# MRI Physics II: Pulse Sequences and Contrast Mechanisms

## Spin-Echo (SE) Sequences

### Conventional Spin-Echo

The conventional spin-echo sequence uses a 90-degree excitation pulse followed by one or more 180-degree refocusing pulses. By refocusing the dephasing caused by static field inhomogeneities, spin-echo sequences eliminate T2' effects and produce true T2-weighted images. However, the original implementation acquires only a single echo per TR period, resulting in long acquisition times that limit its practical utility. Conventional spin-echo has been largely replaced by fast spin-echo in routine clinical practice.

### Fast Spin-Echo (FSE) / Turbo Spin-Echo (TSE)

Fast spin-echo dramatically accelerates acquisition by applying multiple 180-degree refocusing pulses per TR period, generating an echo train. The echo train length (ETL), typically 8 to 32, determines the speed gain: each echo in the train fills a different line of k-space, reducing scan time by a factor equal to the ETL. The effective TE, which determines image contrast, is defined by which echo is assigned to the center of k-space. FSE provides excellent T2 contrast, fast acquisition, and reduced susceptibility artifacts compared to gradient-echo techniques. One notable characteristic is that fat remains relatively bright on T2-weighted FSE images due to J-coupling effects, a departure from what would be expected based on pure T2 decay. Very long echo train lengths can cause blurring from T2 decay along the echo train. FSE is the clinical workhorse for T2-weighted imaging of the brain, spine, abdomen, and musculoskeletal system.

### Single-Shot FSE (SSFSE / HASTE)

Single-shot fast spin-echo acquires an entire image after a single 90-degree excitation pulse by using a very long echo train. This makes it ultrafast (less than 1 second per slice) and highly resistant to motion artifacts. The tradeoff is limited spatial resolution and inherent blurring due to T2 decay along the extended echo train. It is particularly useful for magnetic resonance cholangiopancreatography (MRCP), fetal MRI, rapid abdominal surveys, and imaging uncooperative patients.

## Gradient-Echo (GRE) Sequences

### Basic Gradient-Echo

Gradient-echo sequences use a single RF pulse at a flip angle typically less than 90 degrees, followed by gradient reversal to form an echo. Unlike spin-echo sequences, there is no 180-degree refocusing pulse, so the signal retains T2* weighting rather than pure T2 weighting. The absence of a refocusing pulse allows for shorter TR and TE values, enabling faster acquisitions. Gradient-echo sequences are inherently sensitive to magnetic susceptibility effects, making them valuable for detecting blood products, calcification, iron, and artifacts at air-tissue interfaces. The flip angle determines the balance between T1 and PD/T2* weighting, and the Ernst angle represents the flip angle that maximizes signal for a given TR and T1.

### Spoiled Gradient-Echo (SPGR / FLASH / T1-FFE)

In spoiled gradient-echo sequences, residual transverse magnetization from each RF pulse is deliberately eliminated ("spoiled") before the next pulse using gradient or RF spoiling techniques. This prevents residual transverse magnetization from contributing to the signal, producing clean T1-weighted contrast when combined with a short TR and appropriate flip angle. Spoiled gradient-echo sequences are the backbone of post-contrast T1-weighted imaging of the brain and liver, dynamic contrast-enhanced studies, and 3D volumetric acquisitions. Different vendors use different names: SPGR (GE), FLASH (Siemens), and T1-FFE (Philips).

### Balanced Steady-State Free Precession (bSSFP / TrueFISP / FIESTA)

Balanced SSFP sequences are unique in that both residual transverse and longitudinal magnetization contribute to the steady-state signal. The resulting signal intensity is proportional to the T2/T1 ratio of the tissue, producing bright blood and bright fluid signal without the need for a contrast agent. These sequences offer very fast acquisition with exceptionally high SNR per unit time. However, they are sensitive to B0 field inhomogeneity, which can produce characteristic banding artifacts at off-resonance frequencies. Careful shimming of the magnetic field is essential. Clinical applications include cardiac cine imaging, fetal MRI, and rapid abdominal imaging. Vendor names include FIESTA (GE), TrueFISP (Siemens), and bFFE (Philips).

### Dual-Echo / Multi-Echo GRE (Dixon Technique)

Chemical shift imaging acquires images at specific echo times when fat and water protons are in-phase and opposed-phase. At 1.5T, fat and water protons precess at slightly different frequencies (a chemical shift of approximately 220 Hz), and they cycle between being opposed-phase (at TE of about 2.2 ms) and in-phase (at about 4.4 ms). At 3.0T, these echo times are halved (opposed-phase at about 1.1 ms, in-phase at about 2.2 ms). In voxels that contain both fat and water, signal cancellation occurs on opposed-phase images, producing the characteristic India ink artifact at fat-water interfaces. This technique is a powerful diagnostic tool for characterizing adrenal adenomas (which are lipid-rich and lose signal on opposed-phase images), assessing hepatic steatosis, and separating fat and water signals.

## Pulse Sequence Comparison

| Sequence | Type | Key Feature | Contrast | Vendor Names | Primary Applications |
|----------|------|-------------|----------|-------------|---------------------|
| Conventional SE | Spin-Echo | 90° + 180° refocusing | True T2 | — | Largely replaced by FSE |
| FSE / TSE | Fast Spin-Echo | Echo train (ETL 8-32) | T2 (fat remains bright) | FSE (GE), TSE (Siemens/Philips) | Brain, spine, MSK, abdomen T2 |
| SSFSE / HASTE | Single-Shot FSE | Entire image in <1 s | Heavy T2 | SSFSE (GE), HASTE (Siemens) | MRCP, fetal MRI, uncooperative patients |
| Spoiled GRE | Gradient-Echo (spoiled) | Residual transverse magnetization destroyed | T1 | SPGR (GE), FLASH (Siemens), T1-FFE (Philips) | Post-contrast T1, dynamic CE studies |
| bSSFP | Balanced Steady-State | Both Mxy and Mz contribute | T2/T1 ratio (bright blood/fluid) | FIESTA (GE), TrueFISP (Siemens), bFFE (Philips) | Cardiac cine, fetal MRI |
| Dual-Echo GRE (Dixon) | Chemical Shift | In-phase and opposed-phase TEs | Fat-water separation | LAVA-Flex (GE), Dixon (Siemens), mDIXON (Philips) | Adrenal adenoma, hepatic steatosis |

| Fat Suppression Technique | Mechanism | B0 Inhomogeneity Robustness | Use with Gadolinium? | Best Application |
|--------------------------|-----------|----------------------------|---------------------|-----------------|
| Frequency-Selective (ChemSat) | RF pulse at fat frequency + spoiler | Poor (fails near metal, off-center) | Yes | Post-contrast T1 imaging |
| STIR | Inversion recovery nulling fat T1 | Excellent | No (suppresses enhanced tissue) | MSK bone marrow edema, spine |
| Dixon | Mathematical water-fat separation | Excellent | Yes | Body and MSK imaging |
| SPAIR | Adiabatic inversion of fat | Good (better than ChemSat) | Yes | Body MRI, especially at 3T |

## Inversion Recovery Sequences

### Basic Inversion Recovery

Inversion recovery sequences add a 180-degree inversion pulse before the standard excitation and readout. This pulse inverts longitudinal magnetization from +Mz to -Mz. During the inversion time (TI), magnetization recovers through zero toward +Mz at a rate determined by T1. At the precise moment when a specific tissue's magnetization crosses through zero (the null point), it produces no signal and appears dark on the image. The null point for any tissue can be calculated as TI_null = T1 x ln(2) = 0.693 x T1.

### STIR (Short Tau Inversion Recovery)

STIR uses an inversion time chosen to null fat signal (approximately 150 ms at 1.5T, 220 ms at 3.0T). Its greatest advantage is that it provides uniform fat suppression even in regions where the B0 field is inhomogeneous, such as near metal or at the edges of the field of view, where frequency-selective fat saturation fails. However, STIR is nonselective in that it suppresses all tissues with a similarly short T1, which includes gadolinium-enhanced tissue, melanin, and blood products. For this reason, STIR should not be used with gadolinium contrast, because enhanced lesions may be suppressed along with fat. STIR is widely used in musculoskeletal imaging for bone marrow edema detection, as well as in spine and whole-body imaging.

### FLAIR (Fluid-Attenuated Inversion Recovery)

FLAIR uses a much longer inversion time (approximately 2,400 ms at 1.5T, 2,800 ms at 3.0T) to null CSF signal. The result is T2-like contrast with dark CSF, which dramatically improves the conspicuity of lesions near CSF spaces, particularly periventricular and cortical lesions. FLAIR is essential for detecting multiple sclerosis plaques, identifying subarachnoid hemorrhage (which appears bright on FLAIR), evaluating cortical infarction, and detecting leptomeningeal disease. One limitation is that incomplete CSF suppression can occur due to flow artifacts or incorrect TI selection, which may mimic pathology.

## Diffusion-Weighted Imaging (DWI)

### Principles

Diffusion-weighted imaging measures the random Brownian motion of water molecules in tissue. The technique uses paired diffusion-sensitizing gradients (the Stejskal-Tanner sequence): the first gradient dephases the spins, and the second, applied after a brief interval, rephases them. Stationary water molecules experience equal and opposite phase shifts from the two gradients and retain their signal. Water molecules that have moved between the two gradient applications experience unequal phase shifts and undergo incomplete rephasing, resulting in signal loss. The b-value (measured in s/mm^2) controls the sensitivity to diffusion and is determined by the gradient amplitude, duration, and interval. At b = 0, there is no diffusion weighting and the image is essentially T2-weighted. At b = 1000, which is the standard clinical value, the image is strongly diffusion-weighted. Higher b-values increase sensitivity to slow diffusion but decrease signal-to-noise ratio.

### Apparent Diffusion Coefficient (ADC)

The ADC is a quantitative measure of diffusion calculated from images acquired at two or more b-values. The ADC map is essential because it eliminates T2 shine-through, a confounding effect where tissues with long T2 values appear bright on DWI regardless of their diffusion properties. True restricted diffusion shows as bright on DWI and dark on the ADC map (low ADC value). T2 shine-through shows as bright on DWI but also bright on the ADC map (normal or high ADC value). Typical ADC values (in units of x10^-3 mm^2/s) include approximately 3.0 for CSF, 0.7 to 0.8 for normal brain, 0.3 to 0.4 for acute infarct, 0.3 to 0.6 for abscess, and variable values for tumors depending on their cellularity.

| Tissue / Condition | Typical ADC (×10⁻³ mm²/s) | DWI Signal (b=1000) | ADC Map |
|-------------------|---------------------------|---------------------|---------|
| CSF | ~3.0 | Dark | Bright |
| Normal brain | 0.7-0.8 | Intermediate | Intermediate |
| Acute infarct | 0.3-0.4 | Bright | Dark (true restriction) |
| Abscess | 0.3-0.6 | Bright | Dark (true restriction) |
| Highly cellular tumor | 0.4-0.8 | Bright | Dark |
| T2 shine-through | Normal/high | Bright | Bright (not true restriction) |

### Clinical Applications

DWI has become indispensable across multiple clinical domains. In acute stroke, restricted diffusion appears within minutes of onset due to cytotoxic edema, making DWI the most sensitive sequence for early infarct detection. In infection, restricted diffusion in a ring-enhancing lesion favors abscess (viscous pus) over necrotic tumor (which shows facilitated diffusion). Highly cellular tumors demonstrate restricted diffusion due to the reduced extracellular space available for water molecule movement. Restricted diffusion also helps distinguish cholesteatoma from granulation tissue in the temporal bone. Whole-body DWI is increasingly used for staging metastatic disease and assessing lymphoma.

## Fat Suppression Techniques

### Frequency-Selective Fat Saturation (ChemSat)

This technique applies a narrow-bandwidth RF pulse tuned to the fat resonance frequency (3.5 ppm lower than water) followed by a spoiler gradient to destroy the excited fat signal before the imaging readout begins. It works well in homogeneous B0 fields but fails in regions of inhomogeneity, such as off-center locations, near metal, or at air-tissue interfaces. Frequency-selective fat saturation is the most commonly used technique for post-contrast T1-weighted imaging.

### STIR

As described above, STIR nulls fat based on its T1 relaxation time rather than its resonance frequency, making it robust to B0 inhomogeneity and effective throughout the imaging volume. However, because it is nonselective, it also suppresses other tissues with short T1 values and should be avoided with gadolinium contrast.

### Dixon Technique (Water-Fat Separation)

The Dixon technique acquires in-phase and opposed-phase data and mathematically separates water and fat signals, producing four image sets: water-only, fat-only, in-phase, and opposed-phase. The resulting fat suppression is robust and insensitive to B0 inhomogeneity. The Dixon method can be combined with any pulse sequence type (spin-echo or gradient-echo) and is increasingly the preferred fat suppression approach in body and musculoskeletal imaging.

### Spectral Adiabatic Inversion Recovery (SPAIR)

SPAIR combines elements of both frequency-selective and inversion recovery approaches. An adiabatic inversion pulse selectively inverts fat magnetization, and the imaging readout is timed to occur when fat crosses through the null point. This achieves more uniform fat suppression than frequency-selective techniques alone and is less sensitive to B0 inhomogeneity. SPAIR is widely used in body MRI, particularly at 3T.

<image>A reference chart showing six common MRI pulse sequences and their resulting image contrast on a brain scan. Each column shows a representative axial brain image: (1) T1-weighted spin echo showing white matter brighter than gray matter and dark CSF; (2) T2-weighted fast spin echo showing bright CSF and gray matter brighter than white matter; (3) FLAIR showing T2-like contrast with dark (nulled) CSF and a conspicuous periventricular white matter lesion; (4) STIR showing uniform fat suppression with bright edema; (5) DWI (b=1000) showing an acute infarct as a bright lesion in the left MCA territory; (6) gradient echo/SWI showing dark blooming artifact at sites of hemorrhage or calcification. Each image is labeled with the sequence name and key parameters (TR, TE, TI where applicable).</image>

<image>A diagram illustrating the Dixon technique for fat-water separation. Top row: timing diagram showing two echoes acquired at in-phase (TE1) and opposed-phase (TE2) echo times, with fat and water signal vectors shown as arrows on a clock diagram at each TE. At in-phase TE, fat and water vectors are aligned (additive signal). At opposed-phase TE, they point in opposite directions (subtractive signal, causing cancellation in mixed voxels). Bottom row: four resulting images from a liver scan: in-phase image, opposed-phase image (showing India ink artifact at organ boundaries and signal drop in fatty liver), water-only image (uniform fat suppression), and fat-only image (showing subcutaneous and visceral fat). Mathematical relationships are annotated: Water = (IP + OP)/2, Fat = (IP - OP)/2.</image>

<image>A step-by-step illustration of diffusion-weighted imaging physics. Panel 1: paired diffusion-sensitizing gradients shown on a pulse sequence timing diagram (first gradient lobe, diffusion time delta, second gradient lobe). Panel 2: stationary water molecules (in restricted tissue) experience equal and opposite phase shifts from the two gradients, resulting in full signal rephasing (bright signal retained). Panel 3: freely moving water molecules (in unrestricted tissue) change position between the two gradient applications, experiencing unequal phase shifts and incomplete rephasing (signal loss, dark). Panel 4: clinical example showing a DWI image with a bright acute stroke lesion, alongside its corresponding ADC map showing the same lesion as dark (confirming true restricted diffusion), and an ADC map of a different case showing T2 shine-through (bright on both DWI and ADC).</image>

## Clinical Pearls

Always check the ADC map before reporting restricted diffusion. T2 shine-through on DWI is the most common source of false-positive interpretation, and the ADC map resolves the ambiguity immediately. STIR should never be combined with gadolinium contrast; for post-contrast fat-suppressed T1-weighted imaging, use frequency-selective fat saturation or Dixon instead. The opposed-phase signal drop in chemical shift imaging is pathognomonic for intravoxel fat and is the basis for adrenal adenoma diagnosis, where greater than 20% signal loss on opposed-phase compared to in-phase images strongly supports a lipid-rich adenoma. FLAIR is more sensitive than T2-weighted FSE for periventricular and juxtacortical MS plaques, but it can miss posterior fossa lesions where CSF pulsation causes incomplete nulling. At 3T, chemical shift effects double compared to 1.5T: in-phase and opposed-phase echo times are halved, susceptibility artifacts are amplified, and SAR is quadrupled. Balanced SSFP sequences (FIESTA/TrueFISP) provide the highest SNR per unit time of any sequence but are prone to banding artifacts from B0 inhomogeneity, making careful shimming essential before acquisition.

## References

- McRobbie DW, et al. *MRI From Picture to Proton*, 3rd edition
- Bernstein MA, et al. *Handbook of MRI Pulse Sequences*, Elsevier
- Bitar R, et al. "MR Pulse Sequences: What Every Radiologist Wants to Know but Is Afraid to Ask." *RadioGraphics*, 2006
- Le Bihan D, et al. "Diffusion MR imaging: clinical applications." *AJR*, 2006
- Ma J. "Dixon techniques for water and fat imaging." *JMRI*, 2008