The Brain's Sense of Movement

Alain Berthoz, Giselle Weiss

• Publisher: Harvard University Press
• Year: 2002
• ISBN:
  9780674971103
• Topic: Biological Sciences
• Subtopic: Neuroscience
• Index: Biological Sciences

The neuroscientist Alain Berthoz experimented on Russian astronauts in space to answer these questions: How does weightlessness affect motion? How are motion and three-dimensional space perceived? In this erudite and witty book, Berthoz describes how human beings on earth perceive and control bodily movement. Reviewing a wealth of research in neurophysiology and experimental psychology, he argues for a rethinking of the traditional separation between action and perception, and for the division of perception into five senses.In Berthoz's view, perception and cognition are inherently predictive, functioning to allow us to anticipate the consequences of current or potential actions. The brain acts like a simulator that is constantly inventing models to project onto the changing world, models that are corrected by steady, minute feedback from the world. We move in the direction we are looking, anticipate the trajectory of a falling ball, recover when we stumble, and continually update our own physical position, all thanks to this sense of movement.This interpretation of perception and action allows Berthoz, in The Brain's Sense of Movement, to focus on psychological phenomena largely ignored in standard texts: proprioception and kinaesthesis, the mechanisms that maintain balance and coordinate actions, and basic perceptual and memory processes involved in navigation.

1

PERCEPTION IS SIMULATED ACTION

Our sensations are purely passive, while all our perceptions or ideas are born out of an active principle which judges.

—J.-J. Rousseau

The Motor Theory of Perception

A major theme of this book is that perception is more than just the interpretation of sensory messages. Perception is constrained by action; it is an internal simulation of action. It is judgment and decision making, and it is anticipation of the consequences of action. This is not an entirely new concept, as a glimpse at some of the ideas that prefigured it will show. There are so many theories of perception that I will not try to cover them all here. I wish simply to provide a taste of what is in store.

A history of the motor theory of perception was recently published by Viviani, who recalls that it was very fashionable before 1940.¹ One of the first modern works to signal the importance of movement in perception was that of Lotze in 1852.² His theory affirmed that spatial organization of visual sensations results from their integration with a muscular sense. The idea that the information that triggers a motor command is used by the brain to recognize movement was proposed by Helmholtz.³ For him, motor control operates by comparing sensations with predictions based on the motor command. In 1890, William James also described a neural circuit that anticipates the sensory consequences of movement (Figure 1.1).⁴ A simple way of conceptualizing the perception of movement, he proposed, would be to assume that a sensory cell S is excited and activates a motor neuron M that induces a muscle contraction. A kinesthetic cell K detects the movement and modifies the motor neuron M. James imagined an additional circuit: a collateral axon of cell S activates the kinesthetic cell K at the same time that the motor neuron M is activated. K is thus activated even before it receives information about the muscle movement, which allows it to anticipate the consequences of movement.

Figure 1.1. William James’s concept of an anticipatory neuronal pathway.

In France, the ideas of Janet were very close to those of the pioneers of the motor theory of perception. He established a hierarchy of actions. The most elementary are reflex actions, followed by actions that are “perceptual, social, simple intellectual; and at the level of verbal and volitional action, thoughtful, rational, experimental, progressive.”⁵ He wrote: “Reflex action is impulsive action as opposed to the considered action that characterizes perceptual behaviors.”⁶ In other words, perception is interrupted action, and especially goal-oriented action. “In all these perceptual actions, the starting point is a complex object, like prey or its burrow, and the action leads to a transformation and use of the object: the action adapts to the object, rather than simply following from superficial stimuli.”⁷ But Janet went further and suggested that perceptual action was predictive: “The action that is triggered by the initial stimulus adapts not only to this stimulus but to all the others that the object induces successively; it adapts to stimuli that do not yet exist, but that will only occur later owing to the action itself. This adaptation to an aggregate of future and potential stimuli is characteristic of perceptual behaviors.”⁸

Thus the dissociation between perception and action must be discarded. Perception is simulated action. A further example, also borrowed from Janet, illustrates this point in a way that is very close to my own interpretation: “When we perceive an object, an armchair, for example, we do not see ourselves as acting at that instant, because we are still standing, unmoving, perceiving the armchair. This is an illusion. In reality, we already have within us the action we associate with the armchair, which we call a perceptual schema—here, the action of sitting in a particular way in the armchair.”⁹ Merleau-Ponty put it very well: “Vision is the brain’s way of touching.”¹⁰

Research on the physiological bases of the relationship between perception and action nevertheless remained limited until just recently. Viviani attributes the post-1940 eclipse of the motor theory of perception to the appearance of the analytical neurophysiology of Sherrington, the influence of Gestalt, and even to the constructivism of Piaget. The compartmentalization of disciplines whose efforts should converge, such as biomechanics, experimental and cognitive psychology, psychophysics, functional neurobiology, and so on, were for a long time another obstacle that today the cognitive sciences are attempting to overcome. After 1950, the motor theory of perception was revived. For example, consider the work of Lashley, Gibson, the Teuber school, and more particularly the work of Held and Hein on the role of action (not just activity) in the development of the visual system. Or consider the work of the psychologist Johansson in Scandinavia and that of Fessard and Piéron in France—Piéron maintaining that perception is awareness of external objects and events that give rise to sensations. In the 1970s the research groups of Imbert and Jeannerod made major discoveries concerning the development and functioning of the visual system and its relationship to the vestibular system, the control of ocular movements and posture.

The Concept of Acceptor of the Results of Action

Though underrated in the West, the Russian Anokhin’s seminal ideas made him a trailblazer.¹¹ In his day, Pavlov’s theories and experiments on conditioned reflexes, which profoundly influenced our century, were predominant. But Anokhin found Pavlov’s definition of reflexes too limiting. He also criticized Descartes for having sidestepped the question of the relevance of the reflex response to the higher brain and thereby influenced future studies of the complex adaptive actions of animals and man for years to come. Anokhin introduced a key idea to the theory; that at the end of the reflex arc there is a reflex action. Although Pavlov’s idea was already old, Western physiologists ignored subsequent contributions of Russian physiologists for fifty years, except to plagiarize their ideas without citing them—intellectual petty thievery that the isolation of the Russians helped to disguise. Western physiologists used the word “response” to refer to the effect of a stimulus in a reflex arc, whereas the Soviet literature emphasized and still emphasizes the concept of reflex action. The crucial difference is that if every reflex requires a complete action, this involves the entire organism and consequently its faculties of invention, creativity, and adaptation.

Anokhin provided no experimental proof, but he constructed a theory he called “acceptor of the results of action” that is worthy of comment. He began with the following line of questioning: If the result of reflex activity is an action, doesn’t execution of this action need to be approved one way or another by some configuration of sensory afferent information? And doesn’t execution of the action also need to be compared with a predicted configuration?

Say that we wish to pick up a cup from a table cluttered with dishes but that, just when we are about to grasp the cup, we are distracted and pick up a pitcher in its place. As we all know from personal experience, in general we correct an error like this one immediately. But what physiological mechanism permits us to notice our error and to correct it? The appearance of the pitcher and the gripping of its handle as well as the appearance of the cup and the gripping of its handle are just an aggregate of afferent signals that differ by a few constituents. Why is it, then, that the latter afferentation is precisely the one we choose to sanction our action?

According to Anokhin, our move with the pitcher was initially satisfactory because the set of sensory signals constituting its grasp matched a configuration that was predicted, expected, specified before making the gesture. Mathematicians would say that the sets of afferents corresponding to the cup and the pitcher contain an intersection sufficient to be allowed by Anokhin’s acceptor of action. He states that this ready-made excitatory complex, which precedes the reflex action, must in some way be an afferent control apparatus that determines how closely the return afferentation from the central nervous system corresponds to it.

Thus, Anokhin developed a concept equivalent to what is now called an internal model of a group of preselected elements. He was careful not to use the word “representation.” He referred to Pavlov’s earlier observation that the chemical composition of saliva from a conditioned dog is related precisely to the kind of food used for reinforcement and thus to the character of salivation.

Anokhin next turned to the neural basis of this concept. He used the expression “acceptor of the results of action” to refer to a cortical system specialized in the analysis of complex afferents (sensory information) resulting from reflex action. This analyzer determines how the afferent inputs it receives relate to the planned action as a function of the past experience of the animal. Noting that he could as well have called his apparatus “the acceptor of the afferent results of a completed reflex action,” Anokhin explained that he had chosen the word “acceptor,” from the Latin acceptare, because it combined two ideas: accept and approve. Consequently, he incorporated a clear role for the acceptor of the result of action in decision processes. For example, if a person sitting in the living room decides for one reason or another to go into the dining room, at the precise moment of the decision, the set of afferents from all the stimulatory cues he has received in the dining room in the past (acceptor of action) is reproduced in his cerebral cortex. If, after the person has entered the dining room, the cues coincide perfectly with what the acceptor of the result of action predicts, the person then moves on to the next element of intended behavior. But if the acceptor of the result of action detects an error, an incongruity with what is predicted, the brain produces an orientation reaction, as described in the Soviet literature—that is, it reacts by analyzing new events.

Bernstein’s Comparator

Another master of modern physiology who had an even more profound influence on our generation was the Russian physiologist Bernstein. Persecuted for his overly original ideas in the time of Pavlovian hegemony, he studied natural movement and inferred general rules of cerebral functioning from it.¹² To avoid a naïve linear description of the regulation of the coordination of movement as a succession of phases—prediction, preparation, execution, and control—Bernstein proposed a circular schema (Figure 1.2) that introduced the concept of the action-perception cycle. The basic element is a “comparator” that establishes the “required value.”

This required value fulfils at least three different functions, all equally important. The first is detecting an error between accomplished movement and predicted movement that triggers a correction (from a cybernetic vantage). The second is recognizing that an action has been accomplished, which allows progression to the next action in sequence. “This aspect of function,” says Bernstein, “is mainly reminiscent of what Anokhin has termed ‘sanctional afferentation.’” The third function is the adaptation itself. Indeed, when action encounters surprises, it is impossible, or irrelevant, for corrective impulses to reestablish the initial plan of action. In this case, the receptor of information acts as an initiator (and not as a regulator) of adaptive changes in the program being executed. It does this by introducing either small technical changes into the movement or taking an adjacent trajectory. And it does it until the program has been completely reorganized, even down to the repertoire of consecutive elements and the staging of motor action—in other words, it adopts a new tactical approach to the task.

Figure 1.2. Schematic diagram showing the cerebral organization for control of movement proposed by Bernstein.

I think that the highest cognitive functions are the result of an evolutionary thrust toward developing this ability to reorganize action according to unforeseen events. This ability requires developing a memory of the past, the faculty for predicting and simulating the future, and the metafaculty, in a way, to mobilize all these capabilities rapidly, because they must integrate with a perception-action cycle that sometimes lasts only a tenth or twentieth of a second.

For Bernstein, these corrective processes depend heavily on what he calls the comparator. This neural device occupies a strategic position between the information supplied by the receptors and the elements that will effect the necessary corrections or reorganization. It does not function between two successive or simultaneous receptors to compare two distinct events, but between flowing, continuous reception and an internal guide.

An important property of the comparator is its capacity to detect variations in sensory information owing to the central nervous system’s use of fresh traces. Our bodies, says Bernstein, have no receptive apparatus capable of perceiving velocity directly. This task is resolved in the central nervous system by the comparator. It instantaneously compares the cues about the position of the moving organ with the fresh trace of its position approximately 0.1 second before. The brain thus recognizes two positions with a certain interval of time between the...