Chapter 4

Let us begin our discussion of the neurophysiology of sensation by considering what happens in a receptor when a stimulus is applied to it. In later chapters individual sensory receptors for each sense will be considered separately, but at this point, it is the general properties of receptors that are of concern. We have already seen that when a suprathreshold stimulus (of strength greater than threshold strength) is applied to a receptor, the nerve fiber associated with that receptor discharges. What we would like to know is how the stimulus energy is transformed into action potentials in the normal operation of the system, i.e., what is the nature of the transduction process (the process of converting stimulus energy into spikes)?

Receptor potentials

Fig. 4-1. Relation between impulse response and underlying generator potential. A. Response of a muscle spindle receptor to prolonged stretch, with nerve recording shown in the upper trace and monitor of the stretch of the muscle in the lower trace. B. The same response after bathing the spindle in 0.2% lidocaine (Ottoson D and Shepherd GM: Acta physiol scand 79:423-430, 1970).

The transduction process appears to be similar in all of the various types of receptors that have been studied. As an example of a transducer, let us consider the muscle spindle, a small structure found in skeletal muscle that contains two kinds of receptors both of which can signal the length of the muscle. The anatomy and physiology of the muscle spindle will be considered in more detail in Chapter 11. If a recording electrode is placed on the sensory nerve fiber supplying one of the receptors in the muscle spindle (the primary receptor) at a point very near the spindle itself and then the spindle is stretched, recordings like those in Figure 4-1 result. The upper trace of each pair shows the recording from the fiber, whereas the lower trace shows the monitor of the muscle length (longer being indicated by an upward deflection). In Figure 4-1A, the spike discharge is superimposed upon a slow hypopolarizing potential that develops as the muscle is stretched. If the preparation is bathed in a solution of lidocaine, a local anesthetic agent, or in tetrodotoxin, the action potentials are blocked as in Figure 4-1B, leaving the slow hypopolarizing potential by itself. This potential, called the generator potential, can be shown to originate in part of the spindle receptor, not in the nerve fiber. There are receptors that are actually nerve fibers, part of which has been specialized to be sensitive to stimuli (e.g., receptors in the olfactory bulb and free nerve endings in skin), and there are receptors that are other types of cells in close association with nerve fibers (e.g., Merkel's disks in skin and sensory cells in taste buds). In the latter case, there appears to be a synaptic connection between the nonnerve cell and the nerve cell (the receptor), and the nerve fibers are usually not very sensitive to the same stimulus that excites their receptors.

Some people use the term generator potential synonymously with receptor potential, whereas others prefer to reserve the term receptor potential for those cases where the receptor does not itself generate spikes, as in the receptors of the eye. We will use the term receptor potential here in this latter sense.

Fig. 4-2. Graded responses of a muscle spindle receptor to stretch. Graded stretches are indicated by the stretch monitor in the lower traces; graded generator potentials are shown in the upper traces (Ottoson D and Shepherd GM: Cold Spring Harbor Symp Quant Biol 30:105-114, 1965).

The generator and receptor potentials are local or non-propagated events as shown by the degradation of their amplitudes with increasing distance from the receptor. One need only move the recording electrode along the nerve a few millimeters farther away from the position near the spindle that produced the records in Figure 4-1 before the generator potentials disappear. They are graded responses as opposed to the all-or-none character of the action potential; the amplitude of the generator and receptor potentials increase with increasing stimulus strength. Different amounts of muscle stretch, as shown by the heights of the indicators superimposed in the lower trace, resulted in the graded series of generator potentials superimposed in the upper trace of Figure 4-2. For these records, the action potentials have been blocked by application of lidocaine. The graded increases in the generator potential amplitude with graded stretches of the muscle are clearly illustrated in this figure.

The channels opened during the generator potential are non-specific; they admit all small ions.

The steps in formation of a generator potential are not known for every receptor, but where it has been studied the start of the generator potential usually results from an increase in the permeability of the membrane of the receptor to all small ions, but the ion furthest from its electrochemical equilibrium and in greatest concentration, namely sodium, contributes the greatest current. As during an action potential, an increase in sodium conductance leads to an hypopolarization. The ionic mechanism of the generator potential is similar to that for the action potential but with a longer time constant; however, the restoration of the membrane potential to resting values at the end of the generator potential is a passive process not involving increased potassium conductance as in the action potential.

The transduction process takes place within the receptor region of the cell, that region that is sensitive to adequate stimuli and is responsible for the generation of the generator potential. The membrane of the receptor region is, however, electrically inexcitable; it contains no voltage-gated ionic channels and does not generate spikes. If the receptor region generated action potentials, the graded nature of the generator potential would be destroyed because as soon as the generator potential exceeded the critical firing level an action potential would be initiated, reversing the membrane polarization no matter how large or small the stimulus, i.e., the membrane potential would no longer encode the stimulus intensity. The action potentials themselves are usually initiated in a physically separate region of the cell, the pacemaker or spike-generating region. This may be anatomically separate from or continuous with the receptor region or with the axon itself, the conducting region. For receptors that do not themselves generate spikes, the spike-generating regionis in a different cell. For example, receptor potentials occur in the rods and cones of the eye, but the first spikes in the visual system occur in the ganglion cells. Figure 4-3 shows how a receptor can conceptually be divided into a receptor region, a spike-generating region and a conducting region. As illustrated in the figure, the inward (sodium) current that hypopolarized the active (i.e., with changed conductances) receptor region is concomitant with an outward current (possibly carried by potassium ions) in the spike-generating region of the cell. That this must be true follows from Kirchhoff's current law; current flows only in complete circuits. As anywhere in a nerve cell, outward current though an inactive (i.e., with normal resting conductances) region of membrane causes an hypopolarization of that membrane. An hypopolarization, as we already know, leads to the formation of action potentials, provided that the critical firing level is reached or exceeded. Actually, the spike-generating region simply responds to hypopolarizing current in the same way as any other membrane containing voltage-gated Na⁺ and K⁺ channels. The conducting region of the membrane of the sensory receptor simply "follows" what the spike-generating region does; it simply transmits, without alteration, signals it receives.

Fig. 4-3. A schemtatic diagram of a generaized receptor showing the receptor region, the spike generating region, and the conducting region along with the currents that flow during the generator potential.

Fig. 4-4. Relation between amplitude of the generator potential and frequency of impulse discharge during the dynamic phase of stetch [time when muscle and spindle are actually changing length] (Ottoson D and Shepherd GM: Acta physiol scand 79:423-430, 1970).

In the absence of an anesthetic agent, it is possible to study the relationship between the frequency of discharge in the spindle axon and the amplitude of the generator potential during the dynamic phase of stretch (i.e., the time when the muscle and the muscle spindle are actually changing length). The relationship obtained for various muscle lengths is linear, as plotted in Figure 4-4. Again, recall that the fact that the plot is linear indicates that equal increments in discharge frequency result from equal increments in generator potential amplitude. A linear relationship between the amplitude of the generator potential and the discharge frequency of the nerve has been found for every receptor studied so far. For the muscle spindle, there is also a linear relationship between the amount of stretch applied and the amplitude of the generator potential. The muscle spindle is one of the cases of a receptor in which there is a linear relationship (the exponent of the power function, k=1) between the strength of the stimulus and the response of the cell, i.e., the amount of stretch and the frequency of discharge of the associated nerve fiber. If the relationship between stimulus and response is curvilinear for a sensory receptor, as in those cases where the relationship is a logarithmic or power function (the exponent, k1), then there is a curvilinear relationship between the strength of the stimulus and the amplitude of the generator potential, and, in fact, it is a logarithmic or power function. This must be the case if the relationship between the amplitude of the generator potential and the discharge frequency is always linear, because a linear transformation of any function or relationship always gives the same sort of function or relationship. It is important to bear in mind that the normal sequence of events in transduction is

1. stimulus
2. receptor or generator potential
3. action potentials

The relationship between generator potential amplitude and frequency of discharge is always linear.

Fig. 4-5. A. The setup for recording generator potentials from Pacinian corpusles. B. The stimulus monitor and an example of a Pacinian generator potential.

Summation

Fig. 4-6a. Spatial and temporal summation of generator potentials in a Pacinian corpuscle. Stimuli A and B are applied, with the same stimulator, close together in time. Stimulus C is applied at another place on the receptor at the same time as A in the second set of records.

Fig. 4-6b. The summation that results from varying the interval between stimuli A and C as delivered in Fig. 4-6a.

The generator potential rises from the resting membrane potential, reaches a peak after a given interval, then returns toward the resting membrane potential. If the stimuli are presented so far apart in time that the first generator potential is over before the second stimulus is applied, then there will be no summation. Thus, the amplitude of the summated generator potential depends upon how much the membrane potential of the receptor is altered from the resting potential at the moment the second stimulus is applied. At the bottom of Figure 4-6 are shown a number of examples of how summation is affected by varying the time between the two stimuli, A and C; the maximum summed potential occurs when the stimuli are delivered simultaneously and the amplitude falls off as the second stimulus is delayed. The end result of the summation is an increase in the frequency of discharge of the sensory nerve fiber over the frequency that would be obtained for either stimulus alone. Any number of generator potentials can sum in this same manner, each increment in the amplitude resulting in increased discharge frequency, until a maximum frequency, characteristic of that particular fiber, is reached.

Adaptation

Fig. 4-7. Comparison between rapidly adapting (Pacinian corpuscle) and slowly adapting (muscle spindle) generator potentials.

The reason for having both slowly and rapidly adapting receptors lies in the different kinds of information they signal. Slowly adapting receptors presumably signal the onset and offset, but most importantly the presence of a long sustained stimulus. They do adapt to some extent as indicated by the gradual waning of the sensation of the chair against your posterior as you read this. (This probably does not represent entirely a receptor process, but involves the central nervous system as well, as is indicated by the fact that the sensation returned to your posterior as soon as mention was made of its absence.) In contrast, rapidly adapting receptors signal the start, velocity, and, in some cases, the stop of a stimulus, but do not signal in between. On the other hand, they give very good responses to repeated stimuli, suggesting they may play a role in sensations of vibration. The time over which summation of slowly adapting generator potentials can occur is longer than for rapidly adapting ones for a stimulus of the same duration.

Coding

At one time the nervous system was thought to be a syncytium, a continuous, interconnected, multinucleated mass of protoplasm. The problem of explaining how information was routed from one place to another was most puzzling in such a system, because any stimulus should in theory be able to set the whole mass alive with impulses--all identical. How could any part of the system operate on its own; how could we tell hearing from smelling, high pitches from low? We now know that nervous systems are composed of individual cells strung together and able to communicate with each other in specific patterns. Even with this "simplification," the problem of how the information contained within a stimulus is signaled to distant parts of the nervous system is formidable.

Any stimulus contains within it certain features that are of interest to the organism. Stimuli have

• intensities or strengths
• locations or sites of application
• frequencies of application
• rates of application
• modalities

Some physiologists prefer to define modalities of sensation in terms of the stimuli that produce them. They refer to touch, pressure, and pain as modalities or submodalities of somesthesia with particular reference to the way the skin is stimulated. But, do touch, pressure, and pain represent different forms of stimulation or just different strengths of the same stimulus? With your fingers, lightly grip a fold of skin on your forearm, a stimulus you might consider as touch. A stronger grip yields a sensation of pressure, and an even stronger grip a sensation of pain. Yet the manner of stimulation is the same in all three cases. Taking another example, electrical stimulation of a neuron can lead to the same sensation as a more natural stimulation of the skin. Clearly, electrical energy is not the same form as mechanical energy, still the quality of the sensation is the same.

In our discussion, we use the term modality to mean one of the following:

• vision
• audition
• gustation
• olfaction
• somesthesia (including skin, muscle, position, and visceral senses)

How do neurons recognize and encode the features of a stimulus? The term coding means simply the representation of the facts about the stimulus (information) in terms of neural activity. Most neurons generate action potentials by which they communicate with each other. These action potentials are essentially the same for all neurons. For the neurophysiologist studying coding, the problem is rather like that of a nation that intercepts a secret code sent by a foreign nation--what does the message say? There are quite a few codes possible using neural activity, many or all of which are used somewhere in the nervous system. For the present, the discussion will concentrate on five classes of neural codes, with some specific examples of how they may be used.

Code of specific nerve energies

Careful: it is possible to have a subthreshold adequate stimulus.

Complementary to the principle of the adequate stimulus is a notion formulated by Johannes Müller. Müller noticed that, though the eye is excitable by light, pressure on the eyeball, electrical stimulation of the optic nerve, and some irritative pathological conditions, the sensation experienced is always one of vision--the person "sees light." The Doctrine of Specific Nerve Energies, as formulated by Müller, says that, although a sense organ may be sensitive to many forms of stimulus energy other than its adequate stimulus, the sensation evoked is always like that associated with the adequate stimulus, no matter what kind of energy was applied. With electrical stimulation of the optic nerve, the sensation evoked is one of seeing light, not one of an electrical shock. The doctrine of specific nerve energies implies that the modality or submodality of a sensation is determined not by the stimulus, but by what specific receptor or nerve fiber is stimulated. The doctrine also implies that the subjective qualities of a modality are determined, not in the receptors themselves, but in the central nervous system.

An extension of the doctrine of specific nerve energies is the concept of labeled lines . The concept of labeled lines says that information from a particular receptor travels over particular pathways to particular parts of the nervous system. Thus, the modality of a sensation depends upon which particular cell, pathway, nucleus or lobe is activated by the stimulus. The concept of labeled lines receives some support from the fact that stimulation of the spinothalamic tract in man causes pain or from the observation that stimulation of the visual cortex leads only to a perception of light in a place within the visual world that depends upon where the visual cortex is stimulated.

Fig. 4-8a. An example of a frequency code. The physical setup of the experiment illustrates how the probe is positioned over the receptor structure and how the discharges of the nerve fiber are recorded. At the right, typical response from the fiber (lower trace) to prolonged displacement (upper trace) of the receptor surface (downward deflection of the trace indicates a downward movement of the probe). Entire record is 40 msec in duration.

Fig. 4-8b. A plot of the frequency of discharge of the fiber against the amount of displacement of the receptor. Notice that the amount of displacement is coded as a linear function of the frequency of discharge (Data from Tapper DN: Trans NY Acad Sci 26:697-701, 1964).

The concept of labeled lines implies that between the receptor on the periphery and the place in the central nervous system where the sensation occurs, there is no interaction between the modalities or submodalities. This does not seem to be true. For example, in the cuneate nucleus, a major relay in the somatosensory system, there is interaction between activity arising from hair receptors and activity arising from touch receptors. In addition, there is interaction here between cutaneous, auditory, and visual activity. The nature of this interaction is just beginning to be explored. It is true, however, that some structures within the central nervous system do appear to be associated with particular modalities of sensation.

Even if a specific pathway or labeled line underlies the subjective quality of a sensation, a given neural pathway must use more than one code if other stimulus features are to be signaled. Whereas any activity in the pathway would evoke an experience of a specific quality or modality, the intensity of the sensation is usually coded within the activity itself.

Rate or frequency codes

Interval code

Fig. 4-9. A. The duration of a single nerve impulse. B. The interval between impulses in a train of nerve impulses at 50/sec.

For vibrating stimuli, it is possible to encode the frequency of the stimulus in a linear fashion with one impulse for each stimulus, as a frequency code, or with one impulse for every two or more vibrations of the stimulus, as an interval code. Coding of vibratory frequency is perhaps the most trivial example of a possible interval code and, at the same time, the least different from a frequency code. When the stimulus is not periodic, the interval code is more obvious. It is possible that in some cells, such as pyramidal tract cells, a particular short interval between action potentials, in a train where the other intervals are much longer, may signal (code) the start of some movements. We will examine this in more detail later. Briefly, however, this phenomenon is related to a special event called facilitation that occurs in certain postsynaptic cells only when two presynaptic action potentials arrive at short intervals.

Fig. 4-10. An example of a pattern code. Four different taste receptors: Each responds to each of 4 substances of different tastes. None by itself signals the substance applied, but by looking at the pattern of discharge in all 4 receptors it is possible to deduce what the substance was.

Neuron pattern or ensemble codes

Each column of the figure represents the responses of a single taste receptor to four different solutions, one from each submodality. The height of the bar indicates the magnitude of the response that particular receptor gave when the particular solution was poured over it. Looking at the response of only one receptor, it is impossible to tell what the stimulus was. For example, receptor 1 responds equally to salty and sweet and thus cannot distinguish between the two. By comparing the responses of the four cells, it is possible to detect which stimulus was presented. Salty corresponds to large discharges from receptors 1 and 3, sweet to large discharges from receptors 1 and 4, and so on. There is a unique pattern of discharge across the group of receptors for each different quality, even though no one receptor can signal a given quality uniquely; obviously, there are more than four taste receptors in the tongue, but a similar principle could be applied to a much larger set of receptors.

Nonimpulse codes

However, some cells in the nervous system do not generate action potentials at all. In these cases, the distances traversed are only micrometers, and reduction of potential with distance is no problem. For example, in the eye, information must pass through a minimum of two cells before any action potentials are generated. All of the information in the optic nerve discharges must also be encoded in the hypopolarizations and hyperpolarizations of the receptor cells, horizontal and amacrine cells, and bipolar cells.

Any time the potential across the membrane of a cell changes there is a current flow generated in the extracellular space. These currents are large enough in the cerebral cortex to be recordable on the scalp as the electroencephalogram or EEG. Small amounts of these extracellular currents can penetrate other nearby cells, altering their membrane potentials and thus their excitabilities. This kind of interaction can sometimes be quite powerful. We presume that this is another manner in which information can be coded and transferred, but as yet there is no proof that this is so.

Summary

Transduction is a four-stage process. The stimulus causes:

1. a local change in membrane permeability, which in turn allows
2. the generator current to flow, leading to
3. a local hypopolarization, the generator potential.
4. finally, the generator potential gives rise to the propagated spike in the same or an adjacent region of the membrane

The generator potential is not conducted, but invades the region of spike generation electrotonically. Generator potentials can be summed both temporally and spatially. Receptors differ in their rates of adaptation, depending upon physical properties of the receptor structure and other factors, including accommodation.

There are a number of possible ways that information can be coded in the nervous system. It can be coded in terms of which elements are active--the "doctrine of specific-nerve-energies"--and the labeled line notion. Information can be coded in terms of variations in impulse trains, either their frequencies or the patterns (intervals) of the impulses. It can also be coded as variations in the amount of activity in different elements of an ensemble of neurons. Finally, it can be coded in terms of non-propagated or electrotonic potential changes, as long as transmission distances are very short.

Suggested Reading

1. Aidley DJ: The Physiology of Excitable Cells. Cambridge, Cambridge University Press, 1971.
2. Loewenstein WR [ed]: Principles of Receptor Physiology. Handbook of Sensory Physiology, Vol. I. Berlin, Springer, 1971.
3. Mountcastle VB: Sensory receptors and neural coding: introduction to sensory processes. In Mountcastle VB [ed]: Medical Physiology, 13th ed, Vol. 1. St. Louis, Mosby, 1974.
4. Rushton WAH: Peripheral coding in the nervous system. In Rosenblith WA [ed]: Sensory Communications. Cambridge MA, MIT Press, 1961.
5. Zimmermann, M: Neurophysiology of sensory systems. In Schmidt RF [ed]: Fundamentals of Sensory Physiology. New York, Springer, 1978.

[TOC] [Chapter 4b] [Glossary] [Index] [Abbreviations]