The major cellular component of the human brain is not neurons, at least in numerical terms, but glial cells, especially astrocytes. The astrocyte-to-neuron ratio and astrocyte structural complexity have both increased as the brain has evolved, with the former reaching 10:1 in humans.1, 2 Classically, brain function has been perceived as entirely neuronal, with the glia (derived from the Greek word for glue) playing subordinate roles: oligodendrocytes in myelination, microglia in macrophage function, and astrocytes in supporting and nourishing neurons. However, it is now likely that astrocytes have a more major role in neurological function and disease than previously thought. Astrocytes, like neurons, vary enormously in shape and size in different areas of the brain, but only recently has this (like neurons again) been thought to reflect differing functions. They influence the anatomy of specific brain areas, especially during development and neuronal migration. They closely control brain homeostasis, forming part of the blood–brain barrier and controlling regional blood flow, as well as the ionic and molecular micro-environment.1-4 Disorders directly or indirectly related to astrocytes are increasingly recognized as well including, amongst others, leukodystrophies (Alexander disease, megaloencephalic leukoencephalopathy with subcortical cysts, and vanishing white matter); neurodegenerative conditions such as Huntington and Parkinson diseases; acute and chronic metabolic conditions such as hepatic encephalopathy, Niemann-Pick type C, or acaeruloplasminaemia; chronic pain syndromes; and neuromyelitis optica (aquaporin 4 is mainly expressed in astrocytes). Reactive astrogliosis is an important pathological response to injury that amongst other effects prevents axonal migration. There is a suspected role in epilepsy and migraine as well.1-4 The biggest current controversy concerns gliotransmission and the astrocyte’s role in information transfer. Astrocytes have thousands of processes which connect closely to neurons, in particular to neuronal synapses. These are dynamic as they can change in shape and size over time. The tripartite hypothesis suggests that these astrocyte processes have important influences on synaptic function.3 In vitro astrocytes take up and secrete neurotransmitters (especially glutamate which is the prime excitatory molecule in the brain), express similar ion channels to adjacent neurons, and control calcium signalling. Knockout mice without glutamate uptake channels develop a severe neonatal epileptic encephalopathy.4 Over the longer term, astrocytes also appear to influence synapse formation and pruning. As astrocytes form syncytia with gap junctions that allow cell-to-cell transport (e.g. of calcium), it is also possible that astrocytes influence multiple synapses in contact with the same or adjacent astrocytes, and play a role in coordinated neuronal discharges.3 Such roles could include intrinsic biorhythms such as sleep–wake cycles.5 Again, knockout mice without these gap junctions die, although from probable cardiorespiratory failure.4 Finally, astrocytes secrete agents such as ATP or serine which in vivo can affect adjacent neurons and other astrocytes and are thus putative ‘gliotransmitters’. Until recently the exact contribution all this actually makes to brain function was very uncertain as most had only been demonstrated in cell culture or other rather artificial circumstances. However, last year for the first time it was convincingly demonstrated in an animal model that glia can influence respiratory control. While blood carbon dioxide (pCO2) and pH levels have been known to affect respiratory drive for decades, the exact mechanism has remained unclear. Acid-sensitive neurons in certain brainstem nuclei have been believed to be responsible, although the precise ion channels have not been identified. In fact, connexin 26 hemichannels in brain-stem glial cells have been found to be directly sensitive to pCO2 levels. In response the glia then secrete ATP which in turn affects neurons that control respiration. Blocking these hemichannels reduces the ventilatory response to increased pCO2 by up to 25%.6 More recently, reducing gliotransmitter function in another knockout mouse model has been shown to affect hippocampal memory processes.7 The relevance of gliotransmission to higher brain functions still needs a lot of research, but these new insights have obvious implications for neurodevelopmental disorders. In therapeutic terms, strategies aimed at astrocytes could be important in acquired brain insults such as trauma and stroke, and perhaps in other disorders previously considered entirely neuronal such as epilepsy. The brain is even more complex than previously imagined.