Experiments on the living brain are indispensable, because cognitive functions involve the entire brain and not just a few isolated nerve cells. Any attempt to reproduce the brain with a computer model is doomed to failure, because models can only be as good as the information we put into them.

The brain is incredibly complex. It consists of 80 billion neurons, every one of which is connected through its cellular extensions to some 10,000 other neurons. Several kilometers of these cellular extensions are found in each cubic millimeter of cerebral cortex. All of this creates a giant network of neurons as well as various other types of cells which are not directly involved in information processing. These include matrix cells, immune cells, and cells that, along with the blood vessels, ensure that the neurons receive enough nutrients.

The extensive “circuitry” of this neural network is what allows us humans to perform amazing cognitive tasks – and it is an immense challenge for scientists. Even relatively simple processes such as cognition, attention and memory are so complicated that despite decades of intense research we have still not really solved the puzzle of how they really work. If we were able to understand these brain functions, we would be able to build a robot that could react to certain stimuli much like a human would. Although we are now able to program computers that are far better at playing chess than the human brain, a computer still has trouble doing something as “simple” as recognizing a knocked-over chess piece.

In spite of the complexity of the brain’s structures and processes, however, there are some basic underlying principles and patterns. This gives us hope that one day we will be able to understand complex cognitive processes. A two-fold approach is necessary to study brain functions. On the one hand we must study primitive brains to acquire a systematic understanding of the general principles of how the brain works. On the other hand, there must also be research on complex systems as well, because over the course of evolution these have developed newer working principles and because they alone are capable of higher cognitive functions.

The study of brain function can only contribute to our understanding of the mechanisms behind it if it allows us to draw detailed conclusions about the participating neural structures. In principle, there are two possible ways of studying the activity of individual brain regions:

First, the activity of individual neurons and networks can be measured with microelectrodes. This is done by inserting one or more electrodes the thickness of a hair (approx. 80 µm) into the brain. This is an invasive method, but it provides direct information. It causes no pain, because the brain itself has no pain receptors. In fact, such electrodes are sometimes used on human patients for diagnosis and therapy. The direct measurement of the activity of individual neurons is necessary, for example, in order to validate and interpret measurements made with non-invasive techniques such as functional MRI.

Second, imaging techniques such as functional magnetic resonance imaging (fMRI) allow us to view the entire brain at one time. This method is non-invasive, but it only gives us an indirect look at neural activity. With it, we visualize not the active nerve cells themselves, but the blood perfusion of individual brain regions. For this reason, the measurements have a low spatial resolution. Another disadvantage of such imaging techniques is that they show the active brain regions, but tell us nothing about the functions of signal transmission in the brain. In other words, they let us see which brain regions are active, for example when we are thinking, but tell us nothing about what tasks the neural network actually has to carry out. One of our studies using rhesus monkeys showed that the common assumption that the fMRI signal is equal to nerve cell activity can lead to grave misinterpretations of the data (Logothetis, N. K.: What we can do and what we cannot do with fMRI. Nature 453(7197), 869-878 (June, 2008). DOI: 10.1038/nature06976).

Both of the methods described above can also be combined. Our department, “Physiology of cognitive processes” is currently the world leader in the development of techniques simultaneously employing functional magnetic resonance imaging with electrophysiological recordings. The combination of imaging techniques with invasive studies makes it possible to gather reliable information about how the brain works, and it also helps improve imaging techniques for clinical diagnostics.