next up previous
Next: Up: Signal Model MCMC Sampling Previous: Signal Model MCMC Sampling

$ \lambda _{hi}$

This is sampled from using Metropolis-Hastings. We propose univariate jumps in the six parameter space of $ \lambda _{hi}$ at each voxel. Metropolis-Hastings requires that the following terms are updated:
$\displaystyle \prod_t p(q_{it}\vert\beta,\alpha,q_{i\tilde{t}<t},\phi_{\epsilon_i})$     (48)
$\displaystyle \prod_{t,j\in{\cal N} i} p(q_{jt}\vert\beta,\alpha,q_{i\tilde{t}<t},\phi_{\epsilon_j})$     (49)
$\displaystyle p(\lambda_{hi})$     (50)