Structuring variance of gterms prior in brms

In brms, the gterms have a Gaussian prior. Is it possible to specify the variance for that Guassian prior as a nonlinear function of covariates?