Is it correct that orientation effects in early visual cortex do not produce strong, localized univariate differences, but typically have been studied with multivariate analyses? Is there any paper that decoded orientation effects with univariate methods such as GLM?