Alexandrov, Yu.L, Grechenko, T.N., Gavrilov, V.V., Gorkin, A.G., Shevchenko, D.G., Grinchenko, Yu.V., Aleksandrov, I.O., Maksimova, N.E., Bezdenezhnych, B.N. & Bodunov, M.V. 2000 Formation and realization of individual experience in humans and animals: A psychophysiological approach.
In: R.Miller, A.M.Ivanitsky & P.M.Balaban (eds.) Conceptual advances in brain research. . Vol. 2. Complex brain functions. Conceptual advances in Russian neuroscience, pp. 181-200. Harwood Academic Publishers.
10 Formation and Realization of Individual Experience in Humans and Animals: A Psychophysiological Approach
Yu. I. Alexandrov, T. N. Grechenko, V. V. Gavrilov, A. G. Gorkin, D. G. Shevchenko, Yu. V. Grinchenko, I. O. Aleksandrov, N. E. Maksimova, B. N. Bezdenezhnych and M. V. Bodunov
Laboratory of Neural Basis of Mind, Institute of Psychology, Russian Academy of Sciences, Moscow, Russia nyualex @psychol. ras. ru
A systemic methodological approach to psychophysiology is described. In the framework of this approach a wide range of experimental data are analyzed, including the results of neuronal recordings in vitro, and in awake normal and pathological animals, performing both complex instrumental and simple behavioural acts. Also included are data from experiments with human subjects in tasks involving categorization of words, skilled performance, participation in game activity in groups, and completion of psychodiagnostic questionnaires. On the basis of these analyses, qualitative and quantitative descriptions of the principles of formation and realization of individual experience are suggested within the framework of a unified methodology.
Discovering the principles of organization of behaviour, based on experience accumulated by an individual, and the laws governing the formation of such experience is a multidisciplinary task. This general problem poses the majority of the specific questions of psychology, neurosciences, developmental biology, and genetics. At the same time, the solution of the general problem can be based only on synthesis of the achievements of a wide range of disciplines. Such synthesis is hampered by obstacles resulting from attempts to create a unified description from diverse data relating to humans and animals, of an individual synapse, or a neurone, or a whole organism, complex unlocalized mental processes and local physiological phenomena. The aim of the present article is to suggest a system of views, based on the literature and our own experimental data, within the framework of which such obstacles may be overcome.
In order to describe the cerebral basis of formation and realization of individual experience (IE), we define first the elements of IE (EIE). Today only a few researchers question the conclusion that the "properties ... of a brain are emergent" and are "systemic", not "just the sum ... of properties" of neurones, but a specific quality that emerges as a result of "dynamic interaction" of neurones within the system (Mountcastle, 1995, p. 294).
182 Yu. I. Alexandrov et al.
Analysis of possible levels of behaviour suggests that the level of "a unified group of neurones" is the most elementary level of analysis where the corresponding behaviour may still be described as an emergent function (Bottjer et al, 1994). A cerebral equivalent of EIE, which is established during the formation of a new behaviour, and realized during its subsequent performance, may be defined as an organization of a group of neurones, constituting the corresponding system. The question of what is meant by a "system" must be answered before we can use an understanding of EIE to describe the formation and realization of IE.
From our point of view, the most well-developed and un-contradictory version of the systemic approach to analysis of neuronal basis of behaviour, is the theory of functional systems elaborated by P.K. Anokhin and his school (Anokhin. 1973). The major distinguishing characteristic and advantage of this theory is the definition of a system-creating factor—the result of a system, which is understood as a desired relation between an organism and environment, achieved through the realization of that system. In other words, the principal determinant of a system is an event which is not in the past with respect to behaviour, that is, a stimulus, but which occurs in the future, a result. Thus a system is understood as a dynamic organization of activity of components with different anatomical localization, the interaction of which takes the form of mutual facilitation, in the process of ensuring a result that is adaptive for an organism.
It was demonstrated that the mutual facilitation in achieving any behavioural outcome is ensured by uniting synchronously-activated neurones situated in different brain structures (Shvyrkov, 1990). There is increasing evidence for this suggestion (Bullier and Nowak, 1995). The evidence is also increasingly important for the understanding not only of a specific behaviour, but also of learning. The association of synchronously active cells may ensure the achievement of the result even during the first trial, and may serve as a base for further consolidation: "Neurons wire together if they fire together" (Singer, 1995, p. 760).
In addition to the systemic idea described above, another important premise of the theory of functional systems is the idea of development (Shyleikina and Khayutin, 1989). Both ideas are merged in the concept of systemogeny, which states that, during early ontogeny, those differently localized elements undergo selective and accelerated maturation that is essential for achieving the results of the systems, providing for the survival of an organism at the early stages of individual development (Anokhin, 1973). Nowadays it is commonly accepted that many regularities of modification of functional and morphologic characteristics of neurones, as well as of control of gene expression, serve as a basis for the formation of adaptive behaviour in adults, and are comparable to those found at the early ontogenetic stages (Anokhin and Rose, 1991; Bottjer et al., 1994; Singer, 1995). The idea that systemogeny takes place not only during the early ontogenetic period, but also during adult development was formulated within the framework of the theory of functional systems nearly 20 years ago (Shvyrkov, 1978; Sudakov, 1979). This idea arose because the formation of a new behavioural act is always a formation of a new system. Later it was suggested that an understanding of the role of different neurones in the organization of behaviour depends on the history of behavioural development (Alexandrov, 1989; Alexandrov and Aleksandrov, 1982), or in other words, the history of the successive systemogenies, and the system-selective concept of learning was inferred (Shvyrkov, 1986). The latter concept is in line with the modern idea of "functional specialization" which substituted the idea of "functional localization" (Mountcastle, 1995)
and with the idea of the selective, rather than the instructive, principle of learning (Edelman, 1987). This concept considers the formation of a new system as the fixation of the stage of individual development—the formation of a new EIE during learning. The base of this process is the specialization of some "reserve" of (silent) neurones, but not the change of specialization of previously specialized units. Thus, the new system becomes an "addition" to the existing EIE (Shvyrkov, 1986, 1995). The selection of particular neurones from the reserve is governed by their individual features, that is, by the characteristics of their "metabolic needs" that are genetically determined. Newly formed systems do not substitute previously existing ones, but are "superimposed" over them; the appearance of neurones with new specializations results in the increase of the total number of units activated, whereas the number of neurones with old specializations does not decrease (Gorkin, 1988; Shvyrkov, 1986). The suggestions that the number of active neurones is increased during learning, and that learning involves new neurones rather than "re-learning" of the old ones has recently been confirmed by data from other laboratories (Bradley et til, 1996; Wilson and McNaughton, 1993).
What does it mean—"to superimpose, but not to substitute"? Many experiments in our laboratory have demonstrated that a complex instrumental behaviour is mastered not only through the realization of new systems (Figure 10.1, new systems), that were formed during the process of learning the acts comprising the behaviour, but also by the simultaneous realization of older systems (Figure 10.1. old systems), that had been formed at previous stages of individual development. The latter may be involved in the organization of many behavioural patterns, that is, they belong to EIE that are common to various acts (Figure 10.1). Therefore, it appears that the realization of behaviour is the realization of the history of behavioural development, that is, of many systems, each fixing a certain stage of development of the given behaviour.
These ideas are fundamental for systemic psychophysiology, which suggests the following solution to the psychophysiological (mind-body) problem. The organization of physiological processes into a system is based on specific systemic processes. Their substrate is physiological activity, whereas their informational content is psychical. In other words, psychical and physiological are different aspects of the same systemic processes (Shvyrkov, 1995). From this point of view, mind may be considered as a subjective reflection of the objective relation of an individual to the environment. That is, mind is considered as a structure represented by systems accumulated in the course of evolutionary and individual development. Relations between these systems (intersystem relations) may be described qualitatively, as well as quantitatively. The range of problems of systemic psychophysiology includes studies of formation and actualization of systems (EIE), studies of their taxonomy, and dynamics of intersystem relations in behaviour and activity. Thus, it may be concluded that investigation of the formation and realization of an IE is the task of systemic psychophysiology. It should be carried out at different levels, ranging from cellular and subcellular to complex human activity.
2. INDIVIDUALITY OF A NEURONE
As noted above, the system-selective concept of learning is based on the following suggestion: Neurones are originally diverse in their genetic and, consequently, in their
metabolic properties, and, during learning, only neurones with specific properties are incorporated into a system's organization. The stability of these properties was demonstrated in experiments with completely isolated nerve cells, by using the methods of mechanical and fermentative treatment.
When working with isolated neurones, cells keep the specific properties of background activity that they used to have in the nervous system. Alving (1968) used a mechanical method of isolation to demonstrate that the spontaneous electrical activity of isolated
Formation and Realization of Individual Experience185
nerve cells which she recorded before isolation, stayed similar. Chen et al. (1971), using fermentative treatment, found that completely isolated identified neurones maintained, after isolation, the main electrophysiological characteristics, such as the level of membrane potential, rhythm and patterns of spontaneous and elicited activity. Chemosensitivity was also stable. Isolated neurones were characterized by chemo-sensitivity to the same neurotransmitters that were effective before isolation. Our experiments have been performed on completely isolated neurones of the snail Helix pomatia. The results confirmed the stability of individual electrophysiological characteristics of identified cells. Not only were the background activity and chemosensitivity found to be stable, but also the dynamics of complex forms of neuronal plasticity remained similar prior to and after isolation (Grechenko, 1993). So, from the comparison of these individual characteristics of the same cells in vitro and in vivo it can be concluded that the analyzed properties of neurones in adult animals are stable.
Culturing identified isolated neurones in vitro, and analyzing their properties after involvement in the formation of new neuronal networks, allows us to find out if these properties stay stable in such a new neuronal organization. Syed et al. (1990) have done the experiments by culturing neurones of Aplysia, which formed new interneuronal connections. During these experiments the authors tried to describe the modifications of the neuronal electrophysiological characteristics, but no modifications were elicited by the procedure of culturing or axonotomy, the main parameters of electrophysiological activity remaining constant. It is necessary to note that each neurone formed new synaptic connections similar to those functioning in vivo.
Similar results were obtained in two artificial neuronal nets: respiratory and motor networks. In the latter, the transformation of action potentials was explored in the course of associative learning. These modifications were similar to the experiments in vivo and in vitro. Individual properties of neurones were stable in the neurotransplantation experiments. The stability of structural, intrinsic neurotransmitter, and electrophysiological characteristics of graft transplanted neural tissue have also been shown (Vinogradova, 1994). These findings confirm the stability of individual properties of a neurone, and support views on the regularities of learning suggested by the system-selection concept.
3. RESULTS OF BEHAVIOUR AS A DETERMINANT OF FORMING OF INDIVIDUAL EXPERIENCE
Within the framework of our approach, the specialization of neurones is considered to be a systemic one instead of "sensory" or "motor". Thus, we assume that even in conditions of "sensory deprivation"—for example, cessation of contact with the visual environment—neuronal activity in "visual" structures is necessary for achievement of results of behaviour. Indeed, it was found that the activity of neurones in visual cortex, in retina and lateral geniculate body (Alexandrov and Aleksandrov; 1982; Alexandrov and Jarvilehto, 1993) is related to the realization of food-procuring behavioural acts in animals, both with "open" eyes and with eyes closed with light-tight covers. According to the same logic, it should be assumed that during the formation and realization of behaviour under "motor" deprivation, and even in combination of "sensory and motor" deprivation, the activity of neurones is related to the realization of systems aimed at achieving the
186Yu. I. Alexandrov et al.
results of behaviour as well. This is shown by the fact that if an animal is restricted from moving voluntarily, but is nevertheless able to achieve some behavioural results during passive movement within an experimental arena, then a specific IE is formed, which corresponds to the analogous behaviour in freely moving animals. That is, neurones specialized according to the elements of this IE can be found.
This last assumption was tested in experiments (Gavrilov et al., 1994, 1996) with single unit recordings from CA1 complex-spike cells in awake rats, slightly restrained in a sling, and placed on a computer-driven robot. A rat was moved within a square arena (3 m x 3 m), from one corner to another, along the walls and diagonally. A drop of water was delivered (as a "reward") every time the rat approached one of the corners, this contingency remaining the same throughout the experiment. We found that about a half of the neurones increased their firing rate significantly while the rat was passively transported in particular parts of the arena, although these neurones had "spatial specificity" of low resolution, that is, their "firing fields" were larger compared to the those found in freely moving animals (O'Keefe and Nadel, 1979; O'Keefe and Recce, 1993). Some of these neurones maintained the same spatial selectivity of discharge when the rat was displaced on the robot in total darkness.
These results could be interpreted in terms of the currently-dominant views on the hippocampus as a pivotal structure for forming high-level representations of space on the basis of convergence of multimodal sensory information (O'Keefe and Nadel, 1979). From our point of view, these data support the idea of the determining role of results of behaviour in the formation of elements of IE. Representation of space is considered to be a reflection of the environment divided into elements according to the results achieved in this environment ("space of outcomes") on the base of some sensory "modalities". Formation of this representation is the formation of EIE. This also means that the existence of various spatially selective neurones which are active when the rat approaches one of the corners, irrespective of the direction and speed of passive displacements (i.e. irrespective of different means of attainment of the animal's contact with a particular place of the arena) is due to the fact that this place is in a constant spatial relation to that corner in which the animal was "rewarded" with water. Disappearance of the specific activation of hippocampal neurones when a restrained rat was placed into the "firing fields" of these neurones, that is, into the areas of the arena where these neurones had increased discharge activity when tested in freely moving rats (Foster et al., 1989), appears to be related to the change of behaviour from food-procuring to defense, and hence, to a change in the set of elements of the IE involved in the realization of the behaviour. Context-dependence of behaviour for spatial selectivity of discharges of the hippocampal neurons was shown earlier by Alexandrov et al. (1993) and Wiener et al., (1989).
In sum, the results described above offer good support for our assumptions. Even in restrained animals, passive transportation within the "space of outcomes" results in the activation of neurones in relation to the realization of EIE, which reflect the subjective "division" of the environment according to the results achieved in the environment. The result is similar to the findings for freely moving animals, although the structures of IE (both a set of elements and relationships between the elements) are probably different.
Formation and Realization of Individual Experience187
4. LEARNING HISTORY AND SYSTEMIC ORGANIZATION OF BEHAVIOUR
From the assumption that the structure of IE is determined by the history of its formation, one may suppose that the systemic organization of the same behaviour, formed by different learning strategies, differs between individuals because the different history means the formation of a different IE structure. The role of learning history was demonstrated in our experiments. Rabbits were trained to perform a food-procuring instrumental behaviour in a cage with two feeders and two pedals in the corners (Figure 10.2). At any given moment, only one pedal was effective—pressing that pedal switched on a feeder positioned near the same wall. Two different strategies were used during the training of the animals. The animals of one group were trained to execute the whole behavioural cycle along one wall of the cage (pressing the pedal, coming to the feeder and seizing food, pressing the pedal, and so on), then along the other wall. The animals of the second group were trained to obtain food from one feeder, and then from the other; to press one pedal, and then to press the other one (Gorkin and Shevchenko, 1991a,b, 1995).
The reflection of the learning history in patterns of specialized neurones' activity was studied in experiments by recording the activity of limbic cortex neurones (area retro-splenialis) in rabbits. The averaged frequency of activity and the activation probability were calculated for each behavioural act. Each of two behavioural cycles (along a concrete wall of the experimental cage) was divided into five stages (behavioural acts): seizing food in a feeder, turning a head to a pedal, approaching a pedal, pressing a pedal, approaching a feeder. So, all food-procuring behaviour in the cage turned out to be presented in ten stages: 1st to 5th on the left side of the cage and 6th to 10th on the right side. For each stage we have defined the mean frequency of neuronal activity during the time of its recording, and the distribution of frequencies composed a pattern of neuronal activity in behaviour (Figure 10.2).
For further analysis we selected neurones specialized for new systems of acts of approaching and/or pressing pedals ("pedal" neurones), as well as acts of approaching and/or seizing food in one of the feeders ("feeders" neurones). Neurones that showed activation in relation to different movements of the animal were considered to be specialized relative to old systems. Whether their activation appears or not is related specifically to a certain movement but independent of its behavioral context. Activations always appear during the same movement, which is performed for instance in relation to approaching the feeder or the pedal. Some neurones showed activation in relation to novel behavioural acts established late in individual development, such as during animal's learning in the experimental cage (e.g. approaching the feeder, approaching the pedal, pressing the pedal). Whether their activation appears or not is specifically related to a certain behavioural act but independent of its motor characteristics. For example, similar activity of these neurones is recorded when the animal presses the pedal with the left paw, right paw, or both. It appeared during the behavioural act, for which this neurone was specialized. This activation was usually several times greater than the "nonspecific" activity of the neurone, which was recorded during other behavioural acts and which, unlike the specific one, was much more variable and appeared in fewer than 100% of cases.
Comparison of the activity patterns of neurones with similar specialization showed that their "nonspecific" activity differed greatly (Figure 10.2 C,D). However, the distribution of frequencies was not random. There was additional activation of the neurones, specialized relative to the second pedal (with respect to the order of training), when the rabbit pressed the first one. Supplementary analysis involved normalization of the frequency of nonspecific activity with respect to the maximal frequency of activity during nonspecific acts (Figure 10.2 E,F). The analysis allowed this activation to be related to definite strategies of training. It appeared only when the formation of corresponding acts, related to the first pedal, directly preceded the formation of the acts of approaching and pressing the second pedal in the history of training. Thus, among the systems of behav-
Formation and Realization of Individual Experience189
ioural acts formed during training one after another, and performed by an animal at different sides of the cage, we found facilitating intersystem relations manifested in a raised degree of actualization of the last-formed system while an animal performed the previous one. Achievement of an act's result is ensured by realization of a specific EIE as well as others. Whereas the specific EIE are realized in the act in all cases, the probability and the degree of actualization of nonspecific ones are considerably lower.
A similar phenomenon was found for "feeder" neurones as well. Additional activation was detected in the nonspecific activity of cells, specialized relative to the second feeder, in respect to the order of training. The particular place where this activation appeared— during approaching and pressing a pedal at the other side of the cage, or during seizing of food from the other feeder—depended on the strategy of training, that is, which act preceded the one specific for this neurone in the animal's training.
Earlier we showed that systemic specialization of a neurone is its permanent characteristic (Gorkin and Shevchenko, 1991a,b). That is why neuronal activity can serve as an index of specific EIE actualization (Shvyrkov, 1995), and "nonspecific" activity of a neurone may indicate the retrieval of a specific system from memory during performance of other behavioural acts. Studies of the activity of system-specific neurones during the performance of cyclic food-procuring behaviour may reveal the relations between the specific system and other functional systems of analyzed behaviour. Thus, identification of intersystem relations can reveal the IE structure acquired by learning. The data obtained confirm the assumption that the IE structure and, consequently, the system organization of behaviour in which this IE is actualized, are determined by the developmental history of the behaviour.