“Defined categories of neurons” doesn’t make sense as a reference to neurons that have ion channels, because the drugs are supposed to work by affecting ion channels — this presumes that all the relevant neurons have ion channels. So the possession of ion channels wouldn’t be a way to define different categories of neurons.

“Defined categories of neurons” doesn’t make sense as a reference to neurons that respond to drug treatment, because this reference doesn’t offer a way to distinguish different categories of neurons. The drugs are supposed to work by affecting neurons — so all of the neurons that are relevant to how these drugs work would respond to the drugs. We want to know why certain categories of neurons relate to selectively treating different disorders.

“Defined categories of neurons” doesn’t make sense as a reference to neurons that contain receptor molecules, because all the relevant neurons have receptor molecules. So the possession of receptor molecules wouldn’t be a way to define different categories of neurons.

This best captures the meaning. Drugs might target receptors on “defined categories of neurons” so that the drugs can treat certain brain disorders. How would these drugs work? By targeting neurons associated with particular brain functions. Some neurons, for example, might be associated with mood disorders, others with stroke, others with Alzheimer’s. By targeting the neurons associated with a specific kind of disorder — in other words, a defined category of neuron — the drug might help ameliorate the effects of the disorder.

“Defined categories of neurons” doesn’t make sense as a reference to neurons that react to binding by neurotransmitters, because all the relevant neurons react to neurotransmitters. So the quality of reacting to binding by neurotransmitters wouldn’t be a way to define different categories of neurons.