In order for the brain to use the common amino acid glutamate as a neurotransmitter, it has been necessary to introduce a series of innovations that greatly restrict the availability of glutamate, so that it cannot degrade the signal-to-noise ratio of glutamatergic neurons. The most far-reaching innovations have been: i) to exclude the brain from access to glutamate in the systemic circulation by the blood-brain barrier, thereby making the brain autonomous in the production and disposal of glutamate; ii) to surround glutamatergic synapses with glial cells and endow these cells with much more powerful glutamate uptake carriers than the neurons themselves, so that most released transmitter glutamate is rapidly inactivated by uptake in glial cells; iii) to restrict to glial cells a key enzyme (glutamine synthetase) that is involved in the return of accumulated glutamate to neurons by amidation to glutamine, which has no transmitter activity, and can be safely released to the extracellular space, returned to neurons and deaminated to glutamate; iv) to restrict to glial cells two key enzymes (pyruvate carboxylase and cytosolic malic enzyme) that are involved in, respectively, de novo synthesis (from glucose) of the carbon skeleton of glutamate, and the return of the carbon skeleton of excess glutamate to the metabolic pathway for glucose oxidation. As a consequence of these innovations, neurons constantly require new carbon skeletons from glial to sustain their TCA cycle. When these supplies are withdrawn, neurons are unable to generate amino acid transmitters and their rate of oxidative metabolism is impaired. Given the commensalism that exists between neurons and glia, it may be fruitful to view brain function not just as a series of interactions between neurons, but also as a series of interactions between neurons and their collaborating glial cells.