Abstract
Purpose: Multiband pulses are characterized by highly temporally
modulated waveforms. Rapid phase or frequency modulation
can be extremely demanding on the performance of radiofrequency
(RF) pulse generation, which can lead to errors that can
be avoided if pulses are restricted to amplitude modulation
(AM) only. In this work, three existingmultiband pulse design techniques
are modified to produce AMwaveforms.
Theory and Methods: Multiband refocusing pulses were
designed using phase-optimization, time-shifting, and rootflipping.
Each technique was constrained to produce AM pulses
by exploiting conjugate symmetry in their respective frequency
domain representations. Pulses were designed using the AM and
unconstrained techniques for a range of multiband factors (ie,
number of simultaneously excited slices), time-bandwidth products,
and slice separations. Performance was compared by
examining the resulting effective pulse durations. Phantom and in
vivo experiments were conducted for validation.
Results: Acquired data confirmed that AM pulses can produce
precise results when unconstrained designs may produce artifacts.
The average duration of AM pulses is longer than the
unconstrained versions. Averaged across a range of parameters,
the duration cost for AM pulses was 26, 38, and 20% for phaseoptimizing,
time-shifting and root-flipping, respectively.
Conclusions: Amplitude modulation multiband pulses can be
produced for a relatively small increase in pulse duration.
modulated waveforms. Rapid phase or frequency modulation
can be extremely demanding on the performance of radiofrequency
(RF) pulse generation, which can lead to errors that can
be avoided if pulses are restricted to amplitude modulation
(AM) only. In this work, three existingmultiband pulse design techniques
are modified to produce AMwaveforms.
Theory and Methods: Multiband refocusing pulses were
designed using phase-optimization, time-shifting, and rootflipping.
Each technique was constrained to produce AM pulses
by exploiting conjugate symmetry in their respective frequency
domain representations. Pulses were designed using the AM and
unconstrained techniques for a range of multiband factors (ie,
number of simultaneously excited slices), time-bandwidth products,
and slice separations. Performance was compared by
examining the resulting effective pulse durations. Phantom and in
vivo experiments were conducted for validation.
Results: Acquired data confirmed that AM pulses can produce
precise results when unconstrained designs may produce artifacts.
The average duration of AM pulses is longer than the
unconstrained versions. Averaged across a range of parameters,
the duration cost for AM pulses was 26, 38, and 20% for phaseoptimizing,
time-shifting and root-flipping, respectively.
Conclusions: Amplitude modulation multiband pulses can be
produced for a relatively small increase in pulse duration.
| Original language | English |
|---|---|
| Pages (from-to) | 2185-2193 |
| Journal | Magnetic Resonance in Medicine |
| Volume | 78 |
| Issue number | 6 |
| Early online date | 17 Jan 2017 |
| DOIs | |
| Publication status | Published - Dec 2017 |
Keywords
- Multiband pulse design
- Simultaneous Multislice
- root flipping
- RF pulse design
- excitation stability
- interslice artifact