Exp Brain Res (1990) 79:365-372

9 Springer-Verlag 1990

Errors in force estimation can be explained by tendon organ desensitization S. Thompson, J.E. Gregory, and U. Proske Department of Physiology, Monash University, Clayton, Victoria 3168, Australia

Summary. Here we report observations on the sense

Key words: Tendon organ - Tension - Propriocep-

of muscle tension in human subjects and compare them with responses of tendon organs in cat hindlimb muscles. Human subjects learned under visual guidance to estimate a 4% maximum voluntary contraction (m.v.c.) of elbow flexors of one arm. When they were able to reproduce this force reliably without visual feedback, they repeated the estimation immediately after a 5 second m.v.c or a 5 second period of relaxation. In a second experiment the 4% m.v.c was generated under visual control with one arm, and matched with the other, test arm, without visual feedback. The matching task was then repeated after test arm conditioning. In both experiments subjects reported an accurate match using significantly more than the reference force ("overmatched") after an m.v.c. The overmatching was greatest during the first 5 second period following the conditioning contraction, and during the subsequent 20 seconds it gradually declined to near reference levels. The size of the matching error was directly proportional to the duration of the conditioning contraction. In the first experiment extension of the arm immediately following conditioning increased the error, in the second it slightly decreased it, although tension continued to be overmatched. In a series of experiments on the soleus muscle of anaesthetised cats responses of tendon organs to 10% of maximum contraction were seen to drop sharply when preceded by a conditioning maximum contraction. The time course of recovery was comparable to the decline in matching error in the human experiments. In conclusion, one explanation for the error in force matching seen in human subjects after an m.v.c is that sensitivity of tendon organs has been lowered as a result of the activity generated during the conditioning contraction.

tion - Contraction - Afferent - Human

Offprint requests to: S. Thompson (address see above)

Introduction It is now generally accepted that in addition to a sense of effort or heaviness we have a sense of muscle tension. The sense of effort is thought to be derived from the corollary discharge of motoneurones. The sense of tension probably arises from the afferent signals of tendon organs (McCloskey et al. 1974; Roland and Ladegaard-Pedersen 1977), and it is known from animal experiments that afferents of tendon organs have access to the cerebral cortex ( M d n t y r e et al. 1984). While it is relatively straightforward to produce disturbances in the sense of effort using muscle relaxants, it has proved more difficult to generate errors in the sense of tension. It has recently been reported that human subjects, in the absence of visual feedback, erroneously use too much force when attempting to match a learned, small reference force in elbow flexor muscles after the muscles have been conditioned by a voluntary contraction (Hutton et al. 1987). A brief muscle stretch immediately after the conditioning contraction reduced the size of the errors. The authors suggested that post-contraction potentiation of stretch reflex pathways resulting from increased activity in muscle spindles could account for their findings. A second possible contributing factor was thought to be desensitization of tendon organs. This report was of particular interest to us in view of our own recent observations on the activity of muscle spindles, the size of the stretch reflex and errors in position sense after conditioning contractions (Gregory et al. 1986, 1987, 1988). Furthermore, our findings had led us to conclude that the

366

rise in activity of muscle spindles after a conditioning contraction, rather than having a potentiating effect, had a predominantly inhibitory influence on homonymous motoneurones, probably by raising the prevailing level of presynaptic inhibition of Ia afferent fibres (Mark et al. 1988). It was therefore decided to repeat the experiments of Hutton et al. and to investigate in more detail the possible underlying mechanisms. The use of an inappropriately large force ("overmatching") when attempting to reproduce low levels of force after a conditioning contraction was confirmed, but it was found that the effect of muscle stretch was small and not significant. In a supplementary series of animal experiments it was confirmed that one possible cause of the error in tension matching was a decreased sensitivity of tendon organs. Methods Human experiments The first series of experiments was carried out on 17 female and 23 male, young adult h u m a n subjects. The experimental set up was in large part similar to that used by Hutton et al. (1987). Subjects were seated with their right elbow at a right angle, resting on a padded block. The right forearm was secured with a compression bandage to a hinged, padded board connected to an angle measuring device so that any deviation of the forearm from its set position could be detected. Subjects were instructed to grip a plastic handle which was connected to a strain-gauge by an aluminium rod. They were trained to perform low-level isometric contractions of their elbow flexors, while they were able to monitor the force they were exerting on the screen of an oscilloscope. A r m angle and force values were continuously recorded on a Grass pen-recorder. Most of the equipment was hidden from the subject's view by a screen. At the beginning subjects were instructed simply to match the learned level of force. When it was realised that this could be done in several ways subjects were told to concentrate specifically on the tension they were feeling in elbow flexor muscles. Some subjects reported that they could not sense any tension in biceps when the force produced was 2% of a maximum voluntary contraction, m.v.c, [c.f. Hutton et al. (1987)]. The strength of the test contraction was therefore increased to 4 % m.v.c. Because this was still a very low force, subjects tended to use cues such as taking up the slack in noisy moving parts in the equipment to help them indicate attainment of the required level. All loose connections in the equipment were therefore firmly clamped. Subjects were asked to perform an m.v.c, at least twice, the largest value being used to calculate the 4 % level. Subjects were then taught to reproduce this force by being shown its visual equivalent on the oscilloscope. The task was first carried out under visual feedback and then after view of the screen was blocked. The actual test consisted of maintaining the 4% m.v.c. value with visual feedback followed by a period of 5 or 10 seconds of relaxation. The subject was then instructed to rematch the learned force without visual feedback. The period of re-matching lasted 15 seconds. If the error in matching following the relaxation period was greater than 25%, the subject continued practising. If it was less than 25%, the actual experiment was begun. Most subjects were able to reproduce the 4% of m.v.c, level quite accurately after 10-15 minutes of practice. The

experiment required the subject to re-match the 4% m.v.c for 15 seconds, in the absence of visual feedback, after their elbow flexors had been conditioned by an m.v.c lasting either for 5 seconds in some experiments or for 1,3 or 5 seconds in others. Subjects were instructed to re-match the 4% level immediately after the conditioning contraction, after a conditioning contraction followed by a 5 s rest period, or after a contraction followed by arm extension. Factors such as muscle fatigue and concentration lapses could not be controlled for so it was important that they were monitored. This was done by carrying out a control before each test, and by repeating each conditioning trial three times and calculating the average result. The control consisted of a period of relaxation over the same period as the contraction (Fig. 1). Each post-conditioning estimate of force was expressed as % error of the learned force. Values from the 3 trials were averaged and compared with the average % error in the control trials. A second series of experiments was carried out on 25 subjects, 11 male and 14 female, in which both arms were used in the force matching task. The series was done firstly to eliminate the possibility that the errors in matching were the result of the subject's not remembering accurately the learned, reference force level, and secondly to decide whether the errors were likely to be of central or peripheral origin. Subjects were comfortably seated and their forearms secured to shaped, padded supports with compression bandages. Each support was connected to a strain gauge by means of an aluminium rod attached at the mid point of the support. Handles were therefore not necessary. Subjects were trained to monitor on an oscilloscope the force generated by the elbow flexors of the left arm. No visual cues were available for the force generated by the right arm. The average of left arm and right arm m.v.c's was used to calculate the 4% reference level. The procedure in the experiments concerned with the first point was as follows. Subjects were instructed to maintain the

4% mvc ~

visual .=.=..,.=,~.=~=, r

e

l

a

blind x ~

o D-=

contract

blind

4% rnvc ~ 0

0

5

10

1'5 2'0 2'5 seconds

3'5

Fig. 1. Matching of a learned level of force before and after a maximum voluntary contraction of elbow flexor muscles. Both traces show force recorded by strain gauge on which the subject was pulling. Upper trace recorded during matching of a learned level of force (4% m.v.c) before and after a 5 second period of relaxation. Lower trace, the match attempted after a 5 second duration m.v.c. After the m.v.c the initial overestimate of force of about 100% gradually declined with time. Force calibration applies to both traces

367 4% level under visual control with their left arm. Then they were asked to match this with their right arm, but without visual feedback. They were instructed to concentrate on levels of tension perceived in the left elbow flexors and to reproduce these with the right. Once the task had been learned, matching with the right arm was carried out after three conditioning procedures involving the right elbow flexors: (i) an m.v.c for 5 seconds then matching (ii) an m.v.c for 5 seconds, then 5 seconds rest, before matching (iii) an m.v.c for 5 seconds then arm extension before matching. A number of subjects were unable to match the two sides with an accuracy of better than 33% error. They were therefore not used in the experiment. In the experiments concerned with the question of whether a central effect was involved in tension overmatching, subjects were trained to reproduce with their left arm 4% m.v.c without visual feedback, while generating this force under visual control with the right arm as in the previous experiments. They were then instructed to relax their left arm and to perform an m.v.c with their right arm, after which they re-matched the 4% with their left arm.

Animal experiments Experiments were carried out on adult cats anaesthetised with pentobarbitone sodium (40 mg per kg body weight). Depth of anaesthesia was regularly monitored, and when necessary, subsequent doses of anaesthetic were given. End-tidal CO 2 was monitored by means of a probe in the tracheal tube. Body temperature was maintained at 37~ by a thermostatically controlled electric blanket with rectal probe. The left hindlimb was used. All muscles of the hip and leg were denervated, except for soleus. A marker was placed on the soleus tendon. The ankle joint was maximally flexed and a second marker was placed on nearby tissue, in line with the soleus marker to indicate maximum body length (Ira, x) of the muscle. The soleus tendon was connected to a servo-regulated muscle stretcher, using a chip of calcaneum bone as an anchoring point. The upper and lower ends of the tibia and fibula were fixed to a metal table using rigid clamps. A laminectomy was carried out to expose the lumbosacral spinal cord from segments L1 to $2. Single functional afferents were dissected from filaments of dorsal root. The soleus nerve was freed over a distance of 2 3 cm proximal from its point of entry into the muscle to allow placement of a pair of stimulating electrodes for measurements of conduction velocity. Afferent fibres, were identified as coming from tendon organs by their conduction velocity lying in the G r o u p I range and by their "in series" response during graded muscle contraction. The L7 and Sl ventral roots were divided into four parts, so that when each was stimulated separately it produced about the same tension. Each piece of ventral root was placed on a separate stimulating electrode and the whole electrode array was connected to a distributed stimulator (Rack and Westbury 1969). Responses of tendon organs were recorded during distributed stimulation over a range of frequencies. Afferent responses and muscle tension were recorded by means of a digital oscilloscope and stored on flexible discs.

and then after performing an m.v.c of their elbow flexors. The mean value for 4% m.v.c for the 32 subjects was 7.0 N. Most subjects after a brief training period were able to reproduce this level of force quite accurately, although there was a slight tendency for overmatching in the initial 1-2 seconds of the rematching attempt (Fig. 1). Two subjects were unable to learn the task after 15 minutes of practice and were therefore not used. Of the remaining 30 subjects, most overmatched the learned force after a maximum contraction. Figure 1 shows a typical subject record of 'blind' force matching after an m.v.c and after relaxation. The mean percentage error in force estimation was significantly greater after an m.v.c compared to the control (P

Errors in force estimation can be explained by tendon organ desensitization.

Here we report observations on the sense of muscle tension in human subjects and compare them with responses of tendon organs in cat hindlimb muscles...
738KB Sizes 0 Downloads 0 Views