by Aleksei IVANITSKY, Corresponding Member of Russian Academy of Sciences, Head, Laboratory of Higher Nervous Activity of Man, Institute of Higher Nervous Activity and Neurophysiology, Russian Academy of Sciences

How do brain and consciousness correlate? How is a subjective world of a human being formed on the basis of nerve pulses? For a long time this problem was considered to be the domain of philosophy and other humanitarian disciplines. Only in recent decades the natural scientific approaches to its solution were discovered.

MECHANISMS OF SENSATIONS

Consciousness is the priority function of the brain*. In fact, it is our life, consisting of alternating impressions, thoughts, and sensations. But how rightful is it to explain what we perceive as color, sound, emotion by the motion of nerve pulses? Despite seeming complexity, this problem of nature is not unique in its methodological difficulty and is one in a series of other mysteries of the universe. Two main approaches to search for cerebral mechanisms of consciousness, not ruling out, but supplementing each other, can be distinguished. One is the concept according to which subjective experience emerges on the basis of forward propagation of stimulation from primary zones of the cerebral hemispheric cortex to higher structures, one of which, and a most important one, is the frontal cortex. It has three unique characteristics: capacity for operations with abstract symbols, memorization of the time sequence of real events, and presence of speech centers. These three qualities directly correlate with the signs of consciousness.

The other approach is based on a hypothesis that subjective experience is a result of certain organization of cerebral processes and comparison of new information with the information extracted from memory, in the cortical zones. Due to this, the picture of external events is as if projected to the subject's individual experience, incorporating in the personal context. This hypothesis is shared by many specialists at present. We put it forward for the first time in the 1970s as a result of investigations of cerebral mechanisms of sensations.

In this work we compared the quantitative parameters describing the response to a signal. The volunteers participating in the experiment solved the problem of discriminating between the intensities of two stimulants of similar force (visual in one series and cutaneous in the other). During the experiment we fixed brain activity in the form of the so-called evoked potentials (EP), in other words, electrical reaction to a new signal coming from the sensory organs. These potentials are fluctuations consisting of a series of successive components, complex in shape. It was important to understand which information processes in the brain they reflected. Previous studies found that the early EP components reflected mainly the physical parameters of the stimulus, while the later ones reflected its significance. Quantitative parameters of sensations were evaluated by methods of signal detection theory, regarding perception as a result of interaction between sensory and motivation factors.

* See: A. Ivanitsky, "More About Brain", Science in Russia, No. 3, 2000; A. Ivanitsky, A. Nikolayev, "In Order to Understand Oneself, Science in Russia, No. 5, 1999. - Ed.

стр. 25

The "circle of sensations". The synthesis of information about the qualities of a stimulus on visual cortex neurons leads to emergence of a sensation, which is later recognized with participation of the frontal cortex.

After obtaining appropriate data, we calculated the ratios and correlations between them. The interdependence between intermediate waves of evoked potentials and both perception factors (sensory sensitivity indicator and decision factor) proved to be the most significant. This double correlation reflected the synthesis of information about the physical and signal characteristics of the stimulus on neurons of the so-called projection cortex, receiving signals from sensory organs. The peak latency (time from the moment of stimulation till development of response) of the evoked potential waves was about 150 msec.

It is principally important that this time interval rather accurately coincided with the velocity of sensation emergence, first measured in the 1920s - 1930s in psychophysiological experiments using the "reverse masking" phenomenon. Its gist is as follows: the first weak stimulus is not perceived if it is followed, after a short period, by the second, stronger stimulus. By gradually prolonging the pause between the weak and strong (masking) signals, it is possible to find an interval at which the masking effect disappears, as the sensation in response to the first signal has already been formed. It was found that a sensation manifests itself approximately 150 msec after the stimulus. We obtained the most reliable data (by the way, close to the above) at the beginning of the 1990s by using direct stimulation of the cortex with short magnetic pulse as the masking signal. Importantly, the masking effect developed only after application of magnetic pulse to the projection (in this case visual) cortex, i.e. only at the site of the above-mentioned double correlation of evoked potential waves with the perception indicators. All these data proved that a sensation significantly lagged behind the entry of sensory pulses into the cortex, which takes just about 30 msec. Hence, a sensation is a result of intricate organization of nerve processes, which we investigated.

Based on the data on the physiological origin of evoked potential waves, we described the mechanism providing information synthesis. It included annular motion of stimulation from the projection to associative cortex (temporal for visual stimuli), then to the hippocampal* area and motivation centers of the diencephalon with subsequent return to the cortex. We named this cycle the "circle of sensations". Its gist is that it helps us compare the sensory signal to the information extracted from memory, this process presumably underlying the transition of a physiological process to the level of mental subjective experience. As a result, the emerging sensation not only accurately depicts the physical characteristics of the stimulus, but also is emotionally colored. This concept was called the information synthesis hypothesis. Later it was confirmed in many studies. Similar assumptions were formulated by the American scientist Gerald Edelman (Nobel Prize winner of 1972 for description of antibody structure) independently of our hypothesis; he developed the neurobiological theory of consciousness, with the underlying "repeated entry" idea.

In addition to information synthesis, stimulation return to the cortex also provides integration of individual signs of the stimulus into a single image. Electroencephalogram (EEG) gamma-rhythm with a frequency of about 40 Hz plays an important role in this latter process. Synchronization of the brain biopotentials at a certain rhythm promotes unification of neuron nets in a single system, which is essential for maintaining consciousness.

Sensation is a rather simple mental phenomenon; some authors refer it to the so-called "primary consciousness". Emotions can be referred to the same class; Pavel Simonov** (1926 - 2002), Academy Member, made an important contribution to their investigation. He was the first to suggest an emotion formula, according to which its force is proportional to the requirement multiplied by difference between the data which the individual had and the information needed for satisfying this requirement. Simonov's formula means that, like sensations, emotions emerge as a result of comparison of two information flows. Hence, a certain universal regularity works here.

* Hippocampus is a brain structure in the hemispheric temporal lobe base; is a component of the limbic system; is involved in emotional reactions and memory mechanisms. - Ed.

** See: P. Simonov, "'Ego' and the Brain", Science in Russia. Nos. 5 - 6, 1992. - Ed.

стр. 26

Cortical regions are shown on brain maps in different colors, depending on the standard number of their bonds. The foci are located in the parieto-temporal region in case of descriptive thinking and in the frontal cortex in verbal thinking. Speech perception center in the left temporal cortex is involved in both cases.

THINKING

More intricate mental phenomena, primarily those associated with the emergence of speech, are referred to higher consciousness. Academy Member Natalya Bekhtereva and her disciples made an important contribution to studies of such brain functions as thinking; the studies were begun in the 1960s and are in progress up to the present time.

Modern data indicate that the cortex is highly specialized and its different fields are responsible for different cognitive operations. Therefore, cortical bonds acquire special significance in the thinking process, and coordination of neuronal ensemble working rhythms promotes the development of these bonds. This assumption was put forward by the Russian neurophysiological school. In the 1930s Academy Member Aleksei Ukhtomsky suggested that nerve connection is facilitated when nerve ensembles work on the same frequency. Later (1950 - 1980) Academy Member Mikhail Livanov and Member of Academy of Medical Sciences Vladimir Rusinov showed that synchronization of EEG rhythms can be regarded as a condition for and indicator of cortical bonds development.

We developed a new method for their mapping on the basis of these ideas. Tasks for descriptive, spatial, and abstract verbal thinking were used in the study. Multichannel recording of EEG was carried out during the period between presentation of the task on the monitor display and solution finding. It turned out that the pattern of bonds, symmetrical at rest, was changing during search for the solution: the bonds started to meet at certain cortical fields, forming bond centers, which were called interaction foci. Their topography was different for different types of thinking. In descriptive thinking (recognition of emotions at photographs of faces) the foci were located in the parieto-temporal cortex, in abstract verbal thinking (solution of anagrams or word categorization) in the frontal cortex, and during spatial thinking, including the elements of both descriptive and abstract thinking, they were located in the parietal and frontal cortex.

It was also found that information was delivered to the foci through flexible bonds maintained on different frequencies, each carrying information from specialized cortical compartments. This information is synthesized in the focus where neuron groups are united by rigid bonds, as a result of which the solution is presumably attained. Thus the idea of information synthesis with description of the mechanism of sensations was also transferred to thinking. It is noteworthy that the organization of nervous processes concomitant with them differ: as distinct from two information flows in sensation, there is a much greater number thereof in thinking. They include signals from sensory organs, from operative and long-term memory, and (which is very important!) from motivation centers. Another difference is that during thinking the synthesis centers are located in the associative, and not in the projection cortex, as during emergence of sensations. Interestingly, during the last stage of solution of all types of tasks the foci emerged in

стр. 27

Cortical bonds (colored) in two subranges of beta-rhythm frequencies during solution of a spatial and verbal task. In one case the volunteer was to determine whether two figures shown to him were the same or mirror-symmetrical, in the other he was to find a word belonging to a different sense category than three other words.

the speech left temporal region. Hence, verbalization is an important component of thinking. In this sense we can speak about convergence of two approaches to solution of the "consciousness and brain" problem: the mental emerges on the basis of certain organization of nerve processes, involving some key regions of the frontal cortex.

Important role of relationships between the frontal and left temporal cortex in verbal thinking and understanding of word combinations forming a sentence was detected in our joint research with the Institute of Cognitive and Decision Sciences, University of Oregon, headed by Michael Posner, Member of the U.S. National Academy of Sciences. Our task in this work was to study the picture of cortical bonds. The role of the frontal cortex in deciphering word semantics was shown using positron emission tomography and evoked potentials of the brain. These data appreciably supplemented the classical notions that the meaning of a word was identified in the left temporal cortex (Wernicke zone). On the other hand, it was proven that the relationships between the frontal cortex and temporal zone are involved in connection of several words into a single meaning construction.

DECLARATIVE MEMORY AND ATTENTION

An important feature of consciousness is the ability to memorize and reproduce the sequence of events. Significant progress was attained in the study of human memory mechanisms in recent years. Russian and American neurophysiologists showed in their studies the role of the hippocampal structures, located inside the hemispheric temporal compartments, in operative memory. These formations have extensive bonds both between each other and with sensory and associative compartments of the cortex. It is assumed that during memorization they direct a signal to the associative cortex for long storage, while in reproduction they give an address where the information related to this signal is stored.

Injury to the hippocampus leads to impairment of declarative memory (realized memory about events, information about which can be given to other persons). Patients with this disease can study well at school and have a high intelligence quotient, but they are helpless in everyday life, as they do not remember the sequence of events, do not orient in time, cannot make a plan for the future. This defect manifests itself only from the age of 5 - 6 years, when a normal human starts remembering oneself.

The frontal cortex participates in memorization of a sequence of events together with the hippocampus. Groups of neurons, capable of retaining a trace of signal till the moment when a behavioral response to this signal is needed, are located in the frontal cortex.

Consciousness is closely related to attention: we realize only what we pay attention to. We showed that the role of memory is important for the mechanisms of selective verbal attention, when a human has to perceive and react to just a certain class of signals, singling them out of numerous other signals. For example, one has to keep up a dialogue, singling out the speech of one's interlocutor from other talks, visual sensations, and images. A record of evoked potentials of the brain in response to words simultaneously presented on the monitor display and heard from computer acoustic systems was used in the study. The task of the volunteer taking part in the study was to memorize as many words as possible from one of the channels and to ignore the other channels. Memorization and extraction of verbal information is electrophysiologically expressed as "cognitive" components of an evoked potential at a latency of 400 - 700 msec. An evoked potential to an ignored signal was characterized by a shift in the potential, opposite in polarity to the one observed in memorization, which indi-

стр. 28

Identification of the type of realized thinking by EEC in real time. Verbal logical and spatial problems were presented in a random sequence. The mean duration of solution was about 10 sec.

cated active inhibition of memory processes. We conclude from this that the attention is selective due to the fact that unnecessary information, though perceived (EP components responsible for this process are retained, the individual hears and sees the word), is then blocked and superseded from consciousness.

Let us sum up the notions of the probable mechanisms of consciousness. The basic principle is return of stimulation to sites of primary projections, providing the information synthesis; the frontal cortex plays an important role in the formation of abstract notions and speech; the mediobasal compartments of the hemispheric temporal region are important for maintenance of declarative memory and selective attention processes. Comparison of new information to previous experience determines the content of consciousness as permanent correction of personal experience and what can be called the sense of the inner "ego". Hence, the idea of renewal, rendering the higher sense to life and determining the constant striving of man to novel experience, underlies consciousness.

THE PROBLEM OF ARTIFICIAL INTELLECT

To what measure the knowledge of the mechanisms of thinking and consciousness of man can be needed in creation of an artificial intellect? Its development is possible without utilization of the brain functioning principles, as, for example, the design of an automobile does not imitate the natural methods of human motion, such as walking or running. However, this accumulated knowledge can be useful for one reason. The brain is superior to man-made computing devices by many important parameters: it is more reliable, economical, and easier trained. There are many important differences between the brain and computer. The brain has no central processor operating clear-cut signals and processing them according to preset programs. By contrast, it receives, as a rule, poorly defined signals; their evaluation largely depends on the context, and the brain creates programs by itself as a result of training. Along with rigid bonds, the brain utilizes flexible ones, formed on the basis of synchronization of neuron ensemble activity rhythms, due to which it effectively conducts a search for required information, which underlies the associative memory. This search includes elements of heuristics, unexpected, but useful solutions.

The principle of information return to the site of primary projections is a find of evolution: new data are compared to those stored in memory. The information synthesis mechanism is best of all developed in thinking. As we mentioned above, centers of bonds emerge in the cortex during this process; these are foci of interaction, receiving information from other compartments and synthesizing it. This assumption indicates an important method, not used in the majority of artificial neuron networks: the cerebral neuron nets are not homogeneous, but are constructed according to the hierarchical principle.

Hence, higher functions of the brain emerge on the basis of high degree of integration of specialized components into a single system, representing a higher level of organization and therefore possessing different logic and regularities of self-development. These laws can, in the order of descending determination, modulate the processes in individual components of the systems. For example, speech logic and grammar can regulate the motion of nerve pulses responsible for articulation. The absence of this descending determination from the general to particular is an essential feature distinguishing the computer from living brain. In fact, only the program writer, but not the central processor, can see the total task.

The main difference between brain and computer is, according to the well-known British physicist and mathe-

стр. 29

matician Roger Penrose, that the brain "understands" and the computer cannot understand. Winy an animal, for example, a cat, understands, while a computer, a perfect creation of high intellect, cannot understand? It seems that understanding is a result of support, that is, on the basis of evolutional development of such a basic principle of brain work as conditioned reflex, connecting the outer stimulus and response actions of a subject to satisfaction of a certain requirement. Understanding makes high vital sense. Under natural conditions an animal learns to produce certain actions to meet this or that requirement, in other words, it starts to understand the relationship between external events, its own behavior, and achievement of the desired result. Training of animals is based on this principle: in order to train a dog to perform (understand) certain commands, the trainer uses support (food or punishment). In fact, all this can be also referred to man; virtually all human behavior is based on the same philosophy. For example, a good worker is paid a higher salary, while a bad worker pays fines, a hero is awarded, while a villain is put into prison. Let us note that the market-oriented economy, based on the same principle, proved to be highly effective.

Penrose based his viewpoint on Godel's* theorem on the impossibility to prove by calculations the correctness of the basic arithmetic operations, for example, 1 + 1 = 2. However, any living creature is persuaded of its correctness when he/she gets, for example, two bananas by adding, as a result of this or that action, the second object to the first one. It is noteworthy that the understanding of the essence of doubling (or addition in general) emerges in evolution before the knowledge of counting. A case is described with an indigenous resident of the North: he did not know how many deer he had, but could easily enumerate them by signs. A child can easily recollect all surrounding people or toys, though he/she cannot count yet.

The support shows whether the behavior is correct or erroneous. Both the information constituent proper (calculations) and the indicator of its correctness enter via different channels and are to a certain degree orthogonal. Two components of "understanding" can be represented by a sensory signal and pulses from motivation centers, which, as we said above, are present in transformation of a purely physiological process into mental. Pulses from motivation centers initially directly signal about satisfaction of a requirement. Later they can also be included in the computation system evaluating the efficiency of movement to the purpose and producing a command when the probability of attaining a useful result, allowable in this particular situation, is achieved. The motivation component acts here as an axiom and concludes the estimations. The axiomatic nature of support is based on the sign evaluation of usefulness or harm, understood as not requiring proofs. The brain has to act by the try-and-err method and learn from errors, as the actual conditions of life are so intricate that they can in fact never be estimated to the end.

Hence, the brain largely does not work as a computer. Therefore, the task of organic connection of the brain to a computer, creating an interface between them, usually denoted as Brain-Computer Interface (BCI), is so important. Two approaches to this problem are outlined. One is use of direct signals of the brain for controlling the external devices. For examples, systems for producing a text on the monitor display by using electrical signals of the brain were created.

Igor Shevelyov, Academy Member, works in this direction with good results. In his experiments a volunteer was shown sets of letters, and the brain responded by an amplified wave of evoked potential to the needed letter. The prospects are wheelchair control by disabled patients and later solution of more complex tasks. Another trend of research is creation of systems which will make it possible to penetrate into the content of brain processes by EEG or other parameters. Georgi Ivanitsky, Cand. Sc. (Biol.), developed a technology for recognition of the type of mental operation by means of the EEG pattern. In these experiments the volunteers solved problems of two types: for spatial and verbal logical thinking. EEG was recorded during solving. Artificial neuronetwork (trainable software, as if imitating the nerve network activity) memorized all combinations of EEG spectra by their topography, frequency, and amplitude. It turned out that such combinations of spectral characteristics proved to be rather stable while solving certain types of problems. Therefore a trained neuronetwork could recognize the type of the problem, solved in the mind, in other examples. It is clear that this device could be useful in objective automated monitoring of, for example, activities of operators working with complex systems and of the solutions adopted by them.

It will not be an exaggeration to say that the scientist(s) who will connect brain to a computer, will teach an artificial mind to work together with the brain, enriching it by reliable memory and perfect calculations, will have a decisive advantage in the development of modern technologies, will determine the progress in science and technology in the 21st century, and will provide reliable defense.

We do not yet know many things about brain work, particularly what underlies its higher functions and human consciousness. However, the progress in this sphere is obvious. Recommendations which the science of the brain can give to creators of man-made intellect, are also optimistic. Both the brain and computer are situated in physical world and are governed by its laws. None of the above-listed principles of brain work is beyond this framework, and therefore any of them can be reproduced and perfected in man-made devices.

The Presidium of the Russian Academy of Sciences has awarded a Bekhterev Gold Medal (2007) to RAS Corresponding Member Alexei Ivanitsky for his works on the physiology of human perception, thinking, attention and consciousness.

Illustrations supplied by the author

* Kurt Godel (1906 - 1978) is a logician and mathematician. Born in Austro-Hungary, since 1940 he lived in the USA, - Ed.