Active learning of neural response functions with Gaussian processes