Current deep learning models are able to create images strongly enough to activate specific neurons in the visual cortex. However, researchers say more accurate artificial neural network models should be developed to help produce more accurate control.