Learning to peck

Astronomers have penetrated billions of light-years into space, explained the changing states of stars from their birth to their death, postulated the existence of black holes in which matter disappears, and rightly or wrongly, pinpointed the origins of the universe to a moment in time. The work of neuroscientists demands the same kind of ingenious speculation and theorising and rests on the invention of just as many subtle techniques. For several reasons, however, most people are less interested in what goes on inside their own heads than they are in events in the recesses of space. First, brain biochemists work with tiny structures, many billion times smaller than those with which astronomers deal: there is perhaps something attractively majestic about the distances and time scales over which astronomy ranges. Secondly, most of the conclusions reached by neuroscientists are less secure than those of astronomers. Finally, it is usually impossible to connect findings on the brain with its actual function.

In The Making of Memory Steven Rose recounts his search for the physical changes that occur when anything is committed to memory – a quest that has dominated neuroscience for twenty years. The techniques used to pursue it are remarkably ingenious. In the PET scan, a radioactive isotope of a substance taken up by active nerve cells is injected into an animal’s blood – radioactive glucose is commonly used. The animal then performs some task and by subsequently measuring the distribution of the gamma-rays emitted, the areas of the brain which were most active during the task can be determined. Or a centrifuge can be used to spin material from the brain. Heavier cells move further out than light ones, making it possible to analyse different types of cell independently of each other. Similarly, a mixture of proteins, each with a different weight and charge, can be separated from one another by placing them in a jelly-like substance and passing a small electric current through it (electrophoresis). The electron microscope – as important for the neuroscientist as the radio microscope is for the astronomer – can resolve the details on the membrane of a nerve cell, while the firing of individual nerve cells can be recorded with tiny electrodes. Minute areas of the brain can be burnt out, by passing a current between two small electrodes inserted into it (or by using a laser). And chemicals can be injected into the blood to prevent a given process occurring.

Armed with techniques like these, Steven Rose had to choose an animal and a particular kind of memory. He settled for new-born chicks, who peck naturally at small round objects. If a coloured bead is dipped in a bitter solution, the chick rejects the bead as soon as it has pecked it, shaking its head in disgust. More important, after one such experience, it will never peck a bead of the same colour again. Here is learning with a vengeance. What changes in the chick’s brain mediate it?

In the first of a series of experiments, Rose injected a modified form of radioactive glucose into the chick’s blood. The modification ensured that although active nerve cells would take up the glucose, they would not convert it into other biochemical compounds: it remained, therefore, in the active cells. The chick was then trained to avoid a coloured bead and after an interval it was killed (or ‘sacrificed’ in the neuroscientists’ euphemism). The brain was then cut into slices, which were examined for radioactivity, and two areas were found to be highly radioactive. Although suggestive, this does not prove that the change underlying learning occurred in these areas. The nerve cells might have been active simply because of the bitter taste or the visual stimulus provided by the bead.

It has been known for some time that one nerve cell fires another by releasing a neurotransmitter that binds onto one of a number of receptors on the next cell. It is also widely accepted that the physical change underlying learning occurs at the synapse – that is to say, the junction between the two cells. The change may take the form either of an increase in the amount of neurotransmitter released by the first cell or an increase in the number of receptors on the second cell, which enables it to take up more neurotransmitter molecules. Rose found that in chicks which had learned to avoid a coloured bead, there was an increase in the type of protein needed to make new receptor cells in the areas already identified as being connected with learning, and when he and his colleagues looked through a microscope they found that there was indeed a large increase in one type of receptor site (a spine) in these two areas. This structural change, which they measured 24 hours after the chicks’ training, could well be the basis of memory. Rose went on to confirm this hypothesis in other ways. For example, if an electric shock is given to the chick’s head immediately after it pecks a coloured bead, not only does it subsequently show no sign that it has learned anything about the bead, but there is no increase in the number of receptor sites – which suggests that in the chicks which learned to avoid the coloured beads the increase is connected with memory. Moreover, if the shock is given 24 hours after the training, memory remains intact, as does the increase in receptor sites. Surgical removal of the two relevant areas abolishes the learning, but has no effect on the rest of the chick’s behaviour.

In fact, Rose went much further than this, attempting to trace a whole sequence of biochemical reactions involving the activation of a particular gene that initiates a chain of processes and eventually leads to an increase in the tendency of one cell to fire another. His description of his own work can be read as a detective story. He follows one clue after another, goes down blind alleys, and is assiduous in his efforts to eliminate wrong possibilities (false suspects). He also writes with considerable clarity, though the reader will have to work hard at times in order not to become lost. More seriously, however, one may doubt the meaning and ultimate value of the work. A chick newly dried out from the egg must surely have some in-built mechanism to enable it to stop pecking at indigestible objects. Indeed, at that age it presumably learns very little else except to follow the mother bird. The situation is very different from that confronting a schoolboy who has to learn the dates of kings – an unnatural task if ever there was one. The changes Rose found after learning were massive: there was a 60 per cent increase in one type of receptor site in the relevant brain areas. If this were characteristic of all learning, nerve cells would be so strongly connected that they would be likely to blast off in almost any circumstance and no learning could be revealed. Rose might reply that the newly hatched chick is a particularly good animal on which to study learning precisely because the change in brain state is so large. Then again one has to remember that, despite the similarities in the structure of synapses in birds and mammals, it is dangerous to generalise from one group to the other.

More importantly, one might wonder how much our understanding of memory would be advanced if we knew the full details of the physical change underlying it. Knowing that the storage elements in computers are composed of silicon chips which can be in one of two states will not help you to understand how a computer can play chess at grandmaster level. Indeed the same program could be instantiated in computers that use lasers or fibre optics. To understand memory we need to know the whole circuit in which the memory is embodied, not what and where the physical change is that underlies it. Without such knowledge we cannot begin to answer questions about the way memories are retrieved or how they become distorted by later memories. This is one of several psychological issues over which Rose goes astray. He distinguishes procedural memory (how to ride a bicycle) from declarative memory (remembering that something is the case), and states correctly that procedural memory is not usually forgotten as readily as declarative. But it is likely that these are not two distinct kinds of memory: if you are a tennis-player and then learn squash it will almost certainly interfere with your tennis-playing. The reason we do not forget how to cycle is that, having learned this skill, we do not subsequently engage in similar activities that would interfere with our memory for it. The way in which old memories are distorted by – and to some extent protected from – new ones is an important problem. Except possibly in very young chicks, new memories are always laid down on top of existing ones.

Since his account is highly personal, Rose does not devote much attention to the numerous other research workers engaged in the same search, almost all of whom have their own theories. It is a field that has produced considerable dispute and rancour, with one eminent scientist having to withdraw an interpretation of his own data that he had held for many years in defiance of the opinions of others.

The real successes of brain biochemistry have lain in other areas. We now have a detailed knowledge of how an electrical impulse is propagated down the part of a nerve cell (the axon) that conveys a message to the next cell. We also have a reasonably complete account of how one nerve cell excites or inhibits (that is, prevents the firing of) another. In both cases, the processes are of such enormous complexity that one wonders why they so rarely fail and how they can have evolved. The message is transmitted across the synapse by about forty different neurotransmitters – chemical substances released by the cell that affect the next cell. Different cells release different neurotransmitters and different neurotransmitters predominate in different parts of the brain. It has been established that mental illness can result from too much or too little activity of particular transmitters. Thus, schizophrenia is thought to be caused by too much dopamine activity and can be alleviated, though not cured, by giving drugs that reduce it. Reduced levels of the same substance produce Parkinsonism, which can be ameliorated by drugs that increase dopamine activity. Again, depression is thought to be caused by too little activity in other transmitters, especially serotonin and epinephrine. It can be relieved by drugs (for example, the tricyclics that increase the activity of these two neurotransmitters). Although many details remain to be worked out, all these findings are exciting, comparatively new, reasonably secure and of practical value.

Despite the brain biochemists’ successes, there remains a huge gap in our knowledge. We haven’t the remotest idea why overactivity in the dopamine system should produce the observed symptoms of schizophrenia, such as hearing imaginary voices, believing that one’s thoughts are controlled by outside agencies, or paranoia in all its forms. Nor do we know how fluctuation in the activity of serotonin gives rise to depression and mania, and probably even to normal changes of mood. We know the sites of action in the brain on which opiates operate, but once again there is a huge hiatus between brain and mind. Why do they reduce pain? The gap between the functioning of neurotransmitter systems and the mental illnesses and thought processes and feelings to which they give rise is as wide as ever and nobody has any inkling of how to reduce it.

But Rose’s book is not merely an account of brain biochemistry. He describes how his ideas came to him and, more generally, what it is like to be a scientist. He is particularly good on that curious institution, the scientific conference, and claims that he attends about six a year. He describes registering at the conference hotel, being handed a plastic briefcase marked with the name of a drug company, and throwing away most of its contents apart from the invitations to functions. As to the subsequent proceedings, he rightly says that the most important part of the meeting takes place in the bar, where scientists are less likely to try to cover up their mistakes or hold back ideas for fear that someone else might capitalise on them.

He gives a good account of writing scientific papers, which must stick to a strict format and are designed to show that the omniscient writer had foreseen his results from the outset. He also correctly calls attention to the massive amount of time wasted in preparing grant proposals and the equal waste of time by those refereeing them. Since, in these degenerate days, a scientist’s standing is judged largely by the number of papers he or she produces, everyone, including the administrator, is anxious to have his name included among the authors of an article. In fact, most papers are never referred to by any other scientist: it would be interesting to know what proportion are even read.