The Wall Street Journal has a nice, feel good running opinion from May 19 that raises more than a few common folk ideas about running. Too little time to comment on them all but here is one:
"But running is not just exercise. It's a great stress reliever and an inexpensive source of neurotransmitters like dopamine that wash the body with good feelings."
That is, running makes us high. The general idea is that running long and fast signals the secretion of endogenous opioids, which are our own version of morphine (of which heroine is a derivative) and these endogenous opioids (or endorphins), bind to receptors on cells in brain circuits that make us feel euphoric.
Gina Kolata, a reporter for the New York Times, criticized the endorphin hypothesis of runner's high in her book Ultimate Fitness. I don't have the book but she also had an article in the Milwaukee Journal Sentinal that suggested that runner's high does not come from an "endorphin rush". Most recently, however, she's changed her mind based on a new study that shows (according the the authors and Kolata) (1) endogenous opioid activity increases after a 2hr run compared to before the run and (2) perceived euphoria is greater after the 2hr run than before, and, importantly, (3) the level of opioid activity is positively correlated with perceived euphoria. Here is a lay summary of the paper.
As I understand the original paper, the authors did not actually measure levels of the endogenous opiods but measured the levels of the binding of an radioactive opioid antagonist, which they could detect with PET. Since the antagonist and the endogenous opioids compete for the same binding sites, then if measured levels of the antagonist decrease, the conclusion from the authors is that endogenous opioid activity is increased (they quickly dismissed alternatives, such as downregulation of opioid receptors, but I didn't really spend enough time to evaluate these).
One note - the authors note that runners were both (1) told that they were being treated with opioids ligands and (2) confirmed that they had previously experienced runner's high. Runner's tend to be a well educated group and I'm guessing that more than one figured out what the study was about. It's a nice study but I'm not completely confident that a placebo effect can be ruled out and it still isn't measuring endogenous opioid activity directly.
Also note that there is a group that argues that runner's high is due to endogenous cannibinoids and not opioids. Amby Burfoot has a comment on this.