Abstract

In her review of my book How we remember: Brain mechanisms of episodic memory, Sarah Robins highlights my example of the problem of interference between memories accessed by content-addressable memory. However, she points out the difficulty of solving this problem with index-addressable representations such as time cells or arc length cells. Namely, the index-addressable memory requires knowing the unique index in advance in order to perform effective retrieval. This is a difficult problem, but should be solvable by forming bi-directional associations between an index-addressable sequence of time cells and an array of content-addressable features in the environment