Where Bayes tweaks Gauss: Conditionally Gaussian priors for stable multi-dipole estimation

<p style='text-indent:20px;'>We present a very simple yet powerful generalization of previously described model and algorithm for estimation multiple dipoles from magneto/electro-encephalographic data. Specifically, the consists in introduction log-uniform hyperprior on standard deviation set conditionally linear/Gaussian variables. We use numerical simulations an experimental dataset to show that approximation posterior distribution remains extremely stable under wide range values hyperparameter, virtually removing dependence hyperparameter.</p>