A Functional Neurodynamics

for the Constitution of the Own Body

Jean-Luc PETIT*

Université Marc Bloch (Strasbourg II)

Laboratoire de Physiologie de la Perception et de l’Action (Collège de France)

However little philosopher may as yet be aware of this recent development, the burgeoning field of brain cartography has transformed the traditional dispute between phenomenology and positive science about the adequate treatment of the body into an obsolete quarrelling. Up to now, phenomenology used to dedicate itself to calling attention to the difference (not to say stirring up the conflict) between the fixity of anatomic Körper structure as an object of science, and the free fluidity of the meaning patterns of Leib subjective experience. From now on, one’s inquiry should be whether or not such a contrast is on the verge of vanishing. In fact, neuroscience has resolutely shaken off its former belief in a rigidly somatotopic representation of the peripheral organs of the body within the frontiers of definite somatosensory mapping territories of the centro-parietal cortex and thalamus. Accordingly, a new methodological approach is forcing its way through brain science labs, putting on their common agenda the setting up of a global online recording of constantly moving functional activation patterns (a ‘mental cinema’). These patterns transitorily distribute themselves over varying regions of cerebral tissue at a rate determined by the demands made upon them by the performance of behavioural tasks. Such representational plasticity, far from being genetically predetermined in all its localisational specifics, proves itself to be induced, shaped and modulated to a considerable extent by the unique experience of the organism in its environment. Laying our bet on the chances of a new relationship between phenomenology and objective science, we will take advantage of the opportunities created by these developments. And we will (allowing ourselves some speculation) bring together the flow of functional activity of the the brain and the flow of lived experience of the body in an attempt to bridge (or at least narrow down) the gap between activation patterns and meaning patterns, considering that they are mutually indispensable correlates underlying the auto-affection of the acting person.

Dominated as it is by the paradigm of a brain-machine designed to process information, neuroscience tends to reduce ‘the body’ to one of the representations in the brain alongside representations of other things. And so it becomes the representation of that object by means of which it receives information (mainly tactile) and the muscular movements of which it controls. In one particular branch of the neuroscience, cerebral brain cartography, a branch which has made remarkable progress in the last thirty years, the talk is of ‘somatotopic coding’, regions of ‘cortical representation’, ‘cellular receptor field’, etc. Apparently, this way of talking is inspired by the fairly traditional ideology of representation as an unequivocal correspondence (isomorphism) between the peripheral structure of the body and the central homunculus (or homunculi). However, belief in the rigidity of this projective relation suggested by the expression ‘somatotopic coding’is (at least potentially) contradicted by the discovery of the representational plasticity of the cerebral tissue, a discovery made by this same cerebral cartography. The current generalisation of this phenomenon of plasticity from association to primary areas and to all the sensory modalities, as well as to the motor function, increases the tension between the new intuitions and conceptions and the modes of expression still employed. All the same, the power of the metaphor of the brain-machine upholds the use of the vocabulary of the code and of somatotopy and delays its replacement by a conceptual framework better adapted to the functioning of the brain and to its true relation to the body. With regard to this relation one already suspects (while waiting for the paradigm change which will make it a legitimate claim) that, rather than the representation of a body preconstituted prior to this representation, it will have to take the form of a dynamic interaction between three terms : the body, the brain and also the world (absent from the traditional, representational ideology), terms which cannot be taken to exist prior to this same relation since they bring each other into existence through their mutual interaction.

1. Somatotopic Cartography and Functional Plasticity

A few preliminary remarks are useful to fix the limits of our enquiry. First of all, research into the functional plasticity of the brain does not stop at the representations of the body in the somatosensory (SI) and motor (M1) cortices. It applies equally to the retinotopic representation of visual information in the striate cortex (V1) and to the tonotopic representation of acoustic information in the temporal area (A1). We will restrict our attention to the evidence bearing on the cartography of the body, even though the plasticity of corporeal representations is not isolated from modifications stemming from exteroceptive sensory influences. Second, one of the factors responsible for much of the progress in neuroscience consists in experiments performed on animals and the transfer of hypotheses or concepts developed in connection with mammals or primates to human beings. In particular, the rat is currently an object of intense research, due to the ease with which its sensory system can be manipulated in experiments, a system whose vibrissae are the peripheric organs and the barrel cortex the organ of internal representation. Since evidence relating to a system as specific as this cannot be directly carried over to humans, we won’t go into this any more. On the other hand, restricting ourselves to the human system would put us in a position where we could no longer obtain a global view, not even a view of detail bearing on plasticity and somatotopy, since progress in non-invasive techniques of cerebral imagery have not yet made it possible to reduce the gap between knowledge bearing on the human brain and knowledge already achieved in connection with monkeys (by means of recording techniques based on chronic − i.e. permanent − cerebral electrode implantation). Finally, our interest is in plasticity induced or modulated by experience, understanding by that experience the one that an individual develops through a normal use of his or her body, a use which is evidently enriched and diversified in the course of a learning process. The plasticity that is evidenced by patients that have suffered a stroke or a surgical amputation of a limb and reacted to it by a functional reorganisation of their brain, cannot be described as induced by experience except in a highly extended sense of that word. In particular, we are not going to take into consideration ‘the illusion of phantom limb’, with regard to which the literature tends to be as vast as it is controversial. However, even if we decided not to take the mechanisms brought into play in that case or the other into account, it would be foolish to ignore the knowledge obtained by the study of such reorganisational phenomena in the case of lesions both in humans and animals, because if the word ‘reorganisation’ tends to be employed in this context while the word ‘remodelling’ is more frequently used in the context of normal usage, this verbal difference does not seem to be one which testify of the existence of a distinction in re.

2. Penfield’s Homunculus and its Contemporary ‘Verification’

Penfield himself is remarkably prudent in his statements regarding the value he accords to the ‘sensorial and motor homunculus’ (Penfield and Boldrey, 1937) or to ‘the sensorial homunculus and the motor homunculus’ (Penfield and Rasmussen, 1950) as regards the light it throws on cortical topography of the sensory and motor functional representations. Moreover, the expressions of ‘mapping’ and ‘coding’ have not yet been used. In the first version, ‘this grotesque creature’ is only called in to faithfully represent two features. The first feature is the constant order of succession of the different parts of the body concerned by the movement provoked or the sensation evoked by an electrical stimulus applied bit by bit to the cortex, following the edges of the central sulcus in the medio-lateral direction. These parts are, specifically, the body, decapitated and inverted, then the head from the front, juxtaposed to the thumb, then the tongue out of the mouth, etc. The second feature is the relative vertical extension of that portion of the rolandic cortex devoted to the representation of each part of the body, , , which is carried over to the homunculus as the disproportionate length of the tongue, the face and hands in comparison to the rest of the body. With the result that, with the exception of these two topographical constants, all that the outline could save as representative of a man (‘as though representing a man’, says Penfield, ibid. p. 431) with its specific surface, its size, its precise contours (not to mention hair and skin wrinkles in certain popular illustrations!) had to be treated as arbitrary and misleading. For in fact Penfield does not try to hide the considerable dispersion of the points of stimulation evoking motor or sensory responses ini in different individuals, and in the same individual from one to another surgical intervention. Even though he distinguishes a postcentral sensory cortex and a precentral motor cortex, he admits that he also obtained motor reactions (even though less frequently) by stimulating the postcentral cortex, and sensory reactions (more frequently) by stimulating the precentral cortex. If he proposes a delimitation of the areas responsible for different parts of the body, it is not for making of them ‘the borders of the territory of representation’ relative to these parts, but to underline their mutual interpenetration (fig. 25, p. 430). In the end, he holds back from developing any hypothesis about the correspondence or lack of correspondence between representations, whether this be with the cytoarchitectonic regions of the cerebral tissue or with the distributive density of the sensory captors on the skin of different parts of the body. Hence the notice to the reader in the work of 1950 : “It is a cartoon of representations in which scientific accuracy is impossible (p. 56)”.

The development of a technique of non-invasive cerebral imagery at the end of the 70s and the beginning of the 80s has made possible a certain ‘confirmation’ of this classical description of the somatologic organisation of the functional representation of parts of the body in humans. Measuring the regional blood flow in the cerebral areas through tomographic recording by the emission of positrons (PET), visualisation of the structures of the brain through magnetic nuclear resonance (fMRI), exploration of regions of interest by subtraction of images[1], the addition of images maximalising activations corresponding to each condition in one subject and to one and the same condition in all subjects, without taking into account numerous operations of normalisation, correction, standardisation, redistribution, averaging and calibration, all of the above adds up to a mass of manipulations each of which rests upon a questionable presupposition of neutrality and non-interference with the facts under examination. Since the complexity of the technical apparatus brought into play and the tacit claim of transparency seem to grow at the same pace, the apparatus employed tends to disappear behind the publicly communicated ‘views of the brain’ and their reproduction in works of synthesis. Without going too far into the much needed criticism of such methodology, let us at least ask what in fact the procedure adopted has helped to confirm. Essentially two things : 1) by means of a manual cutaneous vibrator applied successively to the lips, the fingers and the feet, the latter are stimulated in such a way as to evoke responses focused in different regions of the postrolandic cortex (SI) strung out along the central sulcus in a latero-medial order from the parietal opercula to the interhemispheric wall (Fox et al., 1987); 2) chasing a target moving randomly about a video screen, using respectively the big toe, an outstretched arm, the index finger or the tongue, activates precentral zones of the cortex (MI) which follow upon one another from the dorso-lateral edge of the interhemispheric fissure to the neighbourhood of the lateral sulcus, passing across a region of activation where the index finger is superimposed upon that of the arm (Grafton et al., 1991). One notes that only that aspect which Penfield himself considered true remains in accord with Penfield’s homunculus, namely, the sequential order of the functional representations on the medio-lateral axis of the pre- and postrolandic cortices. However, the limits of this agreement tend to be concealed by the expressions employed. The talk is of ‘millimetric localisation’, even though the average difference between the localisations taken two by two in the same subject is of the order of 3mm (Fox et al., p. 39). Or one talks of ‘detailed examination of the somatotopic distribution’ and of ‘localisation to predictable sites’ even though the variability of the gyri and sulci from one individual to another involve displacements in the activations, both in breadth and in depth, which only allow for a range of estimation regarding their probable occurrence (Grafton et al., p. 737; fig. 3, p. 739).

3. Reorganisations of the Functional Structure following a Deafferentation

The existence of a permanent potential for functional reorganisation has been demonstrated at the beginning of the 80s by the research on hand representation in the monkey’s somatosensory cortex (3b) undertaken by Merzenich and his lab (Merzenich et al., 1984; Wall et al., 1986). The consequences of more or less important deafferentations, such as the amputation of one or two fingers, severance of the median nerve innervating the skin of the radial half of the palm and the internal face of the first, second and third fingers, localised crushing of the same nerve, etc., have been controlled by detailed cartographic readings. This has happened on the basis of intracerebral recordings practised at different stages of functional recuperation. These manipulations have proved that the central representations of the body are not subject to a rigid anatomical determinism, attributing rigidly the surface of each part of the body to a well defined cytoarchitectonic area of the brain. In contrast, the representations of the body are far more the expression of a dynamic activity enabling the organism to react in an innovative way to changes in the sensory inputs in order to maintain the integrity of the ‘body image’ (tactile sensitivity, somesthesia, motricity) damaged by the lesion. This activity is displayed at the centre by ‘movements’ in the cortical representations : expansion or contraction of the individual representations of the fingers, displacement of the borders between representations of different fingers, expression of normally latent representations, withdrawal from or reoccupation of deafferented regions. At the periphery, one finds correlated movements in the cutaneous receptor fields (RF[2]) of neurons belonging to the same somatosensory cortex : RF expansion or contraction, the appearance of multiple RFs for one and the same cell, the acquisition of alternative RFs. As points of reference, I will rely only on the most significant evidence.

In response to the amputation of the index and/or major finger, the cortical representations of the adjacent fingers extend into the area where the amputated fingers were represented, and in such a way as to fill the gaps between the represented fingers, thus reestablishing a new borderline. To the extent that this borderline passes between the representations of fingers which are not adjacent in normal anatomy, its emergence appears as a true creation of the functional dynamism. Here and there, the same neurons which until then upheld the representation of the amputated finger are reassigned to the representation of one or the other (but not both of) remaining neighbouring fingers. Since the expansion of the cortical representations (increase in magnification[3]) is coupled with a contraction of the cutaneous receptor fields localised on the same fingers, the intervening reorganisation results in a refinement of the representation of the skin, which can be interpreted as an attempt to compensate for the sensory loss due to the amputation (Merzenich et al., 1984). In response to the severance and suture of the median nerve (an operation favouring a reinnervation of deinnervated skin) the neurons dealing with the cortical representation of the hand begin by losing their receptor fields, which are normally situated on the internal surface of the fingers 1 to 3, and acquire alternative RFs situated on the back of these same fingers. Later on, the regeneration of the median nerve does not result in a centrally diffuse and random reactivation but in a reorganisation of the functional representations, which includes persistent anomalies : discontinuities, delocalisations and superimpositions alongside topographically localised aspects (Wall et al., 1986). After a transitory transfer of the RFs of the dorsal surface of the fingers onto their ventral surface, deafferentation resulting from the localised crushing of the median nerve turns out to be compatible with the reestablishment of correct correspondences between skin and cortex in the context of a normal topographic organisation of the somatosensory cortex (Wall et al., 1983).

The cerebral cartography of the monkey has taken advantage of the accessibility of the somatosensory cortex of the hand in the species under investigation, namely the owl monkey, whose brain has no central sulcus. Consequently, a methodology has been developed making possible the drawing up of veritable maps of the functional topography of the cortical areas, assigning to these areas quite specific borders and allowing by means of objective measuring the demonstration of the occurrence of displacements of these borders. This has been made possible by the chronic implantation of a grid made of many hundreds of microelectrodes, combined with the tactile exploration of the hand by means of a tapered probe designed to make indentations of the skin at the limit of the visible. In that way, minimal RFs for the neurons under examination are defined. In the case of humans, the neurophysiological description of phenomena of functional plasticity, exception made of preoperative and direct explorations, has depended upon the development of techniques of non-invasive functional imagery like PET, fMRI, magneto-encephalography (MEG) and transcranial magnetic stimulation (TMS). Only that these methods can only offer images of more or less diffuse centres of activation, or else curves of motor potentials evoked in the muscles through TMS. Thus, even though the language of ‘maps’ has been retained, the maps in question are very far from delineating the frontiers of the representations with a millimetric precision approximating that achieved with animal experiments. As a result, an evaluation of the amplitude of the reorganisation induced in humans by deafferentations inevitably has a qualitative character, and this whether they are provoked by a local anaesthetic or are are of accidental or pathological origin[4].

4. Remodelling Induced by Experience (1) The Somatosensory cortex

Does this reorganisation of the functional architecture induced by deafferentation (experimental or accidental) depend upon mechanisms essentially different from those of a remodelling linked to a normal use of the perceptive or motor organs ? The question remains controversial. Whatever the outcome, the study of deafferentation has brought to light a principle of plasticity pertaining to the organisational schemata of functional somatotopy. This principle integrates the neurophysiological correlates of the experience of the body with the general dynamism of the functional organisation of the central nervous system. In fact this principle of plasticity is not limited to somato-aesthesia that interests us in connection with the theme of the body image but also concerns exteroceptive sensory modalities, in particular the primary visual and auditory areas (to say nothing of the other senses). Visual and auditory experience are not rigidly predetermined by the anatomical structure of the receptive surfaces and the cortical regions. In spite of the textbooks, it is admitted that the retinotopy of V1 is not the isomorphic (nor deformed) projection of the retina, the tonotopy of A1 is not the isomorphic projection of the cochlea; rather, the projective geometry implemented here and there by the brain has to be incomparably more complex and dynamic. The way in which this happens should be such that the bodily experience draws its significance from the autonomous activity of the organism, which in its effort at a permanently renewed adaptation to the flux of ever renewed experience, finds in itself the resources needed for the emergence, the remodelling, and the persistent renewal of its organisational patterns. This dynamism might eventually prove easier to verify with reference to deafferentations, which are all the more dramatic because the survival of the individual depends upon them. But it ought also to be possible to verify the dynamism in question in the normal circumstances of everyday life. The eminently plastic usage (depending in part if not entirely on a learning process) such as the normal use of the skin as an organ of tactile and somesthetic sensitivity, of the hands as tactilo-kinesthetic organs of action, cannot but bring with it a reorganisation, or at least a modulation, of functional representations. These changes, in turn, condition the improvement (or deterioration) of the behavioural performances.

The cartography of the parietal regions bearing the representation of the hand in the normal adult monkey is highly variable as regards the detailed representation of the hands in individual cases, a fact that has convinced researchers that the maps could not be predetermined with precision by the genes for all the individuals of one and the same species, nor even onto-genetically fixed at a precocious stage of development. In contrast, they have to be formed by the particular use made of the hands by each animal in the course of its individual history. Here are some of the differences from one individual to another : the global form of the area responsible for representing the hand, the total surface of this representation, the magnification of the representation of the regions of the skin, the surface of the representations of the different fingers, the disposition of the representations of the dorsal surface of the fingers, in islets or at the lateral and medial margins, the topological boundaries between representations – continuity, discontinuity, interpolation, proximity of the boundaries (Merzenich et al., 1987). Training to detect a difference between an initial vibratory stimulus applied to a finger and a stimulus of a higher frequency in a series of stimuli of variable frequency produces a topographical complication of the representation of the hand, including an extension of the skin zone stimulated and a shattering of the representation of the stimulated phalanx. This representational change is correlated with the animal’s progress in the realisation of the task, that is, in a reduction of its tactile threshold of detection of the vibratory frequencies, a reflection of the localised improvement in perceptual discrimination of the skin. All the same, and contrary to what one might have expected on the basis of the rule of inverse proportionality between the extension of cortical representations and the extension of the cutaneous receptor fields, this representational change does not correspond to a shrinking of these receptor fields. On the contrary, one notes an extension, a multiplication and a mutual overlapping of RFs of the neurons dealing with the representations of the hand subject to training, numerous RFs being displaced in order to be recentred and superimposed upon the zone of the stimulated skin. This reorganisation does not take place on the occasion of a passive stimulation of the finger, but only when the stimulation is a part of the task, which suggests that it is under the control of attention (Recanzone et al., 1992)[5].

In humans, the representational plasticity of the somatosensory cortex induced by practise has been confirmed for the more complex tasks of professional life before being confirmed again by artificial tasks controlled in the laboratory.

Violinists and other players of string instruments continually make use of the second to fifth fingers of the left hand to press the strings onto the fingerboard while the thumb of the same hand holds the shaft of the instrument with frequent changes of position and variations in the pressure exerted. Since the aim is to ensure a very rapid identification of the right notes with the tips of the fingers along the entire length of the four strings from the fingerboard to the bridge, this movement becomes automatic with practise in all gifted musicians. On the other hand, the movement of the right hand which holds the bow between the thumb and the index (and middle) finger, by blocking (albeit with great flexibility) the individual movements of the fingers is less muscular and constantly calls for the sort of considered decisions in which the artistic personality of the musician is expressed. A practise of this kind, normally initiated at an early age and continued throughout an entire life-time for several hours a week brings with it a considerable disequilibrium in the sensory input of the two hemispheres of the brain. Researchers are interested in the remodelling of the cortical maps of the hand induced by this intensive and highly differentiated use of the fingers.

A tactile stimulation from a (painless) pressure applied with a pneumatic stimulator either sometimes on the thumb and at other times on the little finger of each hand evokes cortical responses which can be recorded on MEG. The representative vectors of the equivalent current dipoles which summate the contributions of the flows of dendrite currents registered in different subjects are transferred on an fMRI image of the cortex of a control subject. It is observed that these vectors, which represent the localisation and the average intensity of the foci of cortical activity corresponding to the individual stimulation of the thumb or little finger are extended and displaced towards the median plane with musicians, and this all he more so when the practise of the instrument begins at an earlier age. The authors infer from this that the size of the cortical representations is not genetically determined but rather modified by practise, and that the expansion of the representation of the fingers of the left hand induced by learning the string instrument can afford the musician a decisive advantage in responding to the demands of this art, to the extent that this expansion reflects the enlistment of a more extensive neuronal network for the processing of a larger flux of tactile information with the musician than the non-musician (Elbert et al., 1995). The focal dystonia of musicians[6], which can be correlated with a fusion (without topographic disorganisation) of the representations of the different fingers on the map of the affected hand, proves that this remodelling by practise can be converted into a handicap when this usage is overdone and a lesion is brought about in the central sensorimotor system by the synchronically abnormal, repetitive and prolonged movements of the hands (Elbert et al., 1998)[7].

In a recent experiment, subjects had to recognise as quickly as possible the orientation of tactile stimuli consisting of three little pins arranged in an arrow pointed at random towards the right or left, and this by pressing a button with the right hand. These stimuli were applied simultaneously on the last phalanx of the thumb and the little finger of the left hand for 50 msecs in a massive and repetitive way for 1 hour a day over 4 weeks. A high resolution electro-encephalogram shows that the passive tactile stimulation of a finger elicits on the scalp an electric field at a latency of 50 to 60 msecs and that the source of this field can be modelled with an electric dipole situated at the level of the somatosensory cortex. An electroencephalogram recorded at the beginning and at the end of the period of training makes it possible to establish, by projecting it on a MRI image of the brain of each subject, the occurrence of any displacement of the localisation of the source of the electric field induced by a new stimulation of the trained thumb and little finger. With regard to the localisation of the cortical representations of these fingers, it appears that their simultaneous stimulation in the context of the task of discrimination produces an effect contrary to their individual and passive stimulation. The representations of the thumb and the little finger of the right hemisphere (contralateral to the trained hand) move away from each other in the medio-lateral axis as a result of the training, denoting an expansion of the areas of representation and from there a disassociation of those neuronal groups activated by each representation. When, on the contrary, the stimulation is applied separately to the two fingers with a random orientation of the stimuli which the subject does not have to identify, the representations of the thumb and little finger get closer to each the other, to the point of superimposition, translating an overlapping of the areas of representation under the effect of a passive stimulation. In their interpretation, the authors do not decide between two hypotheses: 1) a unique map of the hand whose activation is differentiated as a function of the different ways in which the stimulus is processed; 2) multiple maps coexisting in the same cortical area and whose activation is a function of the context (Braun et al., 2000).

5. Remodelling Induced by Experience (2) The Motor Cortex

What direct electrical stimulations of the precentral cortex evoke are bodily movements; what Penfield and the first mappers of the brain sketched out in the form of the homunculus are parts of the body : the fingers of the hand contralateral to the stimulated hemisphere which, in anatomical order, are represented in the latero-medial plane. But movements are rarely evoked in one part of the body without being evoked in the neighbouring parts. The mastery of the independence of the hand with conductors, that of the fingers of pianists or typists, requires a difficult learning process that most probably draws upon important cerebral resources. This inconsistency has probably only been noticed quite late on. A somatotopic organisation of the cortical representation of the hand suggests the existence of a neuron (or several) for the index finger, that is, of a neuronal group exclusively dedicated to the control of a particular finger, alongside other neuronal groups devoted to the control of each of the other fingers. However, nothing of the kind is found. The recording of neurons of the motor cortex during the carrying out of flexional movements and of movements extending different fingers with the monkey shows that the movement of each finger mobilises neurons distributed throughout the entire area of the hand and that the map of the cortical representations of the movements of the fingers is not somatotopic (Schieber, 1993). However, just as an unequivocal correspondence between representations on a somatotopic map and the parts of the body would exclude any possibility of reorganisation, in that way the activations distributed throughout the totality of a neuronal network according to a certain given configuration would lend itself to a functional reorganisation due to the varying usages of the body[8].

The methods of human cerebral imagery (measuring the cerebral blood flow in PET) which proceed by averaging the results obtained with several subjects and which identify regions of interest by subtraction of images are disadvantaged for the examination of phenomena of plasticity linked to a motor learning process. That is due to the fact that the procedure adopted to arrive at the mastery of a new task is not necessarily uniform from one subject to another and to the fact that the non super-imposable activation sites are automatically erased from the resulting image. To get around this difficulty a technique of individualised imagery has been developed which suggests the existence in each subject of a relation which is not that of a simple correspondence movement-cortical area, but that of a complex relation between a particular schema of adaptation to the task and a type of change in the schemas of cerebral activation distributed over varied regions. The task is to carry through blindfolded, as fast as possible and without mistakes, a complex series of movements involving an opposition between the thumb and each of the other fingers of the right hand. Progress over one hour of training differs largely according to the criterion employed: acceleration of the process or correction of the mistakes. Despite an activation of the left primary sensori-motor (and pre-motor) region in all subjects, the authors noted a considerable diversity in the areas of activation from subject to subject, and this no matter the areas in question were cortical (mesio-frontal, parietal, cingular, Broca) or sub-cortical. This is a discovery that raises questions pertaining to the contribution of each of these regions to the particular profile established by the performance of the trained subject (Schlaug et al., 1994).

A longitudinal study of a similar learning task with a training of several weeks adds complementary information resulting from an MRI examination of the regional blood flow in the motor cortex. Starting from an equivalent activation of M1, first with the sequence of learned movements and then with a sequence composed of he same elementary movements in another order, passing a paradoxical though transitory reduction of the area of motor activation corresponding to the sequence of learned movements, one finish with a significant extension of this area in the fourth week, an extension which can be maintained for several months. According to the authors, this durable expansion of the representation of the ordered sequence of learned movements would make of the primary motor cortex a memory of the know-how in the adult (Karni et al., 1995; 1998)[9].

6. Pluralism in the Models of Neurobiological Explanation

In spite of the fact that the interdisciplinary character of the neurosciences makes it possible to hold to the belief in the equal rights of all participating disciplines to their claim for being fundamental, the familiar practise of all these disciplines is still far from being able to risk comparison to any science which is genuinely fundamental, such as quantum mechanics. A fundamental science seeks to develop the paradoxes hidden in its concepts without being afraid of exposing itself to controversy, even on the contrary, seeking controversy. It does not attempt to clothe these concepts with the garb of consensual unanimity, or even to surround the emerging divergences which might menace its dogmas. Those dogmas, moreover, pushed to the limit, might turn out to be contradictory. A truly fundamental science which knows only too well how illusory the irrepressible human tendency toward objectivation, substantialisation and absolutisation of the theoretical models and dominant scientific paradigms of a given epoch (yesterday Lapacian mechanism, today the mechanism of Turing) can be, is not afraid of appearing to progress backwards by systematically referring back its ‘explanatory’ and ‘predictive’ concepts to their conventional and so largely arbitrary principles of construction, the field of its ‘real’ objects to the geometry it makes use of, its ‘exact’ measurements to the limited power of resolution of its instruments. Apparently this is still not the case in neuroscience, where the same dogmatic defenders of the genetic determinism of the cerebral thinking machine with its cognitive programmes also want to present themselves as heralds of epigenesis and of the history of the development of the individual. And the very persons who, in the course of 20 years, have revolutionised cerebral cartography, demonstrated the inanity of its traditional concepts ‘map’, somatotopy’, ‘representation’, ‘coding’, etc., and so laid the basis for the next functional neuro-dynamics, habitually employ a language that preserves and perpetuates the prejudice of a (or even many) humunculi in the brain.

The format of scientific journals which print in small letters the technical account of the cell recordings, the image analysis or the method by which the published ‘maps’ are constructed, leads one to separate these products from their mode of production, thereby incurring the risk of their being envisaged as maps in the brain. But that nothing like such maps is found in the brain is something that can be persuasively upheld. The following items related to maps are evidently not found in the brain : readings obtained from the grids of penetration sites of electrodes in the cytoarchitectonic cortical areas, outlines of the cutaneous neuronal receptor fields, histograms of the neuronal peristimulus action potentials, mosaics of the categories of movement evoked by IMS, electroencephalograms, scintigrams of the rate of consumption of oxygen or glucose by the regional blood flow, the distribution across the scalp of loci of stimulation evocative of motor potentials, dipoles of the sources of the induced electric or magnetic fields, etc. But when one imagines that it might be possible to ‘go further’ (by extrapolating from the available methods of obtaining images or representations) there arises a danger of fixing, objectifying or substantialising the transitory configurations of the functional dynamism of living organisms. That includes that one misses the essential and persistent feature of the potential for reconfiguration and functional reallocation which is not limited to an early age or to the axonal regeneration and functional recuperation of a lesion.

The challenge is to understand neuro-plasticity without trying to situate our conceptual instruments in the brain, by talking of ‘neuronal coding’ or of the ‘genetically programmed’, and without entering into any collusion with a neuronal determinism which conceives of the functioning of the brain as the calculations of a machine that follows a programme that completely specifies in advance all its transitions from state to state.

Even if linguistic habits have not changed greatly, we cannot but concede that this challenge has been met from the time of the first work on cerebral plasticity. In an effort to grasp conceptually the data of Merzernich and his team, Edelman has advanced the idea of a functional and interactional morphogenesis by selective stabilisation of the synaptic connection patterns in conjunction with the activity of the organism (Edelman et al., 1987; Kaas et al., 1983). While avoiding any reductionist explanation, a computer simulation of a simplified model of the neuronal network has made it possible to elucidate analogically and holistically the principles of a dynamic morphogenesis of functional topologic maps, by bringing to light certain of the properties established by deafferentation or amputation of the fingers in the monkey. Without entering into details, we would like to applaud the spirit in which this model has been developed, to the extent that its dynamic approach seems to us to contradict the fixist prejudices conveyed by the language of coding inherited from a mechanistic conception of cerebral functioning[10].

7. Autonomy and Experience in the Constitution of the Own Body

By virtue of its quasi-spontaneous or autonomous character, the correlative emergence of neuronal groups in the cortex of the network of neurons and of neuronal receptor fields in the matrix of the skin captors of the same network is the best analogy that one could find in contemporary naturalistic science for the transcendental constitution of the own body in genetic phenomenology. The use of my hands gives me (in a certain sense) my own body. But in what sense exactly ? According to Husserl’s later manuscripts, the regulated effectuation of tactile kinesthesia (objectifying) and of motor kinesthesia (de-objectifying) is the constitutive operation by means of which alone I acquire the sense of being (and from there consciousness) of my own body, both as a body object, an object among other objects of sensory perception, and as the unique organ of my voluntary movements. In the very course of its functioning, the first group of kinesthesia constitutes a continuous and closed surface which adopts for me the meaning of being ‘my skin’; the second fills this surface with a subjectively animated matter which adopts for me the meaning of being ‘my flesh’. But neither my skin nor my flesh have anything a priori to do with this ‘mass of flesh and bone that I call my body’ (Descartes). They are in essence the products of a constitution, more specifically, of an active auto-constitution on the part of the living organism, a self-organising agent, constantly adapting to its context, moulded by its own history. The organism (as certain eminent physiologists have said in astonishingly phenomenological terms) ‘strives to make sense of itself’ (Bartlett, quoted in Barlow, 1985, p. 121) and ‘chooses from one moment to the next the being it will become’ (Merzenich and deCharmes, 1995, p. 76). Ironically, by adopting the hands, which are both sensitive surfaces and motor organs, as the privileged models for the morphogenesis of the somatotopic maps (for simple reasons of practical convenience I assume), the neuroscientists thereby resuscitated the analyses (developed by Merleau-Ponty and Husserl) of the celebrated example of ‘my right hand touching my left hand, the latter, in turn, passing from being passively touched to actively touching.’

This improbable encounter between a neurodynamic (still in preparation despite the promising perspectives opened up by the ‘mental cinema’) and a genetic phenomenology (unhappily relegated to the field of historical studies) attests to the possibility of at least breaking the magic circle of representation, which still holds neuroscience imprisoned in the paradigm of the mechanical brain and the body representationally intellectualised. What does this opportunity depend on ? On the fact that the emergence of the body schema, on the basis of the functioning of a dynamic system in the brain, and the constitution of our sense of the own body, on the basis of kinesthetic activities of the organism, are (for the one who places himself or herself in the context of the flux of experience and not in the position of an external observer) genuine beginnings, effects without causes, absolute origins. For in fact, for the living organism caught up in the immanence of its own experience, there is no such thing as a physical or anatomical body to be represented, a body which would precede in the order of being its representation, the latter reproducing somewhere in the mind-brain a cartographic image of this same preconstituted body. The signifying form, the sense of being a body, arises from its own operation as self-given sense. The own body is no more the representation of the physical body than the functional body is the representation of the anatomical body. The true relation runs in the reverse direction; first comes the own body, the subjective form of lived experience or the functional configuration of a living organism. As for the anatomical or physical body, it is a later product constituted by a procedure of scientific objectification, and, what is more, a constituted product in the paradigmatic context of yesterday determinist science, a science of permanent objects, the fixed substrates of properties such as physical, functional or mental properties, which can always be precisely located.

References

I. Local anaesthetic. The localisation of the electrical potentials aroused in the posterior wall of the central sulcus by an electrical stimulation of the thumb and the little finger of the right hand under a local anaesthetic injected into the fingers 2, 3 and 4, brings to light an immediate and reversible change in the topography of the representation of the fingers in the left somatosensory cortex. On average, this change consists in a lateral displacement of the dipole (centre of gravity of the electrical field) image of the thumb and a medial displacement of the dipole image of the little finger. Authors have suggested that this modification in the excitation-inhibition equilibrium of the sensorial inputs engenders a hyper-activation of the cortical regions underlying the representation of the fingers, which react by encroaching upon the neighbouring regions (Buchner et al., 1995). It has been established that this reorganisation is subject to the modifying influence of attention. When a stimulus is applied to the back of the hand, the distance between the representation of the thumb and index finger shrinks, while it increases whenever the subject’s attention is normally engaged by the sensation (strong and disagreeable) in the anaesthetised fingers. The hypothesis is that the sensation of deafferentation activates the de-afferentiated cortex normally held responsible for the representation of the fingers 2 to 4, an activation which translates into an expansion of this representation. On the other hand, when the attention is diverted from this sensation of deafferentation, the cortex reverts more easily to the representation of the adjacent fingers (Buchner et al., 1999).

Neuropathies. The extent to which the motor cortex is activated, as measured by fMRI, tends to be greater on average for a voluntary movement of the index finger than for a passive movement. In contradistinction to the fact that passive movement gives no activation of the motor cortex in patients suffering from sensory neuropathy, the volume of activation induced in the motor cortices, both primary and supplementary, and in the somatosensory cortex, undergoes a considerable increase in patients suffering from motor neuropathy in comparison to the control group, and just as much for active as for passive movement. The reorganisation which follows upon a lesion of the motor nerves and an increase in the quantity of neurons engaged in the same movements is supposed to bring along with it an expansion of the representation of the movement of the finger (Reddy et al., 2001).

Syndactyly. In patients suffering from syndactyly, a congenital anomaly which hooks the fingers together, the preoperative examination by MEG shows a typically non-somatotopic map of the right hand with a spatial extension inferior to the normal, with images of the different fingers overlapping, and with the interposition of the little finger between the thumb and the index finger. A postoperative examination starting from the first week after the surgical separation of the fingers and the acquisition of independent use of the fingers shows a considerable reorganisation of the map of the hand. The representations of different fingers are removed into distinct localisations. A somatotopic organisation and a spatial extension which, if not normal is at least closer to the norm, is henceforth established. The authors emphasize that just as the anomaly in the topographical representation of the cortex is the reflection of both a structural and a functional abnormality of the hand (‘consistent with a lifetime of abnormal hand function’ Mogilner et al., 1993, p. 3597), its correction is not to be imputed mechanically to the simple surgical separation of the fingers. On the contrary, ‘(the post-operative topography) clearly reflects the new functional status of the hand, in that each digit was now available to function independently, and a different image was generated in the patient, who felt the fingers as individual entities for the first time in his life’ (ibid., p. 3596).

Amputations. From an anatomical point of view, the most serious deafferentation is the amputation of a limb, to the extent that it gives rise to a retrograde atrophy of the nerve paths as well as to an anarchic axonal regeneration in the stump. However, a body whose hand or arm has been amputated is not simply a body without a hand or an arm; it is a body which has organised its functioning along alternative lines. The existence of just such a functional reorganisation has been demonstrated by TMS of the motor cortex. Applied at random to the positions of the scalp above the sensori-motor areas, this stimulation arouses motor potentials in the muscles which make it possible to establish an approximative motor cortex mapping by way of a transfer on the map of the corresponding excitable positions. The electromyographic recording of the biceps and deltoids show that the muscles of the shoulder and the arm on the same side as the stump are activated from a larger number of positions than the muscles on the opposite side, that the motor potentials recorded in he muscles are more important on that side, and that they are aroused at thresholds of stimulation intensity inferior to those on the other side. The authors interpret this result as proof for an expansion of the cortical representations of the muscles ipsilateral to the amputated arm. This implies, in comparison to the representations of the contralateral muscles, implying a larger percentage of motor neurons enlisted by the former muscles, an expansion which attests to a general reorientation of the motor pathways in response to the disequilibrium suffered by the motor system as a result of the amputation. Therefore, the character of the motor output map is not static, but dynamic, and more generally, of the organising schemata in the cerebral cortex of the adult are dynamic (Cohen et al.,1991).

II. In a series of later experiments closer to ecological conditions, monkeys were trained to fetch grains or food pellets placed in holes of various size in a Klüver Board[11]. This training, carried out for several hours per day and over several weeks, brought to light the ability of monkeys to discover and to finalise in stereotypical form an efficient strategy mobilising, in one rapid and supple movement, a minimum number of fingers. This progress in dexterity is reproduced topographically at the level of the cortex 3b by a representational magnification which moves, this time without ambiguity, in the direction of an increase in the spatial resolution of the cutaneous sensitivity.This is because the extension of the cortical representation of the end of the fingers utilised in the course of the period of training is multiplied by two, while that of the receptor fields of those neurons dealing with the representation of the same fingers is divided by two. Instead of bringing with it any disorganisation, this expansion occurs in a topologically ordered fashion and is accompanied by a diminution of the overlapping of receptor fields. Here, and by emphasizing the fact that this representational change induced by practise in one primary sensorial region is framed within a group of parallel changes in other sensori-motor regions, the authors are more emphatic in attributing to the latter a share of the responsibility in the acquisition of this new motor skill (Xerri et al., 1999).

III. A group of subjects suffering from congenital or accidental blindness and who have been practising Braille reading for at least two hours daily since childhood by using the right index finger alone were subjected to a double series of experiments. 1) A repetitive electrical stimulation applied to the index reading finger by means of electrodes resembling the keys of a Braille keyboard evokes electrical potentials recorded by means of a closed grid of electrodes placed on the scalp above the left sensori-motor cortex (contra-lateral). The analysis of the wave recorded in its typical components yields a peak of negative potential (N20) on the parietal cortex followed by a peak of positive potential (P22) whose pre- or post central localisation is not defined. Once these peaks have been reported for each subject on a topographical map constructed on the basis of the sites recorded, it can be established that their distribution corresponds to a somatotopological organisation which extends to a larger surface when the stimulated finger is the index reading finger with a blind person than when it is the left index or even the right or left index with a seeing person. 2) The transcranial stimulation (TMS) interferes in a relatively focalised way with the functioning of the sensori-motor cortex: applied 50 msecs after the electrical stimulation of the index reading finger the TMS blocks the detection of this stimulus more frequently than when it is applied at the end of a stimulation of the left index finger, and this blockage arises when the magnetic stimulator is placed above positions of the scalp which are more numerous than in the other condition. The convergence of these two results confirms an expansion of the functional representation of the reading finger with blind readers in Braille, an expansion which can no doubt be imputed to the intensity and to the selective character of the sensorial stimulation imposed upon this finger by the rapid and repetitive movements of the tactile detection of letters in Braille (Pascual-Leone et Torres, 1993). A later MEG study has extended these results to blind readers in Braille who habitually employ three fingers of each hand and so verifies a duplication of the extension of the cortical representation of the hand with an increase in the distance between localisations of the representations of the fingers, a change which might be associated with a more intensive use. All the same, this expansion is accompanied by a disorganisation of the somatotopological topography of the representation of the fingers, which no longer follow one another in the normal latero-medial order, a disorganisation which could be related to a difficulty in identifying the finger subject to a minimal tactile stimulation arising from the determination of the threshold of the sensorial sensitivity of these fingers. This observation poses the question of the adaptive value of the functional plasticity induced by use (Sterr et al., 1998).

IV. The plasticity of the functional topography of the motor cortex linked to a motor learning process has been demonstrated in the monkey by a comparison between maps of the representation of movements of the hand and the arm obtained by intracortical electrical microstimulation (IMS) before and after the learning of tasks requiring a movement of the fingers and not of the arm, or a movement of the wrist and of the forearm and not the fingers. The monkey’s ability to acquire an efficient and stereotypic procedure for recuperating a food pellet placed in the smallest of four holes in a Klüver Board has been controlled. In order to oblige the animal to complete successions of supinations – pronations of the forearm, the monkey is trained to turn a key to obtain the food pellets in a plexiglass cylinder which limits the movements around the elbow and shoulder. Cartography under IMS practised on the anaesthetised animal consists in varying the electrical current sent into each implanted microelectrode from 0 to 30 microAmperes (more than 300 implantations on average at 250 micron intervals) in order to determine the threshold of evocation of a barely visible movement (or of more than one). This affords a precise identification of the loci in which manual movements acquired during the learning process are evoked. On the basis of this mosaic of labelled loci, a computer is used to differentiate the regions containing the loci where the stimulus evokes similar movements and, from there, to define the frontiers and to measure the relative surfaces of the representations of the different categories of movements, or of their combinations, and this in order to compare the motor maps before and after the training. The modulatory influence of the type of training is shown by a variation of the representations of the movements when these movements are solicited by the nature of the task to be made in the reverse direction from one condition to another. With animals trained to make finger movements, one notes an extension of representation of the forearm and that of the combined mouvements (flexion of the fingers + extension of the wrist), but a reduction of the representation of the abduction of the wrist. With animals trained for movements of the wrist, one notes a reduction of the surface of representation of the extension/flexion of the fingers but an extension of the surface of representation of the supination of the forearm and of abduction of the wrist (Nudo et al., 1996).

V. If the anatomical architecture of Edelman’s ‘neuronal network’ is initially fixed, the functional properties attributed to the ‘synapses’ are not, but change as a function of the ‘cutaneous’ stimulation, on the one hand, and of the equilibrium established by the exciting and inhibiting influences that ‘the cells’ exert on each other, on the other hand. The operative concept is that of the ‘neuronal group’. Neuronal groups are not anatomical entities but rather purely functional entities, stabilised patterns of cellular activations distributed throughout the network. Their process of formation depends uniquely upon the flux of stimulation of the captors and on the local equilibrium between excitation and inhibition. By hypothesis, the network is deprived of initial organisation, the connections between cells being left to chance. In accordance with the theory of ‘selection of neuronal groups’, three principles of synaptic functioning make possible the ‘spontaneous’ emergence of these neuronal groups. 1) The mutual overlapping of the divergent ‘thalamo-cortical connections’ which take over the stimulation from the captors to the cortex and the intersection of short, excitatory and long, inhibitory cortico-cortical connections conjoin their effects. The result is the initiation of a tendency toward the local confinement of the activations in the network. 2) ‘Selection’: the neuronal groups whose activation is more powerful stabilise their internal connections and refine their receptor fields (which are focalised sometimes on the ‘palm’ sometimes on ‘the back of the hand’) while the weaker ones tend to dissolve 3) On the borders between groups in the course of differentiation, intervening cells form groups which compete with each other, as a result of which a more precise determination of their mutual frontiers becomes possible. The organisation of a network of neurons on the unique and exclusive basis of these three principles results in neuronal groups whose behaviour simulates some observed functional plasticity phenomena of the somatotopic maps of the hand : their expansion under the impact of an abnormal stimulation of a finger, retraction and substitution of the RFs of the palm by RFs on the back of the hand in response to a deafferentation of the median nerve.

Bibliography

Barlow HB, The Quarterly Journal of Experimental Psychology, 37A, 121-145, 1985.

Barlow HB, The neuron doctrine in perception, Chap. 26 in Gazzaniga MS, ed., The Cognitive Neurosciences, MIT, Cambridge Mss, 1995.

Berthoz A, La Décision, Odile Jacob, Paris, 2002.

Braun C, Schweizer R, Elbert T, Birbaumer N, Taub E, The Journal of Neurosciences, 20(1), 446-450, 2000.

Buchner H, Kauert C, Rademacher I, Neuroscience Letters, 198, 57-59, 1995.

Buchner H, Reinartz U, Waberski TD, Gobbelé R, Noppeney U, Scherg M, Neuroscience Letters, 260, 57-60, 1999.

Cohen LG, Bandinelli S, Findley TW, Halett M, Brain, 114, 615-627, 1991.

Fox PT, Burton H, Raichle ME, The Journal of Neurosurgery, 67, 34-43, 1987.

Grafton ST, Woods RP, Mazziotta JC, Phelps ME, The Journal of Neurophysiology, 66, 735-743, 1991.

Kaas JH, Merzenich MM, Killackey HP, Annual Review of Neuroscience, 6, 325-356, 1983.

Karni A, Meyer G, Jezzard, Adams MM, Turner R, Ungerleider LG, Nature, 377, 155-158, 1995.

Karni A, Meyer G, Rey-Hipolito C, Jezzard P, Adams MM, Turner R, Ungerleider LG, Proceedings of the National Academy of Sciences, 95, 861-868, 1998.

Mogilner A, Grossman JAI, Ribary U, Joliot M, Volkmann J, Rapaport D, Beasley RW, Llinas RR, Proceedings of the National Academy of Sciences, 90, 3593-3597, 1993.

Merzenich MM, Nelson RJ, Kaas JH, Stryker MP, Jenkins WM, Zook JM, Cynader MS, Schoppmann A, The Journal of Comparative Neurology, 258, 281-296, 1987.

Merzenich MM, deCharms RC, in Mind and Brain, Llinas R, Churchland P, eds., 61-81, MIT, Cambridge, MA, 1995.

Pascual-Leone A, Torres F, Brain, 116, 39-52, 1993.

Elbert T, Candia V, Altenmüller E, Rau H, Sterr A, Rockstroh B, Pantev C, Taub E, NeuroReport, 9, 3571-3575, 1998.

Elbert T, Pantev C, Wienbruch C, Rockstroh B, Taub E, Science, 270, 305-307, 1995.

Merzenich MM, Nelson RJ, Stryker MP, Cynader MS, Schoppmann A, Zook JM, The Journal of Comparative Neurology, 224, 591-605, 1984.

Nudo RJ, Milliken GW, Jenkins WM, Merzenich MM, The Journal of Neurosciences, 16(2), 785-807, 1996.

Penfield W, Boldrey E, Brain, 37, 389-443, 1937.

Penfield W, Rasmussen T, The Cerebral Cortex of Man, New York, MacMillan, 1950.

Pascual-Leone A, Grafman J, Hallett M, Science, 263, 1287-1289, 1994.

Pascual-Leone A, Tarazona F, Keenan J, Tormos JM, Hamilton R, Catala MD, Neuropsychologia, 37, 2076-217, 1999.

Pearson JC, Finkel LH, Edelman GM, The Journal of Neurosciences, 7(12), 4209-4223, 1987.

Rausell E, Jones EG, The Journal of Neurosciences, 2995, 15(6), 4270-4288.

Recanzone GH, Merzenich MM, Jenkins WM, Grajski KA, Dinse HR, The Journal of Neurophysiology, 67, 1831-1055, 1992.

Reddy H, Floyer A, Donaghy M, Matthews PM, Experimental Brain Research, 138, 484-491, 2001.

Schlaug G, Knorr U, Seitz R, Experimental Brain Research, 98, 523-534, 1994.

Schieber MH, TINS, 13, 11, 440-445, 1990.

Sterr A, Müller E, Elbert T, Rockstroh B, Pantev C, Taub E, The Journal of Neurosciences, 18(1), 4417-4423, 1998.

Wall JT, Kaas JH, Sur M, Nelson RJ, Felleman DJ, Merzenich MM, The Journal of Neurosciences, 6(1), 218-283, 1986.

Wall JT, Felleman DJ, Kaas JH, Science, 221, 771-773, 1983.

Xerri C, Merzenich MM, Jenkins WM, Santucci S, Cerebral Cortex, 9, 264-276, 1999.