Emergence of Prediction by Reinforcement Learning Using a Recurrent Neural Network

Figure 14

Connection weights to the 12th neuron from the input neurons responsible for the $x$ location of the object. Lateral axis indicates the most responsible point of each input neuron.