The alertness caffeine produces is not a direct stimulation of wake-promoting neurons. It is the removal of an inhibition those neurons were under. In the previous chapter, adenosine binding A1 and A2A receptors suppressed histaminergic neurons of the tuberomammillary nucleus and cholinergic neurons of the basal forebrain, and promoted sleep-active neurons in the ventrolateral preoptic area. Caffeine interrupts that suppression at the receptor.
Trace the chain. Adenosine is present and would normally occupy A1/A2A receptors on wake-promoting cells, reducing their firing. Caffeine occupies those receptors instead, so the inhibitory input is not delivered. With less inhibition arriving, the wake-promoting neurons return toward their unopposed firing rate. Their projections to cortex and arousal systems release more histamine and acetylcholine, and cortical activation rises. At the same time, adenosine's promotion of sleep-active ventrolateral preoptic neurons is also reduced, so the brake on arousal systems is loosened from that side as well.
The net effect is disinhibition: arousal increases because a subtractive signal has been subtracted. This is why the pharmacology feels like stimulation even though no excitatory molecule was added. It also explains the ceiling. Caffeine can only remove inhibition that adenosine was supplying. If adenosine signaling is low — after a full night's sleep — there is little to remove, and caffeine has little to do.