D3 - CT

Sapolsky, R. M. (2004). Why zebras don't get ulcers: The, acclaimed guide to stress, stress- related diseases and coping (3rd ed.). St. Martin's Griffin. ISBN: 9780805073690. Chapter 10

Stress and Memory

I’m old now, very old. I’ve seen a lot of things in my time and by now, I’ve forgotten a lot of them but, I tell you, that was one day that I’ll remember forever like it was yesterday. I was twenty-four, maybe twenty-five. It was a cold spring morning. Raw, gray. Gray sky, gray slush, gray people. I was looking for a job again and not having much luck, my stomach complaining about the bad rooming house coffee that was last night’s dinner and today’s breakfast. I was feeling pretty hungry, and I suspect I was starting to look pretty hungry too, like some half- starved animal that picks through a garbage can, and that couldn’t make much of an impression in an interview. And neither could the shabby jacket I was wearing, that last one I hadn’t hocked. I was plodding along, lost in my thoughts, when some guy comes sprinting around the corner, yelling with excitement, hands up in the air. Before I could even get a good look at him, he was shouting in my face. He was babbling, yelling about something being “classic,” something called “classic.” I couldn’t understand what he was talking about, and then he sprinted off. What the hell, crazy guy, I thought. But round the next corner, I see more people running around, yelling. Two of them, a man and woman, come running up to me and, by now, I tell you, I knew that something was up. They grabbed me by the arms, shouting “We won! We won!! We’re getting it back!” They were pretty excited but at least making more sense than the first guy, and I finally figured out what they were saying. I couldn’t believe it. I tried to speak, but I got all choked up, so I hugged them as if they were my brother and sister. The three of us ran into the street, where a big crowd was forming— people coming out of the office buildings, people stopping their cars, jumping out. Everyone screaming and crying and laughing, people shouting, “We won! We won!” Somebody told me a pregnant woman had gone right into labor, another that some old man had fainted right away. I saw a bunch of Navy guys, and one of them stepped right up and kissed this woman, a total stranger, leaning her way back—someone snapped a picture of them kissing, and I heard it became famous afterward. The weird thing is how long ago this was—the couple who first told me are probably long gone, but I can still see their faces, remember how they were dressed, the smell of the guy’s aftershave, the feel of the breeze that was blowing the confetti that people were tossing out the windows above. Still vivid. The mind’s a funny thing. Well anyway, as I was saying, that’s a day I’ll always remember—the day they brought back the original Coke.

A day to remember!

We’ve all had similar experiences. Your first kiss. Your wedding ceremony. The moment when the war ended. And the same for the bad moments as well. The fifteen seconds when those two guys mugged you. The time the car spun out of control and just missed the oncoming truck. Where you were when the earthquake hit, when Kennedy was shot, on 9/11. All etched forever in your mind, when it’s inconceivable that you can recall the slightest thing about incidents in the twenty-four hours before that life-changing event. Arousing, exciting, momentous occasions, including stressful ones, get filed away readily. Stress can enhance memory. At the same time, we’ve all had the opposite experience. You’re in the middle of the final exam, nervous and frazzled, and you simply can’t remember a fact that would come effortlessly at any other time. You’re in some intimidating social circumstance, and, of course, at the critical moment, you can’t remember the name of the person you have to introduce. The first time I was “brought home” to meet my future wife’s family, I was nervous as hell; during a frantically competitive word game after dinner, I managed to blow the lead of the team consisting of my future mother-in-law and me by my utter inability at one critical juncture to remember the word casserole. And some of these instances of failed memory revolve around infinitely greater traumas—the combat vet who went through some unspeakable battle catastrophe, the survivor of childhood sexual abuse—for whom the details are lost in an amnesiac fog. Stress can disrupt memory. By now, this dichotomy should seem quite familiar. If stress enhances some function under one circumstance and disrupts it under another, think time course, think 30-second sprints across the savanna versus decades of grinding worry. Short-term stressors of mild to moderate severity enhance cognition, while major or prolonged stressors are disruptive. In order to appreciate how stress affects memory, we need to know something about how memories are formed (consolidated), how they are retrieved, how they can fail.

A Primer on How Memory Works

To begin, memory is not monolithic, but instead comes in different flavors. One particularly important dichotomy distinguishes short-term versus long-term memories. With the former, you look up a phone number, sprint across the room convinced you’re about to forget it, punch in the number. And then it’s gone forever. Short-term memory is your brain’s equivalent of juggling some balls in the air for 30 seconds. In contrast, long-term memory refers to remembering what you had for dinner last night, the name of the U.S. president, how many grandchildren you have, where you went to college. Neuropsychologists are coming to recognize that there is a specialized subset of long-term memory. Remote memories are ones stretching back to your childhood—the name of your village, your native language, the smell of your grandmother’s baking. They appear to be stored in some sort of archival way in your brain separate from more recent long-term memories. Often, in patients with a dementia that devastates most long-term memory, the more remote facets can remain intact. Another important distinction in memory is that between explicit (also known as declarative) memory and implicit (which includes an important subtype called procedural memory) memory. Explicit memory concerns facts and events, along with your conscious awareness of knowing them: I am a mammal, today is Monday, my dentist has thick eyebrows. Things like that. In contrast, implicit procedural memories are about skills and habits, about knowing how to do things, even without having to think consciously about them: shifting the gears on a car, riding a bicycle, doing the fox-trot. Memories can be transferred between explicit and implicit forms of storage. For example, you are learning a new, difficult passage from a piece of piano music. Each time that stretch approaches, you must consciously, explicitly remember what to do—tuck your elbow in, bring your thumb way underneath after that trill. And one day, while playing, you realize you just barreled through that section flawlessly, without having to think about it: you did it with implicit, rather than explicit, memory. For the first time, it’s as if your hands remember better than your brain does. Memory can be dramatically disrupted if you force something that’s implicit into explicit channels. Here’s an example that will finally make reading this book worth your while—how to make neurobiology work to your competitive advantage at sports. You’re playing tennis against someone who is beating the pants off of you. Wait until your adversary has pulled off some amazing backhand, then offer a warm smile and say, “You are a fabulous tennis player. I mean it; you’re terrific. Look at that shot you just made. How did you do that? When you do a backhand like that, do you hold your thumb this way or that, and what about your other fingers? And how about your butt, do you scrunch up the left side of it and put your weight on your right toes, or the other way around?” Do it right, and the next time that shot is called for, your opponent/victim will make the mistake of thinking about it explicitly, and the stroke won’t be anywhere near as effective. As Yogi Berra once said, “You can’t think and hit at the same time.” Imagine descending a flight of stairs in an explicit manner, something you haven’t done since you were two years old—okay, bend my left knee and roll the weight of my toes forward while shifting my right hip up slightly—and down you go down the stairs.

Just as there are different types of memory, there are different areas of the brain involved in memory storage and retrieval. One critical site is the cortex, the vast and convoluted surface of the brain. Another is a region tucked just underneath part of the cortex, called the hippocampus. (That’s Latin for “sea horse,” which the hippocampus vaguely resembles if you’ve been stuck inside studying neuroanatomy for too long instead of going to the seashore. It actually looks more like a jelly roll, but who knows the Latin term for that?) Both of these are regions vital to memory—for example, it is the hippocampus and cortex that are preferentially damaged in Alzheimer’s disease. If you want a totally simplistic computer metaphor, think of the cortex as your hard drive, where memories are stored, and your hippocampus as the keyboard, the means by which you place and access memories in the cortex. There are additional brain regions relevant to a different kind of memory. These are structures that regulate body movements. What do these sites, such as the cerebellum, have to do with memory? They appear to be relevant to implicit procedural memory, the type you need to perform reflexive, motor actions without even consciously thinking about them, where, so to speak, your body remembers how to do something before you do. The distinction between explicit and implicit memory, and the neuroanatomical bases of that distinction, was first really appreciated because of one of the truly fascinating, tragic figures in neurology, perhaps the most famous neurological patient of all time. This man, known in the literature only by his initials, is missing most of his hippocampus. As an adolescent in the 1950s, “H.M.” had a severe form of epilepsy that was centered in his hippocampus and was resistant to drug treatments available at that time. In a desperate move, a famous neurosurgeon removed a large part of H.M.’s hippocampus, along with much of the surrounding tissue. The seizures mostly abated, and in the aftermath, H.M. was left with a virtually complete inability to turn new short-term memories into long-term ones—mentally utterly frozen in time.* Zillions of studies of H.M. have been carried out since, and it has slowly become apparent that despite this profound amnesia, H.M. can still learn how to do some things. Give him some mechanical puzzle to master day after day, and he learns to put it together at the same speed as anyone else, while steadfastly denying each time that he has ever seen it before. Hippocampus and explicit memory are shot; the rest of the brain is intact, as is his ability to acquire a procedural memory. This shifts us to the next magnification of examining how the brain handles memories and how stress influences the process—what’s going on at the level of clusters of neurons within the cortex and hippocampus? A long-standing belief among many who studied the cortex was that each individual cortical neuron would, in effect, turn out to have a single task, a single fact that it knew. This was prompted by some staggeringly important work done in the 1960s by David Hubel and Torstein Wiesel of Harvard on what was, in retrospect, one of the simpler outposts of the cortex, an area that processed visual information. They found a first part of the visual cortex in which each neuron responded to one thing and one thing only, namely a single dot of light on the retina. Neurons that responded to a sequence of adjacent dots of light would funnel their projections to one neuron in the next layer. And thus, what was this neuron responding to? A straight line. A series of these neurons would project to the next level in a way that each neuron in that cortical level would respond to a particular moving line of light. This led people to believe that there would be a fourth level, where each neuron responded to a particular collection of lines, and a fifth and sixth layer, all the way up until, at the umpteenth layer, there would be a

neuron that responded to one thing and one thing only, namely your grandmother’s face at a particular angle (and next to it would be a neuron that recognized her face at a slightly different angle, and then the next one…). People went looking for what were actually called “grandmother” neurons—neurons way up in the layers of the cortex that “knew” one thing and one thing only, namely a complexly integrated bit of sensory stimulation. With time, it became apparent that there could be very few such neurons in the cortex, because you simply don’t have enough neurons to go around to allow each one to be so narrow-minded and overspecialized.

A highly hypothetical neural network involving a neuron that “knows” about Impressionist

paintings. Rather than memory and information being stored in single neurons, they are stored in the patterns of excitation of vast arrays of neurons—in trendy jargon, in neuronal “networks.” How does one of these work? Consider the wildly simplified neural network shown in the diagram above. The first layer of neurons (neurons 1, 2, and 3) are classical Hubel and Wiesel type neurons, which is to say that each one “knows” one fact for a living. Neuron 1 knows how to recognize Gauguin paintings, 2 recognizes van Gogh, and 3 knows Monet. (Thus, these hypothetical neurons are more “grandmotherly”—specializing in one task—than any real neurons in the brain, but help illustrate well what neural networks are about.) Those three neurons project—send information to—the second layer in this network, comprising neurons A to E. Note the projection pattern: 1 talks to A, B, and C; 2 talks to B, C, and D; 3 talks to C, D, and E. What “knowledge” does neuron A have? It gets information only from neuron 1 about Gauguin paintings. Another grandmotherly neuron. Similarly, E gets information only from neuron 3 and knows only about Monet. But what about neuron C; what does it know about? It knows about Impressionism, the features that these three painters had in common. It’s the neuron that, metaphorically, says, “I can’t tell you the painter, certainly not the painting, but it’s one of those Impressionists.” It has knowledge that does not come from any single informational input, but emerges from the convergence of information feeding into it. Neurons B and D are also

Impressionism neurons, but they’re just not as good at it as neuron C, because they have fewer examples to work with. Most neurons in your cortex process memory like neurons B through D, not like A or E. We take advantage of such convergent networks whenever we are trying to pull out a memory that is almost, almost there. Continuing our art history theme, suppose you’re trying to remember the name of a painter, that guy, what’s his name. He was that short guy with a beard (activating your “short guy” neural network, and your “bearded guy” network). He painted all those Parisian dancers; it wasn’t Degas (two more networks pulled in). My high school art appreciation teacher loved that guy; if I can remember her name, I bet I can remember his…wow, remember that time I was at the museum and there was that really cute person I tried to talk to in front of one of his paintings…oh, what was the stupid pun about that guy’s name, about the train tracks being too loose. With enough of those nets being activated, you finally stumble into the one fact that is at the intersection of all of them: Toulouse-Lautrec, the equivalent of a neuron C. That’s a rough approximation of how a neural network operates, and neuroscientists have come to think of both learning and storing of memories as involving the “strengthening” of some branches rather than others of a network. How does such strengthening occur? For that, we switch to a final level of magnification, to consider the tiny gaps between the thready branches of two neurons, gaps called synapses. When a neuron has heard some fabulous gossip and wants to pass it on, when a wave of electrical excitation sweeps over it, this triggers the release of chemical messengers—neurotransmitters—that float across the synapse and excite the next neuron. There are dozens, probably hundreds, of different kinds of neurotransmitters, and synapses in the hippocampus and cortex disproportionately make use of what is probably the most excitatory neurotransmitter there is, something called glutamate. Besides being superexcitatory, “glutamatergic” synapses have two properties that are critical to memory. The first is that these synapses are nonlinear in their function. What does this mean? In a run-of-the-mill synapse, a little bit of neurotransmitter comes out of the first neuron and causes the second neuron to get a little excited; if a smidgen more neurotransmitter is released, there is a smidgen more excitation, and so on. In glutamatergic synapses, some glutamate is released and nothing happens. A larger amount is released, nothing happens. It isn’t until a certain threshold of glutamate concentration is passed that, suddenly, all hell breaks loose in the second neuron and there is a massive wave of excitation. This is what learning something is about. A professor drones on incomprehensibly in a lecture, a fact goes in one ear and out the other. It is repeated again—and, again, it fails to sink in. Finally, the hundredth time it is repeated, a lightbulb goes on, “Aha!” and you get it. On a simplistic level, when you finally get it, that nonlinear threshold of glutamate excitation has just been reached. The second feature is even more important. Under the right conditions, when a synapse has just had a sufficient number of superexcitatory glutamate-driven “aha’s,” something happens. The synapse becomes persistently more excitable, so that next time it takes less of an excitatory signal to get the aha. That synapse just learned something; it was “potentiated,” or strengthened. The most amazing thing is that this strengthening of the synapse can persist for a long time. A huge number of neuroscientists flail away at figuring out how this process of “long-term potentiation” works.

There’s increasing evidence that the formation of new memories might also sometimes arise from the formation of new connections between neurons (in addition to the potentiating of pre- existing ones) or, even more radically, the formation of new neurons themselves. This latter, controversial idea is discussed below. For the moment, this is all you need to know about how your brain remembers anniversaries and sports statistics and the color of someone’s eyes and how to waltz. We can now see what stress does to the process.

Improving Your Memory During Stress

The first point, of course, is that mild to moderate short-term stressors enhance memory. This makes sense, in that this is the sort of optimal stress that we would call “stimulation”—alert and focused. This effect has been shown in laboratory animals and in humans. One particularly elegant study in this realm was carried out by Larry Cahill and James McGaugh at the University of California at Irvine. Read a fairly unexciting story to a group of control subjects: a boy and his mother walk through their town, pass this store and that one, cross the street and enter the hospital where the boy’s father works, are shown the X-ray room…and so on. Meanwhile, the experimental subjects are read a story that differs in that the central core of it contains some emotionally laden material: a boy and his mother walk through their town, pass this store and that one, cross the street where…the boy is hit by a car! He’s rushed to the hospital and taken to the X-ray room…. Tested weeks later, the experimental subjects remember their story better than do the controls, but only the middle, exciting part. This fits with the picture of “flashbulb memory,” in which people vividly remember some highly aroused scene, such as a crime they witnessed. Memory for the emotional components is enhanced (although the accuracy isn’t necessarily all that good), whereas memory for the neutral details is not. This study also indicated how this effect on memory works. Hear the stressful story and a stress- response is initiated. As we by now well know, this includes the sympathetic nervous system kicking into gear, pouring epinephrine and norepinephrine into the bloodstream. Sympathetic stimulation appears to be critical, because when Cahill and McGaugh gave subjects a drug to block that sympathetic activation (the beta-blocker propranolol, the same drug used to lower blood pressure), the experimental group did not remember the middle portion of their story any better than the controls remembered theirs. Importantly, it’s not simply the case that propranolol disrupts memory formation. Instead, it disrupts stress-enhanced memory formation (in other words, the experimental subjects did as well as the controls on the boring parts of the story, but simply didn’t have the boost in memory for the emotional middle section). The sympathetic nervous system pulls this off by indirectly arousing the hippocampus into a more alert, activated state, facilitating memory consolidation. This involves an area of the brain that is going to become central to understanding anxiety when we get to chapter 15, namely the amygdala. The sympathetic nervous system has a second route for enhancing cognition. Tons of energy are needed for all that explosive, nonlinear, long-term potentiating, that turning on of light-bulbs in your hippocampus with glutamate. The sympathetic nervous system helps those energy needs to be met by mobilizing glucose into the bloodstream and increasing the force with which blood is being pumped up into the brain.

These changes are quite adaptive. When a stressor is occuring it is a good time to be at your best in memory retrieval (“How did I get out of this mess last time?”) and memory formation (“If I survive this, I’d better remember just what I did wrong so I don’t get into a mess like this again.”). So stress acutely causes increased delivery of glucose to the brain, making more energy available to neurons, and therefore better memory formation and retrieval.

Thus, the sympathetic arousal during stress indirectly fuels the expensive process of remembering the faces of the crowd chanting ecstatically about Classic Coke. In addition, a mild elevation in glucocorticoid levels (the type you would see during a moderate, short-term stressor) helps memory as well. This occurs in the hippocampus, where those moderately elevated glucocorticoid levels facilitate long-term potentiation. Finally, there are some obscure mechanisms by which moderate, short-term stress makes your sensory receptors more sensitive. Your taste buds, your olfactory receptors, the cochlear cells in your ears all require less stimulation to get excited under moderate stress and pass on the information to your brain. In that special circumstance, you can pick up the sound of a can of soda being opened hundreds of yards away.

Anxiety: Some Foreshadowing What we’ve just seen is how moderate and transient stress can enhance the sort of explicit memories that are the purview of the hippocampus. It turns out that stress can enhance another type of memory. This is one relevant to emotional memories, a world apart from the hippocampus and its dull concern with factoids. This alternative type of memory, and its facilitation by stress, revolves around that brain area mentioned before, the amygdala. The response of the amygdala during stress is going to be critical to understanding anxiety and post- traumatic stress disorder in chapter 15.

And When Stress Goes on for Too Long

With our “sprinting across the savanna” versus “worrying about a mortgage” dichotomy loaded and ready, we can now look at how the formation and retrieval of memories goes awry when stressors become too big or too prolonged. People in the learning and memory business refer to this as an “inverse-U” relationship. As you go from no stress to a moderate, transient amount of stress—the realm of stimulation—memory improves. As you then transition into severe stress, memory declines. The decline has been shown in numerous studies with lab rats, and with an array of stressors— restraint, shock, exposure to the odor of a cat. The same has been shown when high levels of glucocorticoids are administered to rats instead. But this may not tell us anything interesting. Lots of stress or of glucocorticoids may just be making for a generically messed-up brain. Maybe the rats would now be lousy at tests of muscle coordination, or responsiveness to sensory information, or what have you. But careful control studies have shown that other aspects of brain function, such as implicit memory, are fine. Maybe it’s not so much that learning and memory are impaired, as much as the rat being so busy paying attention to that cat smell, or so agitated by it, that it doesn’t make much headway solving whatever puzzle is in front of it. And within that realm of explicit memory problems, the retrieval of prior memories seems more vulnerable to stress than the formation of new ones. Similar findings have been reported with nonhuman primates.

Hard-charging businessman Billy Sloan is about to learn that continued stress does inhibit

one’s memory. What about humans? Much the same. In a disorder called Cushing’s syndrome, people develop one of a number of types of tumors that result in secretion of tons of glucocorticoids. Understand what goes wrong next in a “Cushingoid” patient and you understand half of this book—high blood pressure, diabetes, immune suppression, reproductive problems, the works. And it’s been known for decades that they get memory problems, specifically explicit memory problems, known as Cushingoid dementia. As we saw in chapter 8, synthetic glucocorticoids are often administered to people to control autoimmune or inflammatory disorders. With prolonged treatment, you see explicit memory problems as well. But maybe this is due to the disease, rather than to the glucocorticoids that were given for the disease. Pamela Keenan of Wayne State University has studied individuals with these inflammatory diseases, comparing those treated with steroidal anti-inflammatory compounds (that is, glucocorticoids) and those getting

nonsteroidals; memory problems were a function of getting the glucocorticoids, not of the disease. As the clearest evidence, just a few days of high doses of synthetic glucocorticoids impairs explicit memory in healthy volunteers. As one problem in interpreting these studies, these synthetic hormones work a bit differently from the real stuff, and the levels administered produce higher circulating glucocorticoid levels than the body normally produces, even during stress. Importantly, stress itself, or infusion of stress levels of the type of glucocorticoid that naturally occurs in humans, disrupts memory as well. As with the nonhuman studies, implicit memory is fine, and it’s the recall, the retrieval of prior information, that is more vulnerable than the consolidation of new memories. There are also findings (although fewer in number) showing that stress disrupts something called “executive function.” This is a little different from memory. Rather than this being the cognitive realm of storing and retrieving facts, this concerns what you do with the facts—whether you organize them strategically, how they guide your judgments and decision making. This is the province of a part of the brain called the prefrontal cortex. We’ll be returning to this in considerable detail in chapter 16 when we consider what stress may have to do with decision making and impulse control.

Neurons of the hippocampus of a rat. On the left, healthy neurons; on the right, neurons with

their projections atrophied by sustained stress. In the previous section, I noted that the brain region called the amygdala plays a central role in the types of emotional memories involved in anxiety. But the amygdala is relevant here as well. The amygdala gets highly activated during major stressors and sends a large, influential neuronal projection to the hippocampus. Activation of this pathway seems to be a prerequisite for stress to disrupt hippocampal function. Destroy a rat’s amygdala, or sever its connection to the hippocampus, and stress no longer impairs the kind of memory that the hippocampus mediates, even amid the usual high glucocorticoid levels. This explains a finding that harks back to the subject of stress “signatures,” and also demonstrates that some activities can represent a challenge to physical allostasis without being psychologically aversive. For example, sex raises glucocorticoid levels in a male rat—without activating the amygdala and without disrupting hippocampal function.

Second, neural networks get disconnected. If you look back at the diagram on the “Impressionism neuron”, you’ll see that there are symbols indicating how one neuron talks to another, “projects” to it. As mentioned a few paragraphs after that, those projections are quite literal—long multibranched cables coming out of neurons that form synapses with the multibranched cables of other neurons. These cables (known as axons and dendrites) are obviously at the heart of neuronal communication and neuronal networks. Bruce McEwen has shown that, in a rat, after as little as a few weeks of stress or of exposure to excessive glucocorticoids, those cables begin to shrivel, to atrophy and retract a bit. Moreover, the same can occur in the primate brain. When that happens, synaptic connections get pulled apart and the complexity of your neural networks declines. Fortunately, it appears that at the end of the stressful period, the neurons can dust themselves off and regrow those connections. This transient atrophy of neuronal processes probably explains a characteristic feature of memory problems during chronic stress. Destroy vast acres of neurons in the hippocampus after a massive stroke or late terminal stage Alzheimer’s disease, and memory is profoundly impaired. Memories can be completely lost, and never again will these people remember, for example, something as vital as the names of their spouses. “Weaken” a neural network during a period of chronic stress by retracting some of the complex branches in those neuronal trees, and the memories of Toulouse-Lautrec’s name are still there. You simply have to tap into more and more associative cues to pull it out, because any given network is less effective at doing its job. Memories are not lost, just harder to access. Third, the birth of new neurons is inhibited. If you learned your introductory neurobiology any time in the last thousand years, one fact that would be hammered in repeatedly is that the adult brain doesn’t make new neurons. In the last decade, it has become clear that this is utterly wrong.* As a result, the study of “adult neurogenesis” is now, arguably, the hottest topic in neuroscience. Two features about such neurogenesis are highly relevant to this chapter. First, the hippocampus is one of only two sites in the brain where these new neurons originate.* Second, the rate of neurogenesis can be regulated. Learning, an enriched environment, exercise, or exposure to estrogen all increase the rate of neurogenesis, while the strongest inhibitors identified to date are, you guessed it, stress and glucocorticoids—as little as a few hours of either in a rat. Two key questions arise. First, when the stress stops, does neurogenesis recover and, if so, how fast? No one knows yet. Second, what does it matter that stress inhibits adult neurogenesis? Intrinsic in this question is the larger question of what adult neurogenesis is good for. This is incredibly controversial, an issue that has adversaries practically wrestling each other on the podium during scientific conferences. At one extreme are studies that suggest that under the right conditions, there are tons of neurogenesis in the adult hippocampus, that these new neurons form connections with other neurons, and that these new connections, in fact, are needed for certain types of learning. At the other extreme, every one of these findings is questioned. So the jury’s out on this one. Fourth, hippocampal neurons become endangered. As noted, within seconds of the onset of stress, glucose delivery throughout the brain increases. What if the stressor continues? By about

thirty minutes into a continuous stressor, glucose delivery is no longer enhanced, and has returned to normal levels. If the stressor goes on even longer, the delivery of glucose to the brain is even inhibited, particularly in the hippocampus. Delivery is inhibited about 25 percent, and the effect is due to glucocorticoids.* Decreasing glucose uptake to this extent in a healthy, happy neuron is no big deal. It just makes the neuron a little queasy and lightheaded. But what if the neuron isn’t healthy and happy, and is instead in the middle of a neurological crisis? It’s now more likely to die than usual. Glucocorticoids will compromise the ability of hippocampal neurons to survive an array of insults. Take a rat, give it a major epileptic seizure, and the higher the glucocorticoid levels at the time of the seizure, the more hippocampal neurons will die. Same thing for cardiac arrest, where oxygen and glucose delivery to the brain is cut off, or for a stroke, in which a single blood vessel in the brain shuts down. Same for concussive head trauma, or drugs that generate oxygen radicals. Disturbingly, same thing for the closest there is to a rat neuron’s equivalent of being damaged by Alzheimer’s disease (exposing the neuron to fragments of an Alzheimer’s-related toxin called beta-amyloid). Same for a rat hippocampus’s equivalent of having AIDS-related dementia (induced by exposing the neuron to a damaging constituent of the AIDS virus called gp120).* My lab and others have shown that the relatively mild energy problem caused by that inhibition of glucose storage by glucocorticoids or stress makes it harder for a neuron to contain the eleventy things that go wrong during one of these neurological insults. All of these neurological diseases are ultimately energy crises for a neuron: cut off the glucose to a neuron (hypoglycemia), or cut off both the glucose and oxygen (hypoxia-ischemia), or make a neuron work like mad (a seizure) and energy stores drop precipitously. Damaging tidal waves of neurotransmitters and ions flood into the wrong places, oxygen radicals are generated. If you throw in glucocorticoids on top of that, the neuron is even less able to afford to clean up the mess. Thanks to that stroke or seizure, today’s the worst day of that neuron’s life, and it goes into the crisis with 25 percent less energy in the bank than usual. Finally, there is now evidence that truly prolonged exposure to stress or glucocorticoids can actually kill hippocampal neurons. The first hints of this came in the late 1960s. Two researchers showed that if guinea pigs are exposed to pharmacological levels of glucocorticoids (that is, higher levels than the body ever normally generates on its own), the brain is damaged. Oddly, damage was mainly limited to the hippocampus. This was right around the time that Bruce McEwen was first reporting that the hippocampus is loaded with receptors for glucocorticoids and no one really appreciated yet how much the hippocampus was the center in the brain for glucocorticoid actions. Beginning in the early 1980s, various researchers, including myself, showed that this “glucocorticoid neurotoxicity” was not just a pharmacological effect, but was relevant to normal brain aging in the rat. Collectively, the studies showed that lots of glucocorticoid exposure (in the range seen during stress) or lots of stress itself would accelerate the degeneration of the aging hippocampus. Conversely, diminishing glucocorticoid levels (by removing the adrenals of the rat) would delay hippocampal aging. And as one might expect by now, the extent of glucocorticoid exposure over the rat’s lifetime not only determined how much hippocampal degeneration there would be in old age, but how much memory loss as well.

Where do glucocorticoids and stress get off killing your brain cells? Sure, stress hormones can make you sick in lots of ways, but isn’t neurotoxicity going a bit beyond the bounds of good taste? A dozen years into studying the phenomenon, we’re not yet sure.

What About Damage to the Human Hippocampus?

We know from earlier in this chapter that an excess of stress and/or glucocorticoids can disrupt functioning of the hippocampus. Is there any evidence that this can include the sort of overt damage to the hippocampus that we’ve been discussing? That is, can it disconnect neural networks by the atrophying of processes, inhibit the birth of new neurons, worsen the neuron death caused by other neurological insults, or overtly kill neurons?

To date, six sets of findings in humans should raise some worries:

1. Cushing’s syndrome. As discussed above, Cushing’s involves any of a number of types of tumors that produce a vast, damaging excess of glucocorticoids, where the consequences include impairment of hippocampal-dependent memory. Monica Starkman at the University of Michigan has used brain imaging techniques on Cushing’s patients to look at the overall size of the brain, and the sizes of various subsections. She reports that there is a selective decrease in the volume of the hippocampus in these individuals. Moreover, the more severe the glucocorticoid excess, the greater the loss of hippocampal volume and the greater the memory problems.

2. Post-traumatic stress disorder (PTSD). As will be discussed in more detail in chapter 15,

this anxiety disorder can arise from a variety of types of traumatic stressors. Work pioneered by Douglas Bremner of Emory University, replicated by others, shows that people with PTSD from repeated trauma (as opposed to a single trauma)—soldiers exposed to severe and repeated carnage in combat, individuals repeatedly abused as children—have smaller hippocampi. Again, the volume loss appears to be only in the hippocampus, and in at least one of those studies, the more severe the history of trauma, the more extreme the volume loss.

3. Major depression. As will be detailed in chapter 14, major depression is utterly

intertwined with prolonged stress, and this connection includes elevated glucocorticoid levels in about half the people with major depression. Yvette Sheline of Washington University and others have shown that prolonged major depression is, once again, associated with a smaller hippocampus. The more prolonged the history of depression, the more volume loss. Furthermore, it is in patients with the subtype of depression that is most associated with elevated glucocorticoid levels where you see the smaller hippocampus.

4. Repeated jet lag. Chapter 11 will consider a single but intriguing study examining airline

flight attendants with long careers of shifting time zones on intercontinental flights. The

shorter the average time allowed to recover from each large bout of jet lag over a career, the smaller the hippocampus and the more memory problems.

5. Normative aging. Work by Sonia Lupien of McGill University, and replicated by others,

has examined healthy elderly people. Check out what their resting glucocorticoid levels are, the size of their hippocampus, and the quality of their hippocampal-dependent memory. Then come back some years later and retest them. As will be discussed in chapter 12, on aging, there is somewhat of a rise in resting glucocorticoid levels with age in humans, although there is a lot of variability in this. What is seen is that those whose glucocorticoid levels have been rising over the years since the study began are the ones who have had the most severe loss of hippocampal volume and the greatest decline in memory.

6. Interactions between glucocorticoids and neurological insults. A handful of studies report

that for the same severity of a stroke, the higher the glucocorticoid levels in a person at the time they come into an emergency room, the more ultimate neurological impairment.

So these studies collectively demonstrate that glucocorticoids damage the human hippocampus. Well, let’s hold on a second. There are some problems and complications: First, there have been some studies suggesting that PTSD involves lower than normal levels of glucocorticoids. Thus it can’t be the case that an excess of the hormones is damaging the hippocampus. However, it looks as if in those PTSD patients with the low levels, there is excessive sensitivity to the glucocorticoids. So the hormones are still plausible culprits. As a next issue, it isn’t clear whether the loss of hippocampal volume in PTSD is caused by the trauma itself, or by the post-traumatic period; amid that uncertainty, there has been at least one excellent study upending both of those ideas. It suggested instead that having a small hippocampus comes before the PTSD and, in fact, makes you more likely to develop PTSD when exposed to trauma. Finally, it should be remembered that the aging studies present a relationship that is merely correlative. In other words, yes, it could be that increasing glucocorticoid levels with age lead to hippocampal atrophy. But there are at least as good reasons to think that it is the other way around, that progressive hippocampal atrophy leads to the rising glucocorticoid levels (as will be explained more fully in chapter 12, this is because the hippocampus also helps to inhibit glucocorticoid release, such that an atrophied hippocampus isn’t very good at that task). In other words, no one is quite sure yet what is going on. One of the biggest problems is a lack of studies of brains like these after people have died. Phenomenally obsessive research could be carried out that would tell us whether the hippocampus is smaller because there are fewer of the millions of hippocampal neurons or because neurons have fewer and shorter cables connecting them to other neurons. Or both. If it turned out that there were fewer neurons, you might even be able to tell whether it is because more of them have died than usual, or because fewer of them were born. Or, again, both.

Actually, even without the postmortem studies, there are a few hints about the sources of the volume loss. Intriguingly, when the tumor that gave rise to the Cushing’s syndrome is removed and glucocorticoid levels revert to normal, the hippocampus slowly comes back to normal size. As noted before, when glucocorticoids cause the cables connecting neurons to shrivel up, it is not a permanent process—stop the glucocorticoid excess and the processes can slowly regrow. Thus, the best guess is that the volume loss in Cushing’s is based on the retraction of processes. In contrast, the volume losses in PTSD and major depression appear to be something approaching permanent, in that the loss persists in the former case decades after the trauma, and, in the latter, years to decades after the depression has been gotten under control with medication. So in those cases, the volume loss in the hippocampus probably can’t be due to shriveling processes of neurons, given that the shriveling can reverse. Beyond that, no one knows at this point why the hippocampus winds up being smaller in these disorders and situations. It is the knee-jerk reflex of all scientists to say, “More research is needed,” but more research really is needed in this case. For the moment, I think it is fair to say that there is decent but not definitive evidence that stress and/or prolonged exposure to glucocorticoids can cause structural, as well as functional, changes in the hippocampus, that these are changes that you probably wouldn’t want to have happen to your hippocampus, and that these changes can be long-lasting. What are some of the disturbing implications of these findings? The first concerns the use by neurologists of synthetic versions of glucocorticoids (such as hydrocortisone, dexamethasone, or prednisone) after someone has had a stroke. As we know from our introduction to glands and hormones in chapter 2, glucocorticoids are classic anti-inflammatory compounds and are used to reduce the edema, the damaging brain swelling that often occurs after a stroke. Glucocorticoids do wonders to block the edema that occurs after something like a brain tumor, but it turns out that they don’t do much for post-stroke edema. Worse, there’s increasing evidence that those famously anti-inflammatory compounds can actually be pro-inflammatory, worsening inflammation in the injured brain. Yet tons of neurologists still prescribe the stuff, despite decades-old warnings by the best people in the field and findings that the glucocorticoids tend to worsen the neurological outcome. So these recent findings add a voice to that caution—clinical use of glucocorticoids tends to be bad news for neurological diseases that involve a precarious hippocampus. (As a caveat, however, it turns out that huge doses of glucocorticoids can occasionally help reduce damage after a spinal cord injury, for reasons having nothing to do with stress or with much of this book.) Related to this is the concern that physicians may use synthetic glucocorticoids to treat problems outside the nervous system and, in the process, might endanger the hippocampus. A scenario that particularly disturbs me concerns the ability of these hormones to worsen gp120 damage to neurons and its relevance to AIDS-related dementia. (Remember?—the gp120 protein is found in the AIDS virus and appears to play a central role in damaging neurons and causing the dementia.) If, many experiments down the line, it turns out that glucocorticoids can worsen the cognitive consequences of HIV infection, this will be worrisome. That isn’t just because people with AIDS are under stress. It’s also because people with AIDS are often treated with extremely high doses of synthetic glucocorticoids to combat other aspects of the disease.

This same logic extends to the use of glucocorticoids in other realms of clinical medicine. About 16 million prescriptions are written annually in the United States for glucocorticoids. Much of the use is benign—a little hydrocortisone cream for some poison ivy, a hydrocortisone injection for a swollen knee, maybe even use of steroid inhalants for asthma (which is probably not a worrisome route for glucocorticoids to get into the brain). But there are still hundreds of thousands of people taking high-dose glucocorticoids to suppress the inappropriate immune responses in autoimmune diseases (such as lupus, multiple sclerosis, or rheumatoid arthritis). As discussed earlier, prolonged glucocorticoid exposure in these individuals is associated with problems with hippocampal-dependent memory. So should you avoid taking glucocorticoids for your autoimmune disease in order to avoid the possibility of accelerated hippocampal aging somewhere down the line? Almost certainly not—these are often devastating diseases and glucocorticoids are often highly effective treatments. Potentially, the memory problems are a particularly grim and unavoidable side effect. An even more disturbing implication of these findings is that if glucocorticoids turn out to endanger the human hippocampus (making it harder for neurons to survive an insult), you’re still in trouble, even if your neurologist doesn’t administer synthetic glucocorticoids to you. This is because your body secretes boatloads of the stuff during many neurological crises—humans coming into ERs after neurological insults have immensely high levels of glucocorticoids in their bloodstreams. And what we know from rats is that the massive outpouring of glucocorticoids at that time adds to the damage—remove the adrenals of a rat right after a stroke or seizure, or use a drug that will transiently shut down adrenal secretion of glucocorticoids, and less hippocampal damage will result. In other words, what we think of as typical amounts of brain damage after a stroke or seizure is damage being worsened by the craziness of our bodies having stress- responses at the time. Consider how bizarre and maladaptive this is. Lion chases you; you secrete glucocorticoids in order to divert energy to your thigh muscles—great move. Go on a blind date, secrete glucocorticoids in order to divert energy to your thigh muscles—probably irrelevant. Have a grand mal seizure, secrete glucocorticoids in order to divert energy to your thigh muscles—and make the brain damage worse. This is as stark a demonstration as you can ask for that a stress- response is not always what you want your body to be having. How did such maladaptive responses evolve? The most likely explanation is that the body simply has not evolved the tendency not to secrete glucocorticoids during a neurological crisis. Stress- induced glucocorticoid secretion works roughly the same in all the mammals, birds, and fish…and it has only been in the last half-century or so that westernized versions of just one of those species had much of a chance of surviving something like a stroke. There simply has not been much evolutionary pressure yet to make the body’s response to massive neurological injury more logical. We are now fifty, sixty years into thinking about ulcers, blood pressure, and aspects of our sex lives as being sensitive to stress. Most of us recognize the ways in which stress can also disrupt how we learn and remember. This chapter raises the possibility that the effects of stress in the nervous system might extend even to damaging our neurons, and the next chapter continues this theme, in considering how stress might well accelerate the aging of our brains. The noted

neuroscientist Woody Allen once said, “My brain is my second-favorite organ.” My guess is that most of us would rank our brains even higher up on a list. Chapter 11

Stress and a Good Night’s Sleep

Then there was the day when my son was about two weeks old. He was our first born, and we had been plenty nervous about how demanding parenting was going to be. It had been a great day—he’d slept well through the night, waking up a few times to nurse, and took some long naps during the day that allowed us to do the same. We’d settled into a schedule. My wife did the nursing, and I fetched the tureens of cranberry juice that she had become obsessed with since giving birth. Our son filled his diapers on cue, and his every gesture was confirming how wondrous he was. Things were calm. In the evening, as he slept and we settled into our old routines, like doing dishes (the first time in days), I indulged myself in some editorializing about the human condition. “You know, this newborn business is really quite manageable if you just stay on top of things. You need to work as a team, be organized, roll with the punches.” I went on fatuously like this for a while. That night, our son woke up to nurse right after we fell asleep. He was fussy, wouldn’t go back to sleep unless I patted him repeatedly, protested each time I tried to stop by waking up. This went on for an insane hour and then he needed to nurse again. Then, after patting him some more, he responded by blowing out his diaper, making a mess of his onesie and me. Then he screamed bloody murder when I washed him off. Finally, he then slept contentedly without patting, for about twenty minutes, before needing to nurse again, another blowout soiling of his fresh onesie, followed by our discovery that we had no clean ones, having neglected to do the laundry. Rather than doing something useful, I orated in a half-psychotic state, “We can’t do this, we’re going to die, I’m serious, people DIE from lack of sleep, it’s not possible to do this, it’s physiologically proven, we’re all going to DIE.” I swung my arms with emphasis, knocking over and loudly breaking a glass of cranberry juice. This woke up our, by then, happily sleeping son, causing all three of us to burst into tears. He eventually settled down and slept like a baby for the rest of the night, while I tossed anxiously, waiting for him to wake up again. Contained in this are the two central features of this chapter. Not getting enough sleep is a stressor; being stressed makes it harder to sleep. Yup, we’ve got a dread vicious cycle on our hands.

The Basics of Sleep All things considered, sleeping is pretty creepy. For a third of your life, you’re just not there, floating in this suspended state, everything slowed down. Except, at points, your brain is more active than when you’re awake, making your eyelids all twitchy, and it’s consolidating memories from the day and solving problems for you. Except when it’s dreaming, when it’s making no

sense. And then you sometimes walk or talk in your sleep. Or drool. And then there’s those mysterious penile or clitoral erections that occur intermittently during the night.* Weird. What’s going on here? To start, sleep is not a monolithic process, a uniform phenomenon. Instead, there are different types of sleep—shallow (also known as stages 1 and 2) sleep, where you are easily awakened. Deep sleep (also known as stages 3 and 4, or “slow wave sleep”). Rapid Eye Movement (REM) sleep, where the puppy’s paws flutter and our eyes dart around and dreams happen. There are not only these different stages, but a structure, an architecture to them. You start off shallow, gradually sleep your way down to slow wave sleep, followed by REM, then back up again, and then repeat the whole cycle about every ninety minutes (and as we’ll see in chapter 14, something goes wrong with the architecture of sleep during a major depression).

Not surprisingly, the brain works differently in different stages of sleep. This can be studied by having people sleep in a brain scanner, while you measure the levels of activity of different brain regions. Take some volunteers, sleep-deprive them for some godawful length of time, stick them in one of these imaging machines, poke them awake a little more while you get a measure of their brains’ activity when they’re awake, and then, snug as a bug in a scanner, let them go to sleep with the scanner running. The picture during slow wave sleep makes lots of sense. Parts of the brain associated with arousal activity slow down. Ditto for brain regions involved in controlling muscle movement. Interestingly, regions involved in the consolidation and retrieval of memories don’t have much of a decrease in metabolism. However, the pathways that bring information to and from those regions shut down dramatically, isolating them. The parts of the brain that first respond to sensory information have somewhat of a metabolic shutdown, but the more dramatic changes are in downstream brain areas that integrate, associate those bytes of sensory information, and give them meaning. What you’ve got is a metabolically quiescent, sleeping brain. This makes sense, as deep slow wave sleep is when energy restoration occurs. This is shown by the fact that the extent of sleep deprivation is not a great predictor of the total amount you will ultimately sleep, but it is a good predictor of how much slow wave sleep there’ll be—a very active brain or a sleep-deprived brain tends to consume a lot of a particular form of energy; the breakdown product of that depleted form of energy is the signal that biases toward slow wave sleep. A very different picture emerges during REM sleep. Overall, there’s an increase in activity. Some brain regions become even more metabolically active than when you’re awake. Parts of the brain that regulate muscle movement, brain stem regions that control breathing and heart rate—all

increase their metabolic rate. In a part of the brain called the limbic system, which is involved in emotion, there is an increase as well. The same for areas involved in memory and sensory processing, especially those involved in vision and hearing. Something particularly subtle goes on in the visual processing regions. The part of the cortex that processes the first bits of visual information does not show much of an increase in metabolism, whereas there is a big jump in the downstream regions that integrate simple visual information.* How can this be, when, on top of it, your eyes are closed? This is dreaming. That tells us something about how dream imagery arises. But something else that happens in the brain tells us something about the content of dreams. There’s a part of the brain, briefly mentioned in the last chapter, called the frontal cortex. It’s the most recently evolved part of the human brain, is disproportionately huge in primates, and is the last part of our brain to fully mature. The frontal cortex is the nearest thing we have to a superego. Starting from toilet training, it helps you to do the harder, rather than easier thing—for example, thinking in a logical, sequential manner, rather than bouncing all over the place cognitively. It keeps you from murdering someone just because you feel like it, stops you from telling someone exactly what you think of their hideous outfit and instead finds something complimentary. The frontal cortex does all this disciplining of you by inhibiting that frothy, emotional limbic system.* If you damage the frontal cortex, someone gets “frontally disinhibited”—doing and saying the things we may think about but would never act upon. During REM sleep, metabolism in the frontal cortex goes way down, disinhibiting the limbic system to come up with the most outlandish ideas. That’s why dreams are dreamlike—illogical, nonsequential, hyperemotional. You breathe underwater, fly in the air, communicate telepathically; you announce your love to strangers, invent languages, rule kingdoms, star in Busby Berkeley musicals.

Alfredo Castañeda, Our Dream (detail), 1999.

So those are the nuts and bolts of sleep. But what is sleep for? You die without it. Even fruit flies do. The most obvious answer is to have a stretch where your brain is going at half speed, in order to build up supplies of energy. Your brain consumes phenomenal amounts of energy to pull off all that calculus and symphony writing that you do—the brain constitutes something like 3 percent of your body weight but needs nearly a quarter of the energy. So stores tend to decline

during the day and some solid slow wave sleep is needed to restock those stores (mostly of a molecule called glycogen, which is also an energy store in liver and muscle).* Others speculate that sleep is for decreasing brain temperature, letting it cool off from all that daytime brainstorming, or for detoxifying the brain. Weirdly, another major reason to sleep is to dream. If you skip a night’s sleep, when you finally get to sleep the next night, you have more REM sleep than normal, suggesting that you’ve built up a real deficit of dreaming. Some extremely difficult studies that make me queasy just to contemplate deprive people or animals of REM sleep preferentially, and the study subjects go to pieces much faster than they do for the equivalent amount of deprivation of other types of sleep. Thus, this begs the question of what dreaming is for. To work out unresolved issues about your mother? To provide a living for surrealists and dadaists? So you can have a sex dream about some unlikely person in your waking life and then act all weird around that person the next morning by the water cooler? Well, maybe. The marked increase in metabolic activity during REM sleep, and in some of the most inhibited areas of the brain during waking, have suggested to some a sort of “use it or lose it” scenario in which dreaming gives some aerobic exercise to otherwise underutilized brain pathways (that is, the oft-neglected Busby Berkeley musical brain circuit). What has become clear is that sleep plays a role in cognition. For example, sleep can facilitate problem solving. This is the realm of “sleeping on a problem,” and then suddenly discovering a solution the next morning while you’re cleaning crud out of the corners of your eyes. The neurobiologist Robert Stickgold of Harvard has emphasized that this type of problem solving is the kind where a morass of unhelpful facts are broken through to get to feelings. As he says, you don’t forget a phone number and then “sleep on it” to remember it. You do it for some complex, ambiguous problem. Both slow wave and REM sleep also seem to play roles in the formation of new memories, the consolidation of information from the previous day, even information that became less accessible to you while awake over the course of the day. One type of evidence supporting this is the fact that if you teach an animal some task and disrupt its sleep that night, the new information isn’t consolidated. While this has been shown in many different ways, the interpretation remains controversial. As we saw in the last chapter, stress can disrupt memory consolidation. As we’re about to see in great detail, sleep deprivation is stressful. Maybe sleep deprivation disrupts memory consolidation merely because of the stress, which wouldn’t prove that sleep normally helps memory consolidation. But the pattern of memory disruption caused by sleep deprivation is different from that caused by stress. Another type of evidence is correlative. Being exposed to lots of new information during the day is associated with more REM sleep that night. Moreover, the amount of certain subtypes of sleep at night predicts how well new information is recalled the next day. For example, lots of REM sleep during the night predicts better consolidation of emotional information from the day before, while lots of stage 2 sleep predicts better consolidation of a motor task, and a combination of lots of REM and slow wave sleep predicts better retention of perceptual information. Others have

taken this further, reporting that it’s not just the amount of some subtype of sleep that predicts some subtype of learning, but whether it occurs early or late in the night. Another style of evidence for the “sleep helps you consolidate memories” story was first obtained by Bruce McNaughton of the University of Arizona. As we saw in chapter 10, the hippocampus has a central role in explicit learning. McNaughton recorded the activity of single hippocampal neurons in rats, identifying ones that became particularly busy while the rat was learning some new explicit information. That night, during slow wave sleep, it would be those same neurons that would be particularly busy. Taking that one step further, he showed that patterns of activation of hippocampal neurons that occur during learning are then repeated when the animal is sleeping. Brain-imaging studies with humans have shown something similar. There’s even evidence that when consolidation is going on during REM, genes are activated that help form new connections between neurons. During slow wave sleep, metabolism remains surprisingly high in areas like the hippocampus. It’s as if sleep is the time when the brain practices those new memory patterns over and over, cementing them into place. Weirdly, amid this general picture of sleep deprivation disrupting cognition, at least one type of learning is facilitated by sleep deprivation, as shown in some recent work by a graduate student of mine, Ilana Hairston. Suppose you have some unlikely task where you have to learn to recite the months of the year backward as rapidly as possible. Why is this going to be hard? Because there will repeatedly be the pull to recite the months in the way that you’ve done your whole life, which is forward; the previous, overlearned version of the task interferes with this new reversal task. Who would excel at this task? Someone who has never learned to do January, February, March, etc., automatically in that direction. If you sleep deprive some rats and give them a rat’s equivalent of a reversal task, they do better than do control animals. Why? Because they can’t remember the prior overlearned version of the task well enough for it to intrude now. So now we have the basics of sleep and what it might be good for. Entrez stress.

Sleep Deprivation as a Stressor

As we glide down into slow wave sleep, some obvious things occur to facets of the stress- response system. For starters, the sympathetic nervous system shuts down, in favor of that calm, vegetative parasympathetic nervous system. In addition, glucocorticoid levels go down. As introduced back in chapter 2, CRH is the hypothalamic hormone that gets the pituitary to release ACTH in order to trigger adrenal release of glucocorticoids. Some of the hypothalamic control of pituitary hormone release consists of an accelerator and a brake—a releasing factor and an inhibiting factor. For years, there’s been evidence floating around for a hypothalamic “corticotropin inhibiting factor” (CIF) that would inhibit the release of ACTH, counteracting the effects of CRH. No one’s sure what CIF is, or if it really exists, but there’s some decent evidence that CIF is a brain chemical that helps bring on slow wave sleep (called “delta sleep-inducing factor”). Thus, sleep deeply, and you turn off glucocorticoid secretion. In contrast, during REM, as you’re mobilizing all that energy to generate that outlandish dream imagery and to move your eyes rapidly, glucocorticoid secretion and the sympathetic nervous

system rev up again. But given that most of what counts as a good night’s sleep consists of slow wave sleep, sleep is predominately a time when the stress-response is turned off. This is seen in species whether they’re nocturnal or diurnal (that is, sleeping during the dark hours, like us). About an hour before you wake up, levels of CRH, ACTH, and glucocorticoids begin to rise. This is not just because merely rousing from slumber is a mini-stressor, requiring mobilization of some energy, but because those rising stress hormone levels play a role in terminating sleep. So deprive yourself of sleep, and the sleep-induced decline in the levels of those stress hormones doesn’t occur. And, no surprise, they rise instead. Glucocorticoid levels increase and the sympathetic nervous system is activated; commensurate with everything that’s been reviewed in previous chapters, down go levels of growth hormone and of various sex hormones. Sleep deprivation definitely stimulates glucocorticoid secretion, although not to a massive extent in most studies (unless the sleep deprivation is really prolonged; however, “it is postulated that these increases [in response to severe sleep deprivation] are due to the stress of dying rather than to sleep loss,” dryly noted one journal article). The elevated glucocorticoid levels during sleep deprivation play a role in breaking down some of the stored forms of energy in the brain. This, along with many of the glucocorticoid effects on memory, could have something to do with why learning and memory are so lousy when you’re sleep-deprived. That’s something we all learned when doing an all-nighter and discovering the next morning during the final exam that we can barely recall what month it was, let alone any of the factoids crammed in our heads the previous night. A recent study beautifully demonstrated one way in which our brains become impaired when we try to think hard on no sleep. Take a normally rested subject, stick her in a brain imager, and ask her to solve some “working memory” problems (holding on to some facts and manipulating them—like adding sequences of three-digit numbers). As a result, her frontal cortex lights up metabolically. Now, take someone who is sleep deprived and he’s awful at the working memory task. And what’s going on in his brain? What you might have guessed is that frontal metabolism would be inhibited, too groggy to get activated in response to the task. Instead, the opposite occurs—the frontal cortex is activated, but so are large parts of the rest of the cortex. It’s as if sleep deprivation has reduced this gleaming computer of a frontal cortex to a bunch of unshaven gibbering neurons counting on their toes, having to ask the rest of their cortical buddies to help out with this tough math problem. So why care if sleep deprivation is a stressor? It’s obvious. We’re accustomed to all sorts of amenities in our modern lives: overnight deliveries of packages, advice nurses who can be called at two in the morning, round-the-clock technical support staff. Therefore, people are required to work under conditions of sleep deprivation. We’re not a nocturnal species and if a person works at night or works swing shifts, regardless of how many total hours of sleep she’s getting, it’s going against her biological nature. People who work those sorts of hours tend to overactivate the stress-response, and there’s little habituation that goes on. Given that an overactive stress- response makes every page of this book relevant, it is not surprising that night work or shift work increases the risk of cardiovascular disease, gastrointestinal disorders, immune suppression, and fertility problems.

A widely reported study a few years back really brought this into focus. Recall how prolonged stress and glucocorticoids can damage the hippocampus and impair hippocampal-dependent explicit memory. Kei Cho of the University of Bristol studied flight attendants working for two different airlines. On one airline, after you worked a transcontinental flight with major jet lag, you’d have a 15-day break until being scheduled for the next transcontinental flight. In contrast, on Airline #2, presumably with a weaker union, you got a 5-day break before the next transcontinental flight.* Cho controlled for total amount of flying time and total number of time zones shifted in the course of flying. Thus, Airline #2’s crews didn’t experience more total jet lag, just less time to recover. Finally, Cho considered only employees who had been doing this for more than five years. He found that Airline #2’s attendants had, on average, impaired explicit memory, higher glucocorticoid levels, and a smaller temporal lobe (the part of the brain that contains the hippocampus). (This study was briefly alluded to in chapter 10). This is obviously not a good thing for the employees working under these conditions. And this may make it less likely that the flight attendant will remember that 17C requested a mixture of ginger ale and skim milk with ice. But it kind of makes one wonder whether the back-to-the-grind-after-5-days pilot is having trouble remembering whether or not this little ol’ switch turns the engine on or off. These worries about sleep deprivation are relevant to even those whose 9-to-5 job is 9-to-5 during daylight hours. We have an unprecedented number of ways to make us sleep deprived, beginning with something as simple as indoor lighting. In 1910, the average American slept nine hours a night, disturbed only by the occasional Model T backfiring. We now average 7.5 and declining. When there’s the lure of 24-hour-a-day fun, activities, and entertainment or, for the workaholic, the knowledge that somewhere, in some time zone, someone else is working while you indulge yourself in sleep, that pull of “just a few more minutes,” of pushing yourself, becomes irresistible. And damaging.*

And Stress as a Disruptor of Sleep

What should happen to sleep during stress? This one’s simple, given a zebra-o-centric view of the world: lion coming, don’t nap (or, as the old joke goes, “The lion and the lamb shall lie down together. But the lamb won’t get much sleep.”). The hormone CRH seems to be most responsible for this effect. As you’ll recall, the hormone not only starts the glucocorticoid cascade by stimulating ACTH release from the pituitary, but it is also the neurotransmitter that activates all sorts of fear, anxiety, and arousal pathways in the brain. Infuse CRH into a sleeping rat’s brain and you suppress sleep—it’s like throwing ice water onto those happily dozing neurons. Part of this is due to the direct effects of CRH in the brain, but part is probably due to CRH activating the sympathetic nervous system. If you go up to high altitude without acclimating, your heart is going to be racing, even when you’re not exerting yourself. This is not because you are stressed or anxious, but simply because your heart has to beat more often to deliver sufficient oxygen. Suddenly you discover that it’s awfully hard to fall asleep with your eyeballs throbbing rhythmically 110 times a minute. So the bodily consequences of sympathetic activation make sleeping hard.

Not surprisingly about 75 percent of cases of insomnia are triggered by some major stressor. Moreover, many (but not all) studies show that poor sleepers tend to have higher levels of sympathetic arousal or of glucocorticoids in their bloodstream. So, lots of stress and, potentially, little sleep. But stress not only can decrease the total amount of sleep but can compromise the quality of whatever sleep you do manage. For example, when CRH infusion decreases the total amount of sleep, it’s predominantly due to a decrease in slow wave sleep, exactly the type of sleep you need for energy restoration. Instead, your sleep is dominated by more shallow sleep stages, meaning you wake up more easily—fragmented sleep. Moreover, when you do manage to get some slow wave sleep, you don’t even get the normal benefits from it. When slow wave sleep is ideal, really restoring those energy stores, there’s a characteristic pattern in what is called the delta power range that can be detected on an EEG (electroencephalogram) recording. When people are stressed presleep, or are infused with glucocorticoids during sleep, you get less of that helpful sleep pattern during slow wave sleep. Glucocorticoids compromise something else that occurs during good quality sleep. Jan Born of the University of Lubeck in Germany has shown that if you infuse glucocorticoids into someone while they’re sleeping, you impair the memory consolidation that would normally be occurring during slow wave sleep.

Jeff Wall Insomnia, transparency in lightbox, 1994.

A Causes B Causes A Causes B Causes…

We have the potential for some real problems here, insofar as lack of sleep or poor-quality sleep activates the stress-response, and an activated stress-response makes for less sleep or lower- quality sleep. Each feeds on the other. Does that mean that experiencing even a smidgen of stress, or staying up late once to see Ted Koppel interview Britney Spears about the evidence for and against global warming, and—that’s it, you’re finished—downward spiral of stress and sleep deprivation?

Obviously not. For one thing, as mentioned, sleep deprivation doesn’t cause all that massive of a stress-response. Moreover, the need to sleep will eventually overcome the most stressful of stressors. Nonetheless, a fascinating study suggests how the two halves might interact, along the lines that the expectation that you’re going to sleep poorly makes you stressed enough to get poor-quality sleep. In the study, one group of volunteers was allowed to sleep for as long as they wanted, which turned out to be until around nine in the morning. As would be expected, their stress hormone levels began to rise around eight. How might you interpret that? These folks had enough sleep, happily restored and reenergized, and by about eight in the morning, their brains knew it. Start secreting those stress hormones to prepare to end the sleep. But the second group of volunteers went to sleep at the same time but were told that they would be woken up at six in the morning. And what happened with them? At five in the morning, their stress hormone levels began to rise. This is important. Did their stress hormone levels rise three hours earlier than the other group because they needed three hours less sleep? Obviously not. The rise wasn’t about them feeling rejuvenated. It was about the stressfulness of anticipating being woken up earlier than desirable. Their brains were feeling that anticipatory stress while sleeping, demonstrating that a sleeping brain is still a working brain. What might be happening, then, if you go to sleep thinking that not only will you be woken up earlier than you would like, but at an unpredictable time? Where any minute could be your last minute of sleep for the night? It’s quite possible that stress hormone levels will be elevated throughout the night, in nervous anticipation of that wake-up call. As we’ve seen, with an elevated stress-response during sleep, the quality of the sleep is going to be compromised. Thus, there is a hierarchy as to what counts as miserable sleep. Continuous, uninterrupted sleep, but too little of it—deadline looming, go to sleep late, get up early, not good. Even worse is too little sleep that is fragmented. As an example, I once did an experiment where every three hours for days I had to take blood samples from some animals. Even though I did next to nothing on these nights and days other than sleep, in fact I got more total sleep per day than was usual for me, I was a wreck. But worst of all is too little sleep that is unpredictably fragmented. You finally get back to sleep, but with the corrosive knowledge that five hours or five minutes from now, another patient will come into the emergency room, or the alarms will go off and it’s back to the fire truck, or someone’s diaper will slowly but surely fill up. This teaches us a lot about what counts as good sleep and how stress can prevent it. But as we’ll see in a couple of chapters, this generalizes beyond sleep. When it comes to what makes for psychological stress, a lack of predictability and control are at the top of the list of things you want to avoid.

Chapter 12

Aging and Death

Predictably, it comes at the most unpredictable times. I’ll be lecturing, bored, telling the same story about neurons I did last year, daydreaming, looking at the ocean of irritatingly young undergraduates, and then it hits, producing almost a sense of wonderment. “How can you just sit there? Am I the only one who realizes that we’re all going to die someday?” Or I’ll be at a scientific conference, this time barely understanding someone else’s lecture, and amid the roomful of savants, the wave of bitterness will sweep over me. “All of you damned medical experts, and not one of you can make me live forever.” It first really dawns on us emotionally sometime around puberty. Woody Allen, once our untarnished high priest of death and love, captures its roundabout assault perfectly in Annie Hall. The protagonist is shown, in flashback, as a young adolescent. He is sufficiently depressed for the worried mother to drag him to the family doctor—“Listen to what he keeps saying, what’s wrong with him, does he have the flu?” The Allenesque adolescent, glazed with despair and panic, announces in a monotone: “The universe is expanding.” It’s all there—the universe is expanding; look how big infinity is and how finite we are—and he has been initiated into the great secret of our species: we will die and we know it. With that rite of passage, he has found the mother lode of psychic energy that fuels our most irrational and violent moments, our most selfish and our most altruistic ones, our neurotic dialectic of simultaneously mourning and denying, our diets and exercising, our myths of paradise and resurrection. It’s as if we were trapped in a mine, shouting out for rescuers, Save us, we’re alive but we’re getting old and we’re going to die.

Morris Zlapo, Gepetto’s Dementia, collage, 1987.

And, of course, before dying, most of us will become old, a process aptly described as not for sissies: wracking pain. Dementia so severe we can’t recognize our children. Cat food for dinner. Forced retirement. Colostomy bags. Muscles that no longer listen to our commands, organs that

betray us, children who ignore us. Mostly that aching sense that just when we finally grow up and learn to like ourselves and to love and play, the shadows lengthen. There is so little time. Oh, it doesn’t have to be that bad. For many years I have spent part of each year doing stress research on wild baboons in East Africa. The people living there, like many people in the nonwesternized world, clearly think differently about these issues than we do. No one seems to find getting old depressing. How could they?—they wait their whole lives to become powerful elders. My nearest neighbors are of the Masai tribe, nomadic pastoralists. I often patch up their various minor injuries and ills. One day, one of the extremely old men of the village (perhaps sixty years old) tottered into our camp. Ancient, wrinkled beyond measure, tips missing from a few fingers, frayed earlobes, long-forgotten battle scars. He spoke only Masai and not Swahili, the lingua franca of East Africa, so he was accompanied by his more worldly, middle-aged neighbor, who translated for him. He had an infected sore on his leg, which I washed and treated with antibiotic ointment. He also had trouble seeing—“cataracts” was my barely educated guess—and I explained that they were beyond my meager curative powers. He seemed resigned, but not particularly disappointed, and as he sat there cross-legged, naked except for the blanket wrapped around him, basking in the sun, the woman stood behind him and stroked his head. In a voice as if describing last year’s weather she said, “Oh, when he was younger, he was beautiful and strong. Soon he will die.” That night in my tent, sleepless and jealous of the Masai, I thought, “I’ll take your malaria and parasites, I’ll take your appalling infant mortality rates, I’ll take the chances of being attacked by buffalo and lions. Just let me be as unafraid of dying as you are.”

An elderly hunter-gatherer shaman in the Kalahari Desert.

Maybe we will luck out and wind up as respected village elders. Perhaps we will grow old with grace and wisdom. Perhaps we will be honored, surrounded by strong, happy children whose health and fecundity will feel like immortality to us. Gerontologists studying the aging process

find increasing evidence that most of us will age with a fair degree of success. There’s far less institutionalization and disability than one might have guessed. While the size of social networks shrink with age, the quality of the relationships improves. There are types of cognitive skills that improve in old age (these are related to social intelligence and to making good strategic use of facts, rather than merely remembering them easily). The average elderly individual thinks his or her health is above average, and takes pleasure from that. And most important, the average level of happiness increases in old age; fewer negative emotions occur and, when they do, they don’t persist as long. Connected to this, brain-imaging studies show that negative images have less of an impact, and positive images have more of an impact on brain metabolism in older people, as compared to young.

So maybe old age is not so bad. The final chapter of this book reviews some of the patterns seen in aged people who are particularly successful in their aging. The purpose of this chapter is to review what stress has to do with the aging process and whether we wind up with the honored village elder model of aging, or the cat food variant.

Aged Organisms and Stress How do aged organisms deal with stress? Not very well, it turns out. In many ways, aging can be defined as the progressive loss of the ability to deal with stress, and that certainly fits our perception of aged individuals as fragile and vulnerable. This can be stated more rigorously by saying that many aspects of the bodies and minds of old organisms work fine, just as they do in young ones, so long as they aren’t pushed. Throw in an exercise challenge, an injury or illness, time pressure, novelty—any of a variety of physical, cognitive, or psychological stressors—and aged organisms don’t do so well. “Not doing so well” in the stress-response department can take at least two forms that should be familiar by now. The first is failing to activate a sufficient stress-response when it is needed. This occurs at many levels during aging. For example, individual cells have a variety of defenses they can mobilize in response to a challenge that can be viewed as a cellular stress-response. Heat a cell to an unhealthy extent and “heat shock proteins” are synthesized to help stabilize cellular function during a crisis. Damage DNA and DNA repair enzymes are activated. Generate oxygen

radicals and antioxidant enzymes are made in response. And all of these cellular stress-responses become less responsive to challenge during aging. A similar theme comes through at the level of how whole organ systems respond to stress. For example, after you eliminate from your study elderly people who have heart disease and look only at healthy subjects of different ages (so as to study aging, instead of inadvertently studying disease), many aspects of cardiac function are unchanged by age. But challenge the system with exercise, for example, and old hearts do not respond as adequately as do young ones, in that the maximal work capacity and the maximal heart rate that can be achieved are nowhere near as great as in a young person.* Similarly, in the absence of stress, old and young rat brains contain roughly the same amount of energy. But when you stress the system by cutting off the flow of oxygen and nutrients, energy levels decline faster in the old brains. Or, as a classic example, normal body temperature, 98.6 degrees, does not change with age. Nevertheless, aged bodies are impaired in mounting a thermoregulatory stress-response, and thus it takes the bodies of the elderly longer to restore a normal temperature after being warmed or chilled. The idea also applies to measures of cognition. What happens to IQ test scores as people get older? (You’ll notice that I didn’t say “intelligence.” What that has to do with IQ test scores is a controversy I don’t want to touch.) The dogma in the field was once that IQ declined with age. Then it was that it did not decline. It depends on how you test it. If you test young and old people and give them lots of time to complete the test, there is little difference. As you stress the system—in this case, by making the subjects race against a time limit—scores fall for all ages, but much further among older people. So sometimes the problem in aging is not enough of a stress-response. Predictably, in some realms, the problem is too much of a stress-response—either one turned on all the time, or one that takes too long to turn off at the end of a stressor. As an example, older individuals are impaired at turning off epinephrine, norepinephrine, or glucocorticoid secretion after a stressor has finished; it takes longer for levels of these substances to return to baseline. Moreover, even in the absence of the stressor, epinephrine, norepinephrine, and glucocorticoid levels are typically elevated in aged rats, nonhuman primates, and humans as well.* Do aged organisms pay a price for having these components of the stress-response turned on too often? This seems to be the case. As one example, which was discussed in the chapter on memory, stress and glucocorticoids inhibit the birth of new neurons in the adult hippocampus and inhibit the growth of new processes in preexisting neurons. Is the birth of new neurons and the elaboration of neuronal processes preferentially inhibited in old rats? Yes, and if their glucocorticoid levels are lowered, neurogenesis and process growth increase to levels seen in young animals. We know by now that, ideally, the hormones of the stress-response should be nice and quiet when nothing bad is happening, secreted in tiny amounts. When a stressful emergency hits, your body needs a huge and fast stress-response. At the end of the stressor, everything should shut off immediately. And these traits are precisely what old organisms typically lack.*

Why You Seldom See Really Old Salmon

We shift over to the other half of the aging-stress relationship—not whether aged organisms can deal well with stress, but whether stress can accelerate aspects of aging. There is some decent evidence that an excess of stress can increase the risk of some of the diseases of aging. Remarkably, it turns out that in more than a dozen species, glucocorticoid excess is the cause of death during aging. Pictures of heroic wild animals, à la Marlin Perkins: penguins who stand all winter amid the Antarctic cold, keeping their eggs warm at their feet. Leopards dragging massive kills up trees with their teeth, in order to eat them free of harassment by lions. Desiccated camels marching scores of miles. And then there’s salmon, leaping over dams and waterfalls to return to the freshwater stream of their birth. Where they spawn a zillion eggs. After which most of them die over the next few weeks. Why do salmon die so soon after spawning? No one is quite sure, but evolutionary biologists are rife with theories about why this and the rare other cases of “programmed die-offs” in the animal kingdom may make some evolutionary sense. What is known, however, is the proximal mechanism underlying the sudden die-off (not “How come they die, in terms of evolutionary patterns over the millennia?” but “How come they die, in the sense of which parts of the body’s functioning suddenly go crazy?”). It is glucocorticoid secretion.

A male sockeye salmon, after the onset of programmed aging.

If you catch salmon right after they spawn, just when they are looking a little green around the gills, you find they have huge adrenal glands, peptic ulcers, and kidney lesions; their immune systems have collapsed, and they are teeming with parasites and infections. Aha, kind of sounds like Selye’s rats way back when.* Moreover, the salmon have stupendously high glucocorticoid concentrations in their bloodstreams. When salmon spawn, regulation of their glucocorticoid secretion breaks down. Basically, the brain loses its ability to measure accurately the quantities of circulating hormones and keeps sending a signal to the adrenals to secrete more of them. Lots of glucocorticoids can certainly bring about all those diseases with which the salmon are festering. But is the glucocorticoid excess really responsible for their death? Yup. Take a salmon right after spawning, remove its adrenals, and it will live for a year afterward.

The bizarre thing is that this sequence of events not only occurs in five species of salmon, but also among a dozen species of Australian marsupial mice. All the male mice of these species die shortly after seasonal mating; cut out their adrenal glands, however, and they too keep living. Pacific salmon and marsupial mice are not close relatives. At least twice in evolutionary history, completely independently, two very different sets of species have come up with the identical trick: if you want to degenerate very fast, secrete a ton of glucocorticoids.

Chronic Stress and the Aging Process in the Mainstream

That is all fine for the salmon looking for the fountain of youth, but we and most other mammals age gradually over time, not in catastrophic die-offs over the course of days. Does stress influence the rate of gradual mammalian aging? Intuitively, the idea that stress accelerates the aging process makes sense. We recognize that there is a connection between how we live and how we die. Around 1900, a madly inspired German physiologist, Max Rubner, tried to define this connection scientifically. He looked at all sorts of different domestic species and calculated things like lifetime number of heartbeats and lifetime metabolic rate (not the sort of study that many scientists have tried to replicate). He concluded that there is only so long a body can go on—only so many breaths, so many heartbeats, so much metabolism that each pound of flesh can carry out before the mechanisms of life wear out. A rat, with approximately 400 heartbeats a minute, uses up its heartbeat allotment faster (after approximately two years) than an elephant (with approximately 35 beats per minute and a sixty- year life span). Such calculations lay behind ideas about why some species lived far longer than others. Soon the same sort of thinking was applied to how long different individuals within a species live—if you squander a lot of your heartbeats being nervous about blind dates when you’re sixteen, there would be that much less metabolic reserve available to you at eighty. In general, Rubner’s ideas about life spans among different species have not held up well in their strictest versions, while the “rate of living” hypotheses about individuals within a species that his ideas inspired have been even less tenable. Nevertheless, they led many people in the field to suggest that a lot of environmental perturbations can wear out the system prematurely. Such “wear and tear” thinking fit in naturally with the stress concept. As we have seen, excessive stress increases the risks of adult-onset diabetes, hypertension, cardiovascular disease, osteoporosis, reproductive decline, and immune suppression. All of these conditions become more common as we age. Moreover, in chapter 4 it was shown that if you have a lot of the indices of allostatic load, it increases your risk of Metabolic syndrome; that same study showed that it increased your mortality risk as well. We return to the tendency of very old rats, humans, and primates to have elevated resting levels of glucocorticoids in the bloodstream. Some aspect of the regulation of normal glucocorticoid secretion is disrupted during aging. To get a sense of why this happens, we must return to chapter 1’s interest about why the water tank on your toilet does not overflow when it’s refilling. Once again, the process of refilling can trigger a sensor—the flotation device—to decrease the amount of water flowing into the tank. Engineers who study this sort of thing term that process negative

feedback inhibition or end-product inhibition: increasing amounts of water accumulating in the tank decrease the likelihood of further release of water. Most hormonal systems, including the CRH/ACTH/glucocorticoid axis, work by this feedback- inhibition process. The brain triggers glucocorticoid release indirectly via CRH and pituitary release of ACTH. The brain needs to know whether to keep secreting more CRH. It does this by sensing the levels of glucocorticoids in the circulation (sampling the hormone from the bloodstream coursing through the brain) to see if levels are at, below, or above a “set point.” If levels are low, the brain keeps secreting CRH—just as when water levels in the toilet tank are still low. Once glucocorticoid levels reach or exceed that set point, there is a negative feedback signal and the brain stops secreting CRH. As a fascinating complication, the set point can shift. In the absence of stress, the brain wants different levels of glucocorticoids in the bloodstream from those required when something stressful is happening. (This implies that the quantity of glucocorticoids in the bloodstream necessary to turn off CRH secretion by the brain should vary with different situations, which turns out to be the case.) This is how the system works normally, as can be shown experimentally by injecting a person with a massive dose of a synthetic glucocorticoid (dexamethasone). The brain senses the sudden increase and says, in effect, “My God, I don’t know what is going on with those idiots in the adrenal, but they just secreted way too many glucocorticoids.” The dexamethasone exerts a negative feedback signal, and soon the person has stopped secreting CRH, ACTH, and her own glucocorticoids. This person would be characterized as “dexamethasone-responsive.” If negative feedback regulation is not working very well, however, the person is “dexamethasone- resistant”—she keeps secreting the various hormones, despite the whopping glucocorticoid signal in the bloodstream. And that is precisely what happens in old people, old nonhuman primates, and old rats. Glucocorticoid feedback regulation no longer works very well. This may explain why very old organisms secrete excessive glucocorticoids (in the absence of stress and during the recovery period after the end of a stressor). Why the failure of feedback regulation? There is a fair amount of evidence that it is due to the degeneration during aging of one part of the brain. The entire brain does not serve as a “glucocorticoid sensor” instead, that role is served by only a few areas with very high numbers of receptors for glucocorticoids and the means to tell the hypothalamus whether or not to secrete CRH. In chapter 10, I described how the hippocampus is famed for its role in learning and memory. As it turns out, it is also one of the important negative feedback sites in the brain for controlling glucocorticoid secretion. It also turns out that during aging, hippocampal neurons may become dysfunctional. When this occurs, some of the deleterious consequences include a tendency to secrete an excessive amount of glucocorticoids—this could be the reason aged people may have elevated resting levels of the hormone, may have trouble turning off secretion after the end of stress, or may be dexamethasone-resistant. It is as if one of the brakes on the system has been damaged, and hormone secretion rushes forward, a little out of control. The elevated glucocorticoid levels of old age, therefore, arise because of a problem with feedback regulation in the damaged hippocampus. Why are neurons damaged in the aging hippocampus? It’s glucocorticoid exposure, as was discussed in chapter 10.

If you’ve read carefully, you will begin to note something truly insidious embedded in these findings. When the hippocampus is damaged, the rat secretes more glucocorticoids. Which should damage the hippocampus further. Which should cause even more glucocorticoid secretion…. Each makes the other worse, causing a degenerative cascade that appears to occur in many aging rats, and whose potential pathological consequences have been detailed throughout virtually every page of this book. Does this degenerative cascade occur in humans? As noted, glucocorticoid levels rise with extreme old age in the human, and chapter 10 outlines the first evidence that these hormones might have some bad effects on the human hippocampus. The primate and human hippocampus appear to be negative feedback regulators of glucocorticoid release, such that hippocampal damage is associated with glucocorticoid excess, just as in the rodent. So the pieces of the cascade appear to be there in the human, raising the possibilities that histories of severe stress, or of heavy use of synthetic glucocorticoids to treat some disease, might accelerate aspects of this cascade.

George Segal, Man in a Chair, wood and plaster, 1969.

Does that mean that all is lost, that this sort of dysfunction is an obligatory part of aging? Certainly not. It was not by chance that two paragraphs above, I described this cascade as occurring in “many” aging rats, rather than in “all.” Some rats age successfully in a way that spares them this cascade, as do many humans—these pleasing stories are part of the final chapter of this book. It is thus not yet clear whether the “glucocorticoid neurotoxicity” story applies to how our brains age. Unfortunately, the answer is not likely to be available for years; the subject is difficult to study in humans. Nevertheless, from what we know about this process in the rat and monkey, glucocorticoid toxicity stands as a striking example of ways in which stress can accelerate aging. Should it turn out to apply to us as well, it will be an aspect of our aging that will harbor a special threat. If we are crippled by an accident, if we lose our sight or hearing, if we are so weakened by heart disease as to be bed-bound, we cease having so many of the things that make

our lives worth living. But when it is our brains that are damaged, when it is our ability to recall old memories or to form new ones that is destroyed, we fear we’ll cease to exist as sentient, unique individuals—the version of aging that haunts us most. Even the most stoic of readers should be pretty frazzled by now, given the detailing in the twelve chapters so far about the sheer number of things that can go wrong with stress. It is time to shift to the second half of the book, which examines stress management, coping, and individual differences in the stress-response. It is time to begin to get some good news. Chapter 13

Why Is Psychological Stress Stressful?

Some people are born to biology. You can spot them instantly as kids—they’re the ones comfortably lugging around the toy microscopes, dissecting some dead animal on the dining room table, being ostracized at school for their obsession with geckos.* But all sorts of folks migrate to biology from other fields—chemists, psychologists, physicists, mathematicians. Several decades after stress physiology began, the discipline was inundated by people who had spent their formative years as engineers. Like physiologists, they thought there was a ferocious logic to how the body worked, but for bioengineers, that tended to mean viewing the body a bit like the circuitry diagram that you get with a radio: input-output ratios, impedance, feedback loops, servomechanisms. I shudder even to write such words, as I barely understand them; but the bioengineers did wonders for the field, adding a tremendous vigor. Suppose you wonder how the brain knows when to stop glucocorticoid secretion—when enough is enough. In a vague sort of way, everyone knew that somehow the brain must be able to measure the amount of glucocorticoids in the circulation, compare that to some desired set point, and then decide whether to continue secreting CRH or turn off the faucet (returning to the toilet tank model). The bioengineers came in and showed that the process was vastly more interesting and complicated than anyone had imagined. There are “multiple feedback domains” some of the time the brain measures the quantity of glucocorticoids in the bloodstream, and sometimes the rate at which the level is changing. The bioengineers solved another critical issue: Is the stress- response linear or all-or-nothing? Epinephrine, glucocorticoids, prolactin, and other substances are all secreted during stress; but are they secreted to the same extent regardless of the intensity of the stressor (all-or-nothing responsiveness)? The system turns out to be incredibly sensitive to the size of the stressor, demonstrating a linear relationship between, for example, the extent of the drop in blood pressure and the extent of epinephrine secretion, between the degree of hypoglycemia (drop in blood sugar) and glucagon release. The body not only can sense something stressful, but it also is amazingly accurate at measuring just how far and how fast that stressor is throwing the body out of allostatic balance. Beautiful stuff, and important. Hans Selye loved the bioengineers, which makes perfect sense, since in his time the whole stress field must have still seemed a bit soft-headed to some mainstream physiologists. Those physiologists knew that the body does one set of things when it is too cold, and a diametrically opposite set when it is too hot, but here were Selye and his crew

insisting that there were physiological mechanisms that respond equally to cold and hot? And to injury and hypoglycemia and hypotension? The beleaguered stress experts welcomed the bioengineers with open arms. “You see, it’s for real; you can do math about stress, construct flow charts, feedback loops, formulas….” Golden days for the business. If the system was turning out to be far more complicated than ever anticipated, it was complicated in a way that was precise, logical, mechanistic. Soon it would be possible to model the body as one big input-output relationship: you tell me exactly to what degree a stressor impinges on an organism (how much it disrupts the allostasis of blood sugar, fluid volume, optimal temperature, and so on), and I’ll tell you exactly how much of a stress-response will occur. This approach, fine for most of the ground that we’ve covered up until now, will probably allow us to estimate quite accurately what the pancreas of that zebra is doing when the organism is sprinting from a lion. But the approach is not going to tell us which of us will get an ulcer when the factory closes down. Starting in the late 1950s, a new style of experiments in stress physiology began to be conducted that burst that lucid, mechanistic bioengineering bubble. A single example will suffice. An organism is subjected to a painful stimulus, and you are interested in how great a stress-response will be triggered. The bioengineers had been all over that one, mapping the relationship between the intensity and duration of the stimulus and the response. But this time, when the painful stimulus occurs, the organism under study can reach out for its mommy and cry in her arms. Under these circumstances, this organism shows less of a stress-response. Nothing in that clean, mechanistic world of the bioengineers could explain this phenomenon. The input was still the same; the same number of pain receptors should have been firing while the child underwent some painful procedure. Yet the output was completely different. A critical realization roared through the research community: the physiological stress-response can be modulated by psychological factors. Two identical stressors with the same extent of allostatic disruption can be perceived, can be appraised differently, and the whole show changes from there. Suddenly the stress-response could be made bigger or smaller, depending on psychological factors. In other words, psychological variables could modulate the stress-response. Inevitably, the next step was demonstrated: in the absence of any change in physiological reality—any actual disruption of allostasis—psychological variables alone could trigger the stress-response. Flushed with excitement, Yale physiologist John Mason, one of the leaders in this approach, even went so far as to proclaim that all stress-responses were psychological stress-responses. The old guard was not amused. Just when the conception of stress was becoming systematized, rigorous, credible, along came this rabble of psychologists muddying up the picture. In a series of published exchanges in which they first praised each other’s achievements and ancestors, Selye and Mason attempted to shred each other’s work. Mason smugly pointed to the growing literature on psychological initiation and modulation of the stress-response. Selye, facing defeat, insisted that all stress-responses couldn’t be psychological and perceptual: if an organism is anesthetized, it still gets a stress-response when a surgical incision is made.

The psychologists succeeded in getting a place at the table, and as they have acquired some table manners and a few gray hairs, they have been treated less like barbarians. We now have to consider which psychological variables are critical. Why is psychological stress stressful?

The Building Blocks of Psychological Stressors

Outlets for frustration You would expect key psychological variables to be mushy concepts to uncover, but in a series of elegant experiments, the physiologist Jay Weiss, then at Rockefeller University, demonstrated exactly what is involved. The subject of one experiment is a rat that receives mild electric shocks (roughly equivalent to the static shock you might get from scuffing your foot on a carpet). Over a series of these, the rat develops a prolonged stress-response: its heart rate and glucocorticoid secretion rate go up, for example. For convenience, we can express the long-term consequences by how likely the rat is to get an ulcer, and in this situation, the probability soars. In the next room, a different rat gets the same series of shocks—identical pattern and intensity; its allostatic balance is challenged to exactly the same extent. But this time, whenever the rat gets a shock, it can run over to a bar of wood and gnaw on it. The rat in this situation is far less likely to get an ulcer. You have given it an outlet for frustration. Other types of outlets work as well—let the stressed rat eat something, drink water, or sprint on a running wheel, and it is less likely to develop an ulcer. We humans also deal better with stressors when we have outlets for frustration—punch a wall, take a run, find solace in a hobby. We are even cerebral enough to imagine those outlets and derive some relief: consider the prisoner of war who spends hours imagining a golf game in tremendous detail. I have a friend who passed a prolonged and very stressful illness lying in bed with a mechanical pencil and a notepad, drawing topographic maps of imaginary mountain ranges and taking hikes through them. A central feature of an outlet being effective is if it distracts from the stressor. But, obviously, more important is that it also be something positive for you—a reminder that there is more to life than whatever is making you crazed and stressed at the time. The frustration-reducing effects of exercise provide an additional layer of benefit, one harking back to my dichotomy, repeated ad nauseam, between the zebra running for its life and the psychologically stressed human. The stress-response is about preparing your body for an explosive burst of energy consumption right now; psychological stress is about doing all the same things to your body for no physical reason whatsoever. Exercise finally provides your body for the outlet that it was preparing for. A variant of Weiss’s experiment uncovers a special feature of the outlet-for-frustration reaction. This time, when the rat gets the identical series of electric shocks and is upset, it can run across the cage, sit next to another rat and…bite the hell out of it. Stress-induced displacement of aggression: the practice works wonders at minimizing the stressfulness of a stressor. It’s a real primate specialty as well. A male baboon loses a fight. Frustrated, he spins around and attacks a subordinate male who was minding his own business. An extremely high percentage of primate aggression represents frustration displaced onto innocent bystanders. Humans are pretty good at it, too, and we have a technical way of describing the phenomenon in the context of stress-related

disease: “He’s one of those guys who doesn’t get ulcers, he gives them.” Taking it out on someone else—how well it works at minimizing the impact of a stressor. Social support An additional way we can interact with another organism to minimize the impact of a stressor on us is considerably more encouraging for the future of our planet than is displacement aggression. Rats only occasionally use it, but primates are great at it. Put a primate through something unpleasant: it gets a stress-response. Put it through the same stressor while in a room full of other primates and…it depends. If those primates are strangers, the stress-response gets worse. But if they are friends, the stress-response is decreased. Social support networks—it helps to have a shoulder to cry on, a hand to hold, an ear to listen to you, someone to cradle you and to tell you it will be okay. The same is seen with primates in the wild. While I mostly do laboratory research on how stress and glucocorticoids affect the brain, I spend my summers in Kenya studying patterns of stress- related physiology and disease among wild baboons living in a national park. The social life of a male baboon can be pretty stressful—you get beaten up as a victim of displaced aggression; you carefully search for some tuber to eat and clean it off, only to have it stolen by someone of higher rank; and so on. Glucocorticoid levels are elevated among low-ranking baboons and among the entire group if the dominance hierarchy is unstable, or if a new aggressive male has just joined the troop. But if you are a male baboon with a lot of friends, you are likely to have lower glucocorticoid concentrations than males of the same general rank who lack these outlets. And what counts as friends? You play with kids, have frequent nonsexual grooming bouts with females (and social grooming in nonhuman primates lowers blood pressure). Social support is certainly protective for humans as well. This can be demonstrated even in transient instances of support. In a number of subtle studies, subjects were exposed to a stressor such as having to give a public speech or perform a mental arithmetic task, or having two strangers argue with them, with or without a supportive friend present. In each case, social support translated into less of a cardiovascular stress-response. Profound and persistent differences in degrees of social support can influence human physiology as well: within the same family, there are significantly higher glucocorticoid levels among stepchildren than among biological children. Or, as another example, among women with metastatic breast cancer, the more social support, the lower the resting cortisol levels.

George Tooker, Landscape with Figures, egg tempera on gesso, 1966.

As noted in chapter 8, people with spouses or close friends have longer life expectancies. When the spouse dies, the risk of dying rises. Recall also from that chapter the study of parents of Israeli soldiers killed in the Lebanon war: in the aftermath of this stressor, there was no notable increase in risk of diseases or mortality, except among those who were already divorced or widowed. Some additional examples concern the cardiovascular system. People who are socially isolated have overly active sympathetic nervous systems. Given the likelihood that this will lead to higher blood pressure and more platelet aggregation in their blood vessels (remember that from chapter 3?), they are more likely to have heart disease—two to five times as likely, as it turns out. And once they have the heart disease, they are more likely to die at a younger age. In a study of patients with severe coronary heart disease, Redford Williams of Duke University and colleagues found that half of those lacking social support were dead within five years—a rate three times higher than was seen in patients who had a spouse or close friend, after controlling for the severity of the heart disease.* Finally, support can exist at the broad community level (stay tuned for chapter 17). If you are a member of an ethnic minority, the fewer members there are of your group in your neighborhood, the higher your risks of mental illness, psychiatric hospitalization, and suicide. Predictability Weiss’s rat studies uncovered another variable modulating the stress-response. The rat gets the same pattern of electric shocks, but this time, just before each shock, it hears a warning bell. Fewer ulcers. Predictability makes stressors less stressful. The rat with the warning gets two pieces of information. It learns when something dreadful is about to happen. The rest of the time, it learns that something dreadful is not about to happen. It can relax. The rat without a warning can always be a half-second away from the next shock. In effect, information that increases predictability tells you that there is bad news, but comforts you that it’s not going to be worse—you are going to get shocked soon, but it’s never going to be sprung on you without warning. We all know a human equivalent of this principle: you’re in the dentist’s chair, no novocaine, the dentist drilling away. Ten seconds of nerve-curling pain, some rinsing, five seconds of drilling, a pause while the dentist fumbles a bit, fifteen seconds of drilling, and so on. In one of the pauses, frazzled and trying not to whimper, you gasp, “Almost done?” “Hard to say,” the dentist mumbles, returning to the intermittent drilling. Think how grateful we are for the dentist who, instead, says, “Two more and we’re done.” The instant the second burst of drilling ends, down goes blood pressure. By being given news about the stressor to come, you are also implicitly being comforted by now knowing what stressors are not coming. As another variant on the helpfulness of predictability, organisms will eventually habituate to a stressor if it is applied over and over; it may knock physiological allostasis equally out of balance the umpteenth time that it happens, but it is a familiar, predictable stressor by then, and a smaller stress-response is triggered. One classic demonstration involved men in the Norwegian military going through parachute training—as the process went from being hair-raisingly novel to something they could do in their sleep, their anticipatory stress-response went from being gargantuan to nonexistent.

The power of loss of predictability as a psychological stressor is shown in an elegant, subtle study. A rat is going about its business in its cage, and at measured intervals the experimenter delivers a piece of food down a chute into the cage; rat eats happily. This is called an intermittent reinforcement schedule. Now, change the pattern of food delivery so that the rat gets exactly the same total amount of food over the course of an hour, but at a random rate. The rat receives just as much reward, but less predictably, and up go glucocorticoid levels. There is not a single physically stressful thing going on in the rat’s world. It’s not hungry, pained, running for its life—nothing is out of allostatic balance. In the absence of any stressor, loss of predictability triggers a stress-response. There are even circumstances in which a stress-response can be more likely to occur in someone despite the reality that the outside world is less stressful. Work by the zoologist John Wingfield of the University of Washington has shown an example of this with wild birds. Consider some species that migrates between the Arctic and the tropics. Bird #1 is in the Arctic, where the temperature averages 5 degrees and where it is, indeed, 5 degrees outside that day. In contrast, Bird #2 is in the tropics, where the average temperature is 80 degrees, but today it has dropped down to 60. Who has the bigger stress-response? Amazingly, Bird #2. The point isn’t that the temperature in the tropics is 55 degrees warmer than in the Arctic (what kind of stressor would that be?). It’s that the temperature in the tropics is 20 degrees colder than anticipated. A human version of the same idea has been documented. During the onset of the Nazi blitzkrieg bombings of England, London was hit every night like clockwork. Lots of stress. In the suburbs the bombings were far more sporadic, occurring perhaps once a week. Fewer stressors, but much less predictability. There was a significant increase in the incidence of ulcers during that time. Who developed more ulcers? The suburban population. (As another measure of the importance of unpredictability, by the third month of the bombing, ulcer rates in all the hospitals had dropped back to normal.) Despite the similarity between the responses of humans and of other animals to a lack of predictability, I suspect that there they are not identical, and in an important way. The warning of impending shocks to a rat has little effect on the size of the stress-response during the shocks; instead, allowing the rat to feel more confident about when it doesn’t have to worry reduces the rat’s anticipatory stress-response the rest of the time. Analogously, when the dentist says, “Only two more times and then we’re done,” it allows us to relax at the end of the second burst of drilling. But I suggest, although I cannot prove it, that unlike the case for the rat, proper information will also lower our stress-response during the pain. If you were told “only two times more” versus “only ten times more,” wouldn’t you use different mental strategies to try to cope? With either scenario, you would pull out the comforting thought of “only one more and then it’s the last one” at different times; you would save your most distracting fantasy for a different point; you would try counting to zero from different numbers. Predictive information lets us know what internal coping strategy is likely to work best during a stressor. We often wish for information about the course of some medical problem because it aids our strategizing about how we will cope. A simple example: you have some minor surgery, and you’re given predictive information—the first post-surgical day, there is going to be a lot of pain, pretty constant, whereas by the second day, you’ll just feel a bit achy. Armed with that

information, you are more likely to plan on watching the eight distracting videos on day one and to devote day two to writing delicate haikus than the other way around. Among other reasons, we wish to optimize our coping strategies when we request the most devastating piece of medical information any of us will ever face: “How much time do I have left?” Control Rat studies also demonstrate a related facet of psychological stress. Give the rat the same series of shocks. This time, however, you study a rat that has been trained to press a lever to avoid electric shocks. Take away the lever, shock it, and the rat develops a massive stress- response. It’s as if the rat were thinking, “I can’t believe this. I know what to do about electric shocks; give me a damned lever and I could handle this. This isn’t fair.” Ulceration city (as well as higher glucocorticoid levels, poorer immune function, and faster tumor growth). Give the trained rat a lever to press; even if it is disconnected from the shock mechanism, it still helps: down goes the stress-response. So long as the rat has been exposed to a higher rate of shocks previously, it will think that the lower rate now is due to its having control over the situation. This is an extraordinarily powerful variable in modulating the stress-response. The identical style of experiment with humans yields similar results. Place two people in adjoining rooms, and expose both to intermittent noxious, loud noises; the person who has a button and believes that pressing it decreases the likelihood of more noise is less hypertensive. In one variant on this experiment, subjects with the button who did not bother to press it did just as well as those who actually pressed the button. Thus, the exercise of control is not critical; rather, it is the belief that you have it. An everyday example: airplanes are safer than cars, yet more of us are phobic about flying. Why? Because your average driver believes that he is a better-than- average driver, thus more in control. In an airplane, we have no control at all. My wife and I tease each other on plane flights, exchanging control: “Okay, you rest for a while, I’ll take over concentrating on keeping the pilot from having a stroke.” The issue of control runs through the literature on the psychology of stress. As will be discussed in the final chapter on coping, exercise can be a great stress reducer, but only so long as it is something that seems even remotely desirable. Amazingly, the same is seen in a rat—let a rat run voluntarily in a running wheel, and it makes it feel great. Force a rat to do the same amount of exercise and it gets a massive stress-response. The issue of control runs through the extensive literature on occupational stress. Sure, there are some jobs where stress comes in the form of someone having too much control and responsibility—that rare occupation where, over the course of an average workday, you might find yourself having to direct the landing pattern of an array of circling jumbo jets at the local airport, personally excise someone’s cerebral aneurysm, and make the final decision as to whether taffeta is going to be in at the fall runway show in Milan. For most, though, occupational stress is built more around lack of control, work life spent as a piece of the machine. Endless studies have shown that the link between occupational stress and increased risk of cardiovascular and metabolic diseases is anchored in the killer combination of high demand and low control— you have to work hard, a lot is expected of you, and you have minimal control over the process. This is the epitome of the assembly line, the combination of stressors that makes for Marx’s alienation of the workers. The control element is more powerful than the demand one—low demand and low control is more damaging to one’s health than high demand and high control.

The stressfulness of lack of control on the job applies in only certain domains, however. For example, there is the issue of what product is made, and lack of control in this realm tends not to be all that stressful—few people are ulcerating because of their deep conviction that all of their capable and motivated fellow workers should be cranking vast numbers of stuffed Snoopys out of this factory instead of ball bearings. Instead, it is stress about lack of control over the process—what work rate is expected and how much flexibility there is about it, what amenities there are and how much control you have over them, how authoritarian the authorities are. These issues can apply just as readily to some less expected workplaces, ones that can be highly prestigious and desirable. For example, professional musicians in orchestras generally have lower job satisfaction and more stress than those in small chamber groups (such as a string quartet). Why? One pair of researchers suggest that this is because of the lack of autonomy in an orchestra, where centuries of tradition hold that orchestras are subservient to the dictatorial whims of the maestro conducting them. For example, it was only in recent years that orchestra unions won the right for regularly scheduled bathroom breaks during rehearsals, instead of having to wait until the conductor cared to note how squirmy the reed players had become.* So the variable of control is extremely important; controlling the rewards that you get can be more desirable than getting them for nothing. As an extraordinary example, both pigeons and rats prefer to press a lever in order to obtain food (so long as the task is not too difficult) over having the food delivered freely—a theme found in the activities and statements of many scions of great fortunes, who regret the contingency-free nature of their lives, without purpose or striving. Loss of control and lack of predictive information are closely related. Some researchers have emphasized this, pointing out that the common theme is that the organism is subjected to novelty. You thought you knew how to manage things, you thought you knew what would happen next, and it turns out you are wrong in this novel situation. The potency of this is demonstrated in primate studies in which merely placing the animal into a novel cage suppresses its immune system. Others have emphasized that these types of stressors cause arousal and vigilance, as you search for the new rules of control and prediction. Both views are different aspects of the same issue. A perception of things worsening Yet another critical psychological variable in the stress- response has been uncovered. A hypothetical example: two rats get a series of electric shocks. On the first day, one gets ten shocks an hour; the other, fifty. Next day, both get twenty-five shocks an hour. Who becomes hypertensive? Obviously, the one going from ten to twenty-five. The other rat is thinking, “Twenty-five? Piece of cheese, no problem; I can handle that.” Given the same degree of disruption of allostasis, a perception that events are improving helps tremendously. The principle often pops up in the realm of human illness. Recall in chapter 9 the scenario where pain is less stressful, can even be welcome, when it means, for example, that the drugs are working, the tumor is shrinking. One classic study demonstrated that in examining parents of children who had a 25 percent chance of dying of cancer. Astonishingly, these parents showed only a moderate rise in glucocorticoid levels in the bloodstream. How could that be? Because the children were all in remission after a period in which the odds of death had been far higher.

Twenty-five percent must have seemed like a miracle. Twenty-five shocks an hour, a certain degree of social instability, a one-in-four chance of your child dying—each can imply either good news or bad, and only the latter seems to stimulate a stress-response. It’s not just the external reality; it’s the meaning you attach to it. A version of this can be observed among the baboons I study in Kenya. In general, when dominance hierarchies are unstable, resting glucocorticoid levels rise. This makes sense, because such instabilities make for stressful times. Looking at individual baboons, however, shows a more subtle pattern: given the same degree of instability, males whose ranks are dropping have elevated glucocorticoid levels, while males whose ranks are rising amid the tumult don’t show this endocrine trait.

Not So Fast Thus, some powerful psychological factors can trigger a stress-response on their own or make another stressor seem more stressful: loss of control or predictability, loss of outlets for frustration or sources of support, a perception that things are getting worse. There are obviously some overlaps in the meaning of these different factors. As we saw, control and predictability are closely aligned; combine them with a perception of things worsening, and you have the situation of bad things happening, out of your control, and utterly unpredictable. The primatologist Joan Silk of UCLA has emphasized how, among primates, a great way to maintain dominance is for the alpha individual to mete out aggression in a randomly brutal way. This is our primate essence of terrorism. Sometimes these different variables conflict and it becomes a question as to which is more powerful. This often involves a dichotomy between control/predictability issues and the perception of whether things are improving or worsening. For example, someone unexpectedly wins the lottery big-time. Is this a stressor? It depends on what is more powerful, the beneficial “perception of things getting better” part or the stressful “lack of predictability” part. Not surprisingly, if the lottery win is big enough, most people’s psyches can handle some unpredictability. Nonetheless, some nonhuman primate studies in which rank was manipulated by the experimenters show that it can go in the other way, that if the change is sufficiently unexpected, it can be stressful, even if it is good change (and psychotherapy often must delve into the reasons why people sometimes find change for the good to be less desirable than persisting with a known misery). Conversely, if a situation is sufficiently awful, the fact that it may have been predictable offers little comfort. These factors play a major role in explaining how we all go through lives full of stressors, yet differ so dramatically in our vulnerability to them. The final chapter of this book examines the bases of these individual differences in greater detail. This will serve as a blueprint for analyzing how to learn to exploit these psychological variables—how, in effect, to manage stress better. The ways in which these different psychological variables can interact brings up a key point, one that will dominate the final chapter. This is that stress management cannot consist merely of the simpleminded solution of “Maximize control. Maximize predictability. Maximize outlets for frustration.” As we will now see, it is considerably more complicated than that. As the most

obvious first pass at this, some lack of control and predictability can be a great thing—a good roller-coaster ride, a superbly terrifying movie, a mystery novel with a great surprise ending, winning a lottery, being subject to a random act of kindness. And sometimes, an overabundance of predictability is a disaster—boredom on the job. The right amounts of loss of control and predictability are what we call stimulation. In chapter 16, we will look at the biology of why stimulation makes us happy, rather than stressed. The goal is never to generate lives in which there is never a challenge to allostasis. And the remainder of this chapter considers when increasing a sense of control and predictability reduces stress.

Some Subtleties of Predictability

We have already seen how predictability can ameliorate the consequences of stress: one rat gets a series of shocks and develops a higher risk for an ulcer than the rat who gets warnings beforehand. Predictability doesn’t always help, however. The experimental literature on this is pretty dense; some human examples of this point make it more accessible. (Remember, in these scenarios, the stressor is inevitable; the warning cannot change the stressor, just the perception of it.) How predictable is the stressor, in the absence of a warning? What if, one morning, an omnipotent voice says, “There is no way out of it; a meteor is going to crush your car while you’re at work today (but it’s the only time it will happen this year).” Not soothing. There’s the good news that it’s not going to happen again tomorrow, but that’s hardly comforting; this is not an event that you anxiously fret over often. At the other extreme, what if one morning an

omnipotent voice whispers, “Today it’s going to be stressful on the freeway—lots of traffic, stops and go’s. Tomorrow, too. In fact, every day this year, except November 9, when there’ll hardly be any traffic, people will wave to each other, and a highway patrol cop will stop you in order to share his coffee cake with you.” Who needs predictive information about the obvious fact that driving to work is going to be stressful? Thus, warnings are less effective for very rare stressors (you don’t usually worry much about meteors) and very frequent ones (they approach being predictable even without the warning). How far in advance of the stressor does the warning come? Each day, you go for a mysterious appointment: you are led into a room with your eyes closed and are seated in a deep, comfortable chair. Then, with roughly even probabilities but no warning, either a rich, avuncular voice reads you to sleep with your favorite childhood stories, or a bucket of ice water is sloshed over your head. Not a pleasing prospect, I would bet. Would the whole thing be any less unsettling if you were told which treatment you were going to get five seconds before the event? Probably not— there is not enough time to derive any psychological benefits from the information. At the other extreme, how about predictive information long in the future? Would you wish for an omnipotent voice to tell you, “Eleven years and twenty-seven days from now your ice-water bath will last ten full minutes”? Information either just before or long before the stressor does little good to alleviate the psychological anticipation. Some types of predictive information can even increase the cumulative anticipatory stressor. For example, if the stressor is truly terrible. Would you be comforted by the omnipotent message: “Tomorrow an unavoidable accident will mangle your left leg, although your right leg will remain in great shape”? Likewise, predictive information can make things worse if the information is vague. As I write this section, we continue to be stressed by the maddening vagueness of predictive information in our post-9/11 world, when we are given warnings that read like horoscopes from hell: “Orange Alert: We don’t know what the threat is, but be extra alert about everything for the next few days.”* Collectively, these scenarios tell us that predictability does not always work to protect us from stress. The much more systematic studies with animals suggest that it works only in a midrange of frequencies and intensities of stressors, and with certain lag times and levels of accurate information.

Subtleties of Control To understand some important subtleties of the effects of control on stress, we need to return to the paradigm of the rat being shocked. It had been previously trained to press a lever to avoid shocks, and now it’s pounding away like crazy on a lever. The lever does nothing; the rat is still getting shocked, but with less chance of an ulcer because the rat thinks it has control. To introduce a sense of control into the experimental design decreases the stress-response because, in effect, the rat is thinking, “Ten shocks an hour. Not bad; just imagine how bad it would be if I wasn’t on top of it with my lever here.” But what if things backfire, and adding a sense of control makes the rat think, “Ten shocks an hour, what’s wrong with me? I have a lever here, I should

have avoided the shocks, it’s my fault.” If you believe you have control over stressors that are, in fact, beyond your control, you may consider it somehow to be your fault that the inevitable occurred. An inappropriate sense of control in the face of awful events can make us feel terrible. Some of our most compassionate words to people experiencing tragedy involve minimizing their perceived sense of control. “It’s not your fault, no one could have stopped in time; she just darted out from between the cars.” “It’s not something you could have done anything about; you tried your best, the economy’s just lousy now.” “Honey, getting him the best doctor in the world couldn’t have cured him.” And some of the most brutally callous of society’s attempts to shift blame attribute more personal control during a stressor than exists. “She was asking for it if she was going to dress that way” (rape victims have the control to prevent the rape). “Your child’s schizophrenia was caused by your mothering style” (this was a destructive belief that dominated psychiatry for decades before the disease was recognized to be neurochemical). “If they’d only made the effort to assimilate, they wouldn’t have these problems” (minorities have the power to prevent their persecution). The effects of the sense of control on stress are highly dependent on context. In general, if the stressor is of a sort where it is easy to imagine how much worse it could have been, inserting an artificial sense of control helps. “That was awful, but think of how bad it would have been if I hadn’t done X.” But when the stressor is truly awful, an artificial sense of control is damaging— it is difficult to conceive a yet-worse scenario that you managed to avoid, but easy to be appalled by the disaster you didn’t prevent. You don’t want to feel as if you could have controlled the uncontrollable when the outcome is awful. People with a strong internal locus of control (in other words, people who think they are the masters of their own ship—that what goes on around them reflects their actions) have far greater stress-responses than do those with external loci when confronted with something uncontrollable. This is a particular risk for the elderly (especially elderly men) as life generates more and more things beyond their control. As we will see in the final chapter, there is even a personality type whose tendency to internalize control in the face of bad, uncontrollable things greatly increases the risk of a particular disease. These subtleties about control and predictability help to explain a confusing feature about the studies of stress. In general, the less control or predictability, the more at risk you are for a stress- induced disease. Yet an experiment conducted by Joseph Brady in 1958 with monkeys gave rise to the view that more control and more predictability cause ulcers. Half the animals could press a bar to delay shocks (“executive” monkeys); the other half were passively yoked to one of the “executives” such that they received a shock whenever the first one did. In this widely reported study, the executive monkeys were more likely to develop ulcers. Out of these studies came the popular concept of the “executive stress syndrome” and associated images of executive humans weighed down with the stressful burdens of control, leadership, and responsibility. Ben Natelson, of the VA Medical Center in East Orange, New Jersey, along with Jay Weiss, noted some problems with that study. First, it was conducted with parameters where control and predictability are bad news. Second, the “executive” and “nonexecutive” monkeys were not chosen randomly; instead, the monkeys that tended to press the bar first in pilot studies were selected to be executives. Monkeys that press sooner have since been shown to be more emotionally reactive animals, so Brady was inadvertently stacking the executive side with the

more reactive, ulcer-prone monkeys. In general, executives of all species are more likely to be giving ulcers than to be getting them, as we will see in chapter 17. To summarize, stress-responses can be modulated or even caused by psychological factors, including loss of outlets for frustration and of social support, a perception of things worsening, and under some circumstances, a loss of control and of predictability. These ideas have vastly expanded our ability to answer the question: Why do only some of us get stress-related diseases? Obviously we differ as to the number of stressors that befall us. After all the chapters on physiology, you can guess that we differ in how fast our adrenals make glucocorticoids, how many insulin receptors we have in our fat cells, the thickness of our stomach walls, and so on. But in addition to those physiological differences, we can now add another dimension. We differ in the psychological filters through which we perceive the stressors in our world. Two people participating in the same event—a long wait at the supermarket checkout, public speaking, parachuting out of an airplane—may differ dramatically in their psychological perception of the event. “Oh, I’ll just read a magazine while I wait” (outlet for frustration); “I’m nervous as hell, but by giving this after-dinner talk, I’m a shoo-in for that promotion” (things are getting better); “This is great—I’ve always wanted to try sky-diving” (this is something I’m in control of). In the next two chapters we will consider psychiatric disorders such as depression and anxiety, and personality disorders, in which there’s a bad match between how stressful the real world is and how stressful the person perceives it to be. As we’ll see, the mismatch between the two can take a variety of forms, but the thing in common is the fact that a potentially considerable price is paid by the sufferer. Following that, in chapter 16, we consider what psychological stress has to do with the process of addiction. Following that is a chapter examining how your place in society, and the type of society it is, can have profound effects on stress physiology and patterns of disease. In the final chapter we will examine how stress-management techniques can aid us by teaching how to exploit these psychological defenses. Chapter 14

Stress and Depression

We are morbidly fascinated with the exotica of disease. They fill our made-for-television movies, our tabloids, and the book reports of adolescents hoping to become doctors someday. Victorians with Elephant Man’s disease, murderers with multiple personality disorders, ten-year- olds with progeria, idiot savants with autism, cannibals with kuru. Who could resist? But when it comes to the bread and butter of human misery, try a major depression. It can be life-threatening, it can destroy lives, demolish the families of sufferers. And it is dizzyingly common—the psychologist Martin Seligman has called it the common cold of psychopathology. Best estimates are that from 5 to 20 percent of us will suffer a major, incapacitating depression at some point in our lives, causing us to be hospitalized or medicated or nonfunctional for a significant length of time. Its incidence has been steadily increasing for decades—by the year 2020, depression is projected to be the second leading cause of medical disability on earth. This chapter differs a bit from those that preceded it in which the concept of “stress” was at the forefront. Initially, that may not seem to be the case in our focus on depression. The two appear

to be inextricably linked, however, and the concept of stress will run through every page of this chapter. It is impossible to understand either the biology or psychology of major depressions without recognizing the critical role played in the disease by stress. To begin to understand this connection, it is necessary to get some sense of the disorder’s characteristics. We have first to wrestle with a semantic problem. Depression is a term that we all use in an everyday sense. Something mildly or fairly upsetting happens to us, and we get “the blues” for a while, followed by recovery. This is not what occurs in a major depression. One issue is chronicity—for a major depression to be occurring, the symptoms have to have persisted for at least two weeks. The other is severity—this is a vastly crippling disorder that leads people to attempt suicide; its victims may lose their jobs, family, and all social contact because they cannot force themselves to get out of bed, or refuse to go to a psychiatrist because they feel they don’t deserve to get better. It is a horrific disease, and throughout this chapter I will be referring to this major, devastating form of depression, rather than the transient blues that we may casually signify with the term “feeling depressed.”

The Symptoms The defining feature of a major depression is loss of pleasure. If I had to define a major depression in a single sentence, I would describe it as a “genetic/neurochemical disorder requiring a strong environmental trigger whose characteristic manifestation is an inability to appreciate sunsets.” Depression can be as tragic as cancer or a spinal cord injury. Think about what our lives are about. None of us will live forever, and on occasion we actually believe it; our days are filled with disappointments, failures, unrequited loves. Despite this, almost inconceivably, we not only cope but even feel vast pleasures. I, for example, am resoundingly mediocre at soccer, but nothing keeps me from my twice-weekly game. Invariably there comes a moment when I manage to gum up someone more adept than I; I’m panting and heaving and pleased, and there’s still plenty more time to play and a breeze blows and I suddenly feel dizzy with gratitude for my animal existence. What could be more tragic than a disease that, as its defining symptom, robs us of that capacity? This trait is called anhedonia: hedonism is “the pursuit of pleasure,” anhedonia is “the inability to feel pleasure” (also often called dysphoria—I’ll be using the terms interchangeably). Anhedonia is consistent among depressives. A woman has just received the long-sought promotion; a man has just become engaged to the woman of his dreams—and, amid their depression, they will tell you how they feel nothing, how it really doesn’t count, how they don’t deserve it. Friendship, achievement, sex, food, humor—none can bring any pleasure. This is the classic picture of depression, and some recent research, much of it built around work of the psychologist Alex Zautra of the University of Arizona, shows that the story is more complex. Specifically, positive and negative emotions are not mere opposites. If you take subjects and, at random times throughout the day, have them record how they are feeling at that moment, the frequencies of feeling good and feeling bad are not inversely correlated. There’s normally not much of a connection between how much your life is filled with strongly positive emotions and how much with strongly negative ones. Depression represents a state where those two independent axes tend toward collapsing into one inverse relationship—too few positive

emotions and too many negative ones. Naturally, the inverse correlation isn’t perfect, and a lot of current research focuses on questions like: Are different subtypes of depression characterized more by the absence of positive emotions or the overabundance of negatives?

George Tooker, Woman at the Wall, egg tempera on gesso, 1974.

Accompanying major depression are great grief and great guilt. We often feel grief and guilt in the everyday sadnesses that we refer to as “depression.” But in a major depression, they can be incapacitating, as the person is overwhelmed with the despair. There can be complex layers of these feelings: not just obsessive guilt, for example, about something that has contributed to the depression, but obsessive guilt about the depression itself—what it has done to the sufferer’s family, the guilt of not being able to overcome depression, a life lived but once and wasted amid this disease. Small wonder that, worldwide, depression accounts for 800,000 suicides per year.* In a subset of such patients, the sense of grief and guilt can take on the quality of a delusion. By this, I do not mean the thought-disordered delusions of schizophrenics; instead, delusional thinking in depressives is of the sort where facts are distorted, over- or underinterpreted to the point where one must conclude that things are terrible and getting worse, hopeless. An example: a middle-aged man, out of the blue, has a major heart attack. Overwhelmed by his implied mortality, the transformation of his life, he slips into a major depression. Despite this, he is recovering from the attack reasonably well, and there is every chance that he will resume a normal life. But each day he’s sure he’s getting worse. The hospital in which he is staying is circular in construction, with a corridor that forms a loop. One day, the nurses walk him once around the hospital before he collapses back in bed. The next day, he does two laps; he is getting stronger. That evening, when his family visits, he explains to them that he is sinking. “What are you talking about? The nurses said that you did two loops today; yesterday you only did one.” No, no, he shakes his head sadly, you don’t understand. He explains that the hospital is being renovated and, um, well, last night they closed off the old corridor and opened a newer, smaller one. And, you see, the distance around the new loop is less than half the distance of the old one, so two laps today is still less than I could do yesterday.

This particular incident occurred with the father of a friend, an engineer who lucidly described radii and circumferences, expecting his family to believe that the hospital had opened up a new corridor through the core of the building in one day. This is delusional thinking; the emotional energies behind the analysis and evaluation are disordered so that the everyday world is interpreted in a way that leads to depressive conclusions—it’s awful, getting worse, and this is what I deserve. Cognitive therapists, like Aaron Beck of the University of Pennsylvania, even consider depression to be primarily a disorder of thought, rather than emotion, in that sufferers tend to see the world in a distorted, negative way. Beck and colleagues have conducted striking studies that provide evidence for this. For example, they might show a subject two pictures. In the first, a group of people are gathered happily around a dinner table, feasting. In the second, the same people are gathered around a coffin. Show the two pictures rapidly or simultaneously; which one is remembered? Depressives see the funeral scene at rates higher than chance. They are not only depressed about something, but see the goings-on around them in a distorted way that always reinforces that feeling. Their glasses are always half empty. Another frequent feature of a major depression is called psychomotor retardation. The person moves and speaks slowly. Everything requires tremendous effort and concentration. She finds the act of merely arranging a doctor’s appointment exhausting. Soon it is too much even to get out of bed and get dressed. (It should be noted that not all depressives show psychomotor retardation; some may show the opposite pattern, termed psychomotor agitation.) The psychomotor retardation accounts for one of the important clinical features of depression, which is that severely, profoundly depressed people rarely attempt suicide. It’s not until they begin to feel a bit better. If the psychomotor aspects make it too much for this person to get out of bed, they sure aren’t going to find the often considerable energy needed to kill themselves. A key point: many of us tend to think of depressives as people who get the same everyday blahs as you and I, but that for them it just spirals out of control. We may also have the sense, whispered out of earshot, that these are people who just can’t handle normal ups and downs, who are indulging themselves. (Why can’t they just get themselves together?) A major depression, however, is as real a disease as diabetes. Another set of depressive symptoms supports that view. Basically, many things in the bodies of depressives work peculiarly; these are called vegetative symptoms. You and I get an everyday depression. What do we do? Typically, we sleep more than usual, probably eat more than usual, convinced in some way that such comforts will make us feel better. These traits are just the opposite of the vegetative symptoms seen in most people with major depressions. Eating declines. Sleeping does as well, and in a distinctive manner. While depressives don’t necessarily have trouble falling asleep, they have the problem of “early morning wakening,” spending months on end sleepless and exhausted from three-thirty or so each morning. Not only is sleep shortened but, as mentioned in chapter 11, the “architecture” of sleep is different as well—the normal pattern of shifting between deep and shallow sleep, the rhythm of the onset of dream states, are disturbed. An additional vegetative symptom is extremely relevant to this chapter, namely that major depressives often experience elevated levels of glucocorticoids. This is critical for a number of reasons that will be returned to, and helps to clarify what the disease is actually about. When

looking at a depressive sitting on the edge of the bed, barely able to move, it is easy to think of the person as energy-less, enervated. A more accurate picture is of the depressive as a tightly coiled spool of wire, tense, straining, active—but all inside. As we will see, a psychodynamic view of depression shows the person fighting an enormous, aggressive mental battle—no wonder they have elevated levels of stress hormones. Chapter 10 reviewed how glucocorticoids can impair aspects of memory that depend on the hippocampus, and the frequently elevated glucocorticoid levels in depression may help explain another feature of the disease, which is problems with hippocampal-dependent memory. The memory problems may reflect, in part, a lack of motivation on the part of the depressed person (why work hard on some shrink’s memory test when everything, everything, is hopeless and pointless?), or an anhedonic inability to respond to the rewards of remembering something in a task. Nonetheless, amid those additional factors, the pure process of storing and retrieving memories via the hippocampus is often impaired. As we’ll see shortly, this fits extraordinarily well with recent findings showing that the hippocampus is smaller than average in many depressives. Another feature of depression also confirms that it is a real disease, rather than merely the situation of someone who simply cannot handle everyday ups and downs. There are multiple types of depressions, and they can look quite different. In one variant, unipolar depression, the sufferer fluctuates from feeling extremely depressed to feeling reasonably normal. In another form, the person fluctuates between deep depression and wild, disorganized hyperactivity. This is called bipolar depression or, more familiarly, manic depression. Here we run into another complication because, just as we use depression in an everyday sense that is different from the medical sense, mania has an everyday connotation as well. We may use the term to refer to madness, as in made-for-television homicidal maniacs. Or we could describe someone as being in a manic state when he is buoyed by some unexpected good news—talking quickly, laughing, gesticulating. But the mania found in manic depression is of a completely different magnitude. Let me give an example of the disorder: a woman comes into the emergency room; she’s bipolar, completely manic, hasn’t been taking her medication. She’s on welfare, doesn’t have a cent to her name, and in the last week she’s bought three Cadillacs with money from loan sharks. And, get this, she doesn’t even know how to drive. People in manic states will go for days on three hours of sleep a night and feel rested, will talk nonstop for hours at a time, will be vastly distractible, unable to concentrate amid their racing thoughts. In outbursts of irrational grandiosity, they will behave in ways that are foolhardy or dangerous to themselves and others— at the extreme, poisoning themselves in attempting to prove their immortality, burning down their homes, giving away their life savings to strangers. It is a profoundly destructive disease. The strikingly different subtypes of depression and their variability suggest not just a single disease, but a heterogeneity of diseases that have different underlying biologies. Another feature of the disorder also indicates a biological abnormality. Suppose a patient comes to a doctor in the tropics. The patient is running a high fever that abates, only to come back a day or two later, abate again, return again, and so on every 48 to 72 hours. The doctor will recognize this instantly as malaria, because of the rhythmicity of the disorder. It has to do with the life cycle of the malarial parasite as it moves from red blood cells to the liver and spleen. The rhythmicity screams biology. In the same way, certain subtypes of depression have a rhythm. A manic-

depressive may be manic for five days, severely depressed for the following week, then mildly depressed for half a week or so, and, finally, symptom-free for a few weeks. Then the pattern starts up again, and may have been doing so for a decade. Good things and bad things happen, but the same cyclic rhythm continues, which suggests just as much deterministic biology as in the life cycle of the malarial parasite. In another subset of depression the rhythm is annual, where sufferers get depressed during the winter. These are called seasonal affective disorders (SADs; “affective” is the psychiatric term for emotional responses), and are thought to be related to patterns of exposure to light; recent work has uncovered a class of retinal cells that respond to light intensity and, surprisingly, send their information directly into the limbic system, the emotional part of the brain. Again, the rhythmicity appears independent of external life events; a biological clock is ticking away in there that has something to do with mood, and something is seriously wrong with its ticking.

The Biology of Depression Neurochemistry and Depression Considerable evidence exists that something is awry with the chemistry of the brains of depressives. In order to appreciate that, it is necessary to learn a bit about how brain cells communicate with one another. The illustration in chapter 14 shows a schematic version of two neurons, the principal type of brain cell. If a neuron has become excited with some thought or memory (metaphorically speaking), its excitement is electrical—a wave of electricity sweeps from the dendrites over the cell body, down the axon to the axon terminals. When the wave of electrical excitation reaches the axon terminal, it releases chemical messengers that float across the synapse. These messengers—neurotransmitters—bind to specialized receptors on the adjacent dendrite, causing the second neuron to become electrically excited. A minor piece of housekeeping, however: What happens to the neurotransmitter molecule after it has done its job and floats off the receptor? In some cases, it is recycled—taken back up by the axon terminal of the first neuron and repackaged for future use. Or it can be degraded in the synapse and the debris flushed out to sea (the cerebrospinal fluid, then to the blood, and then the urine). If these processes of clearing neurotransmitters out of the way fail (reuptake ceases or degradation stops or both), suddenly a lot more neurotransmitter remains in the synapse, giving a stronger signal to the second neuron than usual. Thus, the proper disposal of these powerful messengers is integral to normal neuronal communication. There are trillions of synapses in the brain. Do we need trillions of chemically unique neurotransmitters? Certainly not. You can generate a seemingly infinite number of messages with a finite number of messengers; consider how many words we can form with the mere twenty-six letters in our alphabet. All you need are rules that allow for the same messenger to convey different meanings, metaphorically speaking, in different contexts. At one synapse, neurotransmitter A sends a message relevant to pancreatic regulation, while at another synapse the same neurotransmitter substance may pertain to adolescent crushes. There are many neurotransmitters, probably on the order of a few hundred, but certainly not trillions.

So that’s a primer on how neurons talk to each other with neurotransmitters. The best evidence suggests that depression involves abnormal levels of the neurotransmitters norepinephrine, serotonin, and dopamine. Before reviewing the evidence, it’s important to clear up a point. You are no doubt thinking, “Wasn’t there something about norepinephrine and the sympathetic nervous system many chapters ago?” Absolutely, and that proves the point about the varied roles played by any given neurotransmitter. In one part of the body (the heart, for example), norepinephrine is a messenger concerning arousal and the Four F’s, while in a different part of the nervous system, norepinephrine seems to have something to do with the symptoms of depression.

A neuron that has been excited conveys information to other neurons by means of chemical signals at synapses, the contact points between neurons. When the impulse reaches the axon

terminal of the signaling neuron, it induces the release of neurotransmitter molecules. Transmitters diffuse across a narrow cleft and bind to receptors in the adjacent neuron’s

dendritic spine. Why is it likely that there is something wrong with norepinephrine, serotonin, or dopamine in depression? The best evidence is that most of the drugs that lessen depression increase the amount of signaling by these neurotransmitters. One class of antidepressants, called tricyclics (a reference to their biochemical structure), stops the recycling, or reuptake, of these neurotransmitters into the axon terminals. The result is that the neurotransmitter remains in the synapse longer and is likely to hit its respective receptor a second or third time. Another class of drugs, called MAO inhibitors, blocks the degradation of these neurotransmitters in the synapse by inhibiting the action of a crucial enzyme in that degradation, monoamine oxidase, or MAO. The result, again, is that more of the messenger remains in the synapse to stimulate the dendrite of the receiving neuron. These findings generate a pretty straightforward conclusion: if you use a drug that increases the amount of norepinephrine, serotonin, and dopamine in synapses throughout the brain, and as a result, someone’s depression gets better, there must have been too little of those neurotransmitters in the first place. Case closed. Naturally, not so fast. As a first issue of confusion, is the problem with serotonin, dopamine, or norepinephrine? The tricyclics and MAO inhibitors work on all three neurotransmitter systems, making it impossible to tell which one is critical to the disease. People used to think

norepinephrine was the culprit, when it was thought that those classical antidepressant drugs worked only on the norepinephrine synapse. These days, most of the excitement centers on serotonin, mainly because of the efficacy of reuptake inhibitors that work only on serotonin synapses (selective serotonin reuptake inhibitors, or SSRIs, of which Prozac is the most famous). However, there still remains some reason to think that the other two neurotransmitters remain part of the story, since some of the newest antidepressants appear to work on them more than on serotonin.* A second piece of confusion is actually quite major. Is the defect in depression with these neurotransmitters really one of too little neurotransmitter in the synapse? You would think this was settled—the effective antidepressant drugs increase the amounts of these neurotransmitters in the synapse and alleviate depression; thus, the problem had to be too little of the stuff to begin with. However, some clinical data suggest that this might not be so simple. The stumbling block has to do with timing. Expose the brain to some tricyclic antidepressant, and the amount of signaling with these neurotransmitters in the synapses changes within hours. However, give that same drug to a depressed person, and it takes weeks for the person to feel better. Something doesn’t quite fit. Two theories have arisen in recent years that might reconcile this problem with timing, and they are both extremely complicated. Revisionist theory 1, the “it’s not too little neurotransmitter, it’s actually too much” hypothesis. First, some orientation. If somebody constantly yells at you, you stop listening. Analogously, if you inundate a cell with lots of a neurotransmitter, the cell will not “listen” as carefully—it will “down-regulate” (decrease) the number of receptors for that neurotransmitter, in order to decrease its sensitivity to that messenger. If, for example, you double the amount of serotonin reaching the dendrites of a cell and that cell down-regulates its serotonin receptors by 50 percent, the changes roughly cancel out. If the cell down-regulates less than 50 percent, the net result is more serotonin signaling in the synapse; if more than 50 percent, the result is actually less signaling in the synapse. In other words, how strong the signal is in a synapse is a function both of how loudly the first neuron yells (the amount of neurotransmitter released) and of how sensitively the second neuron listens (how many receptors it has for the neurotransmitter). Okay, ready. This revisionist theory states that the original problem is that there is actually too much norepinephrine, serotonin, and/or dopamine in parts of the brains of depressives. What happens when you prescribe antidepressants that increase signaling of these neurotransmitters even further? At first, that should make the depressive symptoms worse. (Some psychiatrists argue that this actually does occur.) Over the course of a few weeks, however, the dendrites say, “This is intolerable, all this neurotransmitter; let’s down-regulate our receptors a whole lot.” If this occurs and, critical to the theory, more than compensates for the increased neurotransmitter signal, the depressive problem of excessive neurotransmitter signaling goes away: the person feels better. Revisionist theory 2, “It really is too little norepinephrine, serotonin, and/or dopamine after all.” This theory is even more complicated than the first, and also requires orientation. Not only do dendrites contain receptors for neurotransmitters, but it turns out that on the axon terminals of the “sending” neuron as well there are receptors for the very neurotransmitters being released by that

neuron. What possible purpose could these so-called autoreceptors serve? Neurotransmitters are released, float into the synapse, bind to the standard receptors on the second neuron. Some neurotransmitter molecules, however, will float back and wind up binding to the autoreceptors. They serve as some sort of feedback signal; if, say, 5 percent of the released neurotransmitter reaches the autoreceptors, the first neuron can count its toes, multiply by 20, and figure out how much neurotransmitter it has released. Then it can make some decisions—should I release more neurotransmitter or stop now? Should I start synthesizing more? and so on. If this process lets the first neuron do its bookkeeping on neurotransmitter expenditures, what happens if the neuron down-regulates a lot of these autoreceptors? Underestimating the amount of neurotransmitter it has released, the neuron will inadvertently start increasing the amount it synthesizes and discharges. With this as background, here’s the reasoning behind the second theory (that there really is too little norepinephrine, serotonin, or dopamine in a part of the brain of depressives). Give the antidepressant drugs that increase signaling of these neurotransmitters. Because of the increased signaling, over the course of weeks there will be down-regulation of norepinephrine, serotonin, and dopamine receptors. Critical to this theory is the idea that the autoreceptors on the first neuron will down-regulate to a greater extent than the receptors on the second neuron. If that happens, the second neuron may not be listening as well, but the first one will be releasing sufficient extra neurotransmitter to more than overcome that. The net result is enhanced neurotransmitter signaling, and depressive symptoms abate. (This mechanism may explain the efficacy of electroconvulsive therapy, ECT, or “shock therapy.” For decades psychiatrists have used this technique to alleviate major depressions, and no one has quite known why it works. It turns out that among its many effects ECT decreases the number of norepinephrine autoreceptors, at least in experimental animal models.) If you are confused by now, you are in some good company, as the entire field is extremely unsettled. Norepinephrine, serotonin, or dopamine? Too much or too little signaling? If it is, for example, too little serotonin signaling, is it because too little serotonin is being released into synapses, or because there is some defect blunting the sensitivity of serotonin receptors? (To give you a sense of how big a can of worms that one is, there are currently recognized more than a dozen different types of serotonin receptors, with differing functions, efficacies, and distributions in the brain.) Maybe there are a variety of different neurochemical routes for getting to a depression, and different pathways are associated with different subtypes of depression (unipolar versus manic depression, or one that is triggered by outside events versus one that runs with its own internal clockwork, or one dominated by psychomotor retardation versus one dominated by suicidalism). This is a very reasonable idea, but the evidence for it is still scant. Amid all those questions, another good one—why does having too much or too little of these neurotransmitters cause a depression? There are a lot of links between these neurotransmitters and function. For example, serotonin is thought to have something to do with incessant ideation in depression, the uncontrollable wallowing in those dark thoughts. Connected with this, SSRIs are often effective on people with obsessive-compulsive disorder. There is a commonality here: in the depressive case, it is the obsessive sense of failure, of doom, of despair, while in the latter case, it can be obsessive worries that you left the gas on at home when you left, that your hands

are dirty and need to be washed, and so on. Trapped in a mind that just circles and circles around the same thoughts or feelings. Norepinephrine is thought to play a different role in the symptoms of depression. The major pathway that utilizes norepinephrine is an array of projections from a brain region called the locus ceruleus. That projection extends diffusely throughout the brain and seems to play a role in alerting other brain regions—increasing their baseline level of activation, lowering their threshold for responding to outside signals. Thus, a shortage of norepinephrine in this pathway might begin to explain the psychomotor retardation. Dopamine, meanwhile, has something to do with pleasure, a connection that will be reviewed at length in chapter 16. Several decades ago, some neuroscientists made a fundamental discovery. They had implanted electrodes into the brains of rats and stimulated areas here and there, seeing what would happen. By doing so, they found an extraordinary area of the brain. Whenever this area was stimulated, the rat became unbelievably happy. So how can one tell when a rat is unbelievably happy? You ask the rat to tell you, by charting how-many times it is willing to press a lever in order to be rewarded with stimulation in that part of the brain. It turns out that rats will work themselves to death on that lever to get stimulation. They would rather be stimulated there than get food when they are starving, or have sex, or receive drugs even when they’re addicted and going through withdrawal. The region of the brain targeted in these studies was promptly called the “pleasure pathway” and has been famous since. That humans have a pleasure pathway was discovered shortly afterward by stimulating a similar part of the human brain during neurosurgery.* The results are pretty amazing. Something along the lines of “Aaaaah, boy, that feels good. It’s kind of like getting your back rubbed but also sort of like sex or playing in the backyard in the leaves when you’re a kid and Mom calling you in for hot chocolate and then you get into your pajamas with the feet….” Where can we sign up? This pleasure pathway seems to make heavy use of dopamine as a neurotransmitter (and in chapter 16, we’ll see how dopamine signals the anticipation of reward more than it signals reward itself). The strongest evidence for this is the ability of drugs that mimic dopamine, such as cocaine, to act as euphoriants. Suddenly, it seems plausible to hypothesize that depression, which is characterized above all by dysphoria, might involve too little dopamine and, thus, dysfunction of those pleasure pathways. Thus, these are the big three when it comes to the neurotransmitters implicated in depression, with attention these days probably being the most for serotonin and the least for dopamine. All of the leading antidepressant drugs—the SSRIs, and older classes such as tricyclics or MAO inhibitors—work by altering the levels of one or more of these three neurotransmitters. At this point, there is nothing close to resembling a science as to which sort of person will respond best to which type of antidepressants. Naturally, there’s a spate of other neurotransmitters that may be involved. One particularly interesting one is called Substance P. Decades of work have shown that Substance P plays a role in pain perception, with a major role in activating the spinal cord pathways discussed in chapter 9. Remarkably, some recent studies indicate that drugs that block the action of Substance P can

work as antidepressants in some individuals. What’s this about? Perhaps the sense of depression as a disease of “psychic pain” may be more than just a metaphor. Neuroanatomy and Depression I introduce an illustration here of what the brain looks like, to consider a second way in which brain function might be abnormal in depressives, in addition to the neurochemistry discussed. One region regulates processes like your breathing and heart rate. It includes the hypothalamus, which is busy releasing hormones and instructing the autonomic nervous system. If your blood pressure drops drastically, causing a compensatory stress-response, it is the hypothalamus, midbrain, and hindbrain that kick into gear. All sorts of vertebrates have roughly the same connections here.

Layered on top of that is a region called the limbic system, the functioning of which is related to emotion. As mammals, we have large limbic systems; lizards have relatively tiny limbic systems—they are not noted for the complexity of their emotional lives. If you get a stress- response from smelling the odor of a threatening rival, it’s your limbic system that is involved. Above that is the cortex. Everyone in the animal kingdom has some, but it is a real primate specialty. The cortex does abstract cognition, invents philosophy, remembers where your car keys are. The stuff of the previous chapter. Now think for a second. Suppose you are gored by an elephant. You may feel a certain absence of pleasure afterward, maybe a sense of grief. Throw in a little psychomotor retardation—you’re not as eager for your calisthenics as usual. Sleeping and feeding may be disrupted, glucocorticoid levels may be a bit on the high side. Sex may lose its appeal for a while. Hobbies are not as enticing; you don’t jump up to go out with friends; you pass up that all-you-can-eat buffet. Sound like some of the symptoms of a depression?

Now, what happens during a depression? You think a thought about your mortality or that of a loved one; you imagine children in refugee camps, the rain forests disappearing and endless species of life evaporating, late Beethoven string quartets, and suddenly you experience some of the same symptoms as after being gored by the elephant. On an incredibly simplistic level, you can think of depression as occurring when your cortex thinks an abstract negative thought and manages to convince the rest of the brain that this is as real as a physical stressor. In this view, people with chronic depressions are those whose cortex habitually whispers sad thoughts to the rest of the brain. Thus, an astonishingly crude prediction: cut the connections between the cortex and the rest of a depressive’s brain, and the cortex will no longer be able to get the rest of the brain depressed. Remarkably, it actually works sometimes. Neurosurgeons may perform this procedure on people with vastly crippling depressions that are resistant to drugs, ECT, or other forms of therapy. Afterward, depressive symptoms seem to abate.* Obviously, this is a simplified picture—no one actually disconnects the entire cortex from the rest of the brain. After all, the cortex does more than mope around feeling bad about the final chapter of Of Mice and Men. The surgical procedure, called a cingulotomy, or a cingulum bundle cut, actually disconnects just one area toward the front of the cortex, called the anterior cingulate cortex (ACC). The ACC is turning out to have all the characteristics of a brain region you’d want to take offline in a major depression. It’s a part of the brain that is very concerned with emotions. Show people arrays of pictures: in one case, ask them to pay attention to the emotions being expressed by people in the pictures; in another case, ask them to pay attention to details like whether these are indoor or outdoor photographs. In only the former case do you get activation of the ACC. And the emotions that the ACC is involved in seem to be negative ones. Induce a positive state in someone by showing something amusing, and ACC metabolism decreases. In contrast, if you electrically stimulate the ACC in people, they feel a shapeless sense of fear and foreboding. Moreover, neurons in the ACC, including in humans, respond to pain of all sorts. But the ACC response isn’t really about the pain; it more concerns feelings about the pain. As was discussed in chapter 9, give someone a hypnotic suggestion that they will not feel the pain of dipping their hand into ice water. The primary parts of the brain that get pain projections from the spinal cord get just as active as if there were no hypnotic suggestion. But this time, the ACC doesn’t activate. In addition, the ACC and adjacent brain regions activate when you show widows pictures of their lost loved ones (versus pictures of strangers). As another example of this, put a volunteer in a brain-imaging machine and, from inside, ask them to play some game with two other people, via a computer console. Rig up the flow of the game so that, over time, the other two (actually, a computer program) gradually begin just playing with each other, excluding the test subject. Neuronal activity in the ACC lights up, and the more left out the person feels, the more intensely the ACC activates. How do you know this has something to do with that dread junior high school feeling of being picked last for the team? Because of a clever control in the study: set the person up to play with the supposed other two players. Once again, it winds up that the other two only play against each other. The difference, this time, though, is that early on the subject is told

there’s been a technical glitch and that their computer console isn’t working. Excluded because of a snafu in the technology, there’s no ACC activation. Given these functions of the ACC, it is not surprising that its resting level of activity tends to be elevated in people with a depression—this is the fear and pain and foreboding churning away at those neurons. Interestingly, another part of the brain, called the amygdala, seems to be hyperactive in depressives as well. We will hear lots about the role of the amygdala in fear and anxiety in the next chapter. However, in depressives, the amygdala seems to have been recruited into a different role. Show a depressed person a fearful human face and his amygdala doesn’t activate all that much (in contrast to the response you’d see in the amygdala of a control subject). But show him a sad face and the amygdala gets a highly exaggerated activation. Sitting just in front of the ACC is the frontal cortex which, as we saw in chapter 11, is one of the most distinctly human parts of the brain. Work by Richard Davidson of the University of Wisconsin has shown that one subregion called the prefrontal cortex (PFC) seems highly responsive to mood, and in a lateralized way. Specifically, activation of the left PFC is associated with positive moods, and activation of the right PFC, with negative. For example, induce a positive state in someone (by asking him to describe the happiest day of his life), and the left PFC lights up, in proportion to the person’s subjective assessment of his pleasure. Ask him to remember a sad event, and the right PFC dominates. Similarly, separate an infant monkey from its mother and right PFC metabolism rises while left PFC decreases. Thus, not surprisingly, in depressives, there is decreased left PFC activity and elevated activity in the right PFC. There are a few other anatomical changes in the brain in depression, but to make sense of those, we have to consider what hormones have to do with the disease.

Genetics and Depression It is hard to look at the biology of anything these days without genes coming into the picture, and depression is no exception. Depression has a genetic component. As a first observation, depression runs in families. For a long time, that would have been sufficient evidence for some folks that there is a genetic link, but this conclusion is undone by the obvious fact that not only do genes run in families, environment does as well. Growing up in a poor family, an abusive family, a persecuted family, can all increase the risk of depression running through that family without genes having anything to do with it. So we look for a tighter relationship. The more closely related two individuals are, the more genes they share in common and, as it turns out, the more likely they are to share a depressive trait. As one of the most telling examples of this, take any two siblings (who are not identical twins). They share something like 50 percent of their genes. If one of them has a history of depression, the other has about a 25 percent likelihood, considerably higher than would be expected by chance. Now, compare two identical twins, who share all of their genes in common. And if one of them is depressive, the other has a 50 percent chance. This is quite impressive—the more genes in common, the more likelihood of sharing the disease. But there remains a confound: the more genes people share within a family, the more environment they share as well

(starting with the fact that identical twins grow up treated more similarly than are non-identical twins). Tighten the relationship further. Look at children who were adopted at an early age. Consider those whose biological mother had a history of depression, but whose adoptive mother did not. They have an increased risk of depression, suggesting a genetic legacy shared with their biological mother. But the confound there, as we saw in chapter 6, is that “environment” does not begin at birth, but begins much earlier, with the circulatory environment shared in utero with one’s biological mother. For any card-carrying molecular biologist in the twenty-first century, if you want to prove that genes have something to do with depression, you’re going to have to identify the specific genes, the specific stretches of DNA that code for specific proteins that increase the risk for depression. As we’ll see shortly, precisely that has occurred in recent years. Immunology and Depression This subsection did not exist in previous editions of this book. Immunity is about fighting off pathogens, depression is about feeling sad—unrelated subjects. Well, they can be related, but in an idiotically obvious way—like, duh, being sick can be depressing. But it’s more complicated than that. Chronic illness that involves overactivation of the immune system (for example, chronic infections, or an autoimmune disease where the immune system has accidentally activated and is attacking some part of your body) is more likely to cause depression than other equally severe and prolonged illnesses that don’t involve the immune system. Some more threads of interconnection involve the cytokines that act as messengers between immune cells. As you’ll recall from chapter 8, cytokines can also get into the brain, where they can stimulate CRH release. More recently, it’s becoming clear that they also interact with norepinephrine, dopamine, and serotonin systems. Critically, cytokines can cause depression. This is shown in animal models of depression. Furthermore, certain types of cancers are sometimes treated with cytokines (to enhance immune function), and this typically results in depression. So this represents a new branch of study for biological psychiatry—the interactions between immune function and mood. Endocrinology and Depression Abnormal levels of a number of different hormones often go hand in hand with depression. To begin, people who secrete too little thyroid hormone can develop major depressions and, when depressed, can be atypically resistant to antidepressant drugs working. This is particularly important because many people, seemingly with depressions of a purely psychiatric nature, turn out to have thyroid disease. There is another aspect of depression in which hormones may play a role. The incidence of major, unipolar depression differs greatly, with women suffering far more than men. Even when you consider manic depression, where there is no sex difference in its incidence, bipolar women have more depressive episodes than do bipolar men.

Why this female bias? It has nothing to do with the obvious first guess, which is that women are more likely to see a health professional for depression than are men. The difference holds up even when such reporting biases are controlled for. One theory, from the school of cognitive therapy, concentrates on the ways in which women and men tend to think differently. When something upsetting happens, women are more likely to ruminate over it—think about it or want to talk about it with someone else. And men, terrible communicators that they so often are, are more likely to want to think about anything but the problem, or even better, go and do something—exercise, use power tools, get drunk, start a war. A ruminative tendency, the cognitive psychologists argue, makes you more likely to become depressed. Another theory about the sex difference is psychosocial in nature. As we will see, much theorizing about the psychology of depression suggests that it is a disorder of lack of power and control, and some scientists have speculated that because women in so many societies traditionally have less control over the circumstances of their lives than do men, they are at greater risk for depression. In support of this idea, some psychiatrists have produced data suggesting that the elevated rates of depression in women decline to the levels seen in men in some traditional societies in which women don’t have a subordinate role. Yet another theory suggests that men really do have as high a rate of depression as do women, but they are simply more likely to mask it with substance abuse. All of these ideas are reasonable, although they run into trouble when one considers that women and men, as noted, have the same rate of bipolar depression; it is only unipolar depression that is more common among women. These theories seem particularly weak in their failure to explain a major feature of female depressions, namely, that women are particularly at risk for depressions at certain reproductive points: menstruation, menopause, and most of all, the weeks immediately after giving birth. A number of researchers believe such increased risks are tied to the great fluctuations that occur during menstruation, menopause, and parturition in two main hormones: estrogen and progesterone. As evidence, they cite the fact that women can get depressed when they artificially change their estrogen or progesterone levels (for example, when taking birth- control pills). Critically, both of these hormones can regulate neurochemical events in the brain, including the metabolism of neurotransmitters such as norepinephrine and serotonin. With massive changes in hormone levels (a thousandfold for progesterone at the time of giving birth, for example), current speculation centers on the possibility that the ratio of estrogen to progesterone can change radically enough to trigger a major depression. This is a new area of research with some seemingly contradictory findings, but there is more and more confidence among scientists that there is a hormonal contribution to the preponderance of female depressions. Obviously, the next subject in a section on hormones and depression will have to look at glucocorticoids. But given how central this is to the whole venture of this book, the subject requires expansion.

How Does Stress Interact with the Biology of Depression?

Stress, Glucocorticoids, and the Onset of Depression

The first stress-depression link is an obvious one, in that stress and depression tend to go together. This can run in two directions. First, studies of what is called “stress generation” among depressives look at the fact that people who are prone to depression tend to experience stressors at a higher than expected rate. This is even seen when comparing them to individuals with other psychiatric disorders or health problems. Much of this appears to be stressors built around lack of social support. This raises the potential for a vicious cycle to emerge. This is because if you interpret the ambiguous social interactions around you as signs of rejection, and respond as if you have been rejected, it can increase the chances of winding up socially isolated, thereby confirming your sense that you have been rejected…. But the major way in which people think about a link between stress and depression, and the one that concerns us here, has causality running in the other direction. Specifically, people who are undergoing a lot of life stressors are more likely than average to succumb to a major depression, and people sunk in their first major depression are more likely than average to have undergone recent and significant stress. Obviously, not everyone who undergoes major stressors sinks into depression, and what those individual differences are about should be clearer as we proceed through this chapter. As noted, some people have the grave misfortune of suffering from repeated depressive episodes, ones that can take on a rhythmic pattern stretching over years. When considering the case histories of those people, stressors emerge as triggers for only the first few depressions. In other words, have two, three major bouts of depression and, statistically, you are no more at risk for subsequent major depression than anyone else. But somewhere around the fourth depression or so, a mad clockwork takes over, and the depressive waves crash, regardless of whether the outside world pummels you with stressors. What that transition is about will be considered below. Laboratory studies also link stress and the symptoms of depression. Stress a lab rat, and it becomes anhedonic. Specifically, it takes a stronger electrical current than normal in the rat’s pleasure pathways to activate a sense of pleasure. The threshold for perceiving pleasure has been raised, just as in a depressive. Critically, glucocorticoids can do the same. A key point in chapter 10 was how glucocorticoids and stress could disrupt memory. Part of the evidence for that came from people with Cushing’s syndrome (as a reminder, that is a condition in which any of a number of different types of tumors wind up causing vast excesses of glucocorticoids in the bloodstream), as well as from people prescribed high doses of glucocorticoids to treat a number of ailments. It has also been known for decades that a significant subset of Cushingoid patients and patients prescribed synthetic glucocorticoids become clinically depressed, independent of memory problems. This has been a bit tricky to demonstrate. First, when someone is initially treated with synthetic glucocorticoids, the tendency is to get, if anything, euphoric and even manic, perhaps for a week or so before the depression kicks in. You can immediately guess that we are dealing with one of our dichotomies between short- and long-term stress physiology; chapter 16 will explore in even more detail where that transient euphoria comes from. As a second complication, does someone with Cushing’s syndrome or someone taking high pharmacological doses of synthetic glucocorticoids get depressed because glucocorticoids cause that state, or is it because they

recognize they have a depressing disease? You show it is the glucocorticoids that are the culprits by demonstrating higher depression rates in this population than among people with, for example, the same disease and the same severity but not receiving glucocorticoids. At this stage, there’s also not much of a predictive science to this phenomenon. For example, no clinician can reliably predict beforehand which patient is going to get depressed when put on high-dose glucocorticoids, let alone at what dose, and whether it is when the dose is raised or lowered to that level. Nonetheless, have lots of glucocorticoids in the bloodstream and the risk of a depression increases. Stress and glucocorticoids tangle up with biology in predisposing a person toward depression in an additional, critical way. Back to that business about there being a genetic component to depression. Does this mean that if you have “the gene” (or genes) “for” depression, that’s it, you’re up the creek, it’s inevitable? Obviously not, and the best evidence for this is that factoid about identical twins. One has depression and the other, sharing all the same genes, has about a 50 percent chance of having the disease as well, a much higher rate than in the general population. There, pretty solid evidence for genes being involved. But flip this the other way. Share every single gene with someone who is depressive and you still have a 50 percent chance of not having the disease. Genes are rarely about inevitability, especially when it comes to humans, the brain, or behavior. They’re about vulnerability, propensities, tendencies. In this case, genes increase the risk of depression only in certain environments: you guessed it, only in stressful environments. This is shown in a number of ways, but most dramatically in a recent study by Avshalom Caspi at King’s College, London. Scientists identified a certain gene in humans that increases the risk of depression. More specifically, it is a gene that comes in a few different “allelic versions”—a few different types or flavors that differ slightly in function; have one of those versions, and you’re at increased risk. What that gene is I’m not telling yet; I’m saving it for the end of this chapter, as it is a doozy But the key thing is that having version X of this gene Z doesn’t guarantee you get depression, it just increases your risk. And, in fact, knowing nothing more about someone than which version of gene Z she has doesn’t increase your odds of predicting whether she gets depressed. Version X increases depression risk only when coupled with a history of repeated major stressors. Amazingly, the same has been shown with studies of some nonhuman primate species, who carry a close equivalent of that gene Z. It’s not the gene that causes it. It’s that the gene interacts with a certain environment. More specifically, a gene that makes you vulnerable in a stressful environment. Glucocorticoid profiles once a depression has been established Not surprisingly, glucocorticoid levels are typically abnormal in people who are clinically depressed. A relatively infrequent subtype of depression, called “atypical depression,” is dominated by the psychomotor features of the disease—an incapacitating physical and psychological exhaustion. Just as is the case with chronic fatigue syndrome, atypical depression is characterized by lower than normal glucocorticoid levels. However, the far more common feature of depression is one of an overactive stress-response—somewhat of an overly activated sympathetic nervous system and, even more dramatically, elevated levels of glucocorticoids. This adds to the picture that depressed people, sitting on the edge of their beds without the

energy to get up, are actually vigilant and aroused, with a hormonal profile to match—but the battle is inside them. Research stretching back some forty years has explored why, on a nuts-and-bolts level, glucocorticoid levels are often elevated in depression. The elevated levels appear to be due to too much of a stress signal from the brain (back to chapter 2—remember that the adrenals typically secrete glucocorticoids only when they are commanded to by the brain, via the pituitary), rather than the adrenals just getting some depressive glucocorticoid hiccup all on their own now and then. Moreover, the excessive secretion of glucocorticoids is due to what is called feedback resistance—in other words, the brain is less effective than it should be at shutting down glucocorticoid secretion. Normally, the levels of this hormone are tightly regulated—the brain senses circulating glucocorticoid levels, and if they get higher than desired (the “desired” level shifts depending on whether events are calm or stressful), the brain stops secreting CRH. Just like the regulation of water in a toilet bowl tank. In depressives, this feedback regulation fails— concentrations of circulating glucocorticoids that should shut down the system fail to do so, as the brain does not sense the feedback signal.* What are the consequences of elevated glucocorticoid levels before and during a depression? The first most critical question to ask is, how does an excess of glucocorticoids increase the risk of depression? A preceding section detailed, at great length, the considerable confusion about whether depression is about serotonin or norepinephrine or dopamine. To the extent that this is the case, the glucocorticoid angle fits well, in that the hormones can alter features of all three neurotransmitter systems—the amount of neurotransmitter synthesized, how fast it is broken down, how many receptors there are for each neurotransmitter, how well the receptors work, and so on. Moreover, stress has been shown to cause many of the same changes as well. Sustained stress will deplete dopamine from those “pleasure” pathways, and norepinephrine from that alerting locus ceruleus part of the brain. Moreover, stress alters all sorts of aspects of the synthesis, release, efficacy, and breakdown of serotonin. It is not clear which of those stress effects are most important, simply because it is not clear which neurotransmitter or neurotransmitters are most important. However, it is probably safe to say that whatever neurochemical abnormalities wind up being shown definitively to underlie depression, there is precedent for stress and glucocorticoids causing those same abnormalities. Those elevated glucocorticoid levels appear to have some other consequences as well. They may play a role, for example, in the fact that depressive patients often are at least mildly immunosuppressed, and are more prone to osteoporosis. Moreover, prolonged major depression increases the risk of heart disease about three- to fourfold, even after controlling for smoking and alcohol consumption, and the glucocorticoid excess is likely to contribute to that as well. And there may be more consequences. Think back to chapter 10 and its discussion of the many ways in which glucocorticoids can damage the hippocampus. As that literature emerged in the 1980s, it immediately suggested that there may be problems with the hippocampus in people with major depression. This speculation was reinforced by the fact that the type of memory most often impaired in depression—declarative memory—is mediated by the hippocampus. As was

discussed in chapter 10, there is atrophy of the hippocampus in long-term depression. The atrophy emerges as a result of the depression (rather than precedes it), and the longer the depressive history, the more atrophy and the more memory problems. While no one has explicitly shown yet that the atrophy occurs only in those depressives with the elevated glucocorticoid levels, the atrophy is most common in the subtypes of depression in which the glucocorticoid excess is most common. Chronic depression has also been associated in some studies with decreased volume in the frontal cortex. This was initially puzzling for those of us who view the world through glucocorticoid-tinted glasses, but has recently been resolved. In the rat, the hippocampus is overwhelmingly the target in the brain for glucocorticoid action, as measured by the density of receptors for the hormone; however, in the primate brain, the hippocampus and frontal cortex seem to be equally and markedly sensitive to glucocorticoids. So some pretty decent circumstantial evidence suggests that the glucocorticoid excess of depression may have something to do with the decreased volume of the hippocampus and frontal cortex. Chapter 10 noted an array of bad things that glucocorticoids could do to neurons. Some obsessively careful studies have shown loss of cells in the frontal cortex accompanying the volume loss in depression—as one point of confusion, it is those supportive glial cells rather than neurons that are lost. But in the hippocampus, no one has a clue yet; it could be the killing or atrophying of neurons, the inhibition of the birth of new neurons, or all the above.* Whatever the explanation is at the cellular level, it appears to be permanent; years to decades after these major depressions have been gotten under control (typically with medication), the volume loss is still there. Anti-glucocorticoids as antidepressants The glucocorticoid-depression link has some important implications. When I first introduced that link at the beginning of the chapter, it was meant to give some insight into the flavor of what a depression is like—a person looks like an enervated sea sponge, sitting there motionless on the edge of his bed, but he’s actually boiling, in the middle of an internal battle. Tacit in that description was the idea that undergoing a depression is actually immensely stressful, and, therefore, among other things, stimulates glucocorticoid secretion. The data just reviewed suggest the opposite scenario—stress and glucocorticoid excess can be a cause of depression, rather than merely a consequence. If that is really the case, then a novel clinical intervention should work: take one of those depressives with high glucocorticoid levels, find some drug that works on the adrenals to lower glucocorticoid secretion, and the depression should lessen. And, very exciting, that has been shown. The approach, though, is filled with problems. You don’t want to suppress glucocorticoid levels too much because, umpteen pages into this book, it should be apparent by now that those hormones are pretty important. Moreover, the “adrenal steroidogenesis inhibitors,” as those drugs are called, can have some nasty side effects. Nonetheless, some solid reports have shown them to have antidepressant effects in people with high-glucocorticoid depressions. Another version of the same approach is to use a drug that blocks glucocorticoid receptors in the brain. These exist and are relatively safe, and there’s now decent evidence that they work as well.* A relatively obscure hormone called DHEA, which has some ability to block

glucocorticoid access to its receptor, has been reported to have some antidepressant qualities as well. Thus, these recent studies not only teach us something about the bases of depression, but may open the way for a whole new generation of medications for the disease. Some investigators have built on these observations with a fairly radical suggestion. For those biological psychiatrists concerned with the hormonal aspects of depression, the traditional glucocorticoid scenario is outlined above. In it, depressions are stressful and raise glucocorticoid levels; when someone is treated with antidepressants, the abnormal neurochemistry (related to serotonin, norepinephrine, etc.) is normalized, lessening the depression and, by the way, making life feel less stressful, with glucocorticoid levels returning to normal as a by-product. The new scenario is the logical extension of the inverted causality also just discussed. In this version, for any of a number of reasons, glucocorticoid levels rise in someone (because the person is under a lot of stress, because something about the regulatory control of glucocorticoids is awry in that person), causing changes in the chemistry of serotonin (or norepinephrine, etc.) and a depression. In this scenario, antidepressants work by normalizing glucocorticoid levels, thereby normalizing the brain chemistry and alleviating the depression. For this view to be supported, it has to be shown that the primary mechanism of action of the different classes of antidepressants is to work on the glucocorticoid system, and that changes in glucocorticoid levels precede the changes in brain chemistry or depressive symptoms. A few researchers have presented evidence that antidepressants work to rapidly alter numbers of glucocorticoid receptors in the brain, altering regulatory control of the system and lowering glucocorticoid levels, and these changes precede changes in the traditional symptoms of depression; other researchers have not observed this. As usual, more research is needed. But even if it turns out that, in some patients, depression is driven by elevated glucocorticoid levels (and recovery from depression thus mediated by reduction of those levels), that can’t be the general mechanism of the disease in all cases: only about half of depressives actually have elevated glucocorticoid levels. In the other half, the glucocorticoid system seems to work perfectly normally. Perhaps this particular stress/depression link is relevant only during the first few rounds of someone’s depression (before the endogenous rhythmicity kicks in), or only in a subset of individuals. We have now seen ways in which stress and glucocorticoids are intertwined with the biology of depression. That intertwining is made even tighter when considering the psychological picture of the disease.

Stress and the Psychodynamics of Major Depressions

I have to begin with Freud. I know it is obligatory to dump on Freud, and some of it is deserved, but there is much that he still has to offer. I can think of few other scientists who, nearly a century after their major contributions, are still considered important and correct enough for anyone to want to bother pointing out their errors instead of just consigning them to the library archives.

Freud was fascinated with depression and focused on the issue that we began with—why is it that most of us can have occasional terrible experiences, feel depressed, and then recover, while a few of us collapse into major depression (melancholia)? In his classic essay “Mourning and Melancholia” (1917), Freud began with what the two have in common. In both cases, he felt, there is the loss of a love object. (In Freudian terms, such an “object” is usually a person, but can also be a goal or an ideal.) In Freud’s formulation, in every loving relationship there is ambivalence, mixed feelings—elements of hatred as well as love. In the case of a small, reactive depression—mourning—you are able to deal with those mixed feelings in a healthy manner: you lose, you grieve, and then you recover. In the case of a major melancholic depression, you have become obsessed with the ambivalence—the simultaneity, the irreconcilable nature of the intense love alongside the intense hatred. Melancholia—a major depression—Freud theorized, is the internal conflict generated by this ambivalence. This can begin to explain the intensity of grief experienced in a major depression. If you are obsessed with the intensely mixed feelings, you grieve doubly after a loss—for your loss of the loved individual and for the loss of any chance now to ever resolve the difficulties. “If only I had said the things I needed to, if only we could have worked things out”—for all of time, you have lost the chance to purge yourself of the ambivalence. For the rest of your life, you will be reaching for the door to let you into a place of pure, unsullied love, and you can never reach that door. It also explains the intensity of the guilt often experienced in major depression. If you truly harbored intense anger toward the person along with love, in the aftermath of your loss there must be some facet of you that is celebrating, alongside the grieving. “He’s gone; that’s terrible but…thank god, I can finally live, I can finally grow up, no more of this or that.” Inevitably, a metaphorical instant later, there must come a paralyzing belief that you have become a horrible monster to feel any sense of relief or pleasure at a time like this. Incapacitating guilt. This theory also explains the tendency of major depressives in such circumstances to, oddly, begin to take on some of the traits of the lost loved/hated one—and not just any traits, but invariably the ones that the survivor found most irritating. Psychodynamically, this is wonderfully logical. By taking on a trait, you are being loyal to your lost, beloved opponent. By picking an irritating trait, you are still trying to convince the world you were right to be irritated—you see how you hate it when I do it; can you imagine what it was like to have to put up with that for years? And by picking a trait that, most of all, you find irritating, you are not only still trying to score points in your argument with the departed, but you are punishing yourself for arguing as well. Out of the Freudian school of thought has come one of the more apt descriptions of depression—“aggression turned inward.” Suddenly the loss of pleasure, the psychomotor retardation, the impulse to suicide all make sense. As do the elevated glucocorticoid levels. This does not describe someone too lethargic to function; it is more like the actual state of a patient in depression, exhausted from the most draining emotional conflict of his or her life—one going on entirely within. If that doesn’t count as psychologically stressful, I don’t know what does. Like other good parts of Freud, these ideas are empathic and fit many clinical traits; they just feel “right.” But they are hard to assimilate into modern science, especially biologically oriented

psychiatry. There is no way to study the correlation between serotonin receptor density and internalization of aggression, for example, or the effects of estrogen-progesterone ratios on love- hate ratios. The branch of psychological theorizing about depression that seems most useful to me, and is most tightly linked to stress, comes from experimental psychology. Work in this field has generated an extraordinarily informative model of depression.

Stress, Learned Helplessness, and Depression

In order to appreciate the experimental studies underlying this model, recall that in the preceding chapter on psychological stress, we saw that certain features dominated as psychologically stressful: a loss of control and of predictability within certain contexts, a loss of outlets for frustration, a loss of sources of support, a perception of life worsening. In one style of experiment, pioneered by the psychologists Martin Seligman and Steven Maier, animals are exposed to pathological amounts of these psychological stressors. The result is a condition strikingly similar to a human depression. Although the actual stressors may differ, the general approach in these studies always emphasizes repeated stressors with a complete absence of control on the part of the animal. For example, a rat may be subjected to a long series of frequent, uncontrollable, and unpredictable shocks or noises, with no outlets. After awhile, something extraordinary happens to that rat. This can be shown with a test. Take a fresh, unstressed rat, and give it something easy to learn. Put it in a room, for example, with the floor divided into two halves. Occasionally, electricity that will cause a mild shock is delivered to one half, and just beforehand, there is a signal indicating which half of the floor is about to be electrified. Your run-of-the-mill rat can learn this “active avoidance task” easily, and within a short time it readily and calmly shifts the side of the room it sits in according to the signal. Simple. Except for a rat who has recently been exposed to repeated uncontrollable stressors. That rat cannot learn the task. It does not learn to cope. On the contrary, it has learned to be helpless. This phenomenon, called learned helplessness, is quite generalized; the animal has trouble coping with all sorts of varied tasks after its exposure to uncontrollable stressors. Such helplessness extends to tasks having to do with its ordinary life, like competing with another animal for food, or avoiding social aggression. One might wonder whether the helplessness is induced by the physical stress of receiving the shocks or, instead, the psychological stressor of having no control over or capacity to predict the shocks. It is the latter. The clearest way to demonstrate this is to “yoke” pairs of rats—one gets shocked under conditions marked by predictability and a certain degree of control, the other rat gets the identical pattern of shocks, but without the control or predictability. Only the latter rat becomes helpless. Seligman argues persuasively that animals suffering from learned helplessness share many psychological features with depressed humans. Such animals have a motivational problem—one of the reasons that they are helpless is that they often do not even attempt a coping response when they are in a new situation. This is quite similar to the depressed person who doesn’t even

try the simplest task that would improve her life. “I’m too tired, it seems overwhelming to take on something like that, it’s not going to work anyway….” Animals with learned helplessness also have a cognitive problem, something awry with how they perceive the world and think about it. When they do make the rare coping response, they can’t tell whether it works or not. For example, if you tighten the association between a coping response and a reward, a normal rat’s response rate increases (in other words, if the coping response works for the rat, it persists in that response). In contrast, linking rewards more closely to the rare coping responses of a helpless rat has little effect on its response rate. Seligman believes that this is not a consequence of helpless animals somehow missing the rules of the task; instead, he thinks, they have actually learned not to bother paying attention. By all logic, that rat should have learned, “When I am getting shocked, there is absolutely nothing I can do, and that feels terrible, but it isn’t the whole world; it isn’t true for everything.” Instead, it has learned, “There is nothing I can do. Ever.” Even when control and mastery are potentially made available to it, the rat cannot perceive them. This is very similar to the depressed human who always sees glasses half empty. As Beck and other cognitive therapists have emphasized, much of what constitutes a depression is centered around responding to one awful thing and overgeneralizing from it—cognitively distorting how the world works. The learned helplessness paradigm produces animals with other features strikingly similar to those in humans with major depressions. There is a rat’s equivalent of dysphoria—the rat stops grooming itself and loses interest in sex and food. The rat’s failure even to attempt coping responses suggests that it experiences an animal equivalent of psychomotor retardation.* In some models of learned helplessness, animals mutilate themselves, biting at themselves. Many of the vegetative symptoms appear as well—sleep loss and disorganization of sleep architecture, elevated glucocorticoid levels. Most critically, these animals tend to be depleted of norepinephrine in certain parts of the brain, while antidepressant drugs and ECT speed up their recovery from the learned helplessness state. Learned helplessness has been induced in rodents, cats, dogs, birds, fish, insects, and primates, including humans. It takes surprisingly little in terms of uncontrollable unpleasantness to make humans give up and become helpless in a generalized way. In one study by Donald Hiroto, student volunteers were exposed to either escapable or inescapable loud noises (as in all such studies, the two groups were paired so that they were exposed to the same amount of noise). Afterward, they were given a learning task in which a correct response turned off a loud noise; the “inescapable” group was significantly less capable of learning the task. Helplessness can even be generalized to nonaversive learning situations. Hiroto and Seligman did a follow-up study in which, again, there was either controllable or uncontrollable noise. Afterward the latter group was less capable of solving simple word puzzles. Giving up can also be induced by stressors far more subtle than uncontrollable loud noises. In another study, Hiroto and Seligman gave volunteers a learning task in which they had to pick a card of a certain color according to rules that they had to discern along the way. In one group, these rules were learnable; in the other group, the rules were not (the card color was randomized). Afterward, the latter group was less capable of coping with a simple and easily solved task. Seligman and colleagues have also demonstrated that unsolvable tasks induced helplessness afterward in social coping situations.

Thus humans can be provoked into at least transient cases of learned helplessness, and with surprising ease. Naturally, there is tremendous individual variation in how readily this happens— some of us are more vulnerable than others (and you can bet that this is going to be important in considering stress management in the final chapter). In the experiment involving inescapable noise, Hiroto had given the students a personality inventory beforehand. Based on that, he was able to identify the students who came into the experiment with a strongly “internalized locus of control”—a belief that they were the masters of their own destiny and had a great deal of control in their lives—and, in contrast, the markedly “externalized” volunteers, who tended to attribute outcomes to chance and luck. In the aftermath of the uncontrollable stressor, the externalized students were far more vulnerable to learned helplessness. Transferring that to the real world, with the same external stressors, the more that someone has an internal locus of control, the less the likelihood of a depression. Collectively, these studies strike me as extremely important in forming links among stress, personality, and depression. Our lives are replete with incidents in which we become irrationally helpless. Some are silly and inconsequential. Once in the African camp that I shared with Laurence Frank, the zoologist whose hyenas figured in chapter 7, we managed to make a disaster of preparing macaroni and cheese over the campfire. Inspecting the mess, we ruefully admitted that it might have helped if we had bothered to read the instructions on the box. Yet we had both avoided doing that; in fact, we both felt a formless dread about trying to make sense of such instructions. Frank summed it up: “Face it. We suffer from learned cooking helplessness.” But life is full of more significant examples. If a teacher at a critical point of our education, or a loved one at a critical point of our emotional development, frequently exposes us to his or her own specialized uncontrollable stressors, we may grow up with distorted beliefs about what we cannot learn or ways in which we are unlikely to be loved. In one chilling demonstration of this, some psychologists studied inner-city school kids with severe reading problems. Were they intellectually incapable of reading? Apparently not. The psychologists circumvented the students’ resistance to learning to read by, instead, teaching them Chinese characters. Within hours they were capable of reading more complex symbolic sentences than they could in English. The children had apparently been previously taught all too well that reading English was beyond their ability. A major depression, these findings suggest, can be the outcome of particularly severe lessons in uncontrollability for those of us who are already vulnerable. This may explain an array of findings that show that if a child is stressed in certain ways—loss of a parent to death, divorce of parents, being a victim of abusive parenting—the child is more at risk for depression years later. What could be a more severe lesson that awful things can happen that are beyond our control than a lesson at an age when we are first forming our impressions about the nature of the world? As an underpinning of this, Paul Plotsky and Charles Nemeroff of Emory University have shown that rats or monkeys exposed to stressors early in life have a lifelong increase in CRH levels in their brain. “According to our model,” writes Seligman, “depression is not generalized pessimism, but pessimism specific to the effects of one’s own skilled actions.” Subjected to enough uncontrollable stress, we learn to be helpless—we lack the motivation to try to live because we

assume the worst; we lack the cognitive clarity to perceive when things are actually going fine, and we feel an aching lack of pleasure in everything.*

Attempting an Integration Psychological approaches to depression give us some insight into the nature of the disease. According to one school, it is a state brought about by pathological overexposure to loss of control and outlets for frustration. In another psychological view, the Freudian one, it is the internalized battle of ambivalences, aggression turned inward. These views contrast with the more biological ones—that depression is a disorder of abnormal neurotransmitter levels, abnormal communication between certain parts of the brain, abnormal hormone ratios, genetic vulnerability. There are extremely different ways of looking at the world, and researchers and clinicians from different orientations often don’t have a word to say to one another about their mutual interest in depression. Sometimes they seem to be talking radically different languages—psychodynamic ambivalence versus neurotransmitter autoreceptors, cognitive overgeneralization versus allelic variants of genes. What I view as the main point of this chapter is that stress is the unifying theme that pulls together these disparate threads of biology and psychology. We have now seen some important links between stress and depression: extremes of psychological stress can cause something in a laboratory animal that looks pretty close to a depression. Moreover, stress is a predisposing factor in human depression as well, and brings about some of the typical endocrine changes of depression. In addition, genes that predispose to depression only do so in a stressful environment. Tightening the link further, glucocorticoids, as a central hormone of the stress-response, can bring about depression-like states in an animal, and can cause depression in humans. And finally, both stress and glucocorticoids can bring about neurochemical changes that have been implicated in depression. With these findings in hand, the pieces begin to fit together. Stress, particularly in the form of extremes of lack of control and outlets, causes an array of deleterious changes in a person. Cognitively, this involves a distortive belief that there is no control or outlets in any circumstance-learned helplessness. On the affective level, there is anhedonia; behaviorally, there is psychomotor retardation. On the neurochemical level, there are likely disruptions of serotonin, norepinephrine, and dopamine signaling—as will be shown in chapter 16, prolonged stress can deplete dopamine in the pleasure pathways. Physiologically, there are alterations in, among other things, appetite, sleep patterns, and sensitivity of the glucocorticoid system to feedback regulation. We call this array of changes, collectively, a major depression. This is terrific. I believe we have a stress-related disease on our hands. But some critical questions remain to be asked. One concerns why it is that after three or so bouts of major depression the stress-depression link uncouples. This is the business about depressive episodes taking on an internal rhythm of their own, independent of whether the outside world is actually

pummeling you with stressors. Why should such a transition occur? At present, there’s a lot of theorizing but very little in the way of actual data. But the most basic question remains, why do only some of us get depressed? An obvious answer is because some of us are exposed to a lot more stressors than others. And, when factoring in development, that can be stated in a way that also includes history—not only are some of us exposed to more stressors than others, but if we are exposed to some awful stressors early in life, forever after we will be more vulnerable to whatever subsequent stressors are thrown at us. This is the essence of allostatic load, of wear and tear, where exposure to severe stress produces rents of vulnerability. So differential incidences of depression can be explained by differences in the amount of stress, and/or in stress histories. But even for the same stressors and the same history of stress, some of us are more vulnerable than others. Why should some of us succumb more readily? To begin to make sense of this, we have to invert that question, to state it in a more world-weary way. How is it that any of us manage to avoid getting depressed? All things considered, this can be an awful world, and at times it must seem miraculous that any of us resist despair. The answer is that we have built into us a biology of recovering from the effects of stress that provoke depression. As we’ve seen, stress and glucocorticoids can bring about many of the same alterations in neurotransmitter systems that have been implicated in depression. One of the best documented links is that stress depletes norepinephrine. No one is sure exactly why the depletion occurs, although it probably has something to do with norepinephrine being consumed faster than usual (rather than its being made more slowly than usual). Critically, not only does stress deplete norepinephrine, but it simultaneously initiates the gradual synthesis of more norepinephrine. At the same time that norepinephrine content is plummeting, shortly after the onset of stress, the brain is starting to make more of the key enzyme tyrosine hydroxylase, which synthesizes norepinephrine. Both glucocorticoids and, indirectly, the autonomic nervous system play a role in inducing the new tyrosine hydroxylase. The main point is that, in most of us, stress may cause depletion of norepinephrine, but only transiently. We’re about to see there are similar mechanisms related to serotonin. Thus, while everyday stressors bring about some of the neurochemical changes linked to depression along with some of the symptoms—we feel “blue”—at the same time, we are already building in the mechanisms of recovery. We get over it, we put things behind us, we get things in perspective, we move on with our lives…we heal and we recover. So, given the same stressors and stress histories, why do only some of us get depressed? There is increasing evidence for a reasonable answer, which is that the biology of vulnerability to depression is that you don’t recover from stressors very well. Back to that finding of the different versions of “gene Z,” where one version increases your risk for depression, but only when coupled with a history of major stressors. The gene turns out to code for a protein called the serotonin transporter (also known as 5-HTT, derived from the fact that the chemical abbreviation for serotonin is “5-HT”). In other words, the pump that causes the reuptake of serotonin from the synapse. Whose actions are inhibited by drugs like Prozac, which are SSRIs—selective serotonin

reuptake inhibitors. Aha. A whole bunch of pieces here are teetering on the edge of falling into place. The different allelic versions of the 5-HTT gene differ as to how good they are at removing serotonin from the synapse. And where does stress fit in? Glucocorticoids help regulate how much 5-HTT is made from the gene. And, critically, glucocorticoids differ in how good they are at doing that, depending on which allelic version of the 5-HTT gene you have. This allows us to come up with a working model of depression risk. It is a simplistic one, and a more realistic version must incorporate the likelihood of scads more examples of interactions among genes and stressors than simply this stress/glucocorticoids/5-HTT story.* Nonetheless, maybe what occurs is something like this: a major stressor comes along and produces some of the neurochemical changes of depression. The more prior history of stress you have, especially early in life, the less of a stressor it takes to produce those neurochemical changes. But the same stress signal, namely glucocorticoids, alters norepinephrine synthesis, serotonin trafficking, and so on, starting you on the road toward recovery. Unless your genetic makeup means that those recovery steps don’t work very well. This is the essence of the interaction between biology and experience. Take a sufficiently severe stressor and, as studies suggest, virtually all of us will fall into despair. No degree of neurochemical recovery mechanisms can maintain your equilibrium in the face of some of the nightmares that life can produce. Conversely, have a life sufficiently free of stress, and even with a genetic predisposition, you may be safe—a car whose brakes are faulty presents no danger if it is never driven. But in between those two extremes, it is the interaction between the ambiguous experiences that life throws at us and the biology of our vulnerabilities and resiliencies that determines which of us fall prey to this awful disease.