I have been reading Loneliness, a new book by medical researchers John T. Cacioppo and William Patrick. It's a round-up of psychological, physiological, and neurological studies done on the subject of human loneliness and its sub-concerns: isolation, affiliation, empathy, altruism, and evolutionary genetics. The authors' main point is that human beings are genetically wired to live in groups, and that there are definite physiological and neurological consequences when we don't do that. And the authors can write; this is a lively tome, not a collection of dry journal articles.
Obviously, being alone is not the same as being lonely, and different people have different needs and tolerances for "alone time." That said, there are still traceable patterns among those who self-rate as "fairly" or "very" or "often" lonely. These include increased stress hormones (cortisol et. al.) But what interests me more than that are the specific, limited studies on reactions to loneliness.
To cite only one: Lonely people pay more attention, not less, to social cues and signals than do non-lonely people. This seems counter-intuitive until you come to the analogy with food: hungry people pay more attention to food than the sated. However, the lonely are less skilled at correctly interpreting others' nonverbal cues. Which is one reason they're lonely in the first place.
I wish I had read this book before I sold my story "Act One" to ASIMOV'S. That story, which comes out next year (I think) could have benefited from the knowledge in this book. At any rate, I recommend Loneliness to anyone curious about why the human animal behaves as it does.