Understanding ASL Physics¶
Arterial Spin Labeling (ASL) is a non-contrast MRI technique for measuring cerebral blood flow. This page covers the underlying physics.
The Core Idea¶
ASL uses the body's own blood water as an endogenous tracer. By magnetically "labeling" arterial blood before it enters the brain, we create a measurable perfusion signal without injecting any contrast agent.
The Labeling Process¶
Magnetic Labeling¶
Blood water has a magnetic moment (like a tiny compass). ASL manipulates this:
- Equilibrium: Blood protons align with the magnetic field (M₀)
- Labeling: RF pulses flip the magnetization (inversion or saturation)
- Flow: Labeled blood flows into the imaging region
- Readout: Measure the difference between labeled and unlabeled images

Label vs Control¶
We acquire two types of images:
- Control: No labeling (or symmetric "sham" labeling)
- Label: Blood is inverted/saturated
The difference signal (ΔM = Control - Label) is proportional to perfusion.
Labeling Schemes¶
PASL (Pulsed ASL)¶
Labels a "slab" of blood with a single RF pulse:
- Advantages: High labeling efficiency (~95-98%)
- Disadvantages: Sensitive to transit time, variable labeled volume
- Key parameter: TI₁ (inversion time)
CASL (Continuous ASL)¶
Continuously labels blood at a specific plane:
- Advantages: Larger labeled volume
- Disadvantages: SAR concerns, magnetization transfer effects
- Key parameters: τ (labeling duration), PLD
pCASL (Pseudo-Continuous ASL)¶
Combines benefits of PASL and CASL using train of RF pulses:
- Advantages: High labeling efficiency (~80-90%), lower SAR
- Most widely used in clinical and research settings
- Key parameters: τ (labeling duration), PLD (post-labeling delay)
The General Kinetic Model¶
Single-PLD Equation¶
For pCASL, CBF is calculated as:
Where:
| Symbol | Description | Typical Value |
|---|---|---|
| CBF | Cerebral blood flow | mL/100g/min |
| λ | Blood-brain partition coefficient | 0.9 ml/g |
| ΔM | Perfusion-weighted signal | a.u. |
| PLD | Post-labeling delay | 1.5-2.0 s |
| T₁b | T1 of blood | 1.65 s (3T) |
| α | Labeling efficiency | 0.85 (pCASL) |
| τ | Labeling duration | 1.8 s |
| M₀ | Equilibrium magnetization | from calibration |
Multi-PLD Model¶
With multiple PLDs, we can also estimate arterial transit time (ATT):
This allows fitting both CBF and ATT from the PLD curve.
Key Parameters¶
Post-Labeling Delay (PLD)¶
The time between labeling and image acquisition:
- Too short: Labeled blood hasn't arrived → underestimate CBF
- Too long: Label has decayed → low SNR
- Optimal: Matches arterial transit time (~1.5-2.0 s for brain)

Labeling Duration (τ)¶
How long blood is labeled:
- Longer τ: More labeled blood → higher signal
- Trade-off: Magnetization transfer effects increase
- Typical: 1.8 s for pCASL
Labeling Efficiency (α)¶
Fraction of blood that is actually inverted:
| Method | Efficiency |
|---|---|
| PASL | 0.95-0.98 |
| CASL | 0.68-0.73 |
| pCASL | 0.80-0.90 |
Measured using: α = 1 - (M_label / M_control)
M₀ Calibration¶
Why M₀ is Needed¶
The perfusion signal ΔM must be normalized by M₀ for absolute quantification:
M₀ Acquisition¶
Acquire a separate M₀ image with:
- Long TR (> 5 s) for full relaxation
- No labeling
- Same coil and geometry as ASL
M₀ Corrections¶
M₀ may need corrections for:
- T1 recovery: If TR < 5×T1
- Coil sensitivity: Spatial B1 variations
- T2* decay: If TE is significant
Signal Model¶
Full Signal Equation¶
The measured ASL signal is:
Where q(t) is the delivery function depending on PLD and τ.
Signal-to-Noise Considerations¶
ASL has inherently low SNR because:
- ΔM is typically 0.5-1.5% of M₀
- Need multiple averages (20-60 pairs typical)
Ways to improve SNR:
- Background suppression
- 3D readout (whole-brain)
- Multiple averages
- Higher field strength (3T > 1.5T)
Physical Assumptions¶
Single-Compartment Model¶
Standard ASL assumes:
- Well-mixed single compartment: Instantaneous exchange between blood and tissue
- No venous outflow: All labeled blood remains in imaging volume
- Uniform ATT: All blood arrives at same time
These may be violated in:
- Pathology (stroke, tumors)
- White matter (longer ATT)
- Large vessels
Blood-Brain Barrier¶
ASL measures flow of water, not contrast agent:
- Water freely crosses BBB
- No permeability limitation
- But: exchange time affects signal
Practical Considerations¶
Background Suppression¶
Reduces static tissue signal for better ΔM detection:
Because T1(tissue) < T1(blood), carefully timed inversion pulses can null the static tissue signal while preserving the labeled blood signal, improving the quality of the control-label subtraction.
Motion Sensitivity¶
ASL is sensitive to motion because:
- Small signal (ΔM ≈ 1% of M₀)
- Subtraction amplifies motion artifacts
Solutions:
- Background suppression
- 3D acquisition
- Motion correction algorithms
Partial Volume Effects¶
At voxel boundaries:
- Gray matter: CBF ≈ 60 mL/100g/min
- White matter: CBF ≈ 25 mL/100g/min
- CSF: CBF = 0
Partial volume correction may be needed for accurate quantification.
CBF Values¶
Expected Ranges¶
| Tissue | CBF (mL/100g/min) |
|---|---|
| Gray matter | 50-80 |
| White matter | 20-30 |
| Whole brain average | 40-60 |
| Stroke (acute) | ≈ 0 |
| Tumor (enhancing) | 50-150+ |
Interpretation¶
- Low CBF: Ischemia, infarct, hypoperfusion
- High CBF: Hyperperfusion, luxury perfusion, tumor
- Asymmetry: Compare hemispheres
Advantages of ASL¶
- Non-invasive: No contrast injection
- Repeatable: Can acquire multiple times
- Quantitative: Absolute CBF in mL/100g/min
- Safe: No nephrotoxicity concerns
- Pediatric-friendly: No IV access needed
Limitations¶
- Low SNR: Requires multiple averages
- Sensitive to transit time: May miss delayed flow
- Coverage: Historically 2D, now 3D available
- No timing information: Unlike DSC (no MTT, Tmax)
References¶
-
Alsop DC et al. "Recommended implementation of arterial spin-labeled perfusion MRI for clinical applications." Magn Reson Med 2015.
-
Buxton RB et al. "A general kinetic model for quantitative perfusion imaging with arterial spin labeling." Magn Reson Med 1998.
-
Dai W et al. "Continuous flow-driven inversion for arterial spin labeling using pulsed radio frequency and gradient fields." Magn Reson Med 2008.