Table 2 The flexibility of FaBiAN is illustrated by the number of sequence parameters and settings available to the user.

From: A Fetal Brain magnetic resonance Acquisition Numerical phantom (FaBiAN)

Simulations

HASTE

SS-FSE

GA (weeks)

21–33

21–35

Magnetic field strength

1.5 T

1.5 T or 3 T

Field inhomogeneities

20% INU level provided by BrainWeb63,64 for T2w images

Acquisition parameters

Contrast

 Effective echo time (ms)

90

116.256–123.60

 Echo spacing (ms)

4.08

10

 Echo train length

224

224

 Excitation flip angle (°)

90

90

 Refocusing pulse flip angle (°)

180

180

Geometry

 Slice orientation

Sagittal, coronal or transverse

Sagittal, coronal or transverse

 Slice thickness (mm)

3

3–4

 Slice gap (mm)

0.3

0

 Number of slices

45–46

37–51

 Phase oversampling (%)

80

0

 Shift of the field-of-view (mm)

\(\pm\,\, 1.6\)

\(\pm\,\, 1.6\)

Resolution

 Field-of-view (mm\(^2\))

\(360 \times 360\)

\(240 \times 240{-}300 \times 300\)

 Base resolution (voxels)

320–327

256

 Phase resolution (%)

70

100

 Reconstruction matrix

\(320 \times 404 {- 327 \times 414}\)

\(512 \times 512\)

 Zero-interpolation filling

Yes

Acceleration technique

 Reference lines

42

 Acceleration factor

2

Amplitude of 3D rigid motion

Little motion

 Translation (mm) in x

\(\pm\, 1\)

\(\pm\, 1\)

 Translation (mm) in y

\(\pm\, 1\)

\(\pm\, 1\)

 Translation (mm) in z

\(\pm\, 1\)

\(\pm\, 1\)

 3D rotation (\(^\circ\))

\(\pm\, 2\)

\(\pm\, 2\)

Moderate motion

 Translation (mm) in x

\(\pm\, 3\)

\(\pm\, 3\)

 Translation (mm) in y

\(\pm\, 3\)

\(\pm\, 3\)

 Translation (mm) in z

\(\pm\, 3\)

\(\pm\, 3\)

 3D rotation (\(^\circ\))

\(\pm\, 5\)

\(\pm\, 5\)

Strong motion

 Translation (mm) in x

\(\pm\, 4\)

\(\pm\, 4\)

 Translation (mm) in y

\(\pm\, 4\)

\(\pm\, 4\)

 Translation (mm) in z

\(\pm\, 4\)

\(\pm\, 4\)

 3D rotation (\(^\circ\))

\(\pm\, 8\)

\(\pm\, 8\)

Noise

 Mean

0

0

 Standard deviation

0.15

0.01

  1. The ranges of values used to simulate fetal brain MR images are presented, in agreement with the clinical protocols respectively in place at CHUV (HASTE sequence) and Kispi (SS-FSE sequence). The differences in the implementation of both sequences mainly rely on the simulation of the GRAPPA acceleration technique for the HASTE, which affects the way the k-space of the simulated images is sampled, and the simulation of the scanner in-line interpolation for the SS-FSE, which requires low-pass filtering before zero-interpolation filling of k-space.