BACON says, Coffee “comforts the head and heart, and helps digestion”; Dr. WILLIS says, “being daily drank, it wonderfully clears and enlightens each part of the soul, and disperses all the clouds of every function.” The celebrated Doctor HARVEY used it often; VOLTAIRE lived almost on it; and the learned and sedentary of every country have recourse to it, to refresh the brain, oppressed by study and contemplation.
—Benjamin Moseley, M.D., A Treatise Concerning the Properties and Effects of Coffee, 1785
The saying goes, “You can’t be too rich or too thin,” to which perhaps could be added, “or too smart,” because, even if each man is correct about how bright he conceives himself to be, he would find it still better to be even a bit sharper. How far would you go to acquire, for example, a drug that would enable you to perform better on an IQ test, an SAT test, or a Bar examination? Or one that would help you to prepare your taxes or balance your checkbook more accurately, solve chess problems or crossword puzzles more readily, make better investments, or program a computer with more acuity, or even drive home more safely? Surprisingly, you might not have to go very far, because caffeine, in many ways, is a “smart pill” that can do just those things.
As demonstrated by scientific evidence and common experience, caffeine is a rare and wonderful substance that safely improves many mental functions, including alertness, memory, learning, and cognition. As early as 1933, one researcher analyzed the effects of caffeine on solving more than 250 chess problems, comparing performance of test subjects with and without caffeine. He observed a consistently remarkable improvement in performance with caffeine.1 Such improvements were reflected in a 1960s advertising campaign that dubbed coffee “The Think Drink.”
However, as to what the nature of this improvement may be or how great its extent, there is little agreement anywhere. Some people are convinced that they can’t think clearly or precisely without caffeine, while others say it makes them jittery and error prone. Naturally, behavioral scientists have been eager to discover the secret of caffeine’s ability to improve the brain’s information processing. Two complementary hypotheses explaining this remarkable power are supported by experimental data. The first hypothesis, sometimes called the “non-specific energetic” theory, attributes caffeine’s enhancement of mental functions to a generalized energizing effect. The second hypothesis, sometimes called the “specific cognitive” theory, attributes these enhancements to specific effects on brain or neural activity. Finally, a “cognitive-energetic” theory, combining the two, has also been formulated and may offer the most complete and best-integrated elucidation of the phenomena.
J.E.Barmack, in a paper published in the Journal of Experimental Psychology in 1940, was one of the first to advance the non-specific energetic theory. Barmack recognized the possibility that caffeine’s overall antihypnotic and antifatigue properties could be part of the story, but, observing that caffeine increased the rate at which people can add numbers, advanced the notion that caffeine acts non-specifically on “some central process or processes concerned with alertness” that “allay the development of a bored attitude to a task.”2 This idea is supported by many studies of continuous performance, over a period of a half-hour or more of what experimental psychologists call a “vigilance task,” one that requires prolonged attention and responsiveness but little physical activity. In real life, caffeine improves long-term performance on vigilance tasks such as solving arithmetic problems, driving a car, or flying an airplane. Its effects are most apparent when people have been working at their tasks for some time and are minimal when tasks are just begun. When people are allowed to take breaks to alleviate boredom and fatigue, no significant benefit from using caffeine is observed. These findings, based solely on studies of vigilance tasks, apparently confirm Barmack’s theory, that caffeine acts by “refreshing” a fatigued person, so that the enhancing effects of caffeine on long-term performance will obtain on any task that is performed repetitively, monotonously, and requires continuous attention.3
The specific cognitive theory, championed by H.Nash in his 1962 book Alcohol and Caffeine,4 asserts that caffeine acts directly on “specific neural capacities” that are intrinsic to a given task and that it enhances performance on these tasks irrespective of whether a person is fatigued. This idea was suggested to Nash by his examination of performances of several different short-term tasks, some of which exhibited improvement after caffeine was ingested, while others remained unaffected. Nash argues that the benefits of caffeine on performance depend not on an improvement in general energy levels, as Barmack had asserted, but instead on specific benefits related to the nature of the task at hand. Abandoning the metaphor of the organism as an energy system, Nash relied on another metaphor, one that became and remains the most generally accepted in cognitive psychology today: that of the human organism as an information-processing system. He observed improvement in the performance of a number of tasks, such as adding numbers, immediate recall, and word fluency. These benefits were realized even on brief tests administered when the subjects were rested and alert, and neither fatigued nor bored. In contrast, he found no improvement in tests of abstract reasoning, using language, deduction, estimating time intervals, or spotting arithmetic mistakes. The overall conclusion from such studies has been that caffeine “facilitates the speed, but not the memory, component of the task.”
If Barmack’s non-specific energizing theory is correct, we should expect caffeine to improve cognitive performance only when a person has become bored or tired. If Nash’s specific cognitive theory is correct, we should expect an improvement even when a person is rested and alert to start with, but this improvement would be observed only in some tasks and not others. However, because these theories are complementary rather than inconsistent, which is to say, they could both be true at the same time, we must also consider the syncretic hypothesis of A.F.Sanders, who argues that the improvements in mental capacity caused by caffeine are a function of both the energy level of the subject and the cognitive nature and demands of the task.5 Aiming to unite the energetic and cognitive models of human information processing, Sanders published his idea in 1983 that caffeine’s effect on performance is best understood as a function of both the energetic state of the person and the cognitive requirements of the task.
Unfortunately, even with the advancement of Sanders’ cognitive-energetic theory, the scientific community remains far from a complete and consistent explanation of caffeine’s sometimes apparently paradoxical effects on human performance. For example, an adverse effect has been observed on the attempt to repeat numbers backward, while a beneficial effect has been observed on the attempt to repeat them in their original order. In addition, caffeine impaired some factors of cognitive intelligence, while improving those related to speed. In other cases, caffeine had a deleterious effect on a given task until that task was practiced, whereupon the use of caffeine resulted in an improvement.6 Another troubling inconsistency is the low level of test-retest reliability. That is, the results of studies of caffeine’s effects, particularly on the performance of complex tasks, vary widely, forcing us to wonder which conclusions are the correct ones.7
Nevertheless, overall, scientific studies have confirmed some specific effects on cognition and learning: Caffeine improves the performance of simple, familiar, routine tasks, and it impairs or fails to affect the accomplishment of complex, novel, unpracticed tasks. Perhaps the reason for this difference is that, by conferring extra energy, caffeine causes a person to work more quickly but possibly less carefully.8
This explanation is in line with experiments that show that caffeine can stimulate fast and strong but incorrect reactions. For example, an experienced computer programmer may report that using caffeine makes well-practiced programming assignments easier to complete, while it appears to sometimes interfere with the successful solution of new and very difficult programming problems. Because real-life problems frequently present both sorts of challenges, and the nature and distribution of these challenges will vary as among different subjects performing the same tasks, caffeine should be expected to produce a complex array of sometimes contrary effects. In other words, although in some ways caffeine may give us a beneficial boost in our capacity to perform certain tasks, in others it may induce us to make precipitous, overeager choices, or to “jump the gun,” interfering with the circumspection necessary for accurate decisions. Perhaps a quick trigger finger is good for the artist’s hand, if the testimony of artists, musicians, and writers such as Balzac, Voltaire, Samuel Johnson, Beethoven, and Goethe about caffeine’s importance in their creative lives is to be accepted. However, psychology, which has yet to attain a coherent understanding of creativity, cannot shed much light on caffeine’s effects on the creative process.
In the laboratory it is easy to create experimental tasks, such as one requiring subjects to identify or remember numbers or colors flashing across a screen, in which few if any of the study’s participants will have been previously practiced. In life, however, most significant tasks are repeated and even systematically studied and practiced with the intent of improving performance. In addition, in many work situations, because people choose the jobs they pursue, they will often be performing those tasks for which they have the greatest innate abilities. For these reasons, it is essential, in predicting the effect caffeine will have on a person’s performance of a given task, to take into account not only the features of the task itself, but also to reckon with the competency of the person performing it.
Is programming a computer difficult or easy? Many people would find even basic programming tasks challenging, complex, novel, and creative. Experienced, well-practiced, and talented programmers might find many of these same programming tasks easy, simple, familiar, and routine. Caffeine might therefore affect performances on a simple programming test in opposite ways: The performance of the person who had little competency with programming, either because of lack of specific experience or specific ability, might well be impaired by caffeine. The performance of the person who was eminently competent in programming, either because of extensive experience or specific ability, might well be given a significant boost.
The effects of caffeine on task performance in real life are complicated still further by the fact that life’s tasks are compound in nature, and, even for the same person, certain elements of a task may be challenging and other elements easy, so that caffeine would exert a variable effect on different stages and parts of the task. We can reasonably speculate that, overall, the more competent you are in performing a task, the more caffeine will help you do even better, while the less able you are in coping with a task, the more likely it is that caffeine will fail to affect or even impair your efforts. If this notion turns out to be true, the use of sufficiently large doses of caffeine, by tending to push lower scores lower and higher scores higher, should serve to flatten the bell curve of an IQ test into a sort of flying saucer.
A curious twist to this question is the possibility that caffeine may affect introverted people differently from the way it affects extroverted people. One study concluded that, when posed with challenging mental tasks, such as proofreading or solving mathematical problems, impulsive, extroverted people get a boost in performance from caffeine, while those who describe themselves as less impulsive and more introverted often suffer marked detriments after caffeine ingestion. Another study of caffeine’s differing effects on extroverts and introverts performing both simple and complex tasks came to similar conclusions. The routine or simple task was to pick out a letter each time it occurred on a page of type. The challenging or complex task was to answer word analogies and sentence completion questions from the Graduate Record Examination. This study seemed to confirm that everybody tends to do better the higher the dose, if the task is extremely simple. However, on the complex task, the extroverts’ performance improved in a dose-dependent correlation with caffeine, while the introverts’ performance worsened.
One additional possibility pertaining to competency comes to mind: If caffeine’s impairment of challenging tasks is a result of prompting us to “jump the gun,” perhaps a person can learn to compensate for this sort of overeager “coupling” even when riding high on caffeine, learning to pause in order to perform the necessary evaluation for a correct choice. If this can be done, then a savvy caffeine consumer might find that he can multiply the number of tasks in which caffeine is helpful and the degree to which it is helpful, and reduce the number and degree to which it causes impairment.
In recent years, in addition to continuing studies of caffeine’s effects on complex mental activities such as reasoning and learning, researchers have paid increasing attention to its effects on short-term memory. Overall, the results show that caffeine improves performance on tasks that require remembering small amounts of information and impairs or leaves unaffected performance on tasks requiring remembering a great deal. An example of a more demanding sort of memory task is a test in which subjects listen to or read long lists of words and are then asked to remember as many as possible. The experimenters note either no effect from caffeine or perhaps even a small impairment. Another way of conceptualizing these effects is provided by the Humphreys-Revelle 9 model, according to which tasks that are primarily dependent on information processing, such as vigilance, simple arithmetic, or reaction time, are improved, because they make relatively small demands on short-term memory, while tasks with a high short-term memory component may be unaffected or adversely affected. Unfortunately, there is much ambiguity in the data that do exist about these effects. When weighing the conclusions of existing research, we would do well to remember a well-designed 1974 memory experiment by researcher V.E.Mitchell and his colleagues, the cautionary results of which were reminiscent of the title of Luigi Pirandello’s play Right You Are, If You Think You Are, because they seem to demonstrate that performance was improved by caffeine when and only when the participants were told that they had ingested the drug.10
Nevertheless, millions of students use caffeine to fuel “all-nighters.” Based on the available scientific evidence, how does this use of caffeine affect studying and test taking? Caffeine helps people to feel less drowsy and less fatigued, be better able to perform some manual or perfunctory tasks, such as typing or calculating, and, under certain circumstances, to be more capable of sustaining rapid thought and to remember more. However, some studies have found that caffeine does not significantly alter numerical reasoning, short-term memory of complex data, or verbal fluency. In other words, caffeine may help you to stay awake, but it won’t necessarily improve your intellectual skills.
Students depending on caffeine to extend their study time should also be aware of its possible adverse effects when taken in large quantities and be prepared for the crash after its stimulating powers subside. As Socrates suggested, the best guide for students is to know themselves: From a couple of Vivarin tablets, the sensitive may experience restlessness, anxiety, nausea, headache, tense muscles, and sleep disturbances, or a subsequent letdown, while others, from a much higher dose, might feel fine.
A recent startling discovery by Menachem Segal, professor of neurosciences at the Weizmann Institute in Rehovot, Israel, and an expert on neuromodulators in the brain, suggests that caffeine causes changes to brain cells that are likely to have profoundly beneficial effects on long-term memory.11 In earlier research, Segal discovered that increasing the amount of calcium absorbed by brain cells is one way of improving longterm memory. Because caffeine augments the ability of these cells to metabolize calcium, Segal studied the effects of adding caffeine directly to the hippocampus, an area of the brain that is critical to learning and long-term memory, to test the hypothesis that the calcium levels inside the cells would rise as a result. The outcome of his experiment confirmed this conjecture, proving that caffeine increased the calcium levels in brain cells. But Segal also observed a more astonishing phenomenon: Caffeine caused existing dendritic spines, the branching extensions at the ends of nerve cells that allow them to make synaptic connections with each other, to grow longer and even caused new spines and branches to develop as well. Although no direct experimental data are available on the actual effects, if any, of caffeine on long-term memory,12 neuroscientists have long believed that an improvement in “wiring” does in fact improve both long-term memory and learning. If this connection is demonstrated in future studies, caffeine would be confirmed as the only known substance that can augment brain functions by altering the physical structure of the brain.
Few people would ever question that caffeine helps to keep you awake and alert during the day. In 1990 this phenomenon was first systematically investigated under laboratory conditions and quantified using sophisticated measuring techniques: Researcher Zwyghuizen-Doorenbos led a group of scientists who orally administered 250 mg of caffeine at 9 A.M. and 1 P.M. on two successive days. Checking repeatedly for objective parameters of wakefulness and alertness, they determined that caffeine does in fact help keep people awake and alert. More surprisingly, the subjects who had been previously given caffeine continued to demonstrate increased alertness over the placebo group, even on the third day of the experiment, on which a placebo containing no caffeine was administered to all participants. In discussing this study, researcher Jack James concludes that this extended effect is psychological rather than pharmacological, asserting the “development of conditioned alerting responses to certain contextual stimuli that had been associated with caffeine (e.g., the coffee beverage vehicle used to administer the drug) during the previous two days.”13 In ordinary parlance, this means that if you think you are drinking a caffeinated beverage, you may wake up just as if you had actually drunk it simply because you expect that you will.
A broad range of studies from the 1930s through the 1990s have produced conflicting findings about caffeine’s effect on reaction time.14 Some of the apparently contradictory results may be explained in terms of the 1987 findings of researcher J.D. Roache and R.R.Griffiths, who found that reaction times were improved more by a dose of 400 mg of caffeine than by doses of 200 mg or 600 mg.15,16 Inother words, there is no simple, linear relationship between dosage and improvement in reaction time of which we could state, “If some is good, more is better.” Other studies have confirmed that varying amounts of caffeine produced similarly differing effects on reaction time, finding, for example, that in young adults 300 mg of caffeine significantly improved scores, while doses of 600 mg left them unaffected.
As in the debate over the nature and extent of improvements in cognitive performance, there has been some controversy over whether caffeine can improve psychomotor performance only if you are fatigued to begin with or whether it improves such performance even in a well-rested person. Most researchers agree that it is now well established that the beneficial effects of caffeine are found both before and after decrements in performance owing to fatigue. For example, it has been demonstrated that real-life and simulated automobile driving performance levels are improved by caffeine, irrespective of whether the person was tired or well rested to start.17
Alertness, the complex of mental capacities that suffer during sleep loss, was studied by Michael H.Bonnet and colleagues in 1994 when he compared the value of caffeine and naps in helping to sustain performance during two days and nights of sleep loss.18 Previous studies had shown that performance during sleep loss is, not surprisingly, “improved by prophylactic naps as a function of varying nap length.” Bonnet compared the improvement conferred by naps with that realizable with either repeated or single-dose administration of caffeine. As would be expected, the results showed that an eight-hour “nap,” which is to say, a good night’s sleep, did more to improve performance, mood, and alertness than any sort of caffeine regimen and that the benefits lasted longer than the effects of caffeine, which peak and lose effect within six hours. The study also found that naps could be combined effectively with small repetitive doses of caffeine to maintain alertness. Nothing lasts forever, though. Bonnet concludes that neither a nap nor repeated doses of caffeine could preserve performance, mood, and alertness past twenty-four hours. Beyond that, caffeine’s effects in these respects approached those of a placebo.
One of the definitions of drugs that produce a physical dependence is that the abrupt cessation of their use will cause people to perform tests poorly and feel listless and generally “blue.” Leading caffeine researcher Jack James advances a theory that a nearly universal yet unacknowledged physical dependence on caffeine may have confounded the results of many studies that purported to show the psychomotor advantages conferred by the drug.19 His argument is plausible: If nearly everyone uses caffeine nearly every day, then when a scientist takes a pool of subjects and administers caffeine to one part of the group and withholds it from the others, those not receiving caffeine will enter varying stages of caffeine withdrawal and therefore will perform more poorly than those who are not suffering from withdrawal. In effect, James is asserting that most people have what among heroin users is called a “chippie,” a lowgrade habit, of which the habituated person is sometimes unaware, that renders them mildly dysphoric and uncomfortable when deprived of the drug. Such heroin users would probably demonstrate better performance if given a small amount of heroin, simply because the heroin would restore them to a normal metabolic balance and remove the impediment of the mild withdrawal syndrome. If James is correct, all the “improvements” worked by caffeine in psychomotor or cognitive performance may simply be artifacts of an unknown caffeine chippie or a hidden need for caffeine that impairs the habitué when he stops using it. The only way to determine if a nearly universal addiction is distorting experimental findings is to make sure that all the subjects in a given study have been weaned off of caffeine for at least a week or two, so that they begin without any taint of addiction.
It may be nothing to worry about, but anxiety, including such symptoms as unwarranted trepidation, apprehension, agitation, turmoil, and uneasiness, is the most common psychological disorder in the United States. In severe cases, it erupts into recurring panic attacks, the symptoms of which include increased heart rate, palpitations, jitters, irritability, perspiration, and rapid breathing. Caffeine is generally recognized by researchers as an anxiogenic substance—that is, one that is productive of anxiety. The pharmacological basis of this effect remains uncertain. One possible contributory mechanism is the process by which caffeine binds to adenosine receptors, interfering with the systems that would otherwise have reduced anxiety. (Binding to benzodiazepine and endorphin receptors have been cited as well, but at ordinary levels of consumption this activity appears too small to be of significance.) Another is that caffeine interferes with the noradrenergic system so as to increase the release of adrenaline.20 Adrenaline, the hormone the adrenal glands excrete in response to excitement, stress, or fear, produces a more rapid and stronger heartbeat and more rapid and deeper breathing, and it can also produce anxiety. Some claim that caffeine in combination with emotional distress causes the release of more adrenaline than emotional distress alone, suggesting that even if it cannot cause anxiety, caffeine may exacerbate it.
In 1971 R.Lynn, a leading British researcher, conducted a monumental study of personal traits and practices and physical and psychological disorders, including psychiatric disorders, cigarette smoking, suicide, coronary heart disease, and anxiety. Observing that the peoples of different nations suffered from different levels of anxiety, Lynn hypothesized that, in those countries in which anxiety levels were high, people would tend to consume less caffeine in order to avoid exacerbating their problems with it. His theory was confirmed by the data. In those countries in which caffeine consumption is low, anxiety is relatively higher, and in those in which it is high, anxiety is lower. Several studies have supported Lynn’s conclusion.21 It has been shown that patients with panic disorder have lower caffeine consumption and that panic attacks can be induced even in normal people challenged with high doses of caffeine.22
People with high caffeine consumption have higher usage rates of anxiety-reducing drugs such as benzodiazepine or meprobamate, when compared to those with moderate to low caffeine use. No one is sure if they consume higher levels of caffeine in order to shake off the sedative effects of the minor tranquilizers, or if they take minor tranquilizers to counteract the anxiety resulting from high doses of caffeine, or even if there is simply a certain population that enjoys using mild psychoactive substances of both the stimulating and sedating varieties.
In any case, the studies dispel one illusion: the image of the nervous, edgy caffeine user. Only excessive caffeine use is correlated with anxiety. If a correlation with normal use can be stated, it is that the people who do not consume caffeine are more likely to have problems with their nerves than those who do. The members of the population who consume caffeine are dispositionally more relaxed than the general population, and, conversely, those who do not consume caffeine are more jittery than most.
As reflected in a new specific and separate diagnosis category in the American Psychiatric Association’s Diagnostic and Statistical Manual IV (DSM-IV), psychiatrists now believe that caffeine can produce a distinct anxiety disorder, over and above the symptom of anxiety that appears as a component of caffeine intoxication and caffeine withdrawal. The DSM-IV states that, as with the anxiety induced by other psychoactive substances such as cocaine, caffeine-induced anxiety disorder can resemble panic disorder, generalized anxiety disorder, social phobia, or even obsessive-compulsive disorder. There have been no studies on the prevalence or incidence of caffeine intoxication anxiety or applying the DSM-IV set of diagnostic criteria.
Just as it is conjectured that Nervous Nellies spontaneously adjust their caffeine intake downward to avoid its anxiolytic effects,23 it is also supposed that some depressives increase their consumption in order to multiply the benefit from caffeine’s euphoric and stimulating powers.24 In other words, they effectively self-medicate with caffeine to dissipate the dark clouds of lassitude, lethargy, and despair that hover around them. The studies remain inconclusive. An example of the ambiguity that still lingers is found in the Tromsø Heart Study, conducted in 1983, of almost 150,000 people, which found significant correlations between high coffee consumption and depression in women (but not in men). However, because this correlation disappears when the results are adjusted for cigarette smoking, it is difficult to draw any conclusions.
A number of studies have found that people who drink at least two portions of caffeinated beverages a day report improved moods, a better social disposition, and more self-confidence and energy. Two large-scale studies, apparently demonstrating lower rates of suicide among coffee drinkers, strongly suggest that caffeine can significantly ameliorate long-term depression and even make life worth living for some people. A 1993 Kaiser Permanente Medical Care Program study of more than one hundred thousand men and women, reported in the Annals of Epidemiology, examined the effects of coffee and tea on mortality and found a lower risk of suicide among people who ingested more caffeine. Under the direction of Arthur Klatsky, M.D., a cardiologist, the study tracked nearly 130,000 Northern California residents, including the records of 4,500 who died during the research, and demonstrated a statistically significant lower rate of suicide among coffee drinkers than coffee abstainers. Klatsky asserted that this was not a fluke finding, because the study was very large, involved a multiracial population of men and women, and examined closely many factors related to mortality such as alcohol consumption and smoking. Another large-scale study, of more than 85,000 female nurses, conducted by Dr. Ichiro Kawachi of Harvard Medical School and Brigham and Women’s Hospital in Boston, funded by the National Institutes of Health and published by the Archives of Internal Medicine in 1996, concluded that women who drink coffee are less likely to commit suicide than those who do not. Even though coffee drinkers appear to engage in the sort of behavior that would increase their risk of depression and suicide (for example, they tend to smoke more and drink more alcohol than non-coffee drinkers and have higher levels of perceived stress), they seem to be highly protected. This study also found that only 100 mg of caffeine a day, about one cup of coffee, could produce increased feelings of well-being, energy, and motivation for work. However, specialists in depression argue that Kawachi failed to control for several factors, including the effects of antidepressants, high-blood-pressure medication, and whether depressed subjects had been told not to drink coffee by their doctors.
There are few studies of the effects of caffeine on human emotional states other than anxiety and depression. However, some researchers have hypothesized that caffeine may operate to decrease aggression, arguing that an increase in benzodiazepine activity increases aggression, and, as we have seen, caffeine inhibits benzodiazepine activity. Studies in 1983 and 1984 by D.R.Cherek 25 examining the effects of caffeine administration on aggressive behavior seem to confirm these findings in humans. Mindful of the aggressive response normally elicited when someone has money taken from him, Cherek designed an experiment in which subjects were promised certain immediate rewards for their performance. He then pretended to renege on his promises, giving the participants, who had received either caffeine tablets, coffee, decaffeinated coffee, or a placebo, the impression that they had been “ripped off.” He found that those who had ingested either coffee or caffeine were more tolerant of being cheated. However, the specter of dubiety that haunts many other caffeinerelated inquiries appears in connection with the issue of mood effects as well. For, in contrast with Cherek’s work, the 1987 study by Roache and Griffiths mentioned earlier, in which subjects consumed between 200 and 600 mg of caffeine, produced small increases in scores on the Profile of Mood States Questionnaire (POMS) for hostility and anger, while also demonstrating an increase in friendliness.26
J.A.Brillat-Savarin, perhaps the most celebrated chef in history, wrote, in The Physiology of Taste (1805), “It is beyond doubt that coffee greatly excites the cerebral powers; also any man who drinks it for the first time is bound to be kept from a share of his natural sleep.” In fact, everywhere caffeine-containing beverages have been consumed, people have recognized that, just as they can help you to stay awake, they can also interfere with a good night’s sleep. The most common sleep disturbance associated with caffeine is insomnia, although there is, oddly enough, a complaint called “hypersomnia,” or too much sleep, also sometimes consequent to its use. Among researchers, it is generally accepted that caffeine is a common cause of sleep disturbances. In fact, according to P.B.Dews, a leading caffeine researcher, the disturbance of nighttime sleep is much more pronounced and dose dependent than caffeine’s daytime effects.27
Despite great variation in the amount of sleep that people need, ranging from as few as three to as many as twelve hours, with an average of seven to nine, certain general observations are possible. For example, scientists divide sleep into dreamless sleep and the sleep during which we dream. REM is an acronym for “rapid eye movement,” and dreaming sleep is called REM sleep because during our dreams we follow the action, as it were, by shifting our “gaze” back and forth. Non-REM sleep usually lasts about an hour or more and then shifts to REM sleep for about thirty minutes, in a pattern called the “ultradian rhythm” that is repeated four to six times throughout the night. Non-REM sleep, which itself comprises four states of progressively deeper dreamless sleep, constitutes about 75 percent of our sleep, and is characterized by lowered heart and respiration rates, while REM sleep is marked by deep muscular relaxation accompanied by increased or irregular heart and respiration rates and vivid dreaming.28 At the beginning of the night’s sleep, deeper non-REM sleep is a greater proportion of the ultradian cycle than toward the end of the night’s sleep, during which the proportion of lighter REM sleep increases. Interrupting or interfering with either REM or non-REM sleep causes a sleep deficit that the body seeks to restore over succeeding nights. A mounting sleep deficit can impair concentration, diminish energy and performance during the day, and increase anxiety and depression. Extreme sleep deprivation can cause paranoia and hallucinations. One in three adults regularly suffers from sleep problems, and medications are resorted to as a remedy by millions.
Caffeine’s effects on sleep depend on a variety of factors, such as dosage, tolerance to caffeine, individual sensitivity to caffeine, the time between caffeine ingestion and the attempt to sleep, and the ingestion of other psychoactive substances. There are studies confirming the common experience that acute doses of caffeine in the evening delay falling asleep and result in poorer sleep quality. For example, a Japanese study showed that it takes four times longer than normal to get to sleep after drinking a strong cup of coffee. But caffeine can do more than interfere with falling asleep: It can also produce alterations in the onset of REM sleep, total sleep time, and certain characteristics of non-REM sleep, such as shortening the deeper phases of nonREM sleep and lengthening the lighter phases. It does not appear that caffeine affects the length of the REM phase of sleep.29 Heavy caffeine users toss and turn more in bed, perhaps because caffeine increases muscle tension and restlessness. Such movements can also cause frequent awakening.30 In addition, people who consume caffeine before bedtime are more easily awakened by sudden noises. Brain-wave studies show that caffeine disturbs sleep during the first three to four hours. A study of subjects over age fifty found sleep was reduced by as much as two hours when caffeine was taken in the evening. Because older people sleep less than younger people anyway, this diminution represents a proportionately greater loss than it would for younger people.31
Just as people vary in the amount of sleep they need, they vary in the effect caffeine has on their sleep. You have almost certainly met people who claim to be able to sleep well after drinking a couple strong cups of coffee immediately before retiring. Even stranger is the fact that some people are not only capable of sleeping well after consuming caffeine but actually sleep too much, experiencing a condition of pathologic sleepiness called “hypersomnia,” as a result of consuming it. As “Pathologic Sleepiness Induced by Caffeine,” a paper published by Quentin R.Regestein in 1989 in the American Journal of Medicine, states:
The aforementioned patients had severe sleepiness that decreased or remitted after they discontinued caffeine. In some individuals, therefore, heavy use of caffeine apparently provokes sleepiness. This is difficult to explain since caffeine is a stimulant.... The unusual magnitude of the sleepiness and the rarity of this apparent association between caffeine and excessive sleepiness, even in sleep clinic patients, suggest an idiosyncratic phenomenon.32
Another strange effect, which might be called the “reverse placebo” effect, was observed by A.Goldstein in a 1964 study. Participants in his experiment were all given caffeine. Those who knew they had taken the drug were less likely to complain of wakefulness than those who were not informed whether they had taken caffeine or a placebo.33 Perhaps this could also be called the “bravado effect,” whereby people are reluctant to confess a disturbance from what is ordinarily considered a mild agent, such as caffeine. Surveys based on subjective responses clearly indicate that how much caffeine people say they use is not related to how much difficulty they say they have sleeping, and insomniacs do not report high caffeine use, defined as three or more cups of coffee a day. In any case, most studies confirm that the closer to bedtime you consume caffeine, the more likely it is to interfere with sleep. However, as we have observed in our discussion of metabolic variation, some people metabolize caffeine much more slowly than others, and their sleep may be disturbed even by caffeine consumed twelve hours or more earlier.
Other studies of delayed sleep onset and poorer sleep quality, as evaluated on objective criteria such as EEG measurements, confirmed the well-recognized large variation among subjects in terms of caffeine’s effect on sleep (intersubject variation) and also documented a similar large variation in the effect on the same subject on different nights (intrasubject variation). These studies have also demonstrated that sleep disturbances due to caffeine are more likely to occur in people who are not regular caffeine consumers and that the regular use of caffeine and a concomitant caffeine tolerance tends to diminish the disruptive effect of caffeine on sleep.
Overall, the leading research projects based on objective criteria demonstrate that caffeine intake near bedtime increases tossing and turning, reduces deep sleep and increases light sleep, has no effect on REM sleep, increases the time it takes to fall asleep up to threefold, decreases total sleep time by nearly two hours, and increases spontaneous awakenings. People who have not consumed caffeine before bedtime will fall back to sleep after being awakened early in the night more slowly than they will after being awakened later on. However, caffeine consumed shortly before bedtime reverses this pattern, creating the shortest delay in falling back to sleep in the first part of the night. Because the average plasma half-life for caffeine, a measure of how long it remains in the bloodstream, is between three and seven hours, a large enough dose would tend to sustain this effect throughout half the night. Although differing rates at which caffeine is metabolized by different people are generally thought to be the basis of the differences of effects among them, another school of thought attributes the variations in caffeine’s effect on sleep among individuals to differences in neural response sensitivity.34