No Way To Treat The Mind
Steven RoseThe idea of memory-boosting pills is appealing. But we should resist the
claim that there is automatically a chemical fix for all our psychological fallings.
Would you like to increase your mental energy, concentration and alertness? Perform better in school or at work? Improve your ability to solve problems? Who could fail to be tempted by these offers adorning the cover of Smart Drugs and Nutrients: How to Improve Your Memory and Increase Your Intelligence Using the Latest Discoveries in Neuroscience? And if you do answer yes, there are, so enthusiasts will tell you, some 140 or more different chemicals, food additives or drugs that will do the job for you.
But the question remains: do any of the substances advertised as so-called ``smart drugs'' work? Moreover, if the pharmaceutical industry does come up with an effective treatment for Alzheimer's disease and other senile dementias, will it be of any value to healthy people? Britain's Consumers' Association recently asked me, as a neuroscientist specializing in the biochemical basis of memory, to review the evidence.
It was not a cheering experience. In magazines, books and newsletters, smart drug enthusiasts cite an impressive string of scientific papers to support their claims. Using these and other papers reporting the results of experiments on prototype smart drugs, I examined well over 100 studies, some on animals, some on people with dementia, some on healthy people. Most of these are either misleadingly quoted by advocates of smart drugs or describe experiments that are poorly controlled or extravagantly overinterpreted by the researchers. Worse, some studies use procedures that would not pass ethical scrutiny in Britain, and the most dramatic claims often appear in unrefereed conference proceedings or obscure journals.
To assess the claims of the loudest advocates of smart drugs, it is vital to sort out what they mean by ``improving memory'' or ``enhancing cognition.'' Current psychological theory distinguishes between so-called ``procedural'' and ``declarative'' memory, or ``remembering how'' and ``remembering that''. Remembering how to ride a bicycle is procedural'; remembering that a two-wheeled vehicle with saddle and handlebar is called a bicycle is declarative. Declarative memory can be split into ``semantic'' and ``episodic'' memory. Remembering that 1 January is New Year's Day is semantic; remembering the New Year party you went to last year is episodic. The most vulnerable form of memory is episodic.
Only the most severe of memory disorders affect procedural memory; even profound amnesiacs can usually recall how to perform daily activities such as putting on clothes and cooking even when they cannot subsequently describe what they have done.
Much of what we perceive is stored only for a short time before being discarded. Read a seven-figure set of numbers to a colleague, and the odds are he or she will be able to repeat them back without error. Ask again half an hour later, and the sequence will probably have been forgotten. Such observations have led to the idea that declarative memory passes through a short-term store into a long-term store. Persisting for periods of minutes to hours, short-term memory is one of the most susceptible to drugs: anęsthesia, concussion or electrical shock treatment will all obliterate it. By contrast, once a memory is in the long-term store, it is very hard to erase, though not always easy to retrieve without an appropriate cue. It is often hard to recall a person's name, for instance, even though one ``knows that one knows it''. Memory is often not so much lost as hard to find.
This brings us to the first fundamental problem with smart drugs. Most people envision them as improving recall. Yet much of the evidence cited by smart drug enthusiasts comes from animal experiments which can, by virtue of their design, test only a drug's effects on learning, not recall. The animals are trained to perform highly specific tasks and the agents given within an hour or so of the training trials. On this basis, the best a smart drug could do would be to increase the likelihood of the transfer of information from the short-term to long-term memory store. There is little evidence than any of today's smart drugs can do even this much in humans.
Indeed, the awesome complexity of human memory makes the smart drug creed look suspiciously simple. It is founded on two key assumptions. First, memory loss is not confined to patients with conditions such as Alzheimer's disease but happens to us all as we age. Second, this normal ``everyday'' deterioration is caused by a failure or weakening of the same kinds of biochemical mechanisms that break down in conditions such as Alzheimer's disease. To believers in the faith, the corollary is obvious: drugs developed for patients with senile dementia ought to be made available to the rest of us so that we can crank our memories up to maximum speed and capacity.
The first of these assumptions is by no means confined to smart drug advocates. In the US, a vociferous school of thought claims there is a clinically definable condition called Age Associated Memory Impairment (AAMI) - an inexorable loss of memory which, the argument goes, many of us suffer as we move through our fifties. It is true that after the age of 40 performance on standard intelligence tests steadily declines for most people - unless enough time si given to solve the problems. And if speed is taken as the main criterion of memory capacity, then it could be argued that memory grows weaker with age. Yet memory and learning are inextricably bound up with a whole range of other mental and physical abilities. As we age our attention spans, emotions and motivation levels all change. Such effects alone could explain why memory seems to falter and age, without there being any weakening of the molecular and cellular mechanisms of memory itself.
Furthermore, decline is not inevitable. A recent Dutch study from the University of Limberg, Maastricht, has found that memory decline in otherwise healthy people as they age is associated with mild head injury, general anęsthesia or ``social drinking'' earlier in life. In people without such predisposing ``biological life events'', memory remained unimpaired even into extreme old age.
It is extremely difficult to design experiments that test the direct effects of drugs on memory when there are so many indirect influences on mental performance. Elderly people living in institutions who cannot remember what they had for breakfast that morning may be able to recall childhood episodes with great clarity. Have they lost memory, or have they lost interest in institutional food, where one breakfast may be just like another? Recall is not some mechanical process like recovering files from a computer disk, but involves active and interested work on the part of the individual.
Smart drugs are supposed to work in one of two main ways: either by increasing blood flow to the brain, or boosting the levels of one or other of the neurotransmitters thought to play a part in learning and memory. Such drugs are sometimes called ``nootropics'', a term coined by the pharmacologist Cornelius Giurgea in the 1970s from the Greek meaning acting on the mind. To qualify as a nootropic, Giurgea argued, a drug had to:
These criteria, though still quoted with zeal by smart drug enthusiasts today, seem absurdly vague by the standards of mainstream neuroscience. I am not sure, for instance, what is meant by ``facilitating information flow between the cerebral hemispheres'', or how to check that it was happening.
- Enhance learning and memory, especially under conditions of disturbed neural metabolism resulting from a lack of oxygen, electroshock or age-related changes
- Facilitate information flow between the cerebral hemispheres
- Enhance the general resistance of the brain to physical and chemical injuries
- Be devoid of any other psychological or physiological effects
The rationale to the belief that agents which affect blood flow or neurotransmitter levels could function as smart drugs came first from research into the effects of stroke. Stroke is one of the common causes of neurological damage in elderly people affecting many aspects of performance including memory and cognition. The damage is often cause by constriction of the arteries supplying blood, and hence oxygen and glucose, to the brain. Therefore, drugs which increase cerebral blood flow and diminish hypertension might be expected to ``improve the performance'' of otherwise energy-starved neurons. This is why propanolol, phenytoin or hydergine are used as smart drugs.
The second reasoning came from work on the neurochemical deficits in Alzheimer's disease. One of the characteristic features of Alzheimer's is progressive loss of memory. Postmortem studies of the brains of Alzheimer's patients show a dramatic destruction of neurons, and particularly neurons which secrete or utilize the neurotransmitter acetylcholine. During the 1970s, research with animals also suggested that acetylcholine might be a key neurotransmitter in memory formation; for instance drugs such as scopolamine, which deplete brain acetylcholine also impair memory. In the early 1980s these observations gave birth to the ``colinergic hypothesis'' of memory loss. And if loss of acetylcholine was responsible for memory loss, why not attempt to increase the brain's supply of acetylcholine? In the wake of this argument came one of the first proposed smart drugs, piracetam, along with food additives which are potential metabolic precursors of acetylcholine, such as choline and acetlcarnosine.
Despite the focus on the role of acetylcholine, nobody seriously believes that any one neurotransmitter dominates the complex processes that underpin cognition and memory. So the search for smart drugs has broadened to include agents that interact with other neurotransmitters, including glutamate, serotonin (5-HT) and dopamine.
Another class of potential smart drugs are the calcium channel blockers. These, such as verapamil, diltiazem, nifedipine, nitrendipine and nimodipine are not only widely used to treat hypertension, and hence might affect cognition by increasing cerebral blood flow, but also block the entry of calcium ions into neurons. A key stage in the molecular cascade of memory formation appears to be the opening of membrane channels through which calcium ions flow. In old laboratory rodents, the mechanisms that open and close these calcium channels seem to become defective, and so calcium accumulates at dangerous levels inside the cells. The calcium channel blockers might improve memory by countering this effect. Ginseng, that all-purpose elixir of life, contains an agent that blocks calcium channels.
There is some evidence that calcium channel blockers, glutamate agonists, 5-HT antagonists and certain piracetam derivatives can help laboratory animals learn highly specific tasks if the agent is injected around the time of training. Yet in many cases, it is clear [sic; not?] how the drugs act in animal tests. Some may work by interfering with the molecular events leading to memory; others by affecting the animal's general level of physical activity. Testing memory in non-humans presents nearly as many ambiguities as studying it in humans.
A popular approach is to measure how effective a drug is in helping rats and mice their way around mazes. In some of these mazes, the animals must run down a series of ``arms'' fanning out from a central area in search of food. Other mazes consist of a large tank of water with a submerged platform, which animals must learn to swim towards and clamber on to to ensure safety. A different kind of test relies on ``passive avoidance.'' If a rat is placed on a narrow shelf above the floor of its cage, its normal tendency is to step down; if the animal is given a mild electric shock every time it does so, however, it will learn to stay put.
Clearly, such tasks involve many other aspects of performance than merely learning and memory. Drugs that alter sensations of hunger or thirst, or reduce sensitivity to pain, or simply make an animal more active could all improve performance. In some cases animals are trained on tasks and then made amnesiac either by electric shock, or, in one popular test, by administering scopolamine to deplete acetylcholine. Something may be considered a potential smart drug if it can alleviate this kind of amnesia in animals. But there is an obvious flaw: the drug may not necessarily work on less artificially induced amnesia.
There are many other problems with animal studies of smart drugs. The drugs are usually administered by injection, sometimes into the brain, and often at very high doses. In some studies reporting a positive effect on learning, drugs were injected at doses equivalent to inject a human with about 8 grams of the drug. Taken orally, as all smart drugs are, this might scale up to as much as 40 grams - scarcely a realistic proposition. Although it is standard practise in reputable pharmacological papers to publish dose-response curves for any proposed agent, the smart drug literature is remarkably coy on this point. At best two or three concentrations of the drug are tested, often without discernible dose-related effects.
Studies of smart drugs on human patients are much rarer but no less contentious. Most involve patients in hospitals or other institutions, frequently diagnosed as suffering from Alzheimer's or cerebrovascular disease. Many of the trials are based on small samples of patients, less than 10 in each of the drug and placebo groups. A sizeable number of the human studies come from trials carried out in the south of Italy where the ethical controls and statistical procedures require in British and American trials are often lacking, or at least unspecified. What is more, effects on the patients are often evaluated using subjective criteria. One paper, reporting a study of aniracetam in elderly patients, simply states that doctors and nurses evaluated the drug according to its ``resocialising and revitalizing effects''.
The better research papers supplement such observations with IQ and memory tests, in which subjects try to recall or recognize names or dates associated with past public episodes. Yet often researchers find a drug improves performance in only one or two of a battery of tests. The proper way to avoid error is to then repeat the trial with a different group of patients using only the tests which were found to be significant the first time. I have seen no such reports.
But even if an experimental drug does help patients with Alzheimer's disease, it may not necessarily be of value to healthy people. In addition to suffering memory loss, elderly patients in hospital, especially those with Alzheimer's disease, are often angry, suspicious, anxious or depressed. A drug that reduced such feelings could easily result in an apparent improvement in memory while having no effect on the actual mechanisms of memory.
Clinical trials have yielded some drugs which may be of help in the early stages of Alzheimer's disease. But the claims made by smart drug advocates go far beyond offering treatments for devastating diseases; they offer instant cognitive enhancement to us all. True, at best the research papers on which the claims are based imply that for most of the smart drugs to work, a person should self-administer a massive dose, preferably by intravenous injection, either just before or after learning a particular problem or revising for an examination. Not the most likely of scenarios.
But, of course, in addition to claims based on research papers, advocates of smart drugs, such as Dean Morgenthaler, offer personal testimonials as evidence. I am not trying to challenge individual experience and testimony. It is well known that the experience of consuming psychotropic agents varies between individuals, is intensely subjective, depends on mood and expectation, and is often hard to verify by objective scientific criteria. What is abundantly clear, however, is that the primary scientific literature does not justify the claim that smart drugs can be of any therapeutic or ``memory-boosting'' value to healthy humans.
Nor should this surprise us. There is no reason to assume that, for most of us at most time, our enzyme and neurotransmitter systems are not working at more or less optimal levels. The brain is well buffered against the effect of arbitrary increases or decreases in circulating chemicals so that simply consuming food additives which are acetylcholine precursors will not normally increase your brain acetylcholine level. And even if it did, increasing neurotransmitter activity is no guarantee of increased mental performance; rather, it can be positively deleterious to throw chemical spanners into the exquisitely balanced biochemical system that is the human brain. More does not mean better. The smart thing to do with smart drugs is to resist the claim that there is automatically a chemical fix to any or all social or psychological problems. For most of us who believe that we suffer from a weaker memory than we would like, the old training skills offered by the mnemotechnics of the ancients probably stand at least as good a chance of helping us as does the most fashionable molecule-of-the-moment.
* * *
Steven Rose is head of the Brain and Behavior Research Group at the Open University.
New smart drugs
Smart and happy?
Viagra for the brain?
Memory like a goldfish?
Growing new brain cells
Does it hurt to be smart?
Exercise and neurogenesis
Smart drugs and aging brains
Smart drugs and mind control
Mood and cognitive performance
Antidepressants and new brain cells
Sleep boosts ability to learn language