Addition of Motion to a Particular Class of Spins in a Phantom

[sabbir-ahmed12]

Hi, I'm trying to add some translational motion in some particular spins of my phantom. In the current implementation, I need to provide a range of spins in the SpinRange() function. However, I want to add motion to spins which have a particular value. For example, in the phantom given below, I want to add motion only to spins belonging to class 153. Is it possible to achieve anyway? I'd be thankful for your help.

# Define the proton density array
ρ = (class.==102)*.86 .+    # Fat
    (class.==153)*.9 .+     # SoftTissue
    (class.==204)*.4 .+     # SpongyBone
    (class.==255)*.2        # CorticalBone

# Define the T1 decay array
T1 = (class.==102)*366 .+   # Fat
    (class.==153)*1200 .+   # SoftTissue
    (class.==204)*381 .+    # SpongyBone
    (class.==255)*100       # CorticalBone

# Define the T2 decay array
T2 = (class.==102)*70 .+    # Fat
    (class.==153)*80 .+     # SoftTissue
    (class.==204)*52 .+     # SpongyBone
    (class.==255)*.3        # CorticalBone

# Define the T2s decay array
T2s = (class.==102)*70 .+   # Fat
    (class.==153)*80 .+     # SoftTissue
    (class.==204)*52 .+     # SpongyBone
    (class.==255)*.3        # CorticalBone

# Define off-resonance array
Δw_fat = -220 * 2π
Δw = (class.==102) * Δw_fat # FAT1

# Adjust with scaling factor
T1 = T1*1e-3
T2 = T2*1e-3
T2s = T2s*1e-3

[cncastillo]

Hi! there's a way to give a mask instead of a range, an example is shown here: https://juliahealth.org/KomaMRI.jl/dev/reference/2-koma-base/#KomaMRIBase.SpinRange, @pvillacorta could confirm.

Nevertheless, I wouldn't recommend you to do that! As the use of a mask instead of a range makes the simulation very inefficient.

What I can recommend is to create a phantom with all the tissues you want to be static obj_static, and another that is dynamic with spin range AllSpins() called obj_dynamic. You can combine both with obj = obj_static + obj_dynamic. This new object will be efficient, and contain motion for only one tissue.

[pvillacorta]

Hi! Yes, in fact, the current implementation would only allow defining motion for spins which are adjacent in id number, that is, they belong to the same range. It would throw an error if that was not the case, precisely, for performance reasons.

The most appropriate way to do this is actually what @cncastillo mentioned. For your specific case:

obj1 = obj[obj.T2 .== 80e-3]
obj2 = obj[obj.T2 .!= 80e-3]
obj1.motion = translate(0.01, 0.0, 0.0, TimeRange(0.0, 1.0))  # for example
obj_sum = obj1 + obj2