This paper reviews several implementations in artificial intelligence that have neuroscience correspondence. The most robotics-related one is the neuro-replay. In RL, especially navigation related tasks, the performance will be better if the agent stores a certain amount of raw data. The data will be accessed during learning together with the arrival of new data. As we are going to solve navigation, this will be an interesting topic.
https://www.cell.com/neuron/fulltext/S0896-6273(17)30509-3