Although general anaesthetics at high enough concentrations can act nonspecifically on a wide variety of neuronal sites, at clinical concentrations they are much more selective and probably exert their primary effects at a relatively small number of CNS targets.

Although much attention has focused on voltage-gated ion channels, as a class they are very resistant to clinical concentrations of general anaesthetics. Some voltage-gated Ca2+ channels are slightly sensitive, and their inhibition may underlie the small reduction of neurotransmitter release that occurs at some synapses.

Evidence is accumulating that neurotransmitter-gated ion channels are important targets for most general anaesthetics. While some agents act predominantly at excitatory receptors (e.g. ketamine, and possibly also the inert gas xenon, at NMDA receptors), anaesthetic potentiation of inhibitory synaptic receptors (mainly GABAA) best matches the pharmacological profile of a wide variety of agents for producing general anaesthesia in mammals. Some uncertainty remains, however, over the extent to which the activity of intact inhibitory synapses is potentiated by anaesthetics. In addition, some agents (e.g. barbiturates) are clearly effective at both inhibitory and excitatory postsynaptic receptors, so the balance between the inhibition of excitatory synapses and the potentiation of inhibitory synapses in causing general anaesthesia still needs to be established. A powerful approach for identifying which CNS targets are important and which are probably not is to use optical isomers of general anaesthetics. In most cases, anaesthetic enantiomers display differences in their animal potencies. A comparison between the degree of stereoselectivity in animals with that observed in vitro with putative molecular targets provides an additional criterion (together with sensitivity) for assessing relevance.

At the cellular level, anaesthetic actions vary from neuron to neuron, due to the presence of differing ion channels and ion-channel subunits as well as second messenger regulatory systems. In some cases, particular subunits have been identified as being necessary for anaesthetic sensitivity. The case for any major involvement of second messenger systems in general anaesthesia is unproven; the few studies performed at clinical levels suggest they may be relatively insensitive to anaesthetics, but future studies should provide more definitive evidence.

At the molecular level, anaesthetics almost certainly act by binding directly to proteins rather than by perturbing lipid bilayers. Anaesthetics bind to pre-existing pockets or clefts on their target proteins. While this causes only minor changes in local protein structure, anaesthetics exert their effects by shifting the equilibria between natural conformational states of the protein