Our identity is in part determined by our memories. Questions relating to how different memories are represented in the brain have been the subject of intense neuroscientific investigation in the past few decades. In this issue we publish the first in a series of articles on memory systems. This series will describe current theories and debate how different types of memory, such as semantic, episodic, implicit, emotional and working memory, are encoded and anatomically represented in the brain.

On page 872, Squire and colleagues discuss the neural substrates of recognition memory. One current model attributes the two components of recognition memory — recollection and familiarity — to distinct areas of the medial temporal lobe. Here, the authors present a different interpretation of data from single-unit recordings, lesions studies and neuroimaging studies. They suggest that the hippocampus and the perirhinal cortex both contribute to recollection and familiarity, working together in a cooperative and complementary way.

Memory research can also benefit from computational models. This month's Highlight section (page 811) describes a recent computational study by Roudi and Latham which suggests that the memory capacity of individual neuronal networks might be smaller than we once thought, implying that the brain must rely on multiple memory-storing modules.

Considerable progress has been made in recent years towards understanding our ability to store and retrieve large amounts of information, but many issues remain unresolved. With the advent of new techniques, expectations are high that we will soon solve some of the remaining mysteries.