Next: f-contrasts
Up: Inference
Previous: Inference
We do not need to initialise the approximate distribution
parameters of
. This is because we can
initialise the other approximate distribution parameters and then
update
first. To allow us to provide
sensible initialisation of the other approximate distributions,
and , we use the model with the
autoregression parameters set to zero () and with the HRF
constraints removed ( and ). This means we can use the
standard ordinary least squares (OLS) voxelwise frequentist
solution to the GLM to get:
and
The approximate distribution,
, is set
using frequentist results. The mean is set to and the
variance to
. This mean and variance gives the
approximate distribution parameters
and
(see appendix A for the
conversion).
We initialise by using the frequentist solution to
equation 2, where is set to the residuals,
. This gives:
where
,
and where:
We also initialise , and
.
Next: f-contrasts
Up: Inference
Previous: Inference