how to fit link logit for binomial error distribution in gradient descent boosted regression trees

Hello,

Could somebody please advise me on how to fit a logit binary response (probability of success/failure) in the context of a boosted regression tree model. In R it is possible to specify error distribution with a logit link function. Is there something similar in Matlab? There is a gbm package in R that makes calculations for deviance, initial values, gradients, and the constants predicted in each terminal node. How can this be done in Matlab?