These models use the Fabber Bayesian model fitting framework1 to implement a number of models for Arterial Spin Labelling MRI (ASL-MRI).
Note
If you have ASL data that you are looking to process you should start with the BASIL toolset or the OXASL both of which use FABBER_ASL
internally.
The ASL models are included in FSL. We stongly recommend version 6.0.1 or later.
If you need an updated version of the model which has not yet been released to FSL, you will either need to build from source using an existing FSL 6.0.1 or later installation, or download the pre-built Fabber bundle which contains the latest ASL release alongside other models in a standalone package.
This is the most common ASL model. In it simplest form it implements the well-known Buxton model, however it can incorporate additional features such as an arterial component, exchange and dispersion models and partial volume correction. It is the main model used in the OXASL and BASIL pipelines.
This model is selected using --model=aslrest
.
This model is designed for processing multiphase ASL data. It performs the equivalent of label-control subtraction, resulting in a data set which is suitable for processing using the aslrest
model. It is used by the multiphase plugin for the OXASL pipeline.
This model is selected using --model=asl_multiphase
.
This model is intended for processing data from the QUASAR ASL sequence.
This model is selected using --model=quasar
.
This model is intended for processing data from the Turbo-QUASAR ASL sequence.
This model is selected using --model=turboquasar
.
The Velocity-selective model ~~~~~~~~~~~~~~~~~~~~~~
This model is intended for processing data from the Velocity Selective ASL sequence.
This model is selected using --model=velocity_selective
.
This model is designed for the saturation recovery curve calibration method.
This model is selected using --model=satrecov
.
This model is for ASL data captured at a series of different TE values. It is used by the multi-TE plugin for the OXASL pipeline.
This model is selected using --model=asl_multite
.
This model implements only the basic Buxton kinetic model. It has been superceded by the more generic aslrest
model and is kept only for historical compatibility.
This model has been superceded by the exchange options in the aslrest
model.
Chappell, M.A., Groves, A.R., Woolrich, M.W., "Variational Bayesian inference for a non-linear forward model", IEEE Trans. Sig. Proc., 2009, 57(1), 223–236.↩