The Neurocritic has a fascinating summary of a recent paper investigating different types of memory in marathon runners. Why marathoners? Because completing a 26.2 mile race is an insanely arduous exercise, and leads to the massive release of stress hormones such as cortisol. Here are the scientists:
Indeed, cortisol levels recorded 30 min after completion of a marathon rival those reported in military training and interrogation (Taylor et al., 2007), rape victims being treated acutely (Resnick, Yehuda, Pitman, & Foy, 1995), severe burn injury patients (Norbury, Herndon, Branski, Chinkes, & Jeschke, 2008), and first-time parachute jumpers (Aloe et al., 1994).
But before we get back to the marathoners, a quick discussion of human memory. There are two distinct types of memory, which rely on different pathways in the brain. Explicit memory involves the recollection of discrete facts, events, names, faces, etc. It's a conscious process and is largely modulated by the hippocampus. (This was first discovered by studying patients with hippocampal lesions who turned into amnesiacs.) The second type of memory is implicit, largely unconscious and allows us to act based on previous experiences without taking the time to recall discrete memories. Think, for instance, of riding a bicycle. You have an implicit, procedural memory of the motor movements required to balance on two skinny wheels. As a result, you don't have to relearn the movements, or even consciously consider them, every time you go for a ride.
It has long been known that stress disrupts explicit memory, which is why it's not good to be too stressed when taking a test. The scientists hypothesized that finishing a marathon would wreak havoc on our explicit memory system, while leaving implicit memory largely intact. (There are probably good evolutionary reasons for this. It's not useful to forget how to throw a Pleistocene spear when being chased by a bear; if stress disrupted our procedural memory, we'd all be dead by now.)
The experiment itself was straightforward: 261 marathoners running in either the New York City Marathon or the Boston Marathon were given two different verbal memory tests, targeting the different memory pathways. 141 of the runners were tested within 30 min of finishing the race (when their cortex was still flush with cortisol) while the other 120 were tested 1-3 days before the race (this was the control group).
The end result? The group that had just finished the marathon showed a significant decline in explicit memory. They were less able to consciously recall a series of words that they had been shown only a few minutes earlier. However, after running 26.2 miles the marathoners actually showed a large improvement in implicit memory. In other words, the extreme stress and utter physical exhaustion sharpened their ability to act on information stored in their unconscious.