When Stress Messes Up Memory
An Interview with Robert Sapolsky, Ph.D.


by Brenda Patoine

April, 2007

Q: Much of your work focuses on the effects of stress on the brain and, in particular, memory processing. How does emotional stress impact what we remember and what we forget?

A: Naturally, in a complex way. The sound bite about the subject is that “Stresses messes up memory.” The more complicated picture is built around a) how severe and prolonged the stressor is, and b) what kind of memory we’re talking about. Let’s start with conventional memory— conscious declarative facts that can be retrieved, as in “I am called a mammal” or “I have a dentist appointment next Tuesday.”

When a stressor is relatively mild and transient, the formation of new memories is improved. So too, although to a lesser extent, is the retrieval of old memories. What’s that about? Well, what is mild and transient stress? It’s what we would call “stimulation.” However, when stress is more severe and prolonged, consolidation and retrieval of declarative memory is impaired.

The impairments are focused in a key area of the brain involved in this type of cognition, the hippocampus. While mild transient stress increases the metabolic rate and excitability of the hippocampus, the more severe scenario involves lower metabolism, less excitability, atrophy, and even death of neurons. A key class of stress hormones secreted by the adrenal gland, called glucocorticoids, mediates much of these effects.

There is a second type of memory that is also relevant, which we call implicit traumatic memory. Examples are a memory of an earthquake’s vibration, or the accent of the person who did the unspeakable thing to you. This type of implicit, non-conscious memory formation and retrieval is centered in the amygdala, and glucocorticoids do something VERY different there. Major traumatic stress causes enhancement of those implicit memories and, as a result of the elevated glucocorticoid levels, increased excitability and growth of neurons in the amygdala. Thus, a major traumatic stressor can facilitate the implicit memory of an event at the same time that it can impair the explicit components.

Q: From a practical perspective, if we really want to be sure we remember something, do we need to learn it in an emotional setting, or otherwise put ourselves under stress?

A: As outlined above, you’d want to do it with the RIGHT kind of stress, namely the circumstance we call stimulation: mild and transient activation of the stress response.

Q: You’ve spent a lot of time studying baboons on the Serengeti. What can baboons teach humans about stress?

A: There are several things that are relevant. For example, we have learned that in a stable dominance hierarchy, low-ranking animals have the most indices of stress-related disease. In contrast, in an unstable, rapidly shifting hierarchy, those traits are seen in high-ranking animals.

These general rules can be greatly modulated by individual personality factors. Among those of the same social rank, individual males are likely to have more indices of stress-related disease if they have trouble recognizing a neutral interaction with a rival as not being threatening, if they exert little social control, and if they have little participation in affiliative behaviors (e.g., grooming) with other animals.

Taken to another level, these individual-specific factors can be modulated by community-wide factors. In one troop of baboons, for example, low-ranking animals may be groomed a lot more, on average, than in another troop, and this will impact their physiology accordingly—that is, by making them less susceptible to stress-related disease.

Q: A recent report from your lab shows that a specific type of gene therapy improves memory in stressed rats. What are the clinical implications of this work, and how long before we might see clinical applications based on it?

A: Both unfortunately and fortunately, things are still far from the clinic. There are a lot of “plumbing” aspects that need to be solved in the gene therapy field, by which I mean that the major challenges are related to how best to deliver genes to neurons, as opposed to what gene to deliver. On one hand, this is unfortunate because gene therapy could theoretically be immensely helpful. At the same time, this is fortunate, because a lot of bioethical work will need to be figured out when/if this should be done. For example, should this be used for someone who, greatly rushed and stressed, is trying to figure out how best to save someone in an Emergency Room? Should this be used for someone who, greatly rushed and stressed, is trying to figure out how to most efficiently ethnically cleanse a village? These are ethical quandaries that society has yet to grapple with.

Q: What gets you most excited about your research?

A: One thing is the business of doing something that might actually help someone, and in the way that biomedical science does best: you could be flailing in the wilderness, and nothing is working, but then if, against all odds, you hit the jackpot of discovering something important, you may wind up helping lots of people as a result. That aspect of research always reminds me of a story: two people are standing on the edge of a fast-moving river. They spot someone being washed down the river. The first of the two people on the bank dives in to save the person; the second one just watches. More time passes, and another person is carried down the river. Again, the first person leaps in, while the second watches. This happens a third, then a fourth time. Finally, the first person yells at the second, “Why don’t you help?” And the second answers, “I am helping. I’m trying to figure out why people keep falling into this river.” The first person is a physician, and the second is a biomedical scientist; it’s that latter way of doing good that appeals to me enormously.

Q: What are the next steps?

A: For our work concerning how stress and stress hormones (glucocorticoids) can have adverse effects in the nervous system, we hope to understand in detail a very counterintuitive finding that we and others in the field have been uncovering—namely, that these hormones, renowned for their anti-inflammatory effects, can occasionally be pro-inflammatory in the injured nervous system. For our gene therapy work, our goal is to help move the field toward more clinical trials by further refining what genes would be most logical to deliver to the injured nervous system in various circumstances.