close

Вход

Забыли?

вход по аккаунту

?

Dependence of the achilles tendon reflex on the excitability of spinal reflex pathways.

код для вставкиСкачать
Dependence of the AcMes Tendon Reflex
on the Excimbdity of Spinal Reflex Pathways
David Burke, MD, Brian McKeon, BEng, and Nevell F. Skuse
Muscle afferent activity from the triceps surae was recorded during experimentally induced alterations in
amplitude of the Achilles tendon jerk. No changes in the neural afferent response to tendon percussion or in the
background level of neural activity occurred when the reflex response was altered by discomfort, distraction,
changes in attention, or changes in the rate of tendon percussion. Reinforcement of the Achilles tendon jerk by
forceful contraction of the forearm muscles did not alter the relationship between intensity of the tendon tap and
amplitude of the evoked neural afferent volley. Nevertheless, such maneuvers lowered the reflex threshold and
raised reflex sensitivity so that a smaller afferent volley was required to produce a tendon jerk, and an increase in
the afferent volley produced a disproportionately greater increase in reflex electromyographic activity than would
have occurred at rest. Reinforcement maneuvers potentiated the H-reflex but did not alter the electrically
induced afferent volley or the background level of neural activity. It is concluded that these changes in reflex
responsiveness occurred through intrinsic spinal mechanisms independent of the fusimotor system.
Burke D , McKeon B, Skuse NF: Dependence of the Achilles tendon reflex o n the excitability of spinal reflex
pathways. Ann Neurol 10:551-556, 1981
Direct recordings of muscle spindle afferent activity
in humans have provided a large body of evidence
that questions conventional views of the fusimotor
system (see [3, 181 for reviews). Assuming that spinal
reflex excitability remains constant, a greater afferent
input to the spinal cord should result in a heightened
reflex response; but on the evidence of the recordings to date, the prime role of the fusimotor system
does not appear to be regulation of the strength of
proprioceptive reflexes. However, most of these data
came from experiments on fascicles of the peroneal
nerve innervating the pretibial flexor muscles. It
is possible that the principles defined for these muscles may not be applicable to “postural” muscles such
as the triceps surae, from which tendon jerks and
H-reflexes can be obtained with greater ease.
The present study was undertaken to determine
whether experimentally induced fluctuations in the
amplitude of the Achilles tendon reflex result from
alterations in fusimotor activity or from changes in
reflex transmission through spinal cord circuits.
Materials and Methods
The results were obtained from eight experiments o n five
normal adult volunteers, all of whom had given informed
consent to the experimental procedures. N o n e of the five
suffered untoward sequelae as a result of the experiments.
From the Unit of Clinical Neurophysiology, Department of
Neurology, The Prince Henry Hospital, and the School of Medicine, University of New South Wales, Sydney 2036, Australia.
Received Nov 5 , 1980, and in revised form Mar 31, 1981. Accepted for publication Apr 3, 1981.
The subjects lay prone on a comfortable bed with the leg
under study supported on cushions, the hip extended, knee
slightly flexed, and foot secured in an isometric myograph.
Neural Recordings
Using techniques described in full elsewhere [11, 181, a
microelectrode was inserted manually through the skin into
fascicles of the tibia1 nerve innervating the soleus muscle or,
on one occasion, the lateral gastrocnemius and soleus. T h e
fascicles were then identified and shown not to be contaminated by cutaneous afferenrs. This involved low-intensity
stimulation through the microelectrode to produce a twitch
contraction of the innervated muscle without radiating
paresthesias, and recording through the microelectrode the
multiunit afferent discharges in response to tendon percussion, muscle palpation, and muscle stretch, but not to
cutaneous stimulation.
T h e microelectrode was then so positioned within the
fascicle that the activity of afferents from dynamically responding muscle mechanoreceptors dominated the recording. No attempt was made to obtain recordings from
single afferents, as a representative sample of the population of spindle endings provides more valid data than the
activity of a single receptor. T h e major contribution to such
multiunit recordings comes from group la afferents [ 11,
191, but there may be smaller contributions from large afferents originating in the Golgi tendon organ or from large
efferents. Under the present circumstances these can be
neglected for the following reasons. First, the triceps surae
Address reprint requests to Dr Burke, Unit of Clinical
Neurophysiology, The Prince Henry Hospital, Little Bay, NSW
2036, Australia.
0364-5134/81/120551-06$01.25 @ 1981 by the American Neurological Association 5 5 1
was relaxed during the experiments so that there was no
background a efferent activity directed to the muscle; in
any case, efferent activity would make no contribution to
the tap-induced afferent volley. Second, in the absence of
efferent activity, the background discharge of tendon organs and their sensitivity to passive stretch are likely to be
minor [13, 171. Third, the intention was to examine the
afferent response to tendon percussion for evidence of a
change in fusimotor drive. Possible contamination of the
afferent volley by tendon organ afferents would require
consideration only if a change occurred in the percussioninduced afferent volley.
Experimen tul Procedures
Subjects were instructed to relax completely the leg being
tested. Sequences in which limb position was disturbed or a
background contraction developed in the muscles of the
leg were excluded from analysis because both can alter
spindle responsiveness and transmission of the percussion
wave through the muscle.
The Achilles tendon was percussed manually using a
tendon hammer or mechanically using a small vibrator
(Ling Altec Model 201) mounted over the tendon. T h e vibrator produced reproducible tendon taps but, because of its
limited power, only the lower part of the range of tendon
jerk amplitudes could be investigated. With the tendon
hammer the full range of reflex amplitudes could be investigated, but it was difficult to reproduce the same percussion precisely. The vibrator was driven by a power amplifier fed by a square-wave pulse 10 msec in duration.
When the tendon hammer was used, contact with conducting paper on the Achilles tendon completed an electrical circuit in the head of the hammer, thus producing a trigger pulse for timing purposes. With either method, the
tendon was usually percussed at a rate of 0.2 Hz.
An accelerometer (Grass Model SPA1) was strapped to
the tendon 5 cm proximal to the site of percussion to provide a measure of percussion intensity. This method was
chosen instead of direct measurement of percussion force
because it would reflect the dynamic aspects of the percussion (the effective stimulus to primary spindle endings)
more accurately than a measure of force applied tangentially to the tendon. To estimate the intensity of the
stimulus resulting from percussion, the peak-to-peak
amplitude of the accelerometer output was measured in arbitrary units. Monitoring the output of the accelerometer
ensured that transmission of the percussion wave to the
triceps surae remained constant during an experiment.
The electromyogram (EMG) of the soleus muscle was
recorded using surface electrodes 2.5 cm apart secured to
the skin over the soleus immediately distal to the insertion
of the medial and lateral gastrocnemius muscles into the
Achilles tendon [14]. The forces operating at the ankle
joint were recorded in some experiments using a four-arm
strain gauge bridge bonded to a steel bar, which formed the
footplate of the isometric myograph. All experimental data
were monitored during the experiment and recorded with
the appropriate trigger pulses o n tape for subsequent analysis [3, 181.
In multiunit recordings the raw neural activity looks like
an interference pattern, in which the spikes of individual
552
Annals of Neurology
A
B
30ms
Fig 1. Fluctuation in the tendon jerk when subject is engaged
in conversation: (A) four superimposed responses obtained when
relaxed; (B) six superimposed responses obtained when conversing. TraceJ are: top, neural afferent volley (rectified and
smoothed, time constant 0.02 second); middle, output of accelerometer (delayed 5 msec using a digital delay line); bottom,
electromyogram of soleus muscle.
afferents can b e difficult to identify. T h e raw data were
therefore full-wave rectified and smoothed using an RC
low-pass filter with the time constant set so that the
amplitude of the resulting “integrated” neural burst
reflected the intensity of the afferent response to tendon
percussion. The effect of these procedures o n the raw
neural data is illustrated in Figure 1 of the companion paper
[5] ( p 548, this issue).
Reinforcement Maneuvers
T h e only part of the experiment to require the subject’s
active participation was reinforcement of tendon jerks by
contraction of remote, unrelated muscles. Maximum grip
strength was first measured using a hand-held dynamometer. During the experiment, subjects were instructed to
make an abrupt contraction of 60 to 70% of maximum grip
strength in response to a cue from one of the experimenters. Each maneuver was timed to begin approximately 0.5
second before tendon percussion and to last approximately
1 second. Between maneuvers, subjects were requested to
relax completely. It was intended that the disturbance
created by the maneuver b e restricted to the contracting
upper limb. However, vigorous performance often resulted in displacement of the body and unintentional tensing of muscles in the limb being examined; such trials were
eliminated.
Results
“Spontaneous” Fluctuations i n Achilles Tendon RQex
During the course of experiments, subjects underwent changes in alertness as a result of fatigue, discomfort, o r conversation. There were often parallel
changes in reflex responsiveness, but the muscle afferent responses to tendon percussion remained the
same in the more and the less reflexly active periods
(Fig 1). The four superimposed traces in Figure 1A
were recorded when the subject was completely relaxed. Conversation then ensued, and the next six re-
Vol 10 No 6 December 1981
sponses are superimposed in Figure 1B. There was
no detectable change in the state of the triceps
surae, transmission of the percussion wave (as monitored by the accelerometer) did not alter, and the afferent response to tendon percussion remained the
same. In addition, the baseline level of neural activity
remained constant throughout the sequences. During
conversation, however, each tendon tap produced a
reflex contraction (Fig 1B); in the absence of conversation no such response occurred under otherwise
identical circumstances (Fig 1A).
It may be concluded that the change in reflex responsiveness was not the result of a fusimotorinduced change in receptor sensitivity.
Rate of Tendon Percussion
The effect of varying the rate of tendon percussion
on the ensuing reflex response was studied using the
vibrator to produce identical tendon taps. In all subjects the amplitude of the tendon jerk was maximal at
the slowest rate (0.2 Hz). At 2 Hz and over, both the
afferent neural volley and the resulting reflex EMG
were variable because at these rates tendon percussion coincided with o r followed relaxation of the preceding reflex contraction. Relaxation of muscle can
alter both muscle spindle sensitivity and transmission
of the percussion wave from the point of impact. The
change in transmission of the percussion wave was
detected by accompanying changes in the accelerometer and force recordings. Such sequences
were therefore disregarded.
The effects of repetition rate were best demonstrated using brief trains of five stimuli at 1 per second, repeated every 9 seconds so that the first
stimulus of the train occurred 5 seconds after the last
stimulus of the preceding train (Fig 2). The reflex response to the first tap of the train was largest, and
the response to subsequent taps was much smaller.
There was usually a progressive decline in amplitude of the EMG burst over the first three taps,
then some fluctuation but little further change with
the fourth and fifth taps. No major change in the afferent response to tendon percussion occurred during the train.
It can be concluded that suppression of the tendon
jerk at percussion rates greater than 0.2 Hz results
from intrinsic spinal mechanisms rather than a
peripheral phenomenon at the muscle spindle level.
Reinforcement of the Achilles Tendon Jerk
In all subjects, reinforcement potentiated the Achilles tendon jerk but did not alter the afferent response
to tendon percussion, provided that the limb under
test remained quiescent.
The stronger the tendon percussion, the greater
was the afferent response. This relationship is prob-
-47C
+
D
F i g 2. Suppression of the tendon jerk during a train offive
taps at 1 /sec recurring every 9 seconds: (A) single sweep showing the responses t o the five taps of the train; ( B - 0 ) averaged
responses to the first, second, and third taps of the train, respectively,for 3 2 presentations of the train. In each panel, the
upper trace is the neural activity (rectified and smoothed, time
constant 0.01 second), and the lower trace is the electromyogram of the soleus. Tendon percussion produced a movement artifact in the E M G traces; the similarity of this artifact to the
different taps of the train indicates that transmission of the
percussion wave to the muscle was the same.
0
15
30
Tap
F i g 3 . Relationship between intensity of tendon percussion (in
arbitrary units) and amplitude of the rectified and smoothed
neural afferent responses to percussion (also in arbitrary units)
for one subject. (Circles = sequences when the subject was relaxed; triangles = those for which the subject performed reinforcement maneuvers; solid symbols = tendon taps that produced a tendon jerk; open symbols = no rejex response.)
ably best described by a power function, as illustrated
for one subject in Figure 3. When tested during
reinforcement, the relationship was the same as when
tested during relaxation. Performance of the maneuvers did not sensitize spindle endings to the tap.
However, the maneuvers produced effective reflex
reinforcement: they lowered the mechanical and
neural threshold for elicitation of the tendon jerk.
When the subject was relaxed in the experiment
shown in Figure 3, a reflex contraction did not occur
unless the intensity of the tap exceeded 9 units and
Burke et al: Spinal Mechanisms and the Tendon Jerk
553
:
'
I
.T
p-001
A
0
A
0
13
Nerve
26
F i g 4. Relationship between amplitude of the rectified and
smoothed afferent volley and peak-to-peak amplitude of the resulting reflex EMG burst for the data i n Figure 3. Taps that
failed t o produce a tendon jerk are shown as open symbols
alongside the appropriate ajjerent volley size. Dashed lines are
linear regression lines for the taps that produced reflex EMG.
The data obtained during reinforcement maneuvers (solid
triangles) dij$r significantly from those obtained during relaxation (solid circles).
produced an afferent neural volley of approximately
16 units. During reinforcement the reflex threshold
was lower, such that a tap of approximately half
the control intensity, producing a proportionately
smaller afferent volley, became adequate to elicit a
reflex contraction.
Reinforcement also resulted in a change in sensitivity of the reflex mechanism. In Figure 4 , tendon
jerk amplitude is plotted against amplitude of the
neural afferent volley for the data in Figure 3. For
completeness, the 15 trials in which tendon percussion failed to elicit a reflex response (12 when
relaxed, 3 when reinforcing) are shown alongside
the appropriate afferent volley size. T h e regression
lines differ significantly @ < 0.01 by analysis of
covariance). From these data it is clear that reinforcement alters the slope of the relationship between afferent input and reflex output such that an
increment in the afferent volley produces a greater
increment in reflex EMG than would have occurred
at rest.
It may be concluded that reinforcement maneuvers
do not alter receptor sensitivity and that they potentiate the reffex response by altering both the
threshold and the sensitivity of central reflex mechanisms.
Reinforcement of the H-Reflex
Reinforcement has been shown to be capable of
potentiating the H-reflex [7-9, 151. This is suggestive evidence for a change in intrinsic spinal cord
554
Annals o f Neurology
Vol 10 No 6
B
F i g 5 . Reinforcement of the H-reflex: (A)seven control responses; ( B ) seven reinforced responses. Upper traces repre.rent
the rectzfied and averaged neural activity, middle traces the
averuged E M G of the soleus; lower traces are superimpositions
of the individual E M G responses making up the average. In
the upper traces, the early burst of activity is the electrically
induced volley in low-threshold muscle afferents. The smaller
burst (vertical arrows) i.s the rejexly induced volley in a
efferent fibers.
mechanisms independent of fusimotor activity, but
the evidence is not conclusive for the following reasons. Fusimotor activity could alter the background
discharge of spindle endings and so alter the background conditioning of the motoneuron pool against
which the H-reflex is tested [ 151. In addition, vigorous maneuvers often disturb the limb, thus altering
the degree of stretch on the triceps surae and the
background spindle discharge. They could also displace the stimulating electrodes slightly. Only submaximal H-reflexes are effectively potentiated by
reinforcement [ 151; with stimulus levels that produce
little or no M wave, the experimenters cannot be
completely certain that stimulus conditions are constant.
These theoretical problems can be avoided by
demonstrating that both the electrically induced afferent volley and the background level of neural activity remain constant during effective reinforcement
maneuvers. I n three subjects, recordings were made
in the popliteal fossa from tibia1 nerve fascicles innervating the soleus while electrical stimuli were delivered to the nerve using needle electrodes inserted
proximally in the thigh distal to the buttock crease.
The stimulating electrodes were positioned at the site
of lowest threshold for a twitch of the soleus. Stimuli
of 0.5 msec duration were delivered at 0.2 H z at a
level sufficient to produce a liminal reflex. In the
upper traces of Figure 5, the averaged rectified neural
recordings contain an early burst of activity following
the stimulus artifact. This is the direct response to the
stimulus and, since there is no M wave in the EMG
traces, it is an antidromically conducted volley in
low-threshold (group I) muscle afferents. The smaller
December 1981
burst of neural activity is the reflexly induced efferent volley responsible for the reflex EMG. The sequences in Figure 5A were obtained with the subject
at rest, those in 5B with the subject reinforcing. The
maneuvers potentiated the H-reflex, as evidenced by
the larger efferent neural burst and the greater reflex
EMG. However, the afferent volley was not enhanced and no change in background neural activity
occurred. There was marked variability in the size of
the individual reinforced H-reflexes, probably because of differences in timing of the electrical
stimulus after the onset of the maneuver. Although
the reinforced H-reflexes varied in amplitude, the
individual neural afferent volleys did not differ.
It may be concluded that reinforcement of the
H-reflex resulted from a change in intrinsic spinal
mechanisms independent of the fusimotor system.
Discussion
The Achilles tendon jerk can vary in amplitude when
tendon percussion remains constant. In the present
study, reproducible tendon percussion evoked a constant afferent response even though the resulting
tendon jerk varied. This finding indicates that any
change in the reflex response did not emanate from
altered receptor sensitivity. The findings therefore
support the importance of central rather than
peripheral mechanisms in determining the excitability of proprioceptive reflexes. In the present examples, the fusimotor system played no role in adjusting
the responsiveness of the reflex pathway.
When a subject voluntarily relaxes a muscle and
keeps it relaxed, there appears to be no demonstrable
background fusimotor activity directed to the spindle
endings in that muscle, be it a flexor [ 3 , 181 or an
extensor [5]. Despite the absence of fusimotor drive,
the response of spindle endings to percussion of a
relaxed muscle may be sufficiently intense to produce
a tendon jerk [5]. Hence, in relaxed muscles,
fusimotor activity is neither necessary for the tendon
jerk nor responsible for its fluctuations. It is commonly held that alertness, reinforcement, and anticipation produce widespread activation of dynamic
fusimotor neurons and that this activity underlies the
changes in reflex responsiveness accompanying these
conditions 12, 10, 167. The present results indicate
that if dynamic fusimotor neurons are so activated,
the activation is weak, insufficient to alter the spindle
response to percussion, and irrelevant to the change
in reflex responsiveness.
The fusimotor neurons supplying a particular muscle may be activated when the subject contracts that
muscle voluntarily (see [ 3 , 181 for reviews), and this
can have a profound effect on spindle discharge if the
contraction is isometric or produces relatively slow
muscle shortening [ 3 , 181. Paradoxically, abrupt
stretch of a contracting muscle often elicits a smaller
spindle discharge than identical stretch of the noncontracting muscle [ 4 , 61, presumably because contraction dampens transmission of the perturbation to
the spindle endings. Despite the reduced afferent
input, the reflex response is usually enhanced.
Hence, whether the subject is active or at rest, reflex
modulation seems to be carried out within the reflex
pathway rather than at the muscle spindle level. It
seems reasonable to question whether the prime role
of the fusimotor system is ever to modulate reflex
responsiveness.
In discussing possible sites of reflex modulation,
the a motoneuron is commonly advanced as if it were
the sole alternative to the y motoneuron. This is not
so. A purely monosynaptic reflex can be modulated
not only by altering the afferent input and the excitability of the a motoneuron pool, but also by changing
the excitability of the afferent terminals presynaptically. Moreover, although the tendon jerk is generally considered a monosynaptic reflex, there is no
proof that is is exclusively or even predominantly
monosynaptic. In the cat, group Ia afferents may
produce disynaptic and trisynaptic excitatory postsynaptic potentials (EPSPs) in the homonymous
motoneuron pool [20].In humans, the afferent volley set up by tendon percussion is dispersed, particularly by the time it reaches the mononeuron pool,
and the rise time of the resulting EPSP in soleus
motoneurons has been estimated at 8.3 ? 2.5 msec
(mean
SD) 1121. Even the more synchronized
group I volley resulting from electrical stimulation of
the tibia1 nerve in the popliteal fossa produces an
EPSP with a rise time of 3.5 to 5.5 msec in soleus
motoneurons [ 11. These rise times are adequate
for disynaptic and trisynaptic pathways to play a
major role in the resulting reflex discharge: the
motoneurons of lowest threshold may require only
monosynaptic excitation from the fastest group Ia
afferents, but motoneurons of higher threshold will
receive disynaptic and trisynaptic inputs from these
afferents at the same time as the monosynaptic input
from slower la afferents. Thus, if one rejects the
fusimotor neuron as the site of a change in reflex responsiveness, one is not forced to choose the a
motoneuron as the sole alternative. The change could
have occurred at the interneuronal level, either at
interneurons capable of affecting presynaptic Ia afferent terminals or in interneurons directly in the reflex pathway.
*
Supported by the National Health & Medical Research Council of
Australia.
The authors are grateful to James W. Lance, MD, and A. Keith
Lethlean, MD, for their support and encouragement.
Burke et al: Spinal Mechanisms and the Tendon Jerk
555
References
1. Ashby P, Labelle K: Effects of extensor and flexor group I
afferent volleys on the excitability of individual soleus mot0
neurones in man. J Neurol Neurosurg Psychiatry 40:910919, 1977
2. Burg D, Szumski AJ, Struppler A, Velho F: Assessment of
fusimotor contribution to reflex reinforcement in humans. J
Neurol Neurosurg Psychiatry 37:1012-1021, 1974
3. Burke D: The activity of human muscle spindles in normal
motor behavior. In Porter R (ed): Neurophysiology IV. International Review of Physiology. Baltimore, University Park
Press, 1981, pp 91-126
4. Burke D, Hagbarth K-E, Lofstedt L: Muscle spindle responses
in man to changes in load during accurate position maintenance. J Physiol (Lond) 276:159-164, 1978
5. Burke D, McKeon B, Skuse NF: The irrelevance of fusimotor
activity to the Achilles tendon jerk of relaxed humans. Ann
Neurol 10:547-550, 1981 (this issue)
6. Burke D, McKeon B, Skuse NF, Westerman RA: Anticipation and fusimotor activity in preparation for a voluntary contraction. J Physiol (Lond) 306:337-348, 1980
7. Bussel B, Morin C, Pierrot-Deseilligny E: Mechanism of
monosynaptic reflex reinforcement during Jendrassik manoeuvre in man. J Neurol Neurosurg Psychiatry 41:40-44,
1978
8. Clare MH, Landau WM: Fusimotor function. Part V. Reflex
reinforcement under fusimotor block in normal subjects.
Arch Neurol 10:123-127, 1964
9. Gassel MM, Diamantopoulos E: The Jendrassik maneuver. I.
The pattern of reinforcement of monosynaptic reflexes in
normal subjects and patients with spasticity or rigidity.
Neurology (Minneap) 14:555-560, 1964
10. Granit R: The functional role of the muscle spindles-facts
and hypotheses. Brain 98:531-556, 1975
11. Hagbarth K-E, Vallbo AB: Discharge characteristics of human
muscle afferents during muscle stretch and contraction. Exp
Neurol%2:674-694, 1968
12. Homma S, Nakajima Y : Coding process in human stretch
reflex analysed by phase-locked spikes. Neurosci Lett
11:19-22, 1979
13. Houk JC, Henneman E: Responses of Golgi tendon organs to
active contractions of the soleus muscle of the cat. J
Neurophysiol 30:466-481, 1967
14. Hugon M: Methodology of the Hoffmann reflex in man. In
Desmedt JE (ed): New Developments in Electromyography
and Clinical Neurophysiology. Basel, Karger, 1973, vol 3, pp
277-293
15. Landau WM, Clare MH: Fusimotor function. Part IV. Reinforcement of the H reflex in normal subjects. Arch Neurol
10:117-122, 1964
16. Murthy KSK Vertebrate fusimotor neurones and their
influences on motor behavior. Prog Neurobiol 11:249-307,
1978
17. Stuart DG, Goslow GE, Mosher CG, Reinking RM: Stretch
responsiveness of Golgi tendon organs. Exp Brain Res
10:463-476, 1970
18. Vallbo AB, Hagbarth K-E, Torebjork HE, Wallin BG:
Somatosensory, proprioceptive, and sympathetic activity in
human peripheral nerves. Physiol Rev 59:919-957, 1979
19. Wallin BG, Hongell A, Hagbarth K-E: Recordings from muscle afferents in parkinsonian rigidity. In Desmedt JE (ed):
New Developments in Electromyography and Clinical
Neurophysiology. Basel, Karger, 1073, vol 3, pp 263272
20. Watt DGD, Stauffer EK, Taylor A, Reinking RM, Stuart DG:
Analysis of muscle receptor connections by spike-triggered
averaging. 1 . Spindle primary and tendon organ afferents. J
Neurophysiol 39: 1375- 1392, 1976
556 Annals of Neurology Vol 10 No 6 D e c e m b e r 1981
Документ
Категория
Без категории
Просмотров
0
Размер файла
586 Кб
Теги
reflexi, spina, tendon, achilles, dependence, pathways, excitability
1/--страниц
Пожаловаться на содержимое документа