Basket cells — inhibitory interneurons that are found in the neocortex, cerebellum and hippocampus, and form 'baskets' of axonal arborizations onto and around the somata of target neurons — were first described in 1911 by Ramon y Cajal. Since then, it has become clear that there are different types of basket cell with diverse structural and functional properties, but researchers have lacked a detailed classification of these cells. Wang and colleagues have investigated the properties of these neurons in the cortex in unprecedented detail, and conclude that there are three distinct classes of cortical basket cell.

It has generally been accepted that there are two 'classical' subtypes of basket cell — small and large (SBCs and LBCs) — which are distinguished by the morphologies of their axonal arborizations. Many studies have described other basket cells that did not fit into either of these classifications and, in 2000, Gupta et al. named these 'nest' basket cells (NBCs). Until now, the number and importance of these cells in the cortex has been unclear, but new data published in Cerebral Cortex show that NBCs are the most common type of basket cell in layers II/III of the somatosensory cortex, and that they are a powerful inhibitory force in these layers.

Wang et al. used a number of different techniques to characterize basket cells, including electrophysiology, microscopy and single-cell polymerase chain reaction (PCR) analysis of messenger RNA expression. They found that NBCs differed from the other two classes of basket cell not only in their morphology, but also in their electrophysiological responses, which were diverse, and in their molecular characteristics. Each class of basket cell showed a different pattern of mRNA expression; for example, NBCs and LBCs could be distinguished from SBCs because SBCs invariably expressed the mRNA for vasoactive intestinal peptide (VIP), whereas NBCs and LBCs never did.

Overall, more than 70% of the basket cells in layers II/III were NBCs. Although their electrophysiological properties were heterogeneous, a common feature was that they seemed to be 'high-threshold' interneurons — many pyramidal neurons are needed to recruit each NBC. The authors conclude, however, that once recruited, NBCs provide most of the perisomatic inhibition in these layers.

It is clear that such a precise and thorough classification of neurons is vital if we are to approach a full understanding of neural circuitry and interactions in any area of the brain. This kind of multidisciplinary approach will be important in future studies of neuronal classification. In particular, it will be interesting to see whether the properties of basket cells in the cerebellum and hippocampus reflect those found in the cortex.