY

(This is where marks are gained or lost)

Image quality depends on three pillars:

1. Signal-to-Noise Ratio (SNR)

2. Contrast Resolution

3. Spatial Resolution

Signal-to-Noise Ratio (SNR)

Noise comes from:

• Patient (thermal motion)

• Electronics

• Coils

• Field inhomogeneity

Fixed vs Controllable Factors

You cannot change:

• Magnet imperfections

• Patient size

• Electronics noise

You can change:

• RF coil

• Voxel size

• NEX

• Bandwidth

• TR / TE / Flip angle

Golden Rule

9.23.2 PIXEL, VOXEL, MATRIX (ABSOLUTE EXAM FAVORITE)

• Pixel → 2D picture element

• Voxel → 3D tissue volume

• Matrix → grid of pixels

Voxel size =

FOVMatrix×Slice Thickness\frac{FOV}{Matrix} \times Slice\ ThicknessMatrixFOV​×Slice Thickness

Smaller voxel

• Higher spatial resolution

• Lower SNR

Larger voxel

• Lower resolution

• Higher SNR

9.23.3 INTER-SLICE GAP (WHY GAPS EXIST)

Slices are not perfectly rectangular.

Problem:

• RF excites adjacent slices

• Causes cross-talk

• Reduces SNR

Solution:

• Inter-slice gap (25–50%)

9.23.4 MATRIX & FOV EFFECTS

Matrix Size

• 128×128 → coarse

• 512×512 → fine

Fine matrix:

• Better detail

• Worse SNR

• Longer scan time

Field of View (FOV)

Doubling FOV:

• Voxel volume ↑ ×4

• SNR ↑ ×4

Halving FOV:

• SNR ↓ ×4

9.23.5 TR, TE, FLIP ANGLE (CONTRAST CONTROLS)

TR (Repetition Time)

• Controls T1 contrast

• Long TR → high SNR

• Short TR → T1 weighting

TE (Echo Time)

• Controls T2 contrast

• Long TE → T2 weighting

• Short TE → high SNR

Flip Angle

• Large flip → higher sig

• nal

• Small flip → faster scan

Classic Rule

• Long TR + Short TE → best SNR

• Short TR + Long TE → worst SNR

9.23.6 NEX (AVERAGES)

• SNR ∝ √NEX

• Double NEX → SNR ×1.41

• Scan time doubles

9.23.7 RECEIVER BANDWIDTH

• Lower bandwidth → higher SNR

• But:

  • Longer readout

  • Longer minimum TE

  • More chemical shift

9.24 MRI CONTRAST AGENTS

Why We Need Them

Sometimes tissues look too similar.

MRI contrast agents:

• Alter relaxation times

• Do NOT “block X-rays” like CT contrast

Two Main Types

Mechanism

Contrast agents:

• Interact with water protons

• Increase relaxation rate

1/T=1/T0+rC1/T = 1/T_0 + rC1/T=1/T0​+rC

9.25 SPECIAL APPLICATIONS (HIGH-YIELD)

MRA / MRV

• Visualize blood flow

• Flowing blood = bright

• Uses GRE

• Can be:

  • Non-contrast

  • Gadolinium-enhanced

MR Myelography

• Heavy T2 weighting

• CSF = bright

• Cord = dark

• Replaced X-ray myelography

MRCP

• Heavily T2-weighted

• Bile & pancreatic ducts = bright

• No contrast needed

Chemical Shift Imaging (CSI)

• Spectrum per pixel

• Combines MRI + spectroscopy

• Very long scan

• Small matrix

Diffusion-Weighted Imaging (DWI)

• Measures Brownian motion

• Restricted diffusion:

  • High DWI

  • Low ADC

⚠ Always confirm on ADC
(T2 shine-through is real)

FINAL OLD-TUTOR SUMMARY

If you remember only this, you will pass:

• GRE → fast, T2*

• SNR ↔ voxel size

• TR → T1

• TE → T2

• FOV ↑ → SNR ↑↑

• Matrix ↑ → detail ↑, SNR ↓

• DWI needs ADC confirmation

MRI is not memorization.
It is cause and effect.