I think we're likely to see some kind of neural interfaces by 2030, and far more advanced versions by 2050. The problem so far (as I understand it) has been our extreme lack of understanding, and ability to model the human brain down to the level of neurons, and even atoms. The technology to do that, in terms of imaging, and processing power is coming along though, and I think once a functional virtual brain exists, progress will be much faster. Currently scientists are basically just applying electric current, here or there, either literally on the scalp, or with implanted electrodes, and are observing the results. A very mess and inconsistent process given differences from person to person (or rat to rat).