- Email: [email protected]
Inhibition of flexor burst generation by loading ankle extensor muscles in walking cats
Brain Research, 187 (1980) 321-332 '(-'~Elsevier/North-Holland Biomedical Press
321
I N H I B I T I O N OF F L E X O R B U R S T G E N E R A T I O N BY L O A D I N G A N K L E E X T E N S O R MUSCLES I N W A L K I N G CATS
J. DUYSENS* and K. G. PEARSON ** Department oJ Phy~tology, University of Alberta, Edmonton, Alberta (Canada)
(Accepted August 30th, 1979)
SUMMARY The role of propnoceptive input from the ankle extensor triceps surae in the control of walking was examined m premammillary cats walking on a treadmill. The left hindhmb was rigidly fixed m one position after denervatmg almost all the leg muscles except the ankle extensor (triceps surae) and ankle flexor (tlbiahs anterior). Rhythmic alternating contractions of the isolated ankle flexor and extensor occurred in the fixed hlndleg during periods of walking in the other three intact limbs. These rhythmic contractions disappeared when the isolated triceps surae was stretched so as to increase the force of the active contractions to beyond 4 kg. With maintained stretch the periodic contractions in the ankle flexor and extensor returned only after the force Jn the stretched triceps surae gradually decreased and fell below approximately 4 kg. Isometric contractions of the triceps surae produced either by stimulation of ventral root S1 or by large amplitude vibrations also led to the sudden disappearance of ankle flexor bursts. Inhibition of the locomotory rhythm could also be produced in all muscles of a single intact hindleg by clamping the ankle joint in a flexed position so as to stretch the ankle extensor. In all these cases, an increased rate of stepping of the contralateral hmdleg was associated with the inhibition of the rhythmic locomotory act~wty. It is suggested that triceps surae propnoceptors slgnalhng the presence of loading of the hindlimb extensor muscles inhibit the central generation of hmdlimb flexion. During normal walking this mechanism could be of major importance during stance to prevent the initiation of the swing phase of a time when hindlimb extension is fully needed to support the weight of the animal. Thus a necessary, but not always
* Present address: 6, Hoge Vesten, 9300 Aalst, Belgmm. ** To whom requests for reprints should be addressed at the Umversity of Alberta.
322 sufficient, condition for the mihatlon of swing may be an unloading of leg extensor muscles.
INTRODUCTION Recently, marked advances have been made m recording from single muscle afferents in freely walking cats14.lg, 20. Yet very httle is known about how this afferent actwity is used by the nervous system to control walking. Although there ~s some dasagreement about the usefulness of propriocept~ve reflexes at hagher speeds of locomotion 6,16, at is generally thought that feedback from muscle afferents must, under most c~rcumstances, play an ~mportant role. Certainly ~t ~s no longer believed that locomotion is based on a chain of reflexes 18 since deafferentatlon does not abohsh the m o t o r program for walking m the spinal cat 9. Rather, ~t as thought that afferent input and reflexes may play a role m the adaptation of the central locomotor program to the changing needs of the envaronmentL Grillner and Rosslgnol 8 have recently demonstrated that one important group of afferents involved m terminating the stance phase arises from the hip. A necessary condition for the swing phase to be anatiated as that the hap be extended beyond a critical angle (about 120° in their experimental arrangement). Although hip extension may be anecessaryconditlon for lmtmtmg of swing, it does not follow that ~t ~s always a sufficient condition. In fact, the present investigation began from the observation that stepping movements in a hindlimb of a premammlllary cat walking on a treadmill could be inhibated by manually holding the ankle in a flexed position even when the hap was extended well beyond the angle at which swing as normally initiated. Th~s initial observation mdacated that afferents in the ankle extensor muscle (triceps surae) could inhibit the stepping generator and suggested the hypothesis that a necessary condition for the swing phase to be lnitmted was that the leg extensor muscles be unloaded. The present experiments were designed to determine more rigorously whether or not afferents from triceps surae can indeed inhibit the spinal stepping generator. A preliminary account of these experiments has been pubhshed iv. METHODS The present experiments were performed on a series of premammillary cats, 28 of which showed spontaneous walking on a motor-driven treadmill tbr periods ranging from 30 min to 8 h. Stereotaxlc decerebrat~on was performed at a precollicular and premammillary level as described prewously 3,~. The left hindhmbs of 25 of the 28 animals were partially denervated by cuttmg the obturator, femoral, d~stal tibial. superficial peroneal and sural nerves as well as the posterior and medial articular nerves at the knee. The aim of this partml denervatlon was two-fold: (1) to deafferent as much as possible the surroundings of the triceps surae so that any lnampulat~on of the triceps surae would not ehcit afferent actwity from receptors outside these muscles, and (2) to leave the innervation of a few proximal muscles intact so that
323 one could check whether events occurring in ankle flexors and extensors were paralleled by similar events m muscles at a different joint of the same hmdhmb. The animals were held in place by two hip and two ear bars while the pamally denervated leg was fixed to a frame over the treadmill so that the hip was extended (135°), the knee slightly bent (140") and the ankle held at more than 90 ~ (see below). During walking the muscles which were not denervated m the fixed hmdhmb showed rhythmic alternating bursts of activity appropriately phased w~th their counterparts m the free right hmdhmb. The ankle was held at an angle between 90: and full extension so that well developed tlbiahs anterior (TA) bursts occurred, these being comparable in duratzon with those occurring m the free hlndhmb EMG signals were recorded from medial gastrocnemms (MG) and T A m the fixed hmdllmb and MG m the contralateral free hmdlimb using two copper wires insulated except for the t~p and inserted into the muscle at an interelectrode d~stance of 0.5-1 cm. The force m the freed Achilles tendon of the fixed hmdhmb was recorded wa a Grass force transducer. To determine the length of the freed triceps surae relative to their length m the intact hindhmb, a technique was followed as described by Nichols and Houk 1~ A screw was inserted m the distal tlbm and the triceps surae tendon was marked with sutures at points hnmg up w~th the screw when the ankle was held at 90 °, at full flexmn and at full extension (the knee angle was kept constant at 140°). The calcaneus bone was then cut and attached to the force transducer by means of a thread through the cut part of the calcaneus bone. The force transducer to which triceps surae was attached could be moved to any desired position, thus stretching the muscle by means of a manually controlled manipulator. The scale on this manipulator was set to zero when the triceps surae suture corresponding to 90 '' ankle angle was hned up with the screw m the tibia (L = 0) Hence, triceps surae lengths corresponding to ankle angles of greater than 90" were negative length values. The ankle angle corresponding to any g~ven triceps surae length could be calculated since ~t was found, m agreement w~th Stuart et al. 21, that a change m muscle length of 2 mm was equivalent to an angle excursion of 10. In 6 experiments the triceps surae was vibrated by attaching a Pye-Ling V47 vibrator in series with the force transducer. The wbrator was driven by a waveform generator which allowed control of the frequency and amplitude of the wbratory stimulus. In 6 cats, a lammectomy was performed, in addition to the usual procedures, to permit recording from single dorsal root filaments (L7 or S1 ) and to enable stimulatmn of the ventral root SI. The romulus strength to the ventral root was determined relative to the threshold of the largest motor axons in thzs root supplying triceps surae. The frequency of stimulation was usually 90 Hz, but similar results were obtained with lower frequencies provided they were high enough to give fused contractions of triceps surae (about 35 Hzl~). In three cats. the 4 limbs were left free but an adjustable metal clamp was attached to the left ankle with a screw through the tibia and a clamp around the deafferented foot. A bolt m a semicircular sht in this clamp (Fig. 4) allowed the ankle to be fixed m any pos,tlon.
324
iE iE iF cE
El
I
iE
~
~
|
I
v
III
IIII
f
iE iF cE
b
I
I sec
Fig. 1. Inlubition of rhythmic motor activity by stretch of the triceps surae in a fixed hmdhmb of a walking premammillary cat. A 10 mm stretch was applied at a constant rate beginning at the start of record a and ending at the end of record b (this stretch was equivalent to a change in ankle angle from 160° to 110°, see Methods). As the triceps surae was stretched, the duration of the tibialis anterior bursts (third trace, iF) gradually shortened until the bursts suddenly disappeared completely when the force m the stretched triceps surae reached approximately 4 kg (top trace, iE) (force calibration ~- 6.5 kg). The magnitude of the EMG bursts recorded from the medial gastrocnemius (second trace, IE) increased as the stretch was increased (the gain of this trace has been halved in b so as not to obscure the EMG record from the tibialis anterior). Near the end of the stretch the EMG burst activity also disappeared in the medial gastrocnemius and the continuous EMG in this muscle was associated with a maintained contraction of the triceps surae. Note in this case that the first flexor burst failure occurred despite the presence of a complete triceps surae relaxation. Despite the inhibition of rhythmic burst activity in the ankle flexors and extensors, the contralateral limb continued to step normally (fourth trace, cE). RESULTS
Rhythm inhibition by triceps surae stretch T h e present results center a r o u n d the p h e n o m e n o n o f E M G b u r s t fadures m the ankle flexor tibialis a n t e r i o r ( T A ) in response to stretch o f triceps surea (Fig. 1). By a p p l y i n g a slow stretch o f 10 m m at a c o n s t a n t rate to the freed t e n d o n o f the triceps surae in a fixed h i n d l i m b o f a s p o n t a n e o u s l y w a l k i n g cat, the d u r a t i o n o f the a n t a g o n i s t i c T A E M G bursts initially shortens until a p o i n t is reached where they d i s a p p e a r a b r u p t l y . T h e T A b u r s t fadures typically occur when the increasing force in the r h y t h m i c a l l y c o n t r a c t i n g triceps surae reaches a level o f a b o u t 4 kg. A tonic c o n t r a c t i o n o f the triceps surae, indicated by c o n t i n u o u s E M G activity, was seen in c o n j u n c t i o n w i t h the T A b u r s t failures, while all r h y t h m i c b u r s t activity d i s a p p e a r e d as well in o t h e r ipsilateral h i n d l i m b muscles which were n o t d e n e r v a t e d (hip flexors a n d extensors).
325
A / IE
__J"
I Im m
Lenglh IE
Ikg
\
Force ~
~//"
EMG
'
[/'~ i
I00
80
I sec
i
Flexo~Burst Failures
~"--o~-o.
o
I "o.. A "-0
/
o
o
,,"
Flexor Burst Durohon
60 % 4O
20
0 -15
-12
-9
-6
Length
Fig. 2. Effect of an increase in triceps surae length on the probablhty of tib]alis antemor burst failures. A sample record showing that stretching the triceps surae from L -- -15 mm to L -- -I 2 mm caused the tiblahs antemor burst (bottom trace, iF EMG (integrated)) to become shorter or to occasionally disappear. The incomplete drop m triceps surae force and the occurrence of EMG bursts m the contralateral medial gastrocnemms muscle made it possible to estimate where a failure occurred and thus to determine the number of failures during the first 15 cycles following the stretch. The failures tended to occur most frequently just after the stretch (see Fig. 3), but if the fadures were rather infrequent, as in the present example, they could appear after a few steps. This illustrates the threshold nature of these failures. B. percentage of failures occurring m the first 15 cycles following stretch (full circles) plotted against the amplitude of the muscle stretch (mm). The effect on the duration of the flexor burst duration is represented on the same plot (open circles) The duration of the flexor bursts at L ~ - l 5 mm was taken as 100 ~ . Arrows in B indicate data points representing the example gwen m A. The ~psdateral hmdhmb was fixed with the knee at 140 ° and the ankle imtmlly at 160 °.
Results similar to those shown in Fig. 1 were obtained in all 12 cats in which the slow stretch experiment was done. However, the amount of triceps surae stretch required to elicit TA burst failures varied considerably from animal to animal, and even within the same animal, depending on the quality o f locomotion. In most cats TA burst failures were seen when the triceps surae force reached the 4 kg level and required stretches o f between 9 and 15 mm. The final length of triceps surae at which
326 ~F )E
Ferce cE
s~.r
A
4
~'9 mm
~lOmm
kq
'~llmm IOsec
F~g 3, Adaptation of rhythm lnhlblt~on with maintained stretch of the triceps surae in a fixed hindlimb of a walking premammillary cat. A. the triceps surae was quickly stretched from L d5 mm to L -2 mm and held at this length (beginning of record) All rhythmic activity in tiblahs anterior (top trace, IF) and triceps surae (middle trace, iE force) was lmtlally inhibited by this stretch Rhythmtcity returned m an irregular manner over a period of 15 sec. During this period each sudden relaxation of the contraction in triceps surae was associated with a burst of activity m tib:alis anterior. The contralateral leg continued to step throughout as indicated by continuous rhythmic EMG burst activity in the contralateral me&al gastrocnemius (bottom trace, cE). B: force records from triceps surae when quickly stretched 9, 10 and 11 mm (arrows) corresponding to changes of 45 ~, 50° and 55" respectively from an initial angle of 160°, and held stretched for the remainder of the record. In each record the rhythmic relaxations were initiallyabolished and these only began to return when the force fell below a value of about 4 kg (indicated by the large brief dechne in force) The small oscillations m the magnitude of the maintained contractions immediately following each stretch were due to active stepping movements of the contralateral hindlimb, and these should not be confused with the small rhythmic contractions preceding the stretch. T A b u r s t fadures were p r o d u c e d corresponded to ankle angles of between 120 ° a n d 90 °. Thus, the stretches which produced T A burst failures were well within the physiological range of the triceps surae. W i t h very forceful walking, as for example d u r i n g periods o f sham rage sometimes seen shortly after decerebration, the triceps surae force level at which T A bursts were blocked could reach 8-10 kg a n d the length o f the stretch was close to the physiological limit. By contrast, a n i m a l s with p o o r stepping could exhibit occasional T A burst failures without a n y afferent m a n i p u l a t i o n . The effect of the a m p l i t u d e of triceps surae stretch was systematically stu&ed by stretching the triceps surae to different lengths a n d c o u n t i n g the n u m b e r o f T A b u r s t
327 failures during the first 15 step cycles after triceps surae reached its new length (Fig. 2). Between each stretch trml the triceps surae was brought back to the rest length (L -15 ram) for a period ranging from 30 to 60 sec, during which no TA burst failures occurred. In the example illustrated at the top of F~g. 2, a ramp and hold stretch of 3 mm (from L = -15 mm to L ~ -12 ram) caused only one TA burst failure during the first 6 step cycles after application of the stretch. Larger stretches increased the probabd~ty of TA burst failures as ~s seen m the plot of Fig. 2 (filled c~rcles) The largest stretch (to L -~ -4 ram), corresponding to an ankle angle of 110n, blocked all rhythmicity m the left hlndlimb (100 oo TA burst fadures). Apart from eliciting TA burst fadures, the step stretch also caused a shortemng of the duration of those TA bursts which did appear during the stretch (Fig. 2A). This TA shortening effect requxred much smaller amplitude stretches than was reqmred to evoke TA burst fadures. As seen in the plot of F~g. 2, stretching the triceps surae from L -- -15 mm to L ~-- -14 mm caused the TA burst to be shortened by 20% (open c~rcles) while causing no TA burst failures at all. Larger triceps surae stretches d~d not produce more TA burst shortening but clearly increased the number of TA burst fadures. Another factor affecting the occurrence of TA burst failures was the speed of stretch apphed to triceps surae. Slower stretches were less effective m evoking flexor burst failures. ThJs was due to a slow adaptation of the inhibitory effect on TA burst generation when triceps surae was stretched Th~s adapt~on was most easdy seen when a large amphtude stretch was quickly apphed and the triceps surae held for a long period at a new length (Fig. 3). In the example in Fig. 3A the triceps surae was quickly stretched from L ~ -15 mm to L ~ -2 mm (corresponding to a change m ankle angle from 165-to 100°) and held at this length. Although the length of the triceps surae was kept constant at the end of the stretch period (which ~s the start of the record m ]Wig. 3A), the number of TA burst failures dechned stea&ly over a period of 15-25 sec, indicating some form of adaptation. The reappearance of rhythmic TA bursts occurred when the triceps surae force fell to about 4 kg. Th~s observation was qu~te consistent, as is ~llustrated in F~g. 3B, taken from an experiment on another cat. In this case the mceps surae was stretched 9, I0 and 11 mm from the resting length (L = -15 m m to L 6, -5 and -4 ram) and held at that new length. Despite the differences in initial force levels at the end of the stretch, and the differences m final length, it ~s quite apparent that the return of the rhythmicity is hnked to the drop of the force below the 4 kg level in each stretch trial. When such a drop m force was prevented, as for example by adding triceps surae stretch when the force levels came close to the 4 kg level, no return of the rhythmicity occurred.
Rhythm mhibmon by clamping ankle To exclude the possibihty that the above described results were due to the fixation of the whole left hindlimb, experiments were done on three cats which could freely move all limbs but whzch had an ankle clamp attached to the left hindlimb (Fig. 4). With the ankle fixed at increasingly smaller angles, the TA bursts became shorter in duration, then faded to occur. Again the TA burst fadure was a discontinuous
328
( ( T
/
/
,
G _
i
III l _
v"rllm'1
2
1 ~
.
. . . . . . . . . .
~,,llllW- •
-,,,~
:
~
-
.
, F
-
G
Fig. 4 Inhibition of lpsilateral hlndhmb rhythmicity produced by clamping the left ankle in a flexed posit~on, a: clamp loose; b: clamp fixed at 70° Note the hyperextension at the hip and knee of the non-stepping hmb with the ankle clamped. The lack of rhythmicity in the clamped leg was associated with continuous EMG activity m the medial gastrocnemius (iE) and the absence of burst activity in tiblalis anterior (1F). The contralateral hmdlimb continued to step when the rhythm was inlubited by the ankle clamp as indicated by rhythmic burst activity in the contralateral medial gastrocnemius muscle (cE).
phenomenon and it showed adaptlon with a time course similar to that observed in Fig. 3. TA burst failures started to appear when the ankle was clamped at angles of about 90 °. For a complete block of the rhythmicity the ankle had to be clamped at 45", 65 ° and 70 ° for the three cats. With complete block of TA burst generation the cats neglected to use the limb with the ankle clamp and instead held this limb in hyperextension (hip angle close to 150 °) while the other three limbs continued stepping (Fig. 4B). The finding that stretch of the triceps surae, induced by clamping the ankle in a flexed position, leads to TA burst failures and to a complete block of all rhythmic activity m the ipsllateral hindlimb, indicates that fixation of the limb was not the factor responsible for the TA burst failures observed in the previously described triceps surae stretch experiments (Figs. 1, 2 and 3).
Rhythm inhibition by ventral root stimulation The occurrence of flexor bursts depends on the force in the triceps surae rather than on the length (Fig. 3B), suggesting that muscle spindle afferents are not responsible for the flexor burst failure phenomenon. The following results obtained when isometric contractions of triceps surae were produced by ventral root stimulation support this conclusion that receptors other than muscle spindles are primarily responsible for causing the flexor burst failures when triceps surae is stretched. In 6 experiments stimulation of part or all of the S1 ventral root was tested for its ability to induce T A burst failures. To create optimum conditions for the appearance of TA burst failures, the triceps surae was first brought to a length wl~ch was 1 m m below the threshold length for evoking TA burst failures. The distal end of part or all of ventral
329
A
Ventral
Root Stlmulat)on A
,E IF IF cE
Isec
B ~E
J,~,[l.~ ~,.~
cEI~ ]+_.
Triceps Surae V b r a t l o n ,,J. ,lJl ~iiPal. Jl~JJ~il [~i
J,~L., ,.~,,.
,.,l~J L ~_,
+,
.,~
"T
+
II~l,l~ll~tl ~LlaLi,
~IalLII
n r,
"'I ....
1 I
SOC
Fig 5 Failure of burst generation in tlblalis anterior produced by (A) strmulatmn of the Sl ventral r o o t a t l 3 T, and (B) vlbratlon of the tnceps surae I E , ~ F a n d c E a s l n F i g I A' the ventral root surnulat~on (a 480 msec tram of pulses at 90 Hz) was applied midway through the record and ~s indicated by the broademng of the basehne in the ~F trace B" the time of apphcatmn of the vibratory stimulus (a 200 msec period at 64 Hz) is indicated by the rippled bar underneath the rE EMG trace. Note that the period between extensor bursts on the contralateral side was temporarily shortened as a result of ventral root stm~ulat~onor the mceps surae wbration. The resulting resetting of the contralateral rhythm paralleled the resetting of the zpsdateral rhythm. Force calibration In A = 6 5 kg.
root SI was then sttmulated for periods r a n g i n g from 100 msec to 5 sec to reduce an Isometric c o n t r a c t i o n m triceps surae. Prior to and following each r u n the threshold stimulus strength was determined as that reqmred to elicit the first detectable triceps surae twitch. With stimulation of the whole ventral root T A burst failures occurred (Fig. 5A) w~th st~muh at 1.3 >< T, which are below threshold for the activation of g a m m a m o t o n e u r o n s TM. This result suggests that T A burst failures do not require the actwatlon of the tnceps surae spindle afferents since the induced shortening of this muscle must have caused a fall, rather t h a n a rise, in activity in triceps surae spindles. Th~s was d e m o n s t r a t e d directly for three triceps surae spindle afferents m dorsal root S 1 by observing a decrease in their discharge rate when the whole o f the S 1 ventral root was stimulated at a strength o f 1.5 / T.
Trweps surae vibratlon Small amplitude vibration is generally recognized as being an effective means o f
330 ehcitlng activity in primary muscle spindle afferents, while larger amplitude wbratlon may also activate Golg~ tendon organs, especially in contracting muscles ~'. In 6 cats the isolated triceps surae was wbrated for periods ranging from 100 msec to 5 sec during walking. As m the experiments w~th ventral root stimulation, the muscle was first brought to a length just subthreshold for producing TA burst failures. When small amplitude vibration was applied, in the range of 6-10 #m (f -- 64 Hz), the effect was to decrease the duration of the TA bursts (by about 40 ~o) but not to produce TA burst failures. T A burst failures did appear when the wbration amplitude was increased to beyond 30 #m. Long periods of high amplitude vibration, lasting several seconds, sometimes completely abolished TA burst actwlty for the duration of the vibration, and often TA burst failures could be ehclted by a short duration (200 msec) high amplitude vibration (Fig. 5B). As was usually found for muscle stretch or ventral root stimulation, the occurrence of TA burst failure was associated with a shortening of the contralateral step cycle (Fig. 5B). DISCUSSION The present results on walking premammdlary cats provide ewdence that propnoceptors in triceps surae can inhibit the generation of rhythmic bursts of EMG activity in ipsllateral flexors such as tibialis anterior. Furthermore, the sudden and complete abolition of bursts in an ankle flexor as well as in flexors at other joints indicates that the described results are due to triceps surae afferents acting on a premotoneuronal center. We will refer to this center as the flexor burst generator, and we presume it is part of, or is tightly coupled to, the central stepping generator itself. We propose that the initiation of the swing phase requires the activation of the flexor burst generator and depends on this generator reaching a certain threshold. There are many factors which appear to play a role in controlling this threshold. We have shown here that afferents from triceps surae can increase the threshold and prevent flexor burst generation, while previously we showed that flexor burst generation can be inhibited by stimulation of large cutaneous afferent~3. Factors such as the general excitability of the animal may also control threshold. For example, in cats exhibiting very vigorous walking it was difficult to induce TA burst fadures, while such failures could occur spontaneously in animals showing weak stepping. The burst failure phenomenon is not umque for the hmdhmb flexors for It has also been reported to occur m hmdlimb extensors a. Hence it seems that there exist at least two burst generator systems, one for the production of rhythmic flexor bursts, the other for the production of locolnotory bursts of activity in the extensor group of hindlimb muscles. This conclusion has also been reached from recent ~tudies on rhythmic scratching movements of the hindhmbs in cats 1. Although the primary aim of this investigation was not to ~dentlfy the group (or groups) of afferents responsible for producing flexor burst failures, some of our data do bear on this matter. For example, with large maintained stretches of different lengths applied to triceps surae the TA burst activity begins to return only when triceps suraeforce declines below a critical level (about 4 kg m most experiments). It follows
331 t h a t afferents signalling triceps surae force, r a t h e r t h a n length, p r o d u c e the tnhlbit~on of the flexor burst generator a n d t h a t Golgi tendon organs from triceps surae are m o s t hkely to be responsible for the flexor burst failures. A c o n d i t i o n which favors the selective activation o f Golgl tendon organs, namely st~mulatron o f the S1 ventral r o o t at 1.3 T, also p r o d u c e d flexor burst failures (F~g. 5A). The p r o p o s e d flexor burst i n h i b i t o r y function o f the triceps lb afferents is in a p p a r e n t confhct w~th prevxous d a t a 5,13 which ascribed a flexor excitatory and extensor inhibitory role to these afferents. The latter data, however, were o b t a i n e d m spinal cats which are k n o w n to show p r e d o m i n a n t l y flexor excitatory reflex actions. Nevertheless, flexor i n h i b i t o r y effects from extensor lb afferents have occasionally been described in spinal cats (ref. 5, Fig. 12) as well as m decerebrate cats ~2. Such conflicting findings m a y be explained by assuming 'alternatzve reflex p a t h w a y ' (ref. 1 1). The present results suggest that some o f these alternative pathways, which a p p e a r e d rather inslgmficant m experiments on motionless ammals, m a y prove to be very i m p o r t a n t in the actually m o w n g cat. In concluston, the most i m p o r t a n t result o f the present investigation ~s that the p r o d u c t i o n o f forces greater than a b o u t 4 kg in the triceps surae leads to inhibition o f the generation o f flexor bursts. Since the triceps surae easily p r o d u c e s 4 kg o f force d u r i n g stance in n o r m a l walking animals 99, this inhibitory influence from triceps surae afferents o n t o the flexor burst g e n e r a t o r p r o b a b l y functions to prevent the initiation o f swing when the extensor muscles are l o a d e d during stance. ACKNOWLEDGEMENTS We wish to t h a n k Drs. R. B. Stein, T. R. Nichols a n d J. D. Steeves for their valuable criticisms, Ms. O White for typing the manuscript, and K. Burt and F. Loeffter for their help w~th the illustrations This w o r k was s u p p o r t e d by an M R C G r a n t to K G P and a Kfllam A w a r d to JD.
REFERENCES i Berkmbht, M B, Dehagma, T. G , Feldman, A G , Gelfand, 1 M. and Orlovsky, G. N , Generation of scratching. II. Nonregular regimes of generation, J Neurophysiol, 41 (1978) 1058 1069 2 Brown, M. C , Engberg, I and Matthews, P. B. C, The relative sensltwlty to vibration of muscle receptors of the cat, J P,~ystol (Lond), 192 (1967) 773-800 3 Duysens, J , Reflex control of locomotion as revealed by stimulation of cutaneous afferents in spontaneously walking premammlllary cats, J Neurophy~iol, 40 (1977) 737-751 4 Duysens, J, and Pearson, K G , The role of cutaneous afferents from the distal hmdhmb m the regulation of the step cycle of thalamlc cats, Exp Brain Res, 24 (1976) 245 255 5 Eccles, J C, Eccles, R M and Lundberg, A , Synaptxc actions on motoneurones caused by Jmpulses in Golgl tendon organ afferents, J Phystol (Lond), 138 (1957) 227-252 6 Grfllner, S.. The role of muscle stiffness in meeting the changing postural and locomotor reqmrements for force development by the ankle extensors, Acta phy,siol 6cand, 86 (1972) 92-108 7 Gmllner, S, Locomotion in vertebrates' central mechamsms and reflex interaction, Phy~iol Rev, 55 (1975) 247-304 8 Gmllner, S and Ross~gnol, S., On the lmtmtlon of the swing phase of locomotion m chromc spinal cats, Blain Re6earch, 146 (1978) 269-277.
332 9 Grdlner, S. and Zangger, P., Locomotor movements generated by the deafferented spinal cord, .4cta physiol, scand., 91 (1974) 38A-39A. 10 Harvey, R. J. and Matthews, P. B. C , Some effects of sumulation of the muscle nerve on afferent endings of muscle spindles, and the classification of their responses into types AI and A2, J Physiol. (Lond.), 156 (1961) 470-497 11 Holmqvist, B. and Lundberg, A., Differential supraspmal control of synapuc actions evoked by volleys in the flexion reflex afferents m alpha motoneurones, Acta physioL scand., 54, Suppl 186 (1961) 1-51. 12 Hongo, T., Jankowska, E. and Lundberg, A., The rubrospmal tract. 1I. Facilitation of an interneuronal transmission m reflex paths to motoneurones, Exp. Brain Res., 7 (1969) 365-391 13 Laporte, Y. and Lloyd, D. P. C., Nature and significance of the reflex connections estabhshed by large afferent fibres of muscular origin, diner. J. PhysioL, 169 (1952) 609-621. 14 Loeb,' G. E. and Duysens, J., Activity patterns m individual hindhmb primary and secondary muscle spindle afferents during normal movements in unrestrained cats, J. Neurophysiol, 42 (1979) 420-440. 15 Matthews, P. B. C., Mammalian Muscle Receptors and Their Central Acttons, Arnold, London. 1972, 630 pp. 16 Nichols, T. R. and Houk, J. C., Improvement m tineanty and regulation of stiffness that results from actions of stretch reflex, J. Neurophysiol., 39 (1976) 119-142. 17 Pearson, K. G. and Duysens, J , Function of segmental reflexes in the control of stepping m cockroaches and cats. In R. E. Herman, S. Grillner, D. Stuart and P. Stein (Eds.), Neural Control of Locomotion, Plenum Press, New York, 1976. 18 Philippson, M., L'autonomie et la centralisation dans le syst6me nerveux des animaux, Tray. Lab Physiol. Inst. Solvay, 7 (1905) 1-208. 19 Prochazka, A., Westerman, R. A. and Ziccone, S. P., Discharges of single hmdhmb afferents m the freely moving cat, J. NeurophysioL, 39 (1976) 1090-1104 20 Prochazka, A., Westerman, R. A and Ziccone, S. P., Ia afferent activity during a variety of voluntary movements in the cat, J. Physiol. (Lond), 268 (1977) 423-448. 21 Stuart, D. G., Mosher, C G., Gerlach, R L. and Reinking, R. M., Selective activation of la afferents by transient muscle stretch, Exp. Brain Res., 10 (1970) 477-487. 22 Walmsley, B., Hodgson, J. A. and Burke, R. E., Forces produced by medial gastrocnemms and soleus muscles during locomotion in freely moving cats, J. NeurophysioL, 41 (1978) 1203-1216
321
I N H I B I T I O N OF F L E X O R B U R S T G E N E R A T I O N BY L O A D I N G A N K L E E X T E N S O R MUSCLES I N W A L K I N G CATS
J. DUYSENS* and K. G. PEARSON ** Department oJ Phy~tology, University of Alberta, Edmonton, Alberta (Canada)
(Accepted August 30th, 1979)
SUMMARY The role of propnoceptive input from the ankle extensor triceps surae in the control of walking was examined m premammillary cats walking on a treadmill. The left hindhmb was rigidly fixed m one position after denervatmg almost all the leg muscles except the ankle extensor (triceps surae) and ankle flexor (tlbiahs anterior). Rhythmic alternating contractions of the isolated ankle flexor and extensor occurred in the fixed hlndleg during periods of walking in the other three intact limbs. These rhythmic contractions disappeared when the isolated triceps surae was stretched so as to increase the force of the active contractions to beyond 4 kg. With maintained stretch the periodic contractions in the ankle flexor and extensor returned only after the force Jn the stretched triceps surae gradually decreased and fell below approximately 4 kg. Isometric contractions of the triceps surae produced either by stimulation of ventral root S1 or by large amplitude vibrations also led to the sudden disappearance of ankle flexor bursts. Inhibition of the locomotory rhythm could also be produced in all muscles of a single intact hindleg by clamping the ankle joint in a flexed position so as to stretch the ankle extensor. In all these cases, an increased rate of stepping of the contralateral hmdleg was associated with the inhibition of the rhythmic locomotory act~wty. It is suggested that triceps surae propnoceptors slgnalhng the presence of loading of the hindlimb extensor muscles inhibit the central generation of hmdlimb flexion. During normal walking this mechanism could be of major importance during stance to prevent the initiation of the swing phase of a time when hindlimb extension is fully needed to support the weight of the animal. Thus a necessary, but not always
* Present address: 6, Hoge Vesten, 9300 Aalst, Belgmm. ** To whom requests for reprints should be addressed at the Umversity of Alberta.
322 sufficient, condition for the mihatlon of swing may be an unloading of leg extensor muscles.
INTRODUCTION Recently, marked advances have been made m recording from single muscle afferents in freely walking cats14.lg, 20. Yet very httle is known about how this afferent actwity is used by the nervous system to control walking. Although there ~s some dasagreement about the usefulness of propriocept~ve reflexes at hagher speeds of locomotion 6,16, at is generally thought that feedback from muscle afferents must, under most c~rcumstances, play an ~mportant role. Certainly ~t ~s no longer believed that locomotion is based on a chain of reflexes 18 since deafferentatlon does not abohsh the m o t o r program for walking m the spinal cat 9. Rather, ~t as thought that afferent input and reflexes may play a role m the adaptation of the central locomotor program to the changing needs of the envaronmentL Grillner and Rosslgnol 8 have recently demonstrated that one important group of afferents involved m terminating the stance phase arises from the hip. A necessary condition for the swing phase to be anatiated as that the hap be extended beyond a critical angle (about 120° in their experimental arrangement). Although hip extension may be anecessaryconditlon for lmtmtmg of swing, it does not follow that ~t ~s always a sufficient condition. In fact, the present investigation began from the observation that stepping movements in a hindlimb of a premammlllary cat walking on a treadmill could be inhibated by manually holding the ankle in a flexed position even when the hap was extended well beyond the angle at which swing as normally initiated. Th~s initial observation mdacated that afferents in the ankle extensor muscle (triceps surae) could inhibit the stepping generator and suggested the hypothesis that a necessary condition for the swing phase to be lnitmted was that the leg extensor muscles be unloaded. The present experiments were designed to determine more rigorously whether or not afferents from triceps surae can indeed inhibit the spinal stepping generator. A preliminary account of these experiments has been pubhshed iv. METHODS The present experiments were performed on a series of premammillary cats, 28 of which showed spontaneous walking on a motor-driven treadmill tbr periods ranging from 30 min to 8 h. Stereotaxlc decerebrat~on was performed at a precollicular and premammillary level as described prewously 3,~. The left hindhmbs of 25 of the 28 animals were partially denervated by cuttmg the obturator, femoral, d~stal tibial. superficial peroneal and sural nerves as well as the posterior and medial articular nerves at the knee. The aim of this partml denervatlon was two-fold: (1) to deafferent as much as possible the surroundings of the triceps surae so that any lnampulat~on of the triceps surae would not ehcit afferent actwity from receptors outside these muscles, and (2) to leave the innervation of a few proximal muscles intact so that
323 one could check whether events occurring in ankle flexors and extensors were paralleled by similar events m muscles at a different joint of the same hmdhmb. The animals were held in place by two hip and two ear bars while the pamally denervated leg was fixed to a frame over the treadmill so that the hip was extended (135°), the knee slightly bent (140") and the ankle held at more than 90 ~ (see below). During walking the muscles which were not denervated m the fixed hmdhmb showed rhythmic alternating bursts of activity appropriately phased w~th their counterparts m the free right hmdhmb. The ankle was held at an angle between 90: and full extension so that well developed tlbiahs anterior (TA) bursts occurred, these being comparable in duratzon with those occurring m the free hlndhmb EMG signals were recorded from medial gastrocnemms (MG) and T A m the fixed hmdllmb and MG m the contralateral free hmdlimb using two copper wires insulated except for the t~p and inserted into the muscle at an interelectrode d~stance of 0.5-1 cm. The force m the freed Achilles tendon of the fixed hmdhmb was recorded wa a Grass force transducer. To determine the length of the freed triceps surae relative to their length m the intact hindhmb, a technique was followed as described by Nichols and Houk 1~ A screw was inserted m the distal tlbm and the triceps surae tendon was marked with sutures at points hnmg up w~th the screw when the ankle was held at 90 °, at full flexmn and at full extension (the knee angle was kept constant at 140°). The calcaneus bone was then cut and attached to the force transducer by means of a thread through the cut part of the calcaneus bone. The force transducer to which triceps surae was attached could be moved to any desired position, thus stretching the muscle by means of a manually controlled manipulator. The scale on this manipulator was set to zero when the triceps surae suture corresponding to 90 '' ankle angle was hned up with the screw m the tibia (L = 0) Hence, triceps surae lengths corresponding to ankle angles of greater than 90" were negative length values. The ankle angle corresponding to any g~ven triceps surae length could be calculated since ~t was found, m agreement w~th Stuart et al. 21, that a change m muscle length of 2 mm was equivalent to an angle excursion of 10. In 6 experiments the triceps surae was vibrated by attaching a Pye-Ling V47 vibrator in series with the force transducer. The wbrator was driven by a waveform generator which allowed control of the frequency and amplitude of the wbratory stimulus. In 6 cats, a lammectomy was performed, in addition to the usual procedures, to permit recording from single dorsal root filaments (L7 or S1 ) and to enable stimulatmn of the ventral root SI. The romulus strength to the ventral root was determined relative to the threshold of the largest motor axons in thzs root supplying triceps surae. The frequency of stimulation was usually 90 Hz, but similar results were obtained with lower frequencies provided they were high enough to give fused contractions of triceps surae (about 35 Hzl~). In three cats. the 4 limbs were left free but an adjustable metal clamp was attached to the left ankle with a screw through the tibia and a clamp around the deafferented foot. A bolt m a semicircular sht in this clamp (Fig. 4) allowed the ankle to be fixed m any pos,tlon.
324
iE iE iF cE
El
I
iE
~
~
|
I
v
III
IIII
f
iE iF cE
b
I
I sec
Fig. 1. Inlubition of rhythmic motor activity by stretch of the triceps surae in a fixed hmdhmb of a walking premammillary cat. A 10 mm stretch was applied at a constant rate beginning at the start of record a and ending at the end of record b (this stretch was equivalent to a change in ankle angle from 160° to 110°, see Methods). As the triceps surae was stretched, the duration of the tibialis anterior bursts (third trace, iF) gradually shortened until the bursts suddenly disappeared completely when the force m the stretched triceps surae reached approximately 4 kg (top trace, iE) (force calibration ~- 6.5 kg). The magnitude of the EMG bursts recorded from the medial gastrocnemius (second trace, IE) increased as the stretch was increased (the gain of this trace has been halved in b so as not to obscure the EMG record from the tibialis anterior). Near the end of the stretch the EMG burst activity also disappeared in the medial gastrocnemius and the continuous EMG in this muscle was associated with a maintained contraction of the triceps surae. Note in this case that the first flexor burst failure occurred despite the presence of a complete triceps surae relaxation. Despite the inhibition of rhythmic burst activity in the ankle flexors and extensors, the contralateral limb continued to step normally (fourth trace, cE). RESULTS
Rhythm inhibition by triceps surae stretch T h e present results center a r o u n d the p h e n o m e n o n o f E M G b u r s t fadures m the ankle flexor tibialis a n t e r i o r ( T A ) in response to stretch o f triceps surea (Fig. 1). By a p p l y i n g a slow stretch o f 10 m m at a c o n s t a n t rate to the freed t e n d o n o f the triceps surae in a fixed h i n d l i m b o f a s p o n t a n e o u s l y w a l k i n g cat, the d u r a t i o n o f the a n t a g o n i s t i c T A E M G bursts initially shortens until a p o i n t is reached where they d i s a p p e a r a b r u p t l y . T h e T A b u r s t fadures typically occur when the increasing force in the r h y t h m i c a l l y c o n t r a c t i n g triceps surae reaches a level o f a b o u t 4 kg. A tonic c o n t r a c t i o n o f the triceps surae, indicated by c o n t i n u o u s E M G activity, was seen in c o n j u n c t i o n w i t h the T A b u r s t failures, while all r h y t h m i c b u r s t activity d i s a p p e a r e d as well in o t h e r ipsilateral h i n d l i m b muscles which were n o t d e n e r v a t e d (hip flexors a n d extensors).
325
A / IE
__J"
I Im m
Lenglh IE
Ikg
\
Force ~
~//"
EMG
'
[/'~ i
I00
80
I sec
i
Flexo~Burst Failures
~"--o~-o.
o
I "o.. A "-0
/
o
o
,,"
Flexor Burst Durohon
60 % 4O
20
0 -15
-12
-9
-6
Length
Fig. 2. Effect of an increase in triceps surae length on the probablhty of tib]alis antemor burst failures. A sample record showing that stretching the triceps surae from L -- -15 mm to L -- -I 2 mm caused the tiblahs antemor burst (bottom trace, iF EMG (integrated)) to become shorter or to occasionally disappear. The incomplete drop m triceps surae force and the occurrence of EMG bursts m the contralateral medial gastrocnemms muscle made it possible to estimate where a failure occurred and thus to determine the number of failures during the first 15 cycles following the stretch. The failures tended to occur most frequently just after the stretch (see Fig. 3), but if the fadures were rather infrequent, as in the present example, they could appear after a few steps. This illustrates the threshold nature of these failures. B. percentage of failures occurring m the first 15 cycles following stretch (full circles) plotted against the amplitude of the muscle stretch (mm). The effect on the duration of the flexor burst duration is represented on the same plot (open circles) The duration of the flexor bursts at L ~ - l 5 mm was taken as 100 ~ . Arrows in B indicate data points representing the example gwen m A. The ~psdateral hmdhmb was fixed with the knee at 140 ° and the ankle imtmlly at 160 °.
Results similar to those shown in Fig. 1 were obtained in all 12 cats in which the slow stretch experiment was done. However, the amount of triceps surae stretch required to elicit TA burst failures varied considerably from animal to animal, and even within the same animal, depending on the quality o f locomotion. In most cats TA burst failures were seen when the triceps surae force reached the 4 kg level and required stretches o f between 9 and 15 mm. The final length of triceps surae at which
326 ~F )E
Ferce cE
s~.r
A
4
~'9 mm
~lOmm
kq
'~llmm IOsec
F~g 3, Adaptation of rhythm lnhlblt~on with maintained stretch of the triceps surae in a fixed hindlimb of a walking premammillary cat. A. the triceps surae was quickly stretched from L d5 mm to L -2 mm and held at this length (beginning of record) All rhythmic activity in tiblahs anterior (top trace, IF) and triceps surae (middle trace, iE force) was lmtlally inhibited by this stretch Rhythmtcity returned m an irregular manner over a period of 15 sec. During this period each sudden relaxation of the contraction in triceps surae was associated with a burst of activity m tib:alis anterior. The contralateral leg continued to step throughout as indicated by continuous rhythmic EMG burst activity in the contralateral me&al gastrocnemius (bottom trace, cE). B: force records from triceps surae when quickly stretched 9, 10 and 11 mm (arrows) corresponding to changes of 45 ~, 50° and 55" respectively from an initial angle of 160°, and held stretched for the remainder of the record. In each record the rhythmic relaxations were initiallyabolished and these only began to return when the force fell below a value of about 4 kg (indicated by the large brief dechne in force) The small oscillations m the magnitude of the maintained contractions immediately following each stretch were due to active stepping movements of the contralateral hindlimb, and these should not be confused with the small rhythmic contractions preceding the stretch. T A b u r s t fadures were p r o d u c e d corresponded to ankle angles of between 120 ° a n d 90 °. Thus, the stretches which produced T A burst failures were well within the physiological range of the triceps surae. W i t h very forceful walking, as for example d u r i n g periods o f sham rage sometimes seen shortly after decerebration, the triceps surae force level at which T A bursts were blocked could reach 8-10 kg a n d the length o f the stretch was close to the physiological limit. By contrast, a n i m a l s with p o o r stepping could exhibit occasional T A burst failures without a n y afferent m a n i p u l a t i o n . The effect of the a m p l i t u d e of triceps surae stretch was systematically stu&ed by stretching the triceps surae to different lengths a n d c o u n t i n g the n u m b e r o f T A b u r s t
327 failures during the first 15 step cycles after triceps surae reached its new length (Fig. 2). Between each stretch trml the triceps surae was brought back to the rest length (L -15 ram) for a period ranging from 30 to 60 sec, during which no TA burst failures occurred. In the example illustrated at the top of F~g. 2, a ramp and hold stretch of 3 mm (from L = -15 mm to L ~ -12 ram) caused only one TA burst failure during the first 6 step cycles after application of the stretch. Larger stretches increased the probabd~ty of TA burst failures as ~s seen m the plot of Fig. 2 (filled c~rcles) The largest stretch (to L -~ -4 ram), corresponding to an ankle angle of 110n, blocked all rhythmicity m the left hlndlimb (100 oo TA burst fadures). Apart from eliciting TA burst fadures, the step stretch also caused a shortemng of the duration of those TA bursts which did appear during the stretch (Fig. 2A). This TA shortening effect requxred much smaller amplitude stretches than was reqmred to evoke TA burst fadures. As seen in the plot of F~g. 2, stretching the triceps surae from L -- -15 mm to L ~-- -14 mm caused the TA burst to be shortened by 20% (open c~rcles) while causing no TA burst failures at all. Larger triceps surae stretches d~d not produce more TA burst shortening but clearly increased the number of TA burst fadures. Another factor affecting the occurrence of TA burst failures was the speed of stretch apphed to triceps surae. Slower stretches were less effective m evoking flexor burst failures. ThJs was due to a slow adaptation of the inhibitory effect on TA burst generation when triceps surae was stretched Th~s adapt~on was most easdy seen when a large amphtude stretch was quickly apphed and the triceps surae held for a long period at a new length (Fig. 3). In the example in Fig. 3A the triceps surae was quickly stretched from L ~ -15 mm to L ~ -2 mm (corresponding to a change m ankle angle from 165-to 100°) and held at this length. Although the length of the triceps surae was kept constant at the end of the stretch period (which ~s the start of the record m ]Wig. 3A), the number of TA burst failures dechned stea&ly over a period of 15-25 sec, indicating some form of adaptation. The reappearance of rhythmic TA bursts occurred when the triceps surae force fell to about 4 kg. Th~s observation was qu~te consistent, as is ~llustrated in F~g. 3B, taken from an experiment on another cat. In this case the mceps surae was stretched 9, I0 and 11 mm from the resting length (L = -15 m m to L 6, -5 and -4 ram) and held at that new length. Despite the differences in initial force levels at the end of the stretch, and the differences m final length, it ~s quite apparent that the return of the rhythmicity is hnked to the drop of the force below the 4 kg level in each stretch trial. When such a drop m force was prevented, as for example by adding triceps surae stretch when the force levels came close to the 4 kg level, no return of the rhythmicity occurred.
Rhythm mhibmon by clamping ankle To exclude the possibihty that the above described results were due to the fixation of the whole left hindlimb, experiments were done on three cats which could freely move all limbs but whzch had an ankle clamp attached to the left hindlimb (Fig. 4). With the ankle fixed at increasingly smaller angles, the TA bursts became shorter in duration, then faded to occur. Again the TA burst fadure was a discontinuous
328
( ( T
/
/
,
G _
i
III l _
v"rllm'1
2
1 ~
.
. . . . . . . . . .
~,,llllW- •
-,,,~
:
~
-
.
, F
-
G
Fig. 4 Inhibition of lpsilateral hlndhmb rhythmicity produced by clamping the left ankle in a flexed posit~on, a: clamp loose; b: clamp fixed at 70° Note the hyperextension at the hip and knee of the non-stepping hmb with the ankle clamped. The lack of rhythmicity in the clamped leg was associated with continuous EMG activity m the medial gastrocnemius (iE) and the absence of burst activity in tiblalis anterior (1F). The contralateral hmdlimb continued to step when the rhythm was inlubited by the ankle clamp as indicated by rhythmic burst activity in the contralateral medial gastrocnemius muscle (cE).
phenomenon and it showed adaptlon with a time course similar to that observed in Fig. 3. TA burst failures started to appear when the ankle was clamped at angles of about 90 °. For a complete block of the rhythmicity the ankle had to be clamped at 45", 65 ° and 70 ° for the three cats. With complete block of TA burst generation the cats neglected to use the limb with the ankle clamp and instead held this limb in hyperextension (hip angle close to 150 °) while the other three limbs continued stepping (Fig. 4B). The finding that stretch of the triceps surae, induced by clamping the ankle in a flexed position, leads to TA burst failures and to a complete block of all rhythmic activity m the ipsllateral hindlimb, indicates that fixation of the limb was not the factor responsible for the TA burst failures observed in the previously described triceps surae stretch experiments (Figs. 1, 2 and 3).
Rhythm inhibition by ventral root stimulation The occurrence of flexor bursts depends on the force in the triceps surae rather than on the length (Fig. 3B), suggesting that muscle spindle afferents are not responsible for the flexor burst failure phenomenon. The following results obtained when isometric contractions of triceps surae were produced by ventral root stimulation support this conclusion that receptors other than muscle spindles are primarily responsible for causing the flexor burst failures when triceps surae is stretched. In 6 experiments stimulation of part or all of the S1 ventral root was tested for its ability to induce T A burst failures. To create optimum conditions for the appearance of TA burst failures, the triceps surae was first brought to a length wl~ch was 1 m m below the threshold length for evoking TA burst failures. The distal end of part or all of ventral
329
A
Ventral
Root Stlmulat)on A
,E IF IF cE
Isec
B ~E
J,~,[l.~ ~,.~
cEI~ ]+_.
Triceps Surae V b r a t l o n ,,J. ,lJl ~iiPal. Jl~JJ~il [~i
J,~L., ,.~,,.
,.,l~J L ~_,
+,
.,~
"T
+
II~l,l~ll~tl ~LlaLi,
~IalLII
n r,
"'I ....
1 I
SOC
Fig 5 Failure of burst generation in tlblalis anterior produced by (A) strmulatmn of the Sl ventral r o o t a t l 3 T, and (B) vlbratlon of the tnceps surae I E , ~ F a n d c E a s l n F i g I A' the ventral root surnulat~on (a 480 msec tram of pulses at 90 Hz) was applied midway through the record and ~s indicated by the broademng of the basehne in the ~F trace B" the time of apphcatmn of the vibratory stimulus (a 200 msec period at 64 Hz) is indicated by the rippled bar underneath the rE EMG trace. Note that the period between extensor bursts on the contralateral side was temporarily shortened as a result of ventral root stm~ulat~onor the mceps surae wbration. The resulting resetting of the contralateral rhythm paralleled the resetting of the zpsdateral rhythm. Force calibration In A = 6 5 kg.
root SI was then sttmulated for periods r a n g i n g from 100 msec to 5 sec to reduce an Isometric c o n t r a c t i o n m triceps surae. Prior to and following each r u n the threshold stimulus strength was determined as that reqmred to elicit the first detectable triceps surae twitch. With stimulation of the whole ventral root T A burst failures occurred (Fig. 5A) w~th st~muh at 1.3 >< T, which are below threshold for the activation of g a m m a m o t o n e u r o n s TM. This result suggests that T A burst failures do not require the actwatlon of the tnceps surae spindle afferents since the induced shortening of this muscle must have caused a fall, rather t h a n a rise, in activity in triceps surae spindles. Th~s was d e m o n s t r a t e d directly for three triceps surae spindle afferents m dorsal root S 1 by observing a decrease in their discharge rate when the whole o f the S 1 ventral root was stimulated at a strength o f 1.5 / T.
Trweps surae vibratlon Small amplitude vibration is generally recognized as being an effective means o f
330 ehcitlng activity in primary muscle spindle afferents, while larger amplitude wbratlon may also activate Golg~ tendon organs, especially in contracting muscles ~'. In 6 cats the isolated triceps surae was wbrated for periods ranging from 100 msec to 5 sec during walking. As m the experiments w~th ventral root stimulation, the muscle was first brought to a length just subthreshold for producing TA burst failures. When small amplitude vibration was applied, in the range of 6-10 #m (f -- 64 Hz), the effect was to decrease the duration of the TA bursts (by about 40 ~o) but not to produce TA burst failures. T A burst failures did appear when the wbration amplitude was increased to beyond 30 #m. Long periods of high amplitude vibration, lasting several seconds, sometimes completely abolished TA burst actwlty for the duration of the vibration, and often TA burst failures could be ehclted by a short duration (200 msec) high amplitude vibration (Fig. 5B). As was usually found for muscle stretch or ventral root stimulation, the occurrence of TA burst failure was associated with a shortening of the contralateral step cycle (Fig. 5B). DISCUSSION The present results on walking premammdlary cats provide ewdence that propnoceptors in triceps surae can inhibit the generation of rhythmic bursts of EMG activity in ipsllateral flexors such as tibialis anterior. Furthermore, the sudden and complete abolition of bursts in an ankle flexor as well as in flexors at other joints indicates that the described results are due to triceps surae afferents acting on a premotoneuronal center. We will refer to this center as the flexor burst generator, and we presume it is part of, or is tightly coupled to, the central stepping generator itself. We propose that the initiation of the swing phase requires the activation of the flexor burst generator and depends on this generator reaching a certain threshold. There are many factors which appear to play a role in controlling this threshold. We have shown here that afferents from triceps surae can increase the threshold and prevent flexor burst generation, while previously we showed that flexor burst generation can be inhibited by stimulation of large cutaneous afferent~3. Factors such as the general excitability of the animal may also control threshold. For example, in cats exhibiting very vigorous walking it was difficult to induce TA burst fadures, while such failures could occur spontaneously in animals showing weak stepping. The burst failure phenomenon is not umque for the hmdhmb flexors for It has also been reported to occur m hmdlimb extensors a. Hence it seems that there exist at least two burst generator systems, one for the production of rhythmic flexor bursts, the other for the production of locolnotory bursts of activity in the extensor group of hindlimb muscles. This conclusion has also been reached from recent ~tudies on rhythmic scratching movements of the hindhmbs in cats 1. Although the primary aim of this investigation was not to ~dentlfy the group (or groups) of afferents responsible for producing flexor burst failures, some of our data do bear on this matter. For example, with large maintained stretches of different lengths applied to triceps surae the TA burst activity begins to return only when triceps suraeforce declines below a critical level (about 4 kg m most experiments). It follows
331 t h a t afferents signalling triceps surae force, r a t h e r t h a n length, p r o d u c e the tnhlbit~on of the flexor burst generator a n d t h a t Golgi tendon organs from triceps surae are m o s t hkely to be responsible for the flexor burst failures. A c o n d i t i o n which favors the selective activation o f Golgl tendon organs, namely st~mulatron o f the S1 ventral r o o t at 1.3 T, also p r o d u c e d flexor burst failures (F~g. 5A). The p r o p o s e d flexor burst i n h i b i t o r y function o f the triceps lb afferents is in a p p a r e n t confhct w~th prevxous d a t a 5,13 which ascribed a flexor excitatory and extensor inhibitory role to these afferents. The latter data, however, were o b t a i n e d m spinal cats which are k n o w n to show p r e d o m i n a n t l y flexor excitatory reflex actions. Nevertheless, flexor i n h i b i t o r y effects from extensor lb afferents have occasionally been described in spinal cats (ref. 5, Fig. 12) as well as m decerebrate cats ~2. Such conflicting findings m a y be explained by assuming 'alternatzve reflex p a t h w a y ' (ref. 1 1). The present results suggest that some o f these alternative pathways, which a p p e a r e d rather inslgmficant m experiments on motionless ammals, m a y prove to be very i m p o r t a n t in the actually m o w n g cat. In concluston, the most i m p o r t a n t result o f the present investigation ~s that the p r o d u c t i o n o f forces greater than a b o u t 4 kg in the triceps surae leads to inhibition o f the generation o f flexor bursts. Since the triceps surae easily p r o d u c e s 4 kg o f force d u r i n g stance in n o r m a l walking animals 99, this inhibitory influence from triceps surae afferents o n t o the flexor burst g e n e r a t o r p r o b a b l y functions to prevent the initiation o f swing when the extensor muscles are l o a d e d during stance. ACKNOWLEDGEMENTS We wish to t h a n k Drs. R. B. Stein, T. R. Nichols a n d J. D. Steeves for their valuable criticisms, Ms. O White for typing the manuscript, and K. Burt and F. Loeffter for their help w~th the illustrations This w o r k was s u p p o r t e d by an M R C G r a n t to K G P and a Kfllam A w a r d to JD.
REFERENCES i Berkmbht, M B, Dehagma, T. G , Feldman, A G , Gelfand, 1 M. and Orlovsky, G. N , Generation of scratching. II. Nonregular regimes of generation, J Neurophysiol, 41 (1978) 1058 1069 2 Brown, M. C , Engberg, I and Matthews, P. B. C, The relative sensltwlty to vibration of muscle receptors of the cat, J P,~ystol (Lond), 192 (1967) 773-800 3 Duysens, J , Reflex control of locomotion as revealed by stimulation of cutaneous afferents in spontaneously walking premammlllary cats, J Neurophy~iol, 40 (1977) 737-751 4 Duysens, J, and Pearson, K G , The role of cutaneous afferents from the distal hmdhmb m the regulation of the step cycle of thalamlc cats, Exp Brain Res, 24 (1976) 245 255 5 Eccles, J C, Eccles, R M and Lundberg, A , Synaptxc actions on motoneurones caused by Jmpulses in Golgl tendon organ afferents, J Phystol (Lond), 138 (1957) 227-252 6 Grfllner, S.. The role of muscle stiffness in meeting the changing postural and locomotor reqmrements for force development by the ankle extensors, Acta phy,siol 6cand, 86 (1972) 92-108 7 Gmllner, S, Locomotion in vertebrates' central mechamsms and reflex interaction, Phy~iol Rev, 55 (1975) 247-304 8 Gmllner, S and Ross~gnol, S., On the lmtmtlon of the swing phase of locomotion m chromc spinal cats, Blain Re6earch, 146 (1978) 269-277.
332 9 Grdlner, S. and Zangger, P., Locomotor movements generated by the deafferented spinal cord, .4cta physiol, scand., 91 (1974) 38A-39A. 10 Harvey, R. J. and Matthews, P. B. C , Some effects of sumulation of the muscle nerve on afferent endings of muscle spindles, and the classification of their responses into types AI and A2, J Physiol. (Lond.), 156 (1961) 470-497 11 Holmqvist, B. and Lundberg, A., Differential supraspmal control of synapuc actions evoked by volleys in the flexion reflex afferents m alpha motoneurones, Acta physioL scand., 54, Suppl 186 (1961) 1-51. 12 Hongo, T., Jankowska, E. and Lundberg, A., The rubrospmal tract. 1I. Facilitation of an interneuronal transmission m reflex paths to motoneurones, Exp. Brain Res., 7 (1969) 365-391 13 Laporte, Y. and Lloyd, D. P. C., Nature and significance of the reflex connections estabhshed by large afferent fibres of muscular origin, diner. J. PhysioL, 169 (1952) 609-621. 14 Loeb,' G. E. and Duysens, J., Activity patterns m individual hindhmb primary and secondary muscle spindle afferents during normal movements in unrestrained cats, J. Neurophysiol, 42 (1979) 420-440. 15 Matthews, P. B. C., Mammalian Muscle Receptors and Their Central Acttons, Arnold, London. 1972, 630 pp. 16 Nichols, T. R. and Houk, J. C., Improvement m tineanty and regulation of stiffness that results from actions of stretch reflex, J. Neurophysiol., 39 (1976) 119-142. 17 Pearson, K. G. and Duysens, J , Function of segmental reflexes in the control of stepping m cockroaches and cats. In R. E. Herman, S. Grillner, D. Stuart and P. Stein (Eds.), Neural Control of Locomotion, Plenum Press, New York, 1976. 18 Philippson, M., L'autonomie et la centralisation dans le syst6me nerveux des animaux, Tray. Lab Physiol. Inst. Solvay, 7 (1905) 1-208. 19 Prochazka, A., Westerman, R. A. and Ziccone, S. P., Discharges of single hmdhmb afferents m the freely moving cat, J. NeurophysioL, 39 (1976) 1090-1104 20 Prochazka, A., Westerman, R. A and Ziccone, S. P., Ia afferent activity during a variety of voluntary movements in the cat, J. Physiol. (Lond), 268 (1977) 423-448. 21 Stuart, D. G., Mosher, C G., Gerlach, R L. and Reinking, R. M., Selective activation of la afferents by transient muscle stretch, Exp. Brain Res., 10 (1970) 477-487. 22 Walmsley, B., Hodgson, J. A. and Burke, R. E., Forces produced by medial gastrocnemms and soleus muscles during locomotion in freely moving cats, J. NeurophysioL, 41 (1978) 1203-1216