“Joel DiPello previously developed a model of the early visual cortex (V1) that, when used as an input to a convolutional neural network (convnet), improved its performance.”