Sleep

1. Historical introduction

A little more than half a century has passed since C. von Economo (v., 1918) reported that in epidemic lethargic encephalitis, or sleeping sickness, the lesions are localized above all in the region of the midbrain that surrounds the Silvio aqueduct and in the 'hypothalamus. The interpretation he gave of his observations (see Von Economo, 1929) was wrong, but he was certainly right when he thought that the nervous structures damaged by the epidemic disease had critical importance in regulating the sleep-wake cycle.

Not many years later WR Hess (v., 1927) found that sleep could be obtained, in the cat without narcosis and free in its movements, with appropriate electrical stimulations of the medial region of the thalamus. Despite the possibility of errors (see chapter 4), this line of research had a profound influence on the development of sleep physiology. It was the first demonstration that physiological sleep can be achieved with localized electrical stimulation of the brain, and the results needed to be repeatedly confirmed by stimulating other regions of the diencephalon and brainstem. Furthermore, Hess's experiments paved the way for a fruitful line of research in which the study of behavior was associated with different types of electrophysiological stimulation and recording. Of course, it was impossible to locate, in those times, any relationship between von Economo's observations and Hess's experiments. Everything felt so different: case reports and animal experiments, injuries and stimulations, lethargy, and physiological sleep.

The discovery by H. Berger (v., 1929) that the electrical activity of the cerebral cortex could be recorded, in man, through cranial integuments (electroencephalogram, EEG) opened the era of electrophysiological studies in man and animals in the absence of narcosis. Berger himself made a major contribution to sleep studies when he showed that the human EEG exhibits slow, large voltage oscillations during this state, thus eliminating from discussion the naïve assumption that higher centers should be silent when there is no consciousness. He also demonstrated that this type of electrical activity is markedly different from the rapid low-voltage waves (β-waves) that characterize active wakefulness; and also by the slow oscillations of potential at 10/s (α waves) that appear during the relaxed vigil. Finally, Berger found that both α-waves and sleep rhythms could be abolished by sensitive or sensory stimulation and replaced by β-waves: two phenomena designated, respectively, Berger's arrest reaction and electroencephalographic awakening. All these effects were generalized and therefore not limited to the cortical area corresponding to the stimulated sensory modality. Their abrupt onset then prevented the recognition of links with the tonic phenomena responsible for maintaining vigil. It will be up to ED Adrian (v., 1934) to explain Berger's observations on the basis of models of activity of cortical neurons. He suggested the

Two lines of research arose from these early works. The line of 'phenomenological investigation', essentially based on observations, became extremely fruitful with the introduction of electroencephalographic and microelectrode recording techniques in Mammals without narcosis, free in their movements. It led to the demonstration that during sleep, in cats (see Klaue, 1937) and in humans (see Dement and Kleitman, 1957), periods characterized by EEG desynchronization are present and to studies on the relationships between this phase of sleep, called desynchronized sleep or paradox, and dreams ( ibid.). The investigations on the behavior of single units of the cerebral cortex must be situated along the same lines of phenomenological research (see Evarts, 1962 and 1964); and also the electrophysiological works on postural tone (see Jouvet, 1962), on spinal reflexes (see Pompeiano, 1965, 1966 and The neurophysiological ..., 1967) and on sensory transmission during desynchronized sleep (see Pompeiano, Sensory inhibition ..., 1967).

The second line of research, 'experimental investigation', is based on the study of changes in the wake-sleep cycle produced by lesions or by stimulation. Many aspects of this field of investigation have been examined by F. Bremer in the article ascending reticular system. Suffice it to recall here the research on the encéphale isolé and on the cerveau isolé(see Bremer, 1935, 1937, and 1938), the work of G. Moruzzi and HW Magoun (see, 1949) on the ascending lattice system, and the experiments of DB Lindsley et al. (see, 1949 and 1950) on the effects of reticular lesions; and finally the demonstration of the existence in the lower part of the brainstem of a deactivating system, probably hypnogenic (see Moruzzi, 1963, for the literature). The ascending reticular system has made it possible to give a unitary explanation of observations that are apparently unrelated to each other, such as those we referred to in the first part of this introduction.

The most recent experimental developments concern the neurochemical study of the brainstem structures that control the wake-sleep cycle and the electrophysiological investigation of the behavior of single nerve cells of the pons and cerebral cortex.

2. Sleep phenomenology

The study of the neurophysiological and neurochemical mechanisms of the sleep-wake cycle, which will be treated in the following chapters, must be preceded by the description of the phenomena that are observed when no attempt is made to modify sleep or wakefulness or their rhythmic alternation. ‟The name of observer is given to one who applies simple or complex investigation procedures to the study of phenomena which he does not modify and who, consequently, collect them as nature offers him" (see Bernard, 1865, p. 29 This is always the first approach in any field of natural sciences.

Half a century ago the only avenue open to researchers was the study of animal behavior and this line of attack is still followed in much research in comparative physiology and in a field of the natural sciences, ethology. Animal behavior is mainly based on movements or positions, therefore on the phasic and tonic activity of skeletal muscles and the corresponding motor neurons. Modern developments in electrophysiological techniques have made it possible to study the behavior not only of large populations of cerebral neurons (electroencephalography), but also of the activity of single nerve cells in the animal without narcosis, free in its movements. However, the aim of the research is always the same, whether we study the behavior of muscle fibers or that of brain neurons. We simply observe what is happening, making no attempt to influence the physiological sleep-wake cycle. Naturally some experimentation, in the strict sense of the word (see chapter 3), is also inevitable in this type of research. Indeed, it can be useful to specifically ask nature a few questions, as we do, for example, when we observe the effects of sensitive or sensory stimuli. The fundamental point, however, is that there is no attempt to modify the physiological rhythm of the alternation between sleep and wakefulness. Our aim is only to carefully describe natural phenomena and their relationships over time. it is unavoidable even in this type of research. Indeed, it can be useful to specifically ask nature a few questions, as we do, for example, when we observe the effects of sensitive or sensory stimuli. The fundamental point, however, is that there is no attempt to modify the physiological rhythm of the alternation between sleep and wakefulness. Our aim is only to carefully describe natural phenomena and their relationships over time. it is unavoidable even in this type of research. Indeed, it can be useful to specifically ask nature a few questions, as we do, for example, when we observe the effects of sensitive or sensory stimuli. The fundamental point, however, is that there is no attempt to modify the physiological rhythm of the alternation between sleep and wakefulness. Our aim is only to carefully describe natural phenomena and their relationships over time.

NERVOUS SYSTEM

General neuron physiology. - The possibility of recording even the smallest bioelectrical manifestations without deformation has led to an unprecedented advance in our knowledge of neuron physiology.

It is appropriate to summarize what were the classical conceptions of neurophysiologists. By now accepted by all the theories of the neuron, it was believed that on the surface of its morphological constituents - the gynophore, the dendrites, the neurite (see nervous, tissue; XXIV, p. 659) - a membrane separated the inside from the outside of the nerve cell or its extensions. This membrane had never been directly observed, but its existence was postulated as the most likely explanation of the different concentrations of ions, and therefore of the potential difference, existing between the inside and the outside of the neuron. It was recognized that this membrane polarization characterized the cell or nerve fiber at rest. It was seen as the cause of the demarcation potential. It was then believed that the onset of the action potential, which accompanies the activity of the neuron and the propagation along the cylinder axis of the nerve impulse, was due to the disappearance of this polarization (depolarization of the membrane). These notions (cf.; App. II, 1, p. 837) have certainly been profoundly modified in the last quarter of a century, but it is a reason for admiration for the great electrophysiologists of the past to note that the central nucleus of their doctrine has stood the test of time well.

The surface membrane of the neuron.

 - Electron microscopy (see histology, nervous tissue, in this App.) revealed that a membrane of uniform thickness (about 50 Å) separates the inside from the outside of the neuron. It is probably formed by a bimolecular layer of phospholipids and cholesterol, supported by a protein skeleton. The whole general physiology of the neuron hinges on the chemical-physical properties of this membrane. Through the use of glass ultramicroelectrodes, which allow entering the interior of the neuron without killing it, JC Eccles and his collaborators have been able to demonstrate that the membrane potential of the intracellular electrode with respect to an indifferent external electrode is of the order of −60 to −80 mV. The membrane potential, whereby the inside of the neuron is electronegative with respect to the outside, it is a general property of all excitable tissues. The polarization of the surface membrane is the cause of the demarcation potential of classical physiology. The membrane represents a barrier to the free diffusion of ions, because it has a low ionic permeability. The specific resistance and the specific capacitance of the membrane are very high: of the order of 500-1000 Ω/cm respectively2 and 3μF/cm 2 , for mammalian motor neurons.

Origin of the membrane potential. - When the passage of ions through the membrane occurs by simple diffusion, it is regulated i) by the permeability of the membrane and ii) by the electrochemical potential gradient. This gradient is zero when the membrane potential is equal to the equilibrium potential for that particular ion. Knowing the concentration of the ion inside and outside the cell we can, as we shall see below, calculate the equilibrium potential; on the other hand, knowing the membrane potential, we can determine for each ion the concentration gradient between the outside and inside of the cell for which there is electrochemical equilibrium. Therefore, if we find that the equilibrium potential of a given ion is different from the membrane potential, we must conclude that the cell or the nerve fiber actively intervenes, producing work and consuming energy, to maintain a distribution that is different from that allowed by the electrochemical balance. It is then admitted that for that particular ion, there is a "pump" which ensures an "uphill" flow of ions, i.e. in the opposite direction to that which would be given by the electrochemical gradients. In summary, a different distribution of ions inside and outside the cell can be the expression of a chemical-physical equilibrium, such as one could have on the two sides of an inert membrane (provided it is electrically polarised), if the gradient electrochemical is zero, but if it is different from zero we must admit the intervention of an active process,

For there to be equilibrium, without any energy consumption by the cell, ECl must be equal to the membrane potential. But when

Thus a membrane potential of −70 mV equilibrium at a concentration gradient for Cl′ corresponding to the ratio

This is precisely the concentration ratio that normally occurs and therefore the electrochemical potential gradient is zero for Cl'. Therefore the difference in concentration in Cl′ is not linked to any active process but is only a consequence of the membrane potential.

Let us now apply the same considerations to the two cations (Na• and K • ), which also appear unequally distributed on the two sides of the membrane. The table shows the data calculated by JC Eccles (1957).

Let's summarize. Concentration imbalances are in part the simple consequence of the existence of the membrane potential, in part they are linked to active metabolic processes, designated under the name of "Na and K pumps". Naturally, the chemical-physical mechanisms which maintain the membrane potential and the chemical reactions which underlie the functioning of the "pumps" remain to be known. It is not possible to answer the second question at the moment. As regards the first, it has been calculated that an excess of anions equal to 2 × 10 -15 within the motor neuron is sufficient to determine a membrane potential of −70 mV. But what is the chemical nature of this anion is unknown to us. We have given the explanation of the membrane potential based on the ion theory of AL Hodgkin and AF Huxley (1952). We will see in the following section that this theory satisfactorily accounts for the action potential as well.

The action potential. 

- According to the theory of J. Bernstein (1902), the action potential was considered the consequence of the transient and localized annulment of the membrane potential, which would have occurred every time a cell or a fiber became active. It has actually been seen that the polarization of the membrane is not canceled, but reversed when the action potential appears (figs. 1, 2). This discovery, made almost simultaneously by AL Hodgkin (1939) and by HJ Curtis and KS Cole (1940), completely changed our conceptions. This is how the birth of the action potential, often referred to as a spike in the Anglo-American electrophysiological literature, is explained today.

The very large electrochemical potential gradient for Na• does not lead to a high flow of ions from the outside (where the concentration is much higher) into the cell or fiber, because the specific membrane conductance for this cation is far lower than that for K • and Cl′ (respectively of the order of 0.01 millimho/cm 2 and 0.5 millimho/cm 2 ). The poor inward flow of Na• is compensated for by the extrusion of ions by the Na• pump. The situation changes dramatically when the action potential arises. The conductivity for Na• increases dramatically, rising to about 15 millimho/cm 2. This leads to massive entry of Na• into the cell or fiber. The process takes about 1 msec. The entry of positive charges first cancels the polarization of the membrane, which reaches zero potential; but the flow continues and the polarization of the membrane is reversed, until almost reaching the equilibrium potential for Na• which is + 50, + 60 mV. This is the reason why the action potential is always greater than the membrane potential. The specific conductivity for K • also increases, but with a delay, and when this happens, there is an increased outflow of K • , also here according to the electrochemical gradient, i.e. from inside (where the concentration of K • is more elevated) outdoors. L'• initiates the descent of the action potential and that set of processes which will then lead to the re-establishment of the membrane potential and a return to the resting distribution of the ions. This refreshment process is related to the activity of the Na and K pump and, indirectly, to the metabolic activity of the cell or nerve.

The propagation of the nerve impulse. 

- When the action potential appears in a point of a nerve fiber, an electrode placed outside the membrane behaves as electronegative with respect to an electrode placed in the same position, but in a stretch of nerve at rest. On the active point it is as if a cathode connected to a direct current source had been placed. In both cases, positive charges leave the adjacent stretch of nerve at rest. In front of the action potential, the membrane of the fiber section still at rest is therefore depolarized, according to the classic mechanism of the electrotone. When the depolarization reaches a critical level, the specific Na• conductivity increases. Then sharply increases the flow of Na• towards the interior of the fiber, and this leads to a further depolarization of the membrane; this in turn increases the conductivity to Na•. This is how the self-regenerative process takes place which leads to the birth of the nerve impulse, i.e. to the inversion of the depolarization of the membrane in the stretch of fiber located immediately downstream, which was at rest a moment before. The reason for this irresistible march forward is that the action potential leaves behind a refractory-phase fiber so that under normal conditions conduction can only occur in one direction. Nerve conduction is therefore linked to the fact that the action potential represents an electrical stimulus for the downstream nerve section (fig. 3). that is, to the inversion of the depolarization of the membrane in the stretch of fiber located immediately downstream, which was at rest an instant before. The reason for this irresistible march forward is that the action potential leaves behind a refractory-phase fiber so that under normal conditions conduction can only occur in one direction. Nerve conduction is therefore linked to the fact that the action potential represents an electrical stimulus for the downstream nerve section (fig. 3). that is, to the inversion of the depolarization of the membrane in the stretch of fiber located immediately downstream, which was at rest an instant before. The reason for this irresistible march forward is that the action potential leaves behind a refractory-phase fiber, so that under normal conditions conduction can only occur in one direction. Nerve conduction is therefore linked to the fact that the action potential represents an electrical stimulus for the downstream nerve section (fig. 3). so under normal conditions conduction can occur in one direction only. Nerve conduction is therefore linked to the fact that the action potential represents an electrical stimulus for the downstream nerve section (fig. 3). so under normal conditions conduction can occur in one direction only. Nerve conduction is therefore linked to the fact that the action potential represents an electrical stimulus for the downstream nerve section (fig. 3).

The saltatory theory applies only to the myelinated peripheral fibers of Vertebrates, in which the myelin sheath is interrupted at regular intervals by the nodes of Ranvier (see nervous, tissue, XXIV, p. 663). The theory, which has the support of many experimental tests (R. Stämpfli, 1952; I. Tasaki, 1959), assumes that the internodal tract has a purely passive function in the propagation of the impulse. When the action potential appears in one node of Ranvier, the next node is depolarized, because positive charges carry themselves to the active node, flowing out of the myelin sheath that lines the internode. When the depolarization in the second node reaches a critical level, a new action potential arises, which will exert the same influence on a third node of Ranvier. In this way the nerve impulse propagates, almost "jumping" from node to node (fig. 4). This process has a considerable margin of safety: it has been shown that an impulse can "jump"

The synapse and the chemical mediation of the nerve impulse

. - The neuron theory assumes the discontinuity between individual neurons. These meet at well-circumscribed points, called synapses. The existence of synapses has been confirmed by electron microscopy. At the synapse, the membranes of two nerve cells - called the presynaptic membrane and the postsynaptic membrane, with reference to the direction of the transmission of the nerve impulse - come into contact, being separated by a granular space 50-500 Å thick. Within the presynaptic membrane are small vesicles of 300-500 Å in diameter, which are assumed to contain the chemical mediator.

The existence of a chemical mediator in central synapses had already been postulated by analogy with the behavior of peripheral synapses. JC Eccles deems this hypothesis necessary in view of the fact that the presynaptic impulse does not generate a flow of ions sufficient to produce a critical depolarization of the postsynaptic membrane. It, therefore, seems that the presynaptic impulse, having reached the last endings, frees the chemical mediator from an inactive precursor. This alters the ionic permeability of the postsynaptic membrane, giving rise to a new nerve impulse. The nature of the chemical mediator is unknown for most central synapses.