Proprioceptors and their contribution to somatosensory mapping:

complex messages require complex processing'

PETERB. C. MATTHEWS
Laboratog?! ofPhy.~io!ogy,University of O.xf~rd~,
$/nil~ersl'Q Parks Road, 0xfi)rd 8 X d dPT, U . K
Received June 29, 1987

P. B . C. 1988. Proprioceptors and their contribution to somatosensory mapping: complex messages r e q u i ~

MATTHEWS,
complex processing . Can. J. Physiol. Pham~acol.66: 430-438.
This review of other people's work concentrates on two matters. First, which of the various receptors are chiefly involved in
creating the central representations or maps of the body that both underlie the conscious perception of our body image and are
needed to control our motor performance. Second, how are these relatively well-charted signals used in sensorimotor mapping,
Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/23/14. For personal use only.

with particular attention paid 80 various human psychophysical observations and illusions that throw light on the central
integrative rnechanisrns involved. Detailed citation is largely restricted to developments since earlier reviews.

MATTHEWS.
P. B . C. 1988. Proprioceptors and their contribution to somatosensory mapping: complex messages require
complex processing. Can. J. Physical. P h m a c o l . 66 : 430-438.
Cette rkvision du travail d'autres chercheurs traite de deux 'sujets en particulier. Premikrement, des divers rkcepteurs
princhpalement impliquks dams la repdsentation centrale ou N mapping >> du c o p s et qui somt sous-jacents a la perception
consciente de notre image coporelle et nkcessaires au contr6le moteur. Beuxi&mement, de la rnanikre ddont ces signaux
relativement bien structurks sont utilisks dans la representation (mapping) scnsorirnotrice; on porte une attention particulikre aux
diverses observations physiologiques humaines et aux illusions qui aident a comprendre les mCcanismes iintkgratifs centraux
impliquCs. On rapporte en detail principalenaent les dkveloppements obtenus depuis Ies dernikres rCvisions.
Can. J. Physiol. Pharmacol. 1988.66:430-438.

(Traduit par la revue]

The old simplicity TABLE

1 . Receptors that transmit signals during movement
Thirty years ago things looked relatively simple. The joint Position sense Possible neural inputs
receptors were in, and everything else was out. Sksglund (1956,
1973) provided the classical picture of their behaviour. On the I)Joint In capsule
basis of his study on the cat knee, it came to be rather widely In ligaments
believed that joints possessed an array sf receptors, each with its
IE)Skin Fast adapting I and EI
own preferred angle of firing and active over only a limited part Slow adapting J and II
of the angular range of a joint. Thus the problem of constructing
a central map seemed to be the familiar one of requiring the 1EI)MuscHe Spindle: primary and secondary
Tendon organ
higher centres simply to recognize which receptors were active
and which were silent. How the body image was represented IV)Motor centres Corollary discharges
- -

was not at that stage a useful question, and people seem not to
have bothered about the difference between mapping the ment and the disturbances the environment imposes upon us
position of a cutaneous stimulus on a pre-existing map of the while we are doing something. The standard neurological test of
body surface, which can be thought of simply in terms of "position sense,)' namely waggling the patient's passive finger,
receptive fields, and the problem of changing the relative does not comespond to any comnlonly occurring natural
coordinates of the component parts of the map of the body image stimulus. The detection of such essentially external stimuli
as the body moves. Minor complications were recognized, such cannot be the prime function of the internally placed propriocep-
as that the firing of the joint receptors was altered by contraction tors. A priori, this would seem to fall equally within the sphere
of the nearby muscles, but this seems to have been a second of action of the cutaneous receptors.
order effect introducing inaccuracy, perhaps, but not upsetting
the picture. This simplicity has now vanished; joint receptors There is now a comprehensive classification of the large
are largely out and muscle receptors are in9raising the question cutaneous afferent receptors, but there is no point in dwelling on
as to how their signals can be read higher centres to the details. A recent shown by recording
produce what is needed. from single fibres in man, is that for the hand virtually all of
Receptors potentially contributing to position sense these receptors are excited whenever we move it (Hulliger et al.
1979). Introspection tells us that the sensation during movement
Table 1 lists the numerous receptors that transmit signals bears no relation to that elicited by an external stimulus.
whenever we move. The question is which of them contribute Although most of the s m e afferent receptors may be presumed
significantly to position sense, using this term in its widest and to be activated, the signals would seem to be handled differently
loosest sense. I should note here that position sense can be in the two cases, with those produced during movement being at
presumed to be concerned primarily with self-induced move- least partly denied access to the sensorium. As with the visual
'This paper was presented at eke IX International Symposium of suppression that occurs when we make a saccadic eye move-
the Centre de recherche en sciences neurologiques, Universitk de ment, two sorts of mechanism could be responsible for the
Montrkal, entitled Spatial Representations and Sensorimotor Trans- gating. First, it could be the result of movement producing a
fornations (Montrkal, Qud., May 28-29, 1987), and has undergone particular pattern of afferent activity, unlike that normally
the Journal's usual peer review. produced by external stimuli, which then of itself triggers the
Printed in Canada I Impfimt au Canada
 reading system to behave in a particular way. Second, and with d.1982). They will only fire tonically on localized compression
classic physiological support, the afferent signal elicited by of the capsule or on raising the intracapsular pressure by the
movement may be operated upon in conjunction with some infusion of saline. However, they fire phasically in response to a
component of the motor command which tells the analysing variety of less specific stimuli, including capsular stretch, and
centres (such as the dorsal column nuclei) that the sensory so presumably respond during phasic muscle contractions. As
signals are self-generated rather than externally generated. In emphasized by Grigg and Greenspan (1973), the various
%pernay>tern the motor centres send a corollary discharge to the excitatory effects sf muscular contraction raise the possibility
sensory centres to allow them to interpret the sensory message that joint receptors may play a larger part in the signalling of
appropriately. These are included at the bottom of Table 1, as active movements than passive movements, including at inter-
there is no doubt on anatomical grounds that there are plenty of mediate joint angles. These also greatly complicate the task of
internal connections available inside the GNS. "reading" the joint signal in terns of absolute position, since it is
not uniquely dependent on joint angle, but depends also upon
Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/23/14. For personal use only.

Joint receptors the behaviour of a variety of muscles. In conclusion, it seems

Joint afferent receptors have long been recognized histologi- unlikely that joint receptors could be a regular source of
cally, lying both in the capsule and on any special ligaments. information for constructing the body image, but it would be
The reason they are now out of favour is that single fibre premature to dismiss them entirely, particularly in the signalling
recordings in animals, especially those by Burgess and Clark of movement.
(1969) in the cat and by Grigg and Greenspan (1977) in the
monkey, show that for the knee they are only effectively excited Muscle receptors
by moving the joint to one or other of its extremes. In the middle This brings us to the muscle receptors, and note that Table 1
of the range most if not all are normally quite silent, including neglects the free sensory endings supplied by the smallest nerve
while the joint is being moved. If there is no afferent signal, fibres; these seem unlikely to be relevant in the present context.
Can. J. Physiol. Pharmacol. 1988.66:430-438.

there can be no contribution to so-called "joint position sense" The tendon organs lie in series with the muscle and have a low
and the body image. Fenell (B880),however, has been fighting enough threshold for some of them to respond whenever a
a rearguard action and believes that, though weak, the joint muscle contracts. Their sensory contribution would be expected
afferent discharge is enough to help signal throughout the full to be in the genesis of sensations of force and of effort rather
range of joint action; recent human experiments support his than of position, and even here corollary discharges seem to be
view for the awareness of finger movement (Ferrell et al. 1987), at least equally involved (Mceloskey 1978, 1981), so their
but even he does not claim that they are exclusively responsible, sensory role remains largely enigmatic. Of undoubted impsr-
as was the view generally held 20 years ago. Some of the smaller tance for the present purposes are the muscle spindle afferents
knee-joint afferents have recently been found to be excited by that lie in parallel with the main muscle and so signal its length;
nonnoxious stimuli (Schaible and Schmidt 1983), but their in addition they can be specifically excited by the fusimotor
pattern of firing was such that, like the large afferents, they were system or gamma efferents.
considered to be "mainly engaged in signalling that the joint is It continues to merit emphasis that there are functionally
about to leave its working range." separable primary and secondary spindle afferents, differing in
A few years ago it appeared as if the hip joint, with its their sensitivity to dynamic stimuli, so the muscle spindle is
complex range of motion, was equipped with large joint providing two channels of information and is thus probably
afferents firing throughout its range of movement. But on capable of simultaneously signalling the values of two separate
reinvestigating the matter, Rossi and Grigg (1982) found that mechanical parameters, perhaps length and velocity (Matthews
these are muscle spindle afferents supplying the inferior 1972, 1981). Both types sf afferent fire more rapidly while a
gernellus muscle and the joint receptors behave once again as muscle is k i n g stretched than when its length is held constant,
"limit detectors." For the cat elbow, however, Baxendale and but the difference in firing is much greater for the primary
Fenell (1983) using multifibre recording have described an ending. Thus the primary ending is much more of a movement
appreciable level of midrange activity, which they adamantly detector, and the secondary more of a length detector. It will be
claim comes from joint rather than muscle afferents. It remains argued later that we have a sense of absolute position,
an open question whether the contribution provided by joint independent of awareness of movement. This could be based on
receptors varies from joint to joint. In addition, it should be the firing of just the spindle secondary endings, for with
remembered that a given receptor will fire phasically during appropriate gamma bias their dynamic sensitivity may be very
joint movement over a wider range than it will fire tonically to small. In addition, we have a quantitatively graded sensation of
maintained displacement. They are thus somewhat better placed joint angular velocity, which is at least partly based on the
to signal the occurrence of movement than to provide a discharge of spindle primary afferents. It is important to note
continuing measure of absolute position. By studying isolated that these do not give a pure velocity signal, but also fire
innervated sections of the joint capsule, Grigg and Hoffman tonically with the muscle at a constant length; thus a single
(1982) showed that the essential stimulus to the important and observation on their firing frequency cannot tell us, or the
numerous Ruffini receptors is stretch of the capsule rather than sensory centres, whether a joint is moving or still. It seems
transverse compression. This "explains" both why the receptors possible, though this is by no means proven, that this difficulty
are predominantly excited by movements to the extremes, for could be overcome, and angular velocity measured independent
only then does the capsule become taut, and why, as known of position, by combining the signals from primary and
since Skoglund (1956), they may be excited on contraction of secondary endings in an appropriate integrative centre. Know-
nearby muscles, for this may help tense the capsule. In contrast, ledge of present position could also be based on a detailed
the Golgi-Mazzoni receptors found on the inner surface of the memory of past movement, as in the integrating accelerometers
capsule are tuned to transverse pressure on the capsule (Grigg et used in navigation. However, the demands that this
 432 CAN. J . PWYSIOL. PHARMACOC. VOC. 66, 1988

the accuracy of instrumentation and computation remove much a l y in relation to those evoked by pulling upon the tendons so as
of its biological appeal. to actually move the finger joints. Second, when muscle pulling
did produce a sensation of movement, he attributed it to
Evidence for spindle contribution excitation of cutaneous receptors overlying the muscle rather
Vibration than of muscle receptors themselves .
With the dethronement of the joint receptors, the spindle This direct conflict over an experiment that few would wish to
signals are now the ones believed to be chiefly responsible for repeat is most unfortunate; but in my view the balance of
position sense and kinaesthesia. This was, of course, Sheuing- evidence, including that of other workers, favours McCloskey
ton's view at the turn of the century, but for a variety of reasons rather than Moberg and one can accept that the muscle afferent
it came to be discarded (see Matthews 1977, 1982). The initial discharge elicited on tendon pulling does lead to sensations of
evidence for its reinstatement was derived from the use of movement of the appropriate joint. This view is fortified by the
100 Hz vibration in human subjects, and this has continued to be more recent experiments of Gandevia (1985) with electrical
Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/23/14. For personal use only.

an important tool. The value of vibration is that it provides a stimulation of the u l n a nerve at the wrist. By careful adjustment
particularly powerful stimulus for the spindle primary afferents. of the stimulus strength and electrode position, he was able to
Because of their high dynamic sensitivity they are tricked, so to elicit sensations of finger movement without eliciting either any
speak, into firing at a high frequency in a way they would overt muscular contraction or the cutaneous paraesthesiae
normally only do on being rapidly stretched. which axe so prominent on stimulation of a cutaneous nerve. It
The subjective effects of applying vibration to human seems likely that the sensations of movement were produced by
muscles and tendons were first analysed in detail by Goodwin, stimulating the Ia afferents from the spindle primary endings,
McCloskey, and Matthews (1972) and are now well known. since the joint afferents are appreciably smaller and would not
The chief phenomenon is particularly well shown by vibrating be expected to be stimulated at below the motor threshold.
the biceps tendon of one a m while the blindfolded sub~ectis
Relative contibutions of skin, joint, and muscle
Can. J. Physiol. Pharmacol. 1988.66:430-438.

told to keep his sther a m aligned with where he perceives the

vibrated a m to be. He then signals that the vibrated arm feels as Sense sf movement
if it is more extended than it actually is; that is, as where it would On the basis of this and other evidence, virtually everybody
be if biceps were to be grossly stretched. The generally accepted concerned now agrees that muscle afferents do contribute to
interpretation is that the vibration-induced afferent firing is position sense. The precise extent to which joint and cutaneous
being treated as if it were due to muscle stretch, and is used to afferents also do so remains unresolved. For the knee, a variety
update the body image; in other words spindle primary afferent of experiments argue that neither provides a significant contri-
signals do contribute to position sense. Various points still merit bution (Burgess et al. 1982). However, for the jnger the
emphasis. First, note that the sensory information derived from experiments of Gandevia et al. (1983) show that one or both of
muscle does not lead to a sensation of something happening in them contribute to the detection of passive imposed movements,
the muscle itself, but is refemed to the joint and used to indicate provided these are applied relatively rapidly. They studied the
its position. Second, as a logical corollary, it can be noted that perception sf a displacement applied at various velocities to the
vibration of the antagonist, triceps, leads to the feeling that the terminal interphalangealjoint of the middle finger. They did this
elbow is being flexed rather than extended. Thus the subjective with the finger in two different positions. When the finger is
awareness sf joint position would seem to be derived from extended, displacing the finger tip potentially excites A1 types of
summing, with appropriate signs, the spindle activity from all afferent. The muscles change their length, joint receptors may
the muscles acting at a given joint. Third, if one questions the or may not be excited depending upon whether their midrange
subject about what he is actually feeling and what he is silence is absolute, and some cutaneous receptors around the
signalling with his refexnce m , he says that he is largely flexure lines must be activated. When the middle finger is fully
following a sensation of movement of the vibrated arm, and that flexed, while its neighbours are extended by external restraint, it
it is rotating at constant velocity, rather than that it has simply becomes impossible to move its teminal joint voluntarily; this
taken up a new constant position. But pointing tests suggest that is because of the way the tendons of the muscles to the different
there is also a systematic distortion of a sense of absolute fingers are interconnected. Passive movement of the joint is then
position. The chief conclusion is that muscle afferents do no longer transmitted to the muscles, which have been function-
contribute to psition sense, as has since been amply supported ally "'disengaged," so that the effects of an applied displacement
by other types of experiment. are confined to skin and joint. The contribution of muscles on
their own can be studied, with the finger extended, by
Tendon pedlking ring-blocking the finger and SO eliminating the skin and joint
The most direct type of experiment has recently been repeated afferents; the muscles, of course, lie safely out of the way of the
by McCloskey et al. (1983) and by Moberg (19831, each of anaesthetic.
whom independently called upon a surgical colleague to acutely Using relatively rapid movements Gandevia et al. (1983)
expose one sf their own tendons so that it could be pulled upon found that displacements were most reliably detected when all
in the laboratory. Ow pulling the tendon of McCloskey9s the peripheral inputs were operative, next best when the muscle
extensor hallucis longus by 0.5 mrn or more in the direction afferents were working on their own, and least well (and
required to stretch the muscle, he reported the subjective especially badly at lower velocities) when only skin and joint
experience of his big toe being moved downwards, as to be afferents were active. These two latter inputs could not then be
expected from the vibration experiments and in accordance with separated, but the recent work sf Ferrell et al. (198'7) suggests
the view that muscle afferents contribute to position sense. that both contributed, because injection of anaesthetic into the
Moberg, however, came to a very different conclusion on the joint impaired performance when the muscles were disengaged
basis of having his finger flexor tendons pulled upon. First, without abolishing d l awareness of movement (though this
sensations of joint movement were very hard to elicit, pmticul- residual sensitivity might perhaps have been due to incomplete
 joint anaesthesia). It should also be noted that cutaneous inputs movement; for the fingers these are probably normally assisted
are important not only for any phasic information they provide by cutaneous and joint affesents, perhaps as a special case.
about movement, but also in setting the excitability of the There is a final complication with these otherwise relatively
central sensory mechanisms. Clark et al. (1986) provided a straightforward conclusions. When Clark et al. (1986) went on
recent example of this Iatter effect by documenting the reduction to study the proximal interphalangeal joint of the index finger, it
in the detectability of movements of the proximal interphalan- was found to lack an absolute position sense, while having a
geal joint of the index finger that can be produced by anaes- normal movement sense, some of which must have been derived
thetizing the skin of the thumb. from muscle. The adjacent metacarpophalangeal joint, how-
ever, does have pure position sense both for flexion and for
Sense of absolute position abduction. They pointed out that the two interphalangeal joints
The above experiments were concerned essentially with the of a given finger cannot be moved entirely independently, and
detection of movement. Recent work by Clark et al. (1985, there may not be a unique relation between the lengths of the
Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/23/14. For personal use only.

1986) strongly fortifies the case for the existence of a sense of various muscles involved and the angle of each sf the joints, so
absolute position, distinct from any sense of movement, and the CNS might be faced with an insoluble set of equations in
which seems likely to arise solely from muscle; the preferred determining the absolute position of both joints on the basis of
candidate for this must be the secondary ending of the muscle muscle afferent signals; but this does not entirely resolve the
spindle. What is normally called position sense would seem to puzzle. From ordinary introspection we would seem to be just as
apply to a combination of sensations of movement and aware where our fingers are, as we are for our other joints, so we
sensations of position, which are normally inextricably inter- would seem to be synthesizing a similar subjective concept
twined. Clark et al. separated these by displacing a joint so (namely position) from a different type of neural operation,
slowly that the subject was quite unaware of any movement since we do not actually have a measure of absolute position.
taking place. The slowest movements, applied to the ankle, The peripheral neural machinery would seem likely to be equally
Can. J. Physiol. Pharmacol. 1988.66:430-438.

were 3.5" applied at 8.25"/min, and took nearly a quarter of an present for all muscles, with both primary and secondary
hour to complete. The subject did not even have to pay attention spindle afferents, but we do not seem to avail ourselves of this
during this time, and at the end was asked how the angle of the information which, even if not entirely unique, would seem able
joint compared with the same one on the other side of the body, to contribute.
or sometimes to locate by memory the initial position of the
ankle. The original experiments were on the knee. Here, as Problem of' decoding spindle messages
elsewhere, they found that subjects were very good at recogniz- Accepting the evidence that muscle receptors are crucially
ing a new position although they were quite unaware of any implicated in position sense leads inevitably to the question:
antecedent movement; they just became aware that the angular How easily can their messages be read? The answer is not at all
position had changed. readily, and we have practically no information as to how it is
In recent experiments, on the ankle and on the metacarpo- done. Indeed, some earlier workers were so appalled at the
phalangeal joint of the index finger, Clark et al. (1985) studied complexity of the task that they simply refused to accept that
the effect of varying the velocity of displacement. They argued muscle receptors could be involved in position sense (for
that on lowering the velocity the performance of a sense of example, Mountcastle and Powell 1959). Let me run through
absolute position would be relatively unaffected, whereas that some of the difficulties, most of which have recently been
of a movement sense should be drastically disturbed. Under emphasized by Burgess et al. ( 1982).
normal conditions abduction at the metacarpophalangeal joint (1) The rate of firing of spindle receptors depends in a
elicited the response expected from a positional system, with complex manner on both the length of the muscle and the way in
100% detection of 2.5" applied at all velocities. However, when which it is changing, and particularly whether it is lengthening
they eliminated the muscular contribution, while leaving the or shortening, so no one receptor gives us a simple measure of
cutaneous and joint contribution, the behaviour changed to that any one physical variable. As already noted, one might perhaps
of a velocity detection system and slow movements were not be able to do something about this by appropriately combining
perceived. This was done by anaesthetizing the ulnar nerve information from both primary and secondary endings.
which paralyses the relevant muscles while leaving the finger (2) The rate of firing when muscle length is held constant
itself unanaesthetized, since it is supplied by the median nerve. varies both with the length of time the muscle has been held at
Injection of local anaesthetic into the joint cavity had no effect the new length (adaptation) and on whether the length was
on normal performance, so there is no question of pure position reached by stretching or releasing the muscle. Such effects
sense being solely due to joint afferents because animal might perhaps be less marked when the spindles are under their
experiments show that these are then paralyzed. Broadly similar normal fusimotor drive rather than being studied in de-
results were obtained with the ankle; in addition, they determin- efferented preparations. In this respect it is interesting that
ed the threshold of the least stimulus that could be detected, Paillard and Brouchon (1968) showed that the accuracy of
whether as an absolute displacement or as a sensation of matching the position of one a m with that of the other was
movement. On reducing the velocity of rotation the threshold better when the first a m was moved actively by the subject
first rose appreciably, but then settled at an approximately rather than passively by the experimenter. In addition, they
constant vdue irrespective of further reduction in velocity. found systematic distortions of position sense with the passage
It may be concluded from all this that muscle afferent input of time from the initial movement and the subsequent matching,
can lead to a sense of absolute position independent of any sense so perhaps not all these receptor errors are fully corrected.
of movement. It must, at least partly, utilize its own peripheral (3) Wei et al. (1986) emphasized that in the absence of
afferent machinery, most plausibly the secondary ending of the contraction the muscles acting across the cat ankle fall slack
rnuscle spindle. In addition, muscle afferents (probably from when the joint is near one or other of its extremes and then no
the primary spindle endings) contribute to a sensation of joint longer change in length with small changes in the angular
 434 CAN. 4. PHYSIBL. PHARMACOL. VOL. 66, 1988

position of the ankle. Their spindles then cease to provide a A complementary sbsemation was provided by Hulliger et al.
signal of joint movement while those in the now well-stretched (1982) when they recorded the discharge of a human spindle
antagonist continue to do so, and the same seems likely to hold afferent from a contracting finger extensor while the subject was
for man. Thus, as already concluded from the effect of moving his finger against a constant load from one position to
vibration, human position sense must be compounded from both another; movement was restricted to the metacarpophalangeal
agonist and antagonist signals. joint. They found that the spindle firing was the same whether
(4) Simple hinge joints with a single degree of freedom are a the muscle was long or short, indicating that the change in
very particular case, and even here a given muscle may act at fusimotor activity had precisely counteracted the normal length
more than one joint. Thus, and even more so with ball and sensitivity of the afferent. Yet the subject was aware that the
socket joints, the activity of receptors in different muscles must position of his finger was different. So here there seems to be a
be combined by means of appropriate algorithms to produce a sensory experience in the absence of the expected afferent signal
unitary position sense and spatial map of the body. (the antagonists, however, were not studied). They were
(5) More fundamentally, the familiar frequency code of obviously thoroughly puzzled by this and after briefly mention-
Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/23/14. For personal use only.

afferent firing is employed to indicate position, that is, ing ways to escape from the conundrum finished by saying
topography, rather than to signal intensity; this contrasts "Thus, in spite of the absence of explicit position monitoring
powerfully with the cutaneous system where we are used to some remote possibilities remain that muscle spindles may
thinking of higher frequency firing as leading to a more intense provide central structures with infomation that might contri-
subjective sensation. This may perhaps be true also for bute to position sense." But there is nothing remote about a
sensations of joint mvement; but, as emphasized by Burgess et spindle contribution, once one accepts the inevitability of
al. (1982), the quality of one's awareness of the absolute corollary discharges contributing to the analysis. Rather, given
position of a limb seems much the same throughout the normal what we h o w about other systems, it is naive to suppose that
range, yet under static conditions the summed discharge of they are not involved and to treat a search for them with the
spindle receptors seems likely to be much greater at the microelectrode as being rather like trying to find the soul.
Can. J. Physiol. Pharmacol. 1988.66:430-438.

extremes. It follows also that if the map of the body image is laid
out in some topographical manner across an m a y of nerve cells, Proven action of cgero%lary discharges in other systems
that is, as a spatial code, then the initial frequency code has to be
It seems pertinent to describe three sets of observations
totally transmuted. testifying to the reality of corollary discharges.
Fusimotor action and need for cors&&ary discharges Oculsrnotor system
The need for neural processing of the raw receptor signals is First, there are those of Guthrie et al. (1983) which show that
shown even m r e clearly on considering the effects of fusimotor the oculomotor saccade control system of a conscious monkey
activity on the afferent discharge. Fusimotor input does two continues to know the position of the eye in the orbit, quite
separate things to the spindle afferents. First, it biases them so independently of visual clues, after deafferentation of the
they fire at a greater rate for a given muscle length. Second, it extraocular muscles. This can then only come from c o r o l l q
changes their sensitivity to externally applied stimuli. In other discharges. Second, again from the oculomotor system, Rich-
words, the spindle afferents have a variable calibration, so their mond and Wurz (1980) recorded from cells in the superior
discharges cannot be interpreted in quantitative mechanical colliculus, which responded quite differently to translation of
terns of joint angle and so on without a knowledge of their the visual image across the retina depending upon whether this
current sensitivity, since this may be varying from moment to was done by the experimenter displacing the external visual
moment. In principle this is readily done by supplying the scene, or by the monkey itself making a saccadic eye move-
reading centres with corollary discharges indicating the current ment. This difference demanded that the coHicular cells
level of fusimotor activity, and hence the correct calibration received signals related to eye movement as well as from the
factor. The two separate systems of fusimotor fibres, the static retina. Because the essentials of such behaviour remained after
and the dynamic axons, seem to be able to change the static and paralysing the eye muscles and preventing overt saccades,
dynamic calibrations sf the endings relatively independently, so corollary discharges are shown to have been involved.
making the appropriate adjustments is no mean task. It would
seem to be made yet more complex by the existence s f beta Electric jsh
axons that supply both the spindles and the extrafusal muscle Third, turning to a quite different part of the animal kingdom,
fibres. there are Bell's (1982) striking findings on the momyrid
Let us turn next to the problem of spindle bias, since this is electric fish. This species both sends out its own electrical pulses
somewhat easier to think about. When a conscious human and, apparently quite independently, has specific ampullaq
subject makes an isometric contraction, his spindles fire faster electroreceptors to pick up external electrical fields preexisting
because the gamma efferents have been co-activated along with around itself. These external signals would, however, be
the main alpha motor fibres. If this Ia discharge had k e n elicited jammed by its own activity, so whenever the momyrid emits an
by stretching the muscle, the subject would have felt that the electric discharge it sends a corollary discharge to its receiving
appropriatejoint was rotating, but as is easily verified when one centres to prevent this happening. Their action on neurones in
es an isometric contraction against a table there is then no the sensory receiving centres can be directly observed with a
experience of movement. Thus here is a ""sensory signal" that microelectrode. The situation is simplified by curarizing the Ash
fails to elicit a sensory experience. The likely reason is that the so that no actual electrical discharge occurs. Instead, an
afferent message is read in conjunction with a corollary artificial discharge is produced by picking up the descending,
discharge from the motor centres signalling that the movement but now ineffective, motor volley to the electric organ with an
has been commanded so that the consequential sensory excita- electrode in the tail, and then using this to trigger an electrical
tion can be discounted. stimulator acting through the water around the fish. This
 MATTHEWS 435

artificial electric input elicits a complex pattern of excitation and he has done so, although of course he knows he has tried. Thus
inhibition from the appropriate central sensory neurone. On there is the interesting situation that a subject can dispatch a
commencing such stimulation there is initially no corollary motor command, and presumably with it the normal corollary
discharge to the newone, since it gives no response when there discharges, and actually make a movement which the sensorium
is a descending motor command to the electric organ without an entirely fails to register. Presumably, ischaemia paralyses
actual electric discharge, whether real or artificial. However, sensory axons slightly in advance of motor axons, thereby
when the artificial discharge is regularly paired with the removing the signal that the sensorium normally relies upon to
"command" of the descending motor volley, then over a period tell it what has happened, and the continuing corollary dis-
of 5 min a corollary discharge gradually develops and the charges cannot take their place. It might be suggested that the
sensory neurone comes to alter its pattern of firing to the corollary discharges were indeed simple efference copies, but
command signal occurring on its own, without the superadded were purely inhibitory discharges requiring subtraction from
electrical stimulus. The purely corollary-evoked response was excitatory signals and that in the absence of these, with afferent
very approximately the mirror image of the direct response to paralysis, the system was just biased into total inactivity. But
Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/23/14. For personal use only.

stimulation, unmodified by a corollary. Moreover, the response this will not explain the similar finding obtained when the
to the stimulus was appreciably reduced when the neurone was subject is curarized, whether locally or generally (McCloskey
simultaneously acted upon by the corollary. When the fish was and Torda 1975; Stevens et al. 1976). Local paralysis can be
left undisturbed for 15 min or so the corollary seems to be obtained by injecting curare into a vein when the circulation is
switched off, since the command motor discharge no longer occluded, while total paralysis is a much more heroic proced-
produces an effect on the sensory neurone. Recording from the ure. In both cases when the subject attempts to move, he has no
primary afferent fibres confirmed that their behaviour was sensation of any movement occurring and indeed is acutely
constant throughout, and that they were not under efferent aware that he is quite unable to move. Stevens (1978)
control. commented that with whole body paralysis "one felt like a solid
Several lessons can be derived from all this. First, a signal can piece of cement. It was very much like being buried alive." In
Can. J. Physiol. Pharmacol. 1988.66:430-438.

occur in a sensory neurone on dispatch of a motor message; that this case, many spindle afferents should have been continuing to
is, we are looking at the direct effect of a corollary discharge. fire tonically, in spite of being deprived of fusimotor drive, so
Second, the genesis of this corollary is highly plastic and in this any inhibitory corollary discharges would have a background of
case it is only produced if the command elicits an unwanted sensory activity to modify.
sensory input. Third, the pattern of the corollary is tailored to One might take this as evidence that corollaries don't exist,
the pattern of the unwanted input; this was demonstrated yet but it seems much more likely that, in the late Donald MacKay's
more fully in other experiments in which the timing, polarity. or tern, they are used to evaluate a sensory message rather than
size of the stimulus was varied. Hence this particular corollary simply suppress it. Thus to experience movement, there has to
has some of the properties of the efference copy postulated be a change in afferent firing for one to be aware of anything
many years earlier by von Holst, which could be used to annul happening, but the responsible afferent signals have to be read
the unwanted signal by simple subtraction. Fourth, one should with the help of corollary discharges. However, it is not yet
note that this electric fish has two other sets of electroreceptors clear whether this failure of corollary discharges to produce an
which are used for quite different purposes and for each of effect on their own applies equally to the sense of absolute
which corollary discharges are also used centrally but in position, and more particularly for motor coordination involv-
different ways, with neither behaving like efference copies ing spatial mapping. This has been emphasized by Stevens
(Zipser and Bennett 1976; Bell 1984). One system, which is (1978) on the basis of his own experiences of total extraocular
used to detect the discharges of other fish, is simply briefly shut paralysis. On attempting to move his eyes he had the bizme
off by the corollary of the emitting fish. The other system, which experience that though the visual world showed no sign of
is used to detect the electrical effects of the fish's own pulses, is movement, it was somehow displaced. This was not primarily a
briefly facilitated by the corollary. Thus corollary discharges visual experience, but seemed to have more to do with the
can be employed in several different ways and be of various spatial mapping required for movement. When the paralysis
kinds. Finally, and most importantly, if this level of the animal was produced by retmbulbar injections of anaesthetic, he
kingdom can produce and use corollary discharges of such systematically missed on pointing at a target when he was trying
variety and sophistication, we have no right to have survived in to look to one side. The analogous experiment of pointing to a
the evolutionary struggle if we cannot do as well. paralysed limb that one is trying to move has yet to be
performed. Thus the precise role of corollary discharges in the
Mode of action of coro llaries in proprioception synthesis of the body image remains an enigma, and some may
Returning to the main theme, the interest now is not in asking well continue to question their existence. In the related question
simply whether corollary discharges are involved in the genesis of the awareness of "heaviness," however, there is strong
of human position sense, for it seems to me that they must be. evidence again that they take part, but can now, of themselves,
The current questions are where do they occur, and how are they elicit a conscious percept (McCloskey 1981; Gandevia 1987).
used? Here the only definite thing to say is that, as far as the
sense of movement is concerned, they do not behave like the Complications ingroduced BPg, tendon compliance
most elementary types of efferent copy and are not capable on Further investigation requires one to remain continually alert
their own of initiating a sensory experience. This was empha- as to how the peripheral machinery behaves, for this underlies
sized by Goodwin et al. (1972) on the basis of their observations all our sensory judgements, and unsuspected peripheral effects
of making one a m ischaemic by inflating a blood pressure cuff may be lurking in the background ready to mislead. A particular
around it and so pardysing both sensory and motor axons. Just observation for which this might be so is the effect of a changing
before paralysis is complete the subject passes through a stage in level of contractile force in distorting perception of the extent of
which he can make finger movements without being aware that a voluntary movement. Three separate groups of workers
 436 CAN. J. PHYSIBL. PHARMACOL. V O L . 66, 1988

(Roland 1978; Wymer and D' Almeida 1980;Watsonet al. 1984) and elongating so as to remain in contact with the apparently
have independently described the phenomenon in apparent moving m. As one said, "I feel like Pinocchio." So once again
ignorance of each other's contribution. It is that when one the body map can be temporarily redrawn on the basis of a
makes a movement requiring an increasing strength sf contrac- fleeting sensory input.
tion as one moves, as against a spring, then the extent of the J. R. Lacher has produced a large number of other
movement is systematically overestimated by an amount illustrations, similar in principje, on vibrating a variety of
increasing with the increase in exerted force. No such effect is muscles in subjects adopting a variety of postures. There was a
found when a constant force is exerted throughout the course of high degree of variability between different subjects as to how
the movement. Without finally deciding the matter, the various their sensory analysing centres responded to this unusual chal-
authors seem to have felt that this should be attributed either to lenge. But the abiding principle was that our subjective map sf
an interaction between the signals from spindles with those from our body is fluid and highly modifiable from moment to
tendon organs or more likely to an interaction with changing moment. It cannot consist simply of a Cartesian doll with its
corollary discharges. joints movable over a known range. From the point of view of
Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/23/14. For personal use only.

A mush more prosaic possibility, which is ripe for study, is the control of movement, however, one suspects that there must
that the error is a simple consequence of the tendon stretching be other maps with the anatomical constraints built in, since it
under the increasing contractile force, SO that the muscle was would be a disaster if we were to use our full muscular power to
shorter than would otherwise have been expected from the overt try to drive our joints beyond the range they are designed for.
psition of the limb or digit. If so the production of an equivalent Perhaps the real job of the joint receptors is indeed to act as
muscle shortening at constant force, which provides the "limit detectors," and to provide this information in a totally
standard of reference, would entail a greater joint movement, unambiguous manner for the motor system, rather than to keep
since none of the muscle shortening would be lost in the tendon. our subjective sensory map updated. This woujd At well with the
Rack (1985; Rack and Ross 1984) has recently emphasized how gross pathological effects seen with denemation of joints.
greatly tendons may stretch under their normal loads and the
Can. J. Physiol. Pharmacol. 1988.66:430-438.

Relation of body to externaH space

inescapable implications of this for motor control; for example,
Next there is the use of vibration to alter one's subjective
the tendon of the long flexor of the human thumb may elongate
experience of the location of objects in external space; this
by an amount equivalent to a joint movement of 7O for an
demonstrates that there are powerful interactions between quite
increase of load from 8 to 25% maximal, and by a similar
different types of neural signal. In an experiment by Lackner
amount again on increasing the contraction to maximal. It seems
and Levine (1978; Levine and Lackner 19'79) the subject sat in
inevitable that such simple mechanical behaviour must contrib-
darkness with a light source attached to his hand and which he
ute to the observed psychophysical effects, but whether it could fixated. The elbow flexors were vibrated, the a m was felt to
be totally responsible remains to be tested. The moral to be
move, and with it the light appeared to move in space. The
drawn is that there is no escape from continuing attention to the
abnormal proprioceptive input had taken precedence over the
peripheral machinery, even when one's chief interest is in
normal cues of the orientation of one's eyes. In related
central mechanisms.
experiments the same sort of thing was found for a localized
Central maps auditory stimulus, suggesting that-it is not simply a matter of
preferring the measure of the length of one set of muscles (the
Piasticity of body image
elbow Wexors) over that of another set (the extraocular muscles)
Finally, there are various recent psychophysical experiments but rather a remapping of one" spatial orientation. Biguer et al.
that illustrate the complexity of the central maps we have of our (9986) applied vibration over the small muscles at the side of the
body image and of their relation to the external world. Most of neck and so, presumably, causing their muscle spindles to send
them use vibration as a way of injecting an abnomal level of the usual false signal of their length, thereby producing a false
proprioceptive input, presumably chiefly from the spindle indication of the orientation of the head on the body. Again,
primary endings. First, Craske (1977) vibrated the wrist flexors small lights viewed in darkness were perceived as moving,
in the forearm and asked subjects to p i n t to where they felt their although the head and the eyes were still. Moreover, when the
wrist to be.As usual they indicated that they felt the joint to be in subject attempted to point at them, he systematically missed
the position it would be in if the flexors had been stretched, and them. In this case there should have been no abnormality in the
the wrist was felt to be extended. The new thing he stressed was signals from the muscles that were being used to move the a m .
that the subjects were quite prepared to indicate that the wrist Again, spindle input would seem to have led to a remapping of
was hyperextended far beyond its normal range of movement. space, both subjectively and for motor commands. All these
In other words, the map is highly labile and does not have built experiments establish that spindles provide information not only
into it a knowledge of what is anatomically possible. Using the for the creation of a subjective map of the body, but also for the
map is not like consulting a pre-existing atlas where one can orientation of this map in the external world.
only find that which has already been charted. Rather one would
seem to be modifying the mapas one uses it, and letting current Increased gravity
observations override all previous experience. Another example It might be suggested that vibration is not quite to be trusted,
of this has very recently been provided by 9 . 8 . Lackner, as yet that not only is it generating an abnomaljy high spindle input,
unpublished. He vibrated the biceps while the subject was but that it is also exciting a mass of extraneous jamming signals
holding onto his nose and who, as usual, felt that his elbow was from a variety of other phasically sensitive receptors, so that we
extending. But since the stationary nose was still being held this are witnessing a failure of the analysing centres rather than
produced a sensory conflict, arising from the contradictory getting a glimpse of how they nomdly handle spindle data.
infomation as t s where his m was. Some subjects solved this Against this are the observations of Lackner and Graybiel
problem very simply and logically, but by entirely unconscious (1981, 1984) on some subjective effects of temporarily increas-
nervous operations acting on their internal body map, and had ing akne apparent force of gravity. This was done by diving a
the clear subjective impression that their nose was growing out large jet aircraft from 32 000 to 24 000 f ( 1 f = 8.385m) and
 then back up again along an appropriate trajectory. In the available to the sensorium. Moreover, this motor map would not
pull-up phase the apparent gravitation force was held at 2 g as the seem to be using information from the spindle primaries or else
plane accelerated upwards. During this time the subject is told to it too would have been distorted. In another experiment the
step on and off a low stool. Not surprisingly he finds the subject is told to move his am,again as fast as possible, from an
stepping up requires unusually great effort; but, in addition and initial position to a target, which of course he can do. The m is
quite unexpectedly, he has the illusion that the stool sinks down then vibrated at the initial position so that its real position and its
as he steps onto it. This seems likely to arise because the normal perceived position come to differ; normally it tended to flex
spindle pattern of firing has been changed as a result of under the action of the tonic vibration reflex but felt as if it had
interference with the normal relation between the rate of not moved. When told to move rapidly to the target the subject
movement, and the amount of alpha motor discharge, and the did so as successfully as before. Electromyographic recording
level of co-activated fusimotor activity. One may conclude, showed that the motor discharge differed in the two cases, so
without getting entangled in the details, that once again muscle there is no question of his achieving the correct target simply by
Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/23/14. For personal use only.

a f f e ~ n activity
t is being used to alter one's perception of the the mass-spring properties of the arm operating upon an
external world, as well as of one's own body image; cutaneous invariant motor discharge. Rather, he behaved as if his motor
afferent activity might equally be involved. The power of such machinery h e w the correct position of the a m even though his
action was shown yet further when the subject did simple knee sensorium did not. The provisional conclusion would seem to be
bends, moving himself up and down during exposure to 2s. In that we have one map of the body which is consciously
addition to experiencing movement of the floor beneath him, he perceived, distorted by vibration, and perhaps used to control
had the visual illusion that the whole plane was moving as well slow movements, and another subconscious map which is not
as himself; when he went down it rose, when he stood up it affected by vibration and which can be used to programme a
usually sank. It may again be concluded that somatic sensory ballistic movement.
activity plays a part in the mapping of external as well as internal Conclusions
Can. J. Physiol. Pharmacol. 1988.66:430-438.

space. M a t final conclusions can be drawn from the varied matters

that have been discussed above?
Mu ktiplicitg?of mops ? First, we do now have a fair idea of the origin of the peripheral
k t me give you a final question: Is there just one map of our messages underlying position sense, with muscle spindle
body image which we use equally for sensory and for motor afferents given pride of place.
purposes? The answer seems likely to be no. First, we have the Second, these signals are by no means easy to read and
example of the visual system before us. As long known, patients probably need to be processed both in relation to each other, and
with damage to the occipital cortex are blind, in that they say in conjunction with corollary discharges from the motor centres.
they cannot see; they have no consciously accessible visual map Third, the maps they feed into are not just maps sf the body
of the external world. However, Weiskrantz et al. (1974) itself but also include its relation to the external environment.
showed that if a light is flashed in a variety of positions these Moreover, these maps have a surprising degree of fluidity to
"blind" subjects can point rather accurately to its position, even them, in that when the proprioceptive input suggests it the
though they have not consciously seen it. Thus their motor sensorium makes little attempt to maintain the normal body
machinery still has access to a map of the external world which shape.
has an input from the retina. This so-called "blindsight" Fourth, we must be prepared to find more than one site of
indicates the existence of a topographically organized map of mapping, with the maps probably being used for different
visual space lying outside the normal visual cortex, and beyond purposes. There is nothing surprising about this as electrophys-
the reach of our normal conscious self. The irrelevance of the iologicd recordings disclose multiple representations of a given
conscious map of visual space for another motor act, namely the input. For those who would study such processing with a
control of saccadic eye movements, was shown by Lackner and microelectrode, the difficulties ape formidable; but given the
Levine (1979) by distorting the conscious map by example of the visual system, the achievement may be great.
The subject had a light attached to his hand which he perceived
as moving to a new position when biceps were vibrated.
However, when told to look at the light he made saccadic BAXENDAEE, R. H . , and P E ~ E L W.
L , R. 1983. Discharge characteris-
movements to its true position rather than to its perceived tics of the elbow joint nerve of the cat. Exp. Brain Res. 261:
position; his eyes were initially directed either at another target 195-203.
light or at his other arm whose location was being accurately BELL,C. C. 1982. Properties of a modifiable efference copy in an
electric fish. J. Neurophysiol. 4'7: 1043- 1056.
reported by his position sense. Thus the saccadic generator was 1984. Effects of motor commands on sensory inflow, with
using an undistorted map of visual space, while the conscious examples from electric fish. In Comparative physiology of sensory
self was being deceived. systems. Edited by L. Bolis, R. D. Keynes, and S . H. P. MaddrelB.
Next, there are two surprising recent observations with Cambidge University Press, Cambridge.
vibration which suggest that we may have more than one map of BIGUER, B., BONAEDSON, I. M. L., HEIN, A . , and JEANNEROD, M.
limb position (Sittig et al. 1985; Sittig 1986). In one experi- 1986. La vibration des muscles de la wuqrne modifie la position
ment, the subject is ask& to track a moving target light with his appafente d'une cible visuelle. C.R.Acad. Sci. Ser. 3,303: 43-48.
arm while it is being vibrated. As usual he thinks he is BURGES~, P. R., and CLARK, F. J . 1969. Characteristics of knee joint
sucwssfu~lydoing this, although he is showing the standad receptors in the cat. 9. Physiol. (London), 203: 3 17-335.
vibration-induced mismatch (he cannot, of course, see his am). BURGESS, P. R., CLARK,F. J., SIMON,J., and WEI, J. W. 1982.
Signalling of kinesthetic information by peripheral sensory reeep-
He is then told to move his rapidly so as to align it properly tors. Annu. Rev. Neurosci. 5: 171-187.
with the target. This makes no sense to him, since he thinks he is CLARK, F. J., BURGESS, R. C., CHAPIN, J. W., and LIPSCOMB. W. T.
already doing so; none the less he makes a movement and brings 1985. Role of intrmuscular receptors in the awareness of limb
the m accurately to the target. This suggests that certain motor psition. J. Neursphysiol. 54: 1529-1 540.
acts may employ a quite different map of limb position from that CLARK, F.J., BURGESS, W. C., and CHAHN,J. W. 1986. Ropriocep-
 438 CAN. B. PHYSHOL. PMARMACBL. VOL. 64, 1988

tion with the proximal interphalangeal joint of the index finger. MCCLOSKEY, D. I. 1978. Kinesthetic sensibility. Physiol. Rev. 58:
Brain, 109, 1195-1208. 763-820.
CRASKE, B . 1977. Perception of impossible limb positions induced by 1988. Corollary discharges and motor commands. In Hand-
tendon vibration. Science (Washington, DC), 896: 7 1-73. book of physiology. The nervous system. Vol. III. Rlotor control.
FERRELL, W. R. 1988. The adequacy sf stretch receptors in the cat Edited by V. B. Brooks. American Physiological Society, Bethesda,
knee joint for signalling joint angle throughout a full range of MD.
movement. J. Physiol. (London), 299: 85-99. MCCLOSKEY, D. I., and TORDA,T. A. G. 1975. Corollary motor
PEWRELL, W. R., GANDEVIA, S. C., and MCCLOSKEY, D. I. 1987. The discharge and kinaesthesia. Brain Res. 100: 467-470.
role of joint receptors in human kinaesthesia when intramuscular MCCLOSKEY, D. I., CROSS,M. J., HONNER,I%., and POTTER,5. K.
meptors cannot contribute. J. Physiol. (London), 386: 63-7 1 . 1983. Sensory effects of pulling or vibrating exposed tendons in
GANDEVIA, S. C. 1985. Illusory movements produced by electrical man. Brain, 106: 21-37.
stimulation of Bow-threshold muscle afferents from the hand. Brain, ~CDBER E.G1983.
, The role sf cutaneous afferents in position sense,
108: 965-98 1 , Enaesthesia, and motor function of the hand. Brain, 106: 1-19.
1987. Roles for perceived voluntary motor commands in motor MOUNTCASTLE, V. B., and POWELL,T. P. S. 1959. Central nervous
Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/23/14. For personal use only.

controlol.Trends NeuroSci. 10: 8 1 -85. mechanisms subserving position sense and kinesthesis. Bull. Johns
GANDEVIA, S. C., WALL,L. A., MCCLOSKEY, D. I., and POTTER,E. Hopkins Hosp. 105: 173-200.
K. 1983. Proprioceptive sensation at the terminal joint of the middle PAILLARD,M e , and BROUCHON,M . 1968. Active and passive
finger. J. Physiol. (London), 335: 507-5 17. movements in the calibration of position sense. In The neuropsy-
GOODWIN,G. M . , MCCLBSKEY, D. I., and MATTHEWS,P. B . C. chology of spatially oriented behaviour . Edtfed by S . J. Freedman.
1972. The contribution of muscle afferents to kinaesthesia shown by Dorsey Press, Homewood, IL. pp. 37-55.
vibration induced illusions of movement and by the effects of RACK,P. M. H. 1985. Stretch reflexes in man: the significance of
paralysing the joint afferents. Brain, 95: 705-748. tendon compliance. In Feedback and motor control in invertebrates
GRIGG,P., and GREENSPAN, B. J. 1977. Response of primate joint and vertebrates. Edited by W. J . P. B m e s and M. H. Gladden.
afferent neurons to mechanical stimulation of knee joint. J . Croom Helm, London. pp. 2 17-229.
Neurophysiol. 40: 1 -8. RACK,P. M. H., and ROSS,H.F. 1984. The tendon of flexor pollicis
Can. J. Physiol. Pharmacol. 1988.66:430-438.

GRIGG,P., and HOFFMAN, A. H. 1982. Properties of Ruffini afferents longus: its effects on the muscular control of force and position at the
revealed by stress analysis of isolated sections of cat knee capsule. J. human thumb. J. Physiol. (London), 351: 99-1 10.
Neurophysiol. 47: 41 -54. RICHMOND, 9. R., and WURZ,R. H. 1980. Vision during saccadic eye
GRIGG,P., HOFFMAN, A. H.,and FOGARTY, K. E. 1982. Rope~tiesof movements. II. A corollary discharge to monkey superior cslliculus.
Golgi-Mazzoni afferents in cat knee joint capsule, as revealed by J. Neurophysiol . 43: 1 1 56- 1 1 67.
mechanical studies of isolated joint capsule. %.Neurophysiol. 47: ROLAND,P. E. 1978. Sensory feedback to the cerebral cortex during
41-40. voluntary movement in man. Behav. Brain Sci. 1: 129- 847.
GUTH~E B., L., PORTER,J. D., and SPARKS, D. L. 1983. Corollary RBSSI,A., and GWIGG,P. 198%.Characteristics of hip joint mechano-
discharge provides accurate eye position information to the oculo- receptors in the cat. J . Neurophysiol. 47: 1029- 1042.
motor system. Science (Washington, DC), 221: 1 193- 1 195. RYMER,W. Z., and B'ALMEIDA,A. 1980. Joint position sense: the
HULLZGER, M . , N o ~ HE.,
, TWELIN, A.-E., and VALLBO,A. B. 1979. effects sf muscle contraction. Brain, 103: 1-22.
The responses of afferent fibres from the glabrous skin of the hand QHAIBLE,W-G., and SCHMIDT, R. F. 1983. Responses of fine medial
during voluntary finger movements in man. J. Physiol. (London), articular nerve afferents to passive movements of knee joint. J.
291: 233-249. Meurophysiol . 49: 1 1 18- 1 126.
HULLIGER, M., N o ~ HE., , and VALLBO,A. B. 1982. The absence of SH~IG A., C. 1986. Kinesthesis and motor control. Doctoral thesis.
position response in spindle afferent units from human finger University of Utrecht.
muscles during accumte position holding. J. Physiol. (London), SITTIG,A.C.,DENIERVANDERGCBN,J. J.,GIELEN,C.C. A. M., and
322: 167-179. VANWIJK,A. J. M. 1985. The attainment of target position during
LACKNER, 9. 9%. ,and GRAYBHEL, A. 198 1. Illusions of postural, visual, step-tracking movements despite a shift of initial position. Exp.
and aircraft motion elicited by deep knee bends in the increased Brain Res. 60: 487-410.
gravitoinertial force phase of parabolic flight. Exp. Brain Res. 44: SKWLUND,S. 8956, Anatomical and physiological studies of knee
312-316. joint innervation in the cat. Acta Physiol. Scand. Suppl. 36: 124.
1984. Perception of body weight and body mass at twice 1973. Joint receptors and kinesthesis. In Handbook of sensory
emh-gravity acceleration levels. Brain, 107: 133-144. physiology. Vol. II. Somatosensory system. Edifed by A. Iggo. pp.
LACKNER, J. R., and LEVHNE, M. S. 1978. Visual direction depends on 1 1 1-136.
the operation of spatial constancy mechanisms: the oculobrachial STEVENS, J. K. 1978. The corollary discharge: Is it a sense of position
illusion. Neurosci. Lett. 7: 207-212. or a sense of space? Behav. Brain Sci. 1: 163- 165.
1979. The guidance sf saccadic eye movements to perceptually STEVENS, J. K., EMERSON, R. C., ~ E W S T E IG.
N , k.,KALLOS,T.,
mislocdized visual and non-visual targets. Aviat. Space, Environ. NEUFELD,6 . R., NICHOLS,C. W., and RBSENQUHST, A. C. 1976.
Med. 50: 46 1-465. Paralysis of the awake human: visual perceptions. Vision Res. 16:
LEVINE,M. S., and LACKNER, J. R. 1979. Some sensory and motor 93-98.
factors influencing the control and appreciation of eye and limb WATSON, J. D. G., COLEBATCH, J. G., and MCCLOSKEY, D. I. 1984.
position. Exp. Brain Res. 36: 275-283. Effects of externally imposed elastic loads on the ability to estimate
MATTHEWS, P. B. C. 1972. Mammalian muscle receptors and their psiton and force. Behav. Brain Res. 13: 267-271.
central actions. Arnold, London. WEI, J. Y . , SIMON,I., RANDIC,M . , and BURGESS, P. R. 1986. Joint
1977. Muscle afferents and kinaesthesia. Br. Med. Bull. 33: angle signalling by muscle spindle receptors. Brain Res. 370:
137-142. 108-1 18.
1981. Muscle spindles: their messages and their fusimotor WEISKRANTZ,L., WARRINGTON, E. K., SANDERS,M. D., and
supply. 8n Handbook of physiology. The nervous system. Vol. IIH. MARSHALL,J. 1974. Visual capacity in the hernianopie field
Motor control. Edited by V. B. Brooks. American Physiological following a restricted cortical ablation. Brain, 97: 709-728.
Society, Bethesda, MD. ZTPSER, B., and BENNETT,M. V. L. 1976. Interaction of electrosen-
1982. Where does Shemington's 66muscularsense" originate'? s o y and electromotor signals in lateral line lobe of a morrnyrid fish.
muscles, joints, corollary discharges? Annu. Rev. Neurosci. 5: % . Neurophysiol . 38: 7 13-72 1 .
189-218.