I did not find an easy-to-use native deep neural network package on R.
IMO darch is actually a good one. But it still takes some effort to pre-process data, tune the parameters, etc.
Below is a quick sample to model the MNIST data set using deep neural network (784 x 400 x 400 x 10, a small network). Here I uses the data downloaded from Kaggle. H2O platform need to be installed beforehand.
It would easily get 97% accuracy, and rank 300+ in Kaggle leader board.
Note the implementation in H2O is rather modern, compared to e.g. darch. It uses ReLU as activation function of hidden unit. It trains with dropout, which contributes to generalization. The implementation utilizes concurrency without locking.
For further knowledge, I recommend A ‘Brief’ History of Neural Nets and Deep Learning as an introduction material.