- Email: [email protected]
On the general theory of neural circuitry
Medical Hypotheses I
I
Mcdicol Hymheses (1994) 42. 291-298 0 Longman Group Ltd 1994
On the General Theory of Neural C‘ircu itry D. J. KINGHAM Box 163, RR # 2 Kettleby, Ontario, Canada LOG lJ0
Abstract - A general theory of neural circuitry is proposed wherein neural impulses travel in a continuous circuit from the brain to the extremities and back to the brain. At the extremities the impulse may be modified by the environment there. At the spinal column the return signal is compared with the outgoing signal and the appropriate motoneuronal ‘reflex’ signal is generated if the difference is sufficiently large. In the thalamus the return signal is again compared with the outgoing signal and the difference between the two generates a sensory impulse which is sent to the cortical regions of the brain for comparison with stored patterns from similar signals of past experience. This theory allows for an explanation of feelings of pain and pleasure, pain remote from an area of trauma, phantom limb pain and the relationship between sensory impulses and motor impulses. New approaches to reducing pain are suggested by this theory.
Introduction It has been generally thought that a body senses the environment around it as a result of the stimulation of nerve endings, generating neural impulses which travel in one direction only - to the brain. This approach does not, however, explain certain phenomena which are now well documented, the most obvious of which is the phenomenon referred to as ‘phantom limb pain’. The hypothesis described in this paper conceives of neural circuitry as continuous, from the brain to the extremities, where neural impulses are modulated, and back to the brain for analysis of the modulated signal. Hypothesis This paper postulates a general theory on how neural circuits function in animals. It proposes that there are Date received 8 July 1993 Date accepted 5 November
at least five interrelated neural circuits which function in similar ways, leading to a normal sense of the comfortable self, coordinated movement, reflex actions and the sensation of pain. The five circuits discussed here are: 1. 2. 3. 4. 5.
the the the the the
neuromatrix to sensory modulator circuit sensory to normal movement circuit pain-modulated sensory circuit reflex movement short-circuit phantom limb open-circuit.
This theory accepts Ronald Melzack’s concept of the neuromatrix (1) as an integral and driving force in neural circuits and modifies as appropriate the gate control theory of Melzack and Wall (2,3). In general, the neuromatrix generates impulses in a specific, personal and complex pattern, like the central computer on a modern aircraft generates sensing signals. In this theory, the neuromatrix transmits its impulses to
1993
291
292 sensory modulators throughout the body in a fashion similar to that of the aircraft computer sending a series of sensing electrical signals to sensors throughout the aircraft. In the biological system the end result is the generation of a cerebral impulse based on the difference between the outgoing neuromatrix signal and the modulated return signal, and a comparison of that cerebral impulse with memory of stored cerebral impulse signatures. It is crucial to this theory that one do away with the concept of nerve endings. In their place the theory requires the existence of neural impulse modulators. Neural circuitry is thus not unlike that of the blood circulatory system, with the neuromatrix as the heart, and the sensory portion of the circuit as the capillaries. At certain points in the circuit the transmitted and return signals are compared for any difference potential between them. The comparison function is like a switching transistor in an electronic circuit, in the sense that when the difference potential exceeds a certain threshold level a new signal is generated, leading to subsequent feelings or actions. The position of the switching nerve cells in the neural circuit is critical: the outgoing and return signals must begin to pass by them, in opposite directions, at precisely the same time so that from the perspective of the switching cell the relative voltage difference is zero in the case where there has been no modulation and no new cerebral or reflex action signal is required. In mathematical terms, the length of the circuit from a reflex action switching nerve cell to its sensory modulator and back must be precisely an integral number of times of the duration of a neural impulse cycle multiplied by the speed of the impulse. A neural impulse cycle, here, is the sum of the time of the impulse plus the time between impulses. 1 = n X Qycle x Vi (where n is an integer and vi is the speed of the impulse) For example, if one assumes an impulse time of 2 milliseconds (43 and an inter-impulse time of 18 milliseconds, for a cycle time of 20 milliseconds, along with an impulse speed of 50 m/s;
MEDICALHYPOTHESES
Similarly, the length of this circuit to the thalamus would be an additional integral number of times the (cycle time x speed) product. Thus the length of the circuit from the thalamus to the sensory modulator and back (L) must also be:
L = n x tcyck x vi (where n is another, larger, integer than for I) These relationships very much simplified,
are shown schematically, on Figure 1.
and
Implh
Qenefeted bythe neuromiltrix
L = 1+ln
Fig. 1 Schematic of the basic sensory neural circuit showing the relationship between the thalamus (T), the point of emergence of the nerve in question at the base of the spinal column (S) and the sensory modulator (SM). The length of the circuit (r) from the ‘switching’ nerve cell at S, to SM and back to S, must be exactly an integral number(n) times the time required for a completed impulse cycle (trycre),being the duration of the impulse plus the inter-impulse time, multiplied by the speed (Vi) of the impulse. The length of the circuit L from the thalamus to the sensory modulator and back must be exactly I plus another n times the time required for a completed impulse cycle multiplied by the speed (Vi) of the impulse.
I= n x 0.02 x 50 m/s = 1 m In this example, chosen for its simplicity rather than its accuracy, an impulse would be about to arrive at the base of the spinal column from the neuromatrix while the preceding impulse is passing through the sensory modulator and the one preceding that is arriving at the spinal column from the sensory modulator.
The function of the switching nerve cell, at either the spinal column for the reflex action short circuit, or in the vicinity of the thalamus for sensation, is to sense the impulses passing by it in opposite directions and generate a signal of its own, which I call the ‘difference signal’. If the signal from the brain and
ON THE GENERAL
the signal from the sensory modulator arrive at the same time and with no change in their amplitudes there is no difference between them and no signal is generated in the switching nerve cell. Thus there is no reflex action or sensation, as appropriate to the switching nerve cell in question. But if the signal is speeded up, slowed down, amplified, or attenuated, or a combination of these, there is a difference in electrical potential which electrically polarizes the nerve cell and generates the appropriate reflex action muscle impulse or cerebral impulse of pain or pleasure. There are thus 8 absolute variations of difference signals and an infinite number of relative variations. The 8 absolutes are: impulse impulse impulse impulse impulse impulse impulse impulse
293
THEORY OF NEURAL CIRCUITRY
speed unchanged, amplitude increased speed unchanged, amplitude attenuated accelerated, amplitude unchanged accelerated, amplitude increased accelerated, amplitude attenuated delayed, amplitude unchanged delayed, amplitude increased delayed, amplitude attenuated
It is this considerable variation in modulation of the impulse in the neural circuit that gives rise to the variation in sensation as measured in the region of the thalamus, perhaps between the Periaqueductal grey (PAG) and the Anterior Pretectal Nucleus (APtN) as there is an acknowledged reciprocal projection between the PAG and the APtN (6,7) and the stimulation of the PAG has been reported to be as effective in blocking pain as high doses of morphine (8). But my theory takes quite a different approach to the role of the APtN in sensory processing than that proposed by Rees and Roberts who hold that a ‘long distance reverberatory circuit maintains descending inhibition of responses to noxious peripheral stimuli’ (9). The essence of my theory is that if there is any ‘reverberation’ it is at the sensory modulator at the periphery, and the detection of the change takes place at the thalamus. Consider for example, the situation which would exist if the return signal were delayed for 1 millisecond by a change to the local environment (hereafter ‘the microenvironment’) at the periphery. One side of the switching nerve cell is sensing the outgoing negatively charged ‘tail’ of an impulse while the other side is just receiving the positively charged ‘head’ of the delayed return signal. The switching cell becomes polarized and generates a new cerebral signal several orders of magnitude smaller (30-40 microvolts) (IO) than either the outgoing or incoming signals. This is shown schematically in Figure 2.
I will now discuss the five principal circuits, or circuit combinations, of greatest importance in animals.
The neuromatrix
to sensory modulator
circuit
The most significant aspect of the theory as it relates to sensory circuits, is that the concept of nerve endings is discarded. The sensory part of the circuit, at the tip of the finger, for example, moduIates the signal received from the brain and allows the modulated signal to continue along a return path to the somatosensory cortex. From the earliest stages of development of the brain the neuromatrix begins its lifelong process of signal generation and transmission. Melzack postulates signal generation (but not transmission to the body) at this early stage: ‘The neural network that underlies the experience of one’s self is genetically determined but can be modified by sensory experience’ (1) and that is consistent with my theory of neural circuitry. The return pattern in the fetus is as it is, and any difference, as measured between the outgoing and return signals near the thalamus, is transmitted to the cortical regions of the brain where it is stored as memory of the normal comfortable pattern of self. This is shown schematically in Figure 3. This normal pattern continues for life. In the absence of any stimulus or trauma, there would be no experience of other than the comfortable self, and no memory. With the first unexpected experience, the sensory part of the circuit is affected accordingly. Warmth changes the microenvironment of the fingertip and this change has its own characteristic modulating effect on the onward transmission of the signal from the neuromatrix. The first experience of something warm is not known to be warm. But in the thalamus the difference signal is unique and it triggers a unique signal to the cortical regions of the brain. When a child is told that what he has felt is warm, he has two parallel sensory inputs to process: one is the new signal from the thalamus, the other is the new word from the auditory circuit. The two are stored as a unique memory of what ‘warm’ is. The same sort of information processing and memory storage occurs when pressure is applied to the fingertip or the ball of the foot.
The sensory to normal movement
circuit
This compound circuit involves both the neuromatrix to sensory modulator circuits and the motoneural circuits. When a child first stands, there is pres-
294
MEDICAL HYPOTHESES
cerebral pot8ntbl for pain I- 35 Jw.1 from brain to ~ensorv modulator
from aenaorv modulator to brain
t
--induced in nerve cell pol& by 1 ms. delay in return impulse
_
T 1 ma.
+2
0
combined action
-2
-4
+2 0 -2 -4 -6 combined action potential 1mv.l transmitted to sensory modulator
-6
paranrlaltmv.1
fromaenaorymodulator
Fig. 2 Polarity induced in switching nerve cell and corresponding cerebral impulse for pain as a result of a one millisecond delay in the return signal from the sensory modulator.
from
Fig. 3 The normal sensing of the comfortable self. The signal generated in the thalamus is caused by the difference, if any, between the outgoing and returning signals from and to the neuromatrix.
sure on the ball of his foot which, if not compensated for by contraction of the correct muscle, will result in the child falling on his face and the corresponding unwanted pain sensations. Through trial and error the child learns that the correct response
to pressure on the ball of the foot is to contract the calf muscle of the leg in question. The sensory return signal is combined with the continuous afferent flow of nerve impulses from the mechanoreceptors to the neuromatrix (11-13). This both avoids the likely pain sensation and achieves the goal of standing. Through time these circuits work in perfect harmony. The child stands without consciously thinking about it, as the normal pattern for standing is reinforced thousands of times a day until it becomes automatic. This is shown schematically in Figure 4. As was the case for the sense of the normal self, there emerges a sense of the normal walk. If, however, while walking across a stretch of smooth pavement a person inadvertently steps upon a stone, the situation changes. Now the anticipated return signal is appropriately modulated within the context of what has become the normal pattern for walking. The motor system makes a correct compensation for this unexpected difference signal, as directed by the neuromatrix, and a large number of signals are sent to the muscle fibres
ON THE GENERAL THEORY OF NEURAL CIRCUlTRY
295 The pain-modulated sensory circuit
Fig. 4 For normal movement, normal differences between the outgoing and return signals lead to normal impulses to the muscle sets responsible for movement. Unexpected difference patterns in the thalamus generate a new pattern of signals to generate a new set of actions.
of the legs and upper body to maintain balance and reduce the unexpected pressure on the ball of the foot by reducing the contraction of the calf muscle and allowing the heel of the foot to bear more weight. This reinforcement of the normal interaction of the neuromatrix and motor circuits proceeds to the highest level of animal behaviour. A pole vaulter, for example, has little time to think about all the coordinated actions that will be required to clear the bar at the moment his body is propelled from the ground. But a repetition of all the actions required to do so successfully leads to the strengthening of the pattern of what is normally required to succeed. The brain need only compensate for the relatively minor variations, and their corresponding difference signals, which relate to unanticipated sensations along the pole vaulter’s path. So it is with a violin player, for example, although a quite different subset of the neuromatrix and motor circuits are called into play here. Auditorv circuits have learned where the correct pitch and intensity lies: slight changes to the position and pressure on the fingertips will return the desired difference signal for the intended note from the thalamus.
The pain-modulated sensory circuit is quite similar to the neuromatrix to sensory modulator circuit with a few significant changes. Where warmth may be sought after, hot may have to be avoided. If the modulation of the outgoing signal from the neuromatrix is proportional to the change in the microenvironment at the sensory part of the circuit, then the strength of the difference signal in the thalamus is much greater and the experience of the sensation proportionately stronger. There are a wider range of pain sensations than the normal self-sensing difference signals. A cut affects the sensory part of the circuit simultaneously across many closely spaced sensory ‘capillaries’ whereas warmth affects many sensory modulator synapses over a wider area with the benefit of some layers of insulating tissue. An itch, a pinprick or the overextension of a ligament will each generate its own modulation of the signal from the neuromatrix, giving its own difference signal at the thalamus. Each difference signal will then be compared with the brain’s library of difference signals to find the one that most closely corresponds to the trauma and the subject will feel pain accordingly. Combination effects generate their own signals: ‘rubbing salt in the wound’ could greatly affect the amplitude of the modulated return signal which has already been distorted by the original wound. Similarly, hot water on an existing injury adds its unique signature to the injury’s existing signal. But pain may appear to exist where the cause of pain seems well removed in space from the sensation. A bulging vertebral disc, for example may put pressure on the outgoing and return neurons passing by it. If either the outgoing or return signals are sufficiently modulated by the pressure, the difference signal read at the thalamus will be interpreted in the cortical regions of the brain as a pain in the sensory part of the circuit in the leg, for example, even though there may have been no trauma to the leg. This is shown schematically in Figure 5. Restoring the microenvironment around the neurons at the vertebrae will restore the normal difference pattern in the thalamus and the leg will ‘feel better’. The reflex movement short-circuit Certain traumas are sufficiently threatening to require action before the nature of the trauma is known. To remove the body from the threat quickly there is a neuronal short-circuit at the spinal column. Here the
296
MEDICAL HYPOTHESES
from neuromatrix to rxtremitirs 1
from
reflex impulse
s-sow modulator
I
w v
HnsW
modulstor
Fig. 5 The ‘pinched’ nerve syndrome alters either or both of the outgoing and returning signals such that the difference between the two signals in the thalamus is interpreted as pain in the limb being sensed by the circuit.
nerve cell, (the difference measuring ‘transistor’) is located at the spinal column and is directly coupled to a motoneuronal axon. If the difference is sufficiently great to trigger this transistor, a signal is generated directly to the associated muscle fibre causing rapid contraction and removal of the body from the threat. It is important to note that even if the difference is great and the reflex reaction is triggered, the original outgoing and return signals are not interfered with by measuring the difference. The situation is rather analogous to measuring the voltage difference between a pair of wires carrying two different electrical currents: the voltage measurement does not detract from the current in either wire (4). This means that once the reflex reaction has occurred and the body has been removed from harm, the difference signal subsequently measured in the thalamus and transmitted to other regions of the brain allows the animal to sense that from which it withdrew, from a somewhat safer perspective. The body did not have to know that it was burning, stabbing or cutting pain from which it withdrew at the time, but it is useful for the brain to be able to analyze the situation afterwards for future reference. This is shown schematically in Figure 6.
switching
I
Fig. 6 The reflex short-circuit. The difference signal at the spinal column is sufficient to trigger an impulse through the motor axon to retract the limb before the brain has received the difference signal from the thalamus and analysed the nature of the threat.
This reflex movement short-circuit explains the observations of Tasker et al (14) who found that injections of lidocaine into the lateral hypothalamus of rats apparently had no effect in the foot-flick ‘pain’ test but significantly decreased pain in the formalin test. I hypothesize that the subject rats felt little or no pain in the foot-flick test either, but the reflex movement short-circuit caused the foot to be withdrawn because the difference signal at the spinal column switching nerve cell was sufficiently large to generate the appropriate signal to motoneurons. The formalin pain test, building its effect more slowly as it does, would not stimulate the reflex action short-circuit but would normally generate a cerebral impulse for pain arising from the difference signal in the thalamus, which signal could be blocked by an appropriate lidocaine injection.
The phantom
limb open-circuit
Somewhat in opposition to the reflex movement short-circuit is the phantom limb open-circuit. It is almost always the case that when a limb is amputated the patient senses the limb after it is gone. The sensation may be one of burning or stabbing pain in the portions of the limb which are clearly no longer there or, often much later, it may be a sensation that the limb is still there, giving normal sensations or feelings of movement or cramps.
ON THE GENERAL THEORY OF NEURAL CIRCUITRY
In this theory of neural circuitry the cause of such sensations can be accounted for. The outgoing signal from the neuromatrix remains; the return signal is non-existent. The difference between the two is nothing like the body has experienced before. The thalamus sends its difference-stimulated cerebral impulse to the cortical regions of the brain for comparison with past experiences. The brain searches for the closest match and, depending on the microenvironment at the site of the amputation, the brain interprets the difference signal as burning, itching, cramps and so on. During healing of the amputated limb a number of possible outcomes exist. The microenvironment could return to something approximating its original state, in which case the patient may sense that the limb is still there. Healing could, on the other hand, lead to a microenvironment and return neural impulse pattern which corresponds to a burning sensation and the patient could ‘feel’ chronic pain as a result. Many other variants are possible, with the concomitant range of sensations reported by amputees. One important and distressing variant is that the microenvironment at the point where the neuromatrix to sensory modulator circuit has been cut may deteriorate over a period of months or years, leading to worse pain than was experienced at the original severing of that circuit. The basic concept and two possible outcomes are shown schematically in Figure 7. Possible applications circuitry
of this theory of neuronal
The most important possible application of this theory is for the treatment of chronic pain, whether phantom limb pain or otherwise. I hypothesize that the cure for pain will be found in one of three areas. First, eliminating the outgoing signal from the neuromatrix through meditation (15), medication or surgery. This has already been proposed, ‘... if the pattern for pain is generated by cyclical processing and synthesis, then it should be possible to block it by injection of a local anesthetic into appropriate discrete areas that are thought to comprise the widespread neuromatrix’ (I), and tested, ‘Injections of lidocaine into the lateral hypothalamus decrease experimentally induced pain in rats...’ (14) but may not prove practical for the long-term treatment of chronic pain. The second approach is to try to stimulate a normal return signal by means of suitable electrical devices or acupuncture. This approach has already yielded some success. The third approach to the cure for pain is to restore the microenvironment at the site of damage to as near normal as is possible by chemical or surgical means. If my theory is correct,
297
(B)
JJ
4
sensory modulator
badly healed
*’ i*
Fig. 7 The phantom limb open-circuit. When a limb is initially amputated there may be no return signal. The difference signal generated in the thalamus between the normal outgoing signal and the absent return is transmitted to cortical regions where it is compared to memory of pain and the closest match leads to the description of feeling given by the patient. Good healing (A) can lead to a sense of well-being in the patient while a significant alteration of the microenvironment in the neuromas at the point of amputation (B) can lead to chronic pain.
the modification of the environment at the sensory modulator part of the circuit, based on an understanding of what the normal modulator microenvironment should be, will be the most fruitful route to pain relief for millions of sufferers. A second possible application is in the understanding of the relationship between the motor circuit and the sensory circuit as it pertains to rehabilitation of patients who have been seriously injured and where such injury may have affected one or both circuits. Training such patients to recognize and accept their new pattern of difference signals may help speed the practical use of such new patterns. References I. 2. 3.
4.
5.
Melzack R. Phantoms and the concept of a neuromatrix, Trends Neurosci 1990; 13: 88. Melzack R, Wall P D. Pain mechanisms: a new theory. Science 1965; 1.50: 971. Melzack R. The gate control theory 25 years later: new perspectives on phantom limb pain. In: Proceedings of the Vlth World Congress on Pain, Elsevier Science Publishers BV 21, 1991. Kuypers P D L, Gielen F L H, Wai R T D, Hovius S E R, Godschalk M, Egeraat J M. A comparison of electric and magnetic action sign& as quantitative assays of peripheral nerve regeneration. Muscle & Nerve June 1993; 634. Uncini A, Santoro M. Corbo M, Lugaresi A, Latov N. Conduction abnormalities induced by sera of patients with mul-
298
6. 7.
8.
9. 10. 11.
MEDICAL HYPOTHESES
tifocal motor neuropathy and anti-GM1 antibodies. Muscle & Nerve June 1993; 610. Berman N. Connections of the pretectum in the cat. J Comp Neurol 1977; 174: 227. Foster G A, Sizer A R, Rees H, Roberts M T H. Afferent projections to the rostra1 anterior pretectal nucleus of the rat: a possible role in the processing of noxious stimuli. Neuroscience 1989; 29: 685. Mayer D J, Liebeskind J C. Pain reduction by focal electrical stimulation of the brain; an anatomical and behavioral analysis. Brain Research 1974; 68: 73. Rees H, Roberts M H T. The anterior pretectal nucleus: a proposed role in sensory processing. Pain 1993; 53: 121. Meier W, Klucken M, Soyka D, Bromm B. Pain 1993; 53. Barker D. The morphology of muscle receptors. In: Hunt
12.
13.
14. 15.
C, ed. Muscle receptors. Handbook of sensory physiology. Berlin: Springer, 1974. Jasza L, Kvist M, Kannus P, Jarvinen M. The effect of tenotomy and immobilisation on muscle spindles and tendon organs of the rat calf muscles: a histochemical and morphometrical study. Acta Neuropathol 1988; 76: 465. Katonis P G, Assimakopoulos A P, Agapitos M V, Exarchou E 1. Mechanoreceptors in the posterior cruciate ligament: histological study on cadaver knees. Acta Orthop Stand 1991; 62: 216. Tasker R A R, Choinitre M, Libman S M, Melzack R. Pain 1987; 31: 237. Melzack R, Perry D. Self-regulation of pain: the use of alpha feedback and hypnotic training for the control of chronic pain. Exp Neurol 1975; 46: 452.
I
Mcdicol Hymheses (1994) 42. 291-298 0 Longman Group Ltd 1994
On the General Theory of Neural C‘ircu itry D. J. KINGHAM Box 163, RR # 2 Kettleby, Ontario, Canada LOG lJ0
Abstract - A general theory of neural circuitry is proposed wherein neural impulses travel in a continuous circuit from the brain to the extremities and back to the brain. At the extremities the impulse may be modified by the environment there. At the spinal column the return signal is compared with the outgoing signal and the appropriate motoneuronal ‘reflex’ signal is generated if the difference is sufficiently large. In the thalamus the return signal is again compared with the outgoing signal and the difference between the two generates a sensory impulse which is sent to the cortical regions of the brain for comparison with stored patterns from similar signals of past experience. This theory allows for an explanation of feelings of pain and pleasure, pain remote from an area of trauma, phantom limb pain and the relationship between sensory impulses and motor impulses. New approaches to reducing pain are suggested by this theory.
Introduction It has been generally thought that a body senses the environment around it as a result of the stimulation of nerve endings, generating neural impulses which travel in one direction only - to the brain. This approach does not, however, explain certain phenomena which are now well documented, the most obvious of which is the phenomenon referred to as ‘phantom limb pain’. The hypothesis described in this paper conceives of neural circuitry as continuous, from the brain to the extremities, where neural impulses are modulated, and back to the brain for analysis of the modulated signal. Hypothesis This paper postulates a general theory on how neural circuits function in animals. It proposes that there are Date received 8 July 1993 Date accepted 5 November
at least five interrelated neural circuits which function in similar ways, leading to a normal sense of the comfortable self, coordinated movement, reflex actions and the sensation of pain. The five circuits discussed here are: 1. 2. 3. 4. 5.
the the the the the
neuromatrix to sensory modulator circuit sensory to normal movement circuit pain-modulated sensory circuit reflex movement short-circuit phantom limb open-circuit.
This theory accepts Ronald Melzack’s concept of the neuromatrix (1) as an integral and driving force in neural circuits and modifies as appropriate the gate control theory of Melzack and Wall (2,3). In general, the neuromatrix generates impulses in a specific, personal and complex pattern, like the central computer on a modern aircraft generates sensing signals. In this theory, the neuromatrix transmits its impulses to
1993
291
292 sensory modulators throughout the body in a fashion similar to that of the aircraft computer sending a series of sensing electrical signals to sensors throughout the aircraft. In the biological system the end result is the generation of a cerebral impulse based on the difference between the outgoing neuromatrix signal and the modulated return signal, and a comparison of that cerebral impulse with memory of stored cerebral impulse signatures. It is crucial to this theory that one do away with the concept of nerve endings. In their place the theory requires the existence of neural impulse modulators. Neural circuitry is thus not unlike that of the blood circulatory system, with the neuromatrix as the heart, and the sensory portion of the circuit as the capillaries. At certain points in the circuit the transmitted and return signals are compared for any difference potential between them. The comparison function is like a switching transistor in an electronic circuit, in the sense that when the difference potential exceeds a certain threshold level a new signal is generated, leading to subsequent feelings or actions. The position of the switching nerve cells in the neural circuit is critical: the outgoing and return signals must begin to pass by them, in opposite directions, at precisely the same time so that from the perspective of the switching cell the relative voltage difference is zero in the case where there has been no modulation and no new cerebral or reflex action signal is required. In mathematical terms, the length of the circuit from a reflex action switching nerve cell to its sensory modulator and back must be precisely an integral number of times of the duration of a neural impulse cycle multiplied by the speed of the impulse. A neural impulse cycle, here, is the sum of the time of the impulse plus the time between impulses. 1 = n X Qycle x Vi (where n is an integer and vi is the speed of the impulse) For example, if one assumes an impulse time of 2 milliseconds (43 and an inter-impulse time of 18 milliseconds, for a cycle time of 20 milliseconds, along with an impulse speed of 50 m/s;
MEDICALHYPOTHESES
Similarly, the length of this circuit to the thalamus would be an additional integral number of times the (cycle time x speed) product. Thus the length of the circuit from the thalamus to the sensory modulator and back (L) must also be:
L = n x tcyck x vi (where n is another, larger, integer than for I) These relationships very much simplified,
are shown schematically, on Figure 1.
and
Implh
Qenefeted bythe neuromiltrix
L = 1+ln
Fig. 1 Schematic of the basic sensory neural circuit showing the relationship between the thalamus (T), the point of emergence of the nerve in question at the base of the spinal column (S) and the sensory modulator (SM). The length of the circuit (r) from the ‘switching’ nerve cell at S, to SM and back to S, must be exactly an integral number(n) times the time required for a completed impulse cycle (trycre),being the duration of the impulse plus the inter-impulse time, multiplied by the speed (Vi) of the impulse. The length of the circuit L from the thalamus to the sensory modulator and back must be exactly I plus another n times the time required for a completed impulse cycle multiplied by the speed (Vi) of the impulse.
I= n x 0.02 x 50 m/s = 1 m In this example, chosen for its simplicity rather than its accuracy, an impulse would be about to arrive at the base of the spinal column from the neuromatrix while the preceding impulse is passing through the sensory modulator and the one preceding that is arriving at the spinal column from the sensory modulator.
The function of the switching nerve cell, at either the spinal column for the reflex action short circuit, or in the vicinity of the thalamus for sensation, is to sense the impulses passing by it in opposite directions and generate a signal of its own, which I call the ‘difference signal’. If the signal from the brain and
ON THE GENERAL
the signal from the sensory modulator arrive at the same time and with no change in their amplitudes there is no difference between them and no signal is generated in the switching nerve cell. Thus there is no reflex action or sensation, as appropriate to the switching nerve cell in question. But if the signal is speeded up, slowed down, amplified, or attenuated, or a combination of these, there is a difference in electrical potential which electrically polarizes the nerve cell and generates the appropriate reflex action muscle impulse or cerebral impulse of pain or pleasure. There are thus 8 absolute variations of difference signals and an infinite number of relative variations. The 8 absolutes are: impulse impulse impulse impulse impulse impulse impulse impulse
293
THEORY OF NEURAL CIRCUITRY
speed unchanged, amplitude increased speed unchanged, amplitude attenuated accelerated, amplitude unchanged accelerated, amplitude increased accelerated, amplitude attenuated delayed, amplitude unchanged delayed, amplitude increased delayed, amplitude attenuated
It is this considerable variation in modulation of the impulse in the neural circuit that gives rise to the variation in sensation as measured in the region of the thalamus, perhaps between the Periaqueductal grey (PAG) and the Anterior Pretectal Nucleus (APtN) as there is an acknowledged reciprocal projection between the PAG and the APtN (6,7) and the stimulation of the PAG has been reported to be as effective in blocking pain as high doses of morphine (8). But my theory takes quite a different approach to the role of the APtN in sensory processing than that proposed by Rees and Roberts who hold that a ‘long distance reverberatory circuit maintains descending inhibition of responses to noxious peripheral stimuli’ (9). The essence of my theory is that if there is any ‘reverberation’ it is at the sensory modulator at the periphery, and the detection of the change takes place at the thalamus. Consider for example, the situation which would exist if the return signal were delayed for 1 millisecond by a change to the local environment (hereafter ‘the microenvironment’) at the periphery. One side of the switching nerve cell is sensing the outgoing negatively charged ‘tail’ of an impulse while the other side is just receiving the positively charged ‘head’ of the delayed return signal. The switching cell becomes polarized and generates a new cerebral signal several orders of magnitude smaller (30-40 microvolts) (IO) than either the outgoing or incoming signals. This is shown schematically in Figure 2.
I will now discuss the five principal circuits, or circuit combinations, of greatest importance in animals.
The neuromatrix
to sensory modulator
circuit
The most significant aspect of the theory as it relates to sensory circuits, is that the concept of nerve endings is discarded. The sensory part of the circuit, at the tip of the finger, for example, moduIates the signal received from the brain and allows the modulated signal to continue along a return path to the somatosensory cortex. From the earliest stages of development of the brain the neuromatrix begins its lifelong process of signal generation and transmission. Melzack postulates signal generation (but not transmission to the body) at this early stage: ‘The neural network that underlies the experience of one’s self is genetically determined but can be modified by sensory experience’ (1) and that is consistent with my theory of neural circuitry. The return pattern in the fetus is as it is, and any difference, as measured between the outgoing and return signals near the thalamus, is transmitted to the cortical regions of the brain where it is stored as memory of the normal comfortable pattern of self. This is shown schematically in Figure 3. This normal pattern continues for life. In the absence of any stimulus or trauma, there would be no experience of other than the comfortable self, and no memory. With the first unexpected experience, the sensory part of the circuit is affected accordingly. Warmth changes the microenvironment of the fingertip and this change has its own characteristic modulating effect on the onward transmission of the signal from the neuromatrix. The first experience of something warm is not known to be warm. But in the thalamus the difference signal is unique and it triggers a unique signal to the cortical regions of the brain. When a child is told that what he has felt is warm, he has two parallel sensory inputs to process: one is the new signal from the thalamus, the other is the new word from the auditory circuit. The two are stored as a unique memory of what ‘warm’ is. The same sort of information processing and memory storage occurs when pressure is applied to the fingertip or the ball of the foot.
The sensory to normal movement
circuit
This compound circuit involves both the neuromatrix to sensory modulator circuits and the motoneural circuits. When a child first stands, there is pres-
294
MEDICAL HYPOTHESES
cerebral pot8ntbl for pain I- 35 Jw.1 from brain to ~ensorv modulator
from aenaorv modulator to brain
t
--induced in nerve cell pol& by 1 ms. delay in return impulse
_
T 1 ma.
+2
0
combined action
-2
-4
+2 0 -2 -4 -6 combined action potential 1mv.l transmitted to sensory modulator
-6
paranrlaltmv.1
fromaenaorymodulator
Fig. 2 Polarity induced in switching nerve cell and corresponding cerebral impulse for pain as a result of a one millisecond delay in the return signal from the sensory modulator.
from
Fig. 3 The normal sensing of the comfortable self. The signal generated in the thalamus is caused by the difference, if any, between the outgoing and returning signals from and to the neuromatrix.
sure on the ball of his foot which, if not compensated for by contraction of the correct muscle, will result in the child falling on his face and the corresponding unwanted pain sensations. Through trial and error the child learns that the correct response
to pressure on the ball of the foot is to contract the calf muscle of the leg in question. The sensory return signal is combined with the continuous afferent flow of nerve impulses from the mechanoreceptors to the neuromatrix (11-13). This both avoids the likely pain sensation and achieves the goal of standing. Through time these circuits work in perfect harmony. The child stands without consciously thinking about it, as the normal pattern for standing is reinforced thousands of times a day until it becomes automatic. This is shown schematically in Figure 4. As was the case for the sense of the normal self, there emerges a sense of the normal walk. If, however, while walking across a stretch of smooth pavement a person inadvertently steps upon a stone, the situation changes. Now the anticipated return signal is appropriately modulated within the context of what has become the normal pattern for walking. The motor system makes a correct compensation for this unexpected difference signal, as directed by the neuromatrix, and a large number of signals are sent to the muscle fibres
ON THE GENERAL THEORY OF NEURAL CIRCUlTRY
295 The pain-modulated sensory circuit
Fig. 4 For normal movement, normal differences between the outgoing and return signals lead to normal impulses to the muscle sets responsible for movement. Unexpected difference patterns in the thalamus generate a new pattern of signals to generate a new set of actions.
of the legs and upper body to maintain balance and reduce the unexpected pressure on the ball of the foot by reducing the contraction of the calf muscle and allowing the heel of the foot to bear more weight. This reinforcement of the normal interaction of the neuromatrix and motor circuits proceeds to the highest level of animal behaviour. A pole vaulter, for example, has little time to think about all the coordinated actions that will be required to clear the bar at the moment his body is propelled from the ground. But a repetition of all the actions required to do so successfully leads to the strengthening of the pattern of what is normally required to succeed. The brain need only compensate for the relatively minor variations, and their corresponding difference signals, which relate to unanticipated sensations along the pole vaulter’s path. So it is with a violin player, for example, although a quite different subset of the neuromatrix and motor circuits are called into play here. Auditorv circuits have learned where the correct pitch and intensity lies: slight changes to the position and pressure on the fingertips will return the desired difference signal for the intended note from the thalamus.
The pain-modulated sensory circuit is quite similar to the neuromatrix to sensory modulator circuit with a few significant changes. Where warmth may be sought after, hot may have to be avoided. If the modulation of the outgoing signal from the neuromatrix is proportional to the change in the microenvironment at the sensory part of the circuit, then the strength of the difference signal in the thalamus is much greater and the experience of the sensation proportionately stronger. There are a wider range of pain sensations than the normal self-sensing difference signals. A cut affects the sensory part of the circuit simultaneously across many closely spaced sensory ‘capillaries’ whereas warmth affects many sensory modulator synapses over a wider area with the benefit of some layers of insulating tissue. An itch, a pinprick or the overextension of a ligament will each generate its own modulation of the signal from the neuromatrix, giving its own difference signal at the thalamus. Each difference signal will then be compared with the brain’s library of difference signals to find the one that most closely corresponds to the trauma and the subject will feel pain accordingly. Combination effects generate their own signals: ‘rubbing salt in the wound’ could greatly affect the amplitude of the modulated return signal which has already been distorted by the original wound. Similarly, hot water on an existing injury adds its unique signature to the injury’s existing signal. But pain may appear to exist where the cause of pain seems well removed in space from the sensation. A bulging vertebral disc, for example may put pressure on the outgoing and return neurons passing by it. If either the outgoing or return signals are sufficiently modulated by the pressure, the difference signal read at the thalamus will be interpreted in the cortical regions of the brain as a pain in the sensory part of the circuit in the leg, for example, even though there may have been no trauma to the leg. This is shown schematically in Figure 5. Restoring the microenvironment around the neurons at the vertebrae will restore the normal difference pattern in the thalamus and the leg will ‘feel better’. The reflex movement short-circuit Certain traumas are sufficiently threatening to require action before the nature of the trauma is known. To remove the body from the threat quickly there is a neuronal short-circuit at the spinal column. Here the
296
MEDICAL HYPOTHESES
from neuromatrix to rxtremitirs 1
from
reflex impulse
s-sow modulator
I
w v
HnsW
modulstor
Fig. 5 The ‘pinched’ nerve syndrome alters either or both of the outgoing and returning signals such that the difference between the two signals in the thalamus is interpreted as pain in the limb being sensed by the circuit.
nerve cell, (the difference measuring ‘transistor’) is located at the spinal column and is directly coupled to a motoneuronal axon. If the difference is sufficiently great to trigger this transistor, a signal is generated directly to the associated muscle fibre causing rapid contraction and removal of the body from the threat. It is important to note that even if the difference is great and the reflex reaction is triggered, the original outgoing and return signals are not interfered with by measuring the difference. The situation is rather analogous to measuring the voltage difference between a pair of wires carrying two different electrical currents: the voltage measurement does not detract from the current in either wire (4). This means that once the reflex reaction has occurred and the body has been removed from harm, the difference signal subsequently measured in the thalamus and transmitted to other regions of the brain allows the animal to sense that from which it withdrew, from a somewhat safer perspective. The body did not have to know that it was burning, stabbing or cutting pain from which it withdrew at the time, but it is useful for the brain to be able to analyze the situation afterwards for future reference. This is shown schematically in Figure 6.
switching
I
Fig. 6 The reflex short-circuit. The difference signal at the spinal column is sufficient to trigger an impulse through the motor axon to retract the limb before the brain has received the difference signal from the thalamus and analysed the nature of the threat.
This reflex movement short-circuit explains the observations of Tasker et al (14) who found that injections of lidocaine into the lateral hypothalamus of rats apparently had no effect in the foot-flick ‘pain’ test but significantly decreased pain in the formalin test. I hypothesize that the subject rats felt little or no pain in the foot-flick test either, but the reflex movement short-circuit caused the foot to be withdrawn because the difference signal at the spinal column switching nerve cell was sufficiently large to generate the appropriate signal to motoneurons. The formalin pain test, building its effect more slowly as it does, would not stimulate the reflex action short-circuit but would normally generate a cerebral impulse for pain arising from the difference signal in the thalamus, which signal could be blocked by an appropriate lidocaine injection.
The phantom
limb open-circuit
Somewhat in opposition to the reflex movement short-circuit is the phantom limb open-circuit. It is almost always the case that when a limb is amputated the patient senses the limb after it is gone. The sensation may be one of burning or stabbing pain in the portions of the limb which are clearly no longer there or, often much later, it may be a sensation that the limb is still there, giving normal sensations or feelings of movement or cramps.
ON THE GENERAL THEORY OF NEURAL CIRCUITRY
In this theory of neural circuitry the cause of such sensations can be accounted for. The outgoing signal from the neuromatrix remains; the return signal is non-existent. The difference between the two is nothing like the body has experienced before. The thalamus sends its difference-stimulated cerebral impulse to the cortical regions of the brain for comparison with past experiences. The brain searches for the closest match and, depending on the microenvironment at the site of the amputation, the brain interprets the difference signal as burning, itching, cramps and so on. During healing of the amputated limb a number of possible outcomes exist. The microenvironment could return to something approximating its original state, in which case the patient may sense that the limb is still there. Healing could, on the other hand, lead to a microenvironment and return neural impulse pattern which corresponds to a burning sensation and the patient could ‘feel’ chronic pain as a result. Many other variants are possible, with the concomitant range of sensations reported by amputees. One important and distressing variant is that the microenvironment at the point where the neuromatrix to sensory modulator circuit has been cut may deteriorate over a period of months or years, leading to worse pain than was experienced at the original severing of that circuit. The basic concept and two possible outcomes are shown schematically in Figure 7. Possible applications circuitry
of this theory of neuronal
The most important possible application of this theory is for the treatment of chronic pain, whether phantom limb pain or otherwise. I hypothesize that the cure for pain will be found in one of three areas. First, eliminating the outgoing signal from the neuromatrix through meditation (15), medication or surgery. This has already been proposed, ‘... if the pattern for pain is generated by cyclical processing and synthesis, then it should be possible to block it by injection of a local anesthetic into appropriate discrete areas that are thought to comprise the widespread neuromatrix’ (I), and tested, ‘Injections of lidocaine into the lateral hypothalamus decrease experimentally induced pain in rats...’ (14) but may not prove practical for the long-term treatment of chronic pain. The second approach is to try to stimulate a normal return signal by means of suitable electrical devices or acupuncture. This approach has already yielded some success. The third approach to the cure for pain is to restore the microenvironment at the site of damage to as near normal as is possible by chemical or surgical means. If my theory is correct,
297
(B)
JJ
4
sensory modulator
badly healed
*’ i*
Fig. 7 The phantom limb open-circuit. When a limb is initially amputated there may be no return signal. The difference signal generated in the thalamus between the normal outgoing signal and the absent return is transmitted to cortical regions where it is compared to memory of pain and the closest match leads to the description of feeling given by the patient. Good healing (A) can lead to a sense of well-being in the patient while a significant alteration of the microenvironment in the neuromas at the point of amputation (B) can lead to chronic pain.
the modification of the environment at the sensory modulator part of the circuit, based on an understanding of what the normal modulator microenvironment should be, will be the most fruitful route to pain relief for millions of sufferers. A second possible application is in the understanding of the relationship between the motor circuit and the sensory circuit as it pertains to rehabilitation of patients who have been seriously injured and where such injury may have affected one or both circuits. Training such patients to recognize and accept their new pattern of difference signals may help speed the practical use of such new patterns. References I. 2. 3.
4.
5.
Melzack R. Phantoms and the concept of a neuromatrix, Trends Neurosci 1990; 13: 88. Melzack R, Wall P D. Pain mechanisms: a new theory. Science 1965; 1.50: 971. Melzack R. The gate control theory 25 years later: new perspectives on phantom limb pain. In: Proceedings of the Vlth World Congress on Pain, Elsevier Science Publishers BV 21, 1991. Kuypers P D L, Gielen F L H, Wai R T D, Hovius S E R, Godschalk M, Egeraat J M. A comparison of electric and magnetic action sign& as quantitative assays of peripheral nerve regeneration. Muscle & Nerve June 1993; 634. Uncini A, Santoro M. Corbo M, Lugaresi A, Latov N. Conduction abnormalities induced by sera of patients with mul-
298
6. 7.
8.
9. 10. 11.
MEDICAL HYPOTHESES
tifocal motor neuropathy and anti-GM1 antibodies. Muscle & Nerve June 1993; 610. Berman N. Connections of the pretectum in the cat. J Comp Neurol 1977; 174: 227. Foster G A, Sizer A R, Rees H, Roberts M T H. Afferent projections to the rostra1 anterior pretectal nucleus of the rat: a possible role in the processing of noxious stimuli. Neuroscience 1989; 29: 685. Mayer D J, Liebeskind J C. Pain reduction by focal electrical stimulation of the brain; an anatomical and behavioral analysis. Brain Research 1974; 68: 73. Rees H, Roberts M H T. The anterior pretectal nucleus: a proposed role in sensory processing. Pain 1993; 53: 121. Meier W, Klucken M, Soyka D, Bromm B. Pain 1993; 53. Barker D. The morphology of muscle receptors. In: Hunt
12.
13.
14. 15.
C, ed. Muscle receptors. Handbook of sensory physiology. Berlin: Springer, 1974. Jasza L, Kvist M, Kannus P, Jarvinen M. The effect of tenotomy and immobilisation on muscle spindles and tendon organs of the rat calf muscles: a histochemical and morphometrical study. Acta Neuropathol 1988; 76: 465. Katonis P G, Assimakopoulos A P, Agapitos M V, Exarchou E 1. Mechanoreceptors in the posterior cruciate ligament: histological study on cadaver knees. Acta Orthop Stand 1991; 62: 216. Tasker R A R, Choinitre M, Libman S M, Melzack R. Pain 1987; 31: 237. Melzack R, Perry D. Self-regulation of pain: the use of alpha feedback and hypnotic training for the control of chronic pain. Exp Neurol 1975; 46: 452.