Contribution of the tonic vibration reflex to muscle stress and muscle fatigue.

B1. Contribution of the tonic vibration reflex to muscle stress and muscle fa tigue. Scand J Health 1993;19:35- -42. The aim of the investigation was to deter mine the influence of vibration displacement amplitude (200, 300 11m peak-to-peak), as opposed to acceleration effects at selected frequencies (40, 80, 100, 120, 150,200 Hz), on a commonly observed but often undesired motor response elicited by local vibratory stimulation, that is, the tonic vibrat ion reflex (TVR). Vibration was applied to the distal tendons of the hand flexor muscles. Changes in the activity of hand flexor and extensor muscles were analyzed as a funetion of both their initial contrac tion level (0, 10, 20% of maximal voluntary contraction) and the vibration parameters. The main re sults indicate that TVR increases with the initial muscle contraction and increases with vibration fre quency up to 100---150 Hz but decreases beyond. High-frequency vibration seems to induce less muscle and tendon stress. This result is particularly important for the design of handheld vibrating tools.

Exposure of whol e or part of the body to mech anical vibration has been shown to be detrimental to hum an sensorimoto r performance (1-4). The most pervasive effects of vibration, other than its dir ect mechanical effects on the biome chanical structure of the human body, result from its ab ility to influence the neurol ogical network by stimulating sensory receptors within the cutaneous, mu scular, and articular structures (5)(6)(7)(8) and its abil ity to affect spinal refl exes (9)(10)(11)(12). The vibration-induced activity of the se receptors is considered a lead ing cause of specific percepti ve and sensorimotor impai rment s in persons exposed to a vibratory environment (2,4,II,12). In addition, exposure to repetiti ve long-duration vibration leads to spec ific pathologies ( 13,14).
A commonly observed motor effect of vibrati on is a phenomenon ca lled the toni c vibration refle x (TVR) ( 15). It is well known that vibrat ion appl ied to a muscle or its tendon elicits a tonic reflex contraction in that muscle (16)(17)(18)(19), or its antagonist (7), dep ending upon the experimental context. Th is motor respon se result s prim aril y from the vib ration-induced activity of the primary endin gs (Ia) in the muscle spindle (20)(21)(22)(23) (activity medi ated by mon osynapt ic and polysynaptic spinal pathways), and it is highly dependent on central influences. Thus, the amplitude of the TVR is dependent on motor neuron accessibilit y by proprioceptive afferents. In addition, a cutaneo us component mediated by cutaneous afferents is likely to contribute to the TVR (23,24) . The ton ic contraction increases progre ssively with exposure time and can persist a few seconds after vibration ends . It appears in both con tractin g and relaxed muscles.
On one hand, this tonic activity superimposed on ongoing voluntary contractions is partly responsible for the alteration of force contro l, which result s in an increase in force exertion, force variability, and tendon stress (3,25). Another component of this alteration is related to the vibration-induced mod ification of the structure of the prop rioceptive afferent messages (26,27 ). On the other hand , the TVR can contribute to muscle fati gue and increase the risk of cumulative trauma di sorders resulting from expos ure to repetitive long-du ration vibration (25,28). Furthermore, in a worke r population that has used vibrating tools, the prev alence of the hand-arm vibration syndrome ranges from 6 to 100% (29).
Thus it appears that the TVR can be one ori gin of motor control perturbation and that it can contribute to specific soft-tiss ue disorders. Further understanding of the mech anisms by which vibration interacts with motor activities should lead to better prev ention of the short-and long-term risks associated with vibration expos ure.
In additi on, vibration transmitt ed by the hand varies with the impedance or stiffness of the hand -arm system and vibration frequency (30). A tight grip is often associated with an increase in impedance (14). Thus the gri p force can influence the harmfulness of vibration.
Most investigat ion s thu s far have studied the frequency dependency of the measured performance at a constant acceleration amplitude. Acco rding to a 35 Scand J Work Environ Health 1993. vol t9 . no t previous study (9). the use of a constant displacement amplitude would be more appropriate than acceleration to describe the neurophysiological effects underlying impairment in sensorimotor performance and vibration-induced motor responses such as TVR.
The objectives of the present study were, first, to quantify the changes in TVR as a function of the level of voluntary muscular activity and the frequency of the vibration applied with a constant displacement amplitude to muscle tendons and, second , to refine the understanding of the underlying neurophysiological mechanisms.

Subje cts
Ten healthy subjects who gave their informed consent were paid for their participation in the experiments. All of the subjects were college students, and their average age was 22.6 years . They were free from any known neurological or musculoskeletal disorder s.

Experimental apparatus
The subject was seated in a comfortable armchair with the right ann supported by a padded armrest , The right hand, in slight extension, gripped a vertical handle fixed to the armrest. An adjustable support helped to maintain the wrist in this position and thus imposed an isometric condition. The handle was equipped with a strain-gage dynamometer (figure I). The height of the armrest and the horizontal position of the handle were adjusted so as to obtain approximately a l20-degree angle of the elbow.
Mechanical vibration was applied perpendicularly to the distal tendons of the hand flexor muscles by means of an electromagnetic vibrator (Ling Dynamic System, type 203) equipped with a specially designed probe ( figure 1). An accelerometer placed

36
inside the vibrating probe provided feedback to a vibration compressor (Trig-tek 8018) driven by a sinewave generator. The compressor, coupled to a vibration monitor (Trig-tek 6108 ) which computed the vibration displacement amplitude, was used to maintain this latter constant in the tested frequency range. The servo-controlled vibration signal was transmitted to the vibrator through a power amplifier.
The electrical activity (electromyography) of a finger flexor muscle (flexor digitorum profundu s) and a wrist flexor muscle (flexor carpi radiali s) was recorded by pairs of small cupular surface electrodes embodied in preamplifier devices to minimize noise and wire artifacts. The respective signals were then amplified, rectified, and integrated to obtain the rootmean-square (nns) values. A ground electrode was attached on the radial styloid of the wrist, and it provided an electrical reference .

Procedure
The maximal voluntary contraction (MVC) of the grip was determin ed before each experiment. Then the subjects were trained to maintain a grip force of 10 and 20% of their MVC for I-min period s, using only the proprioceptive feedback. At first, a visual feedback from a voltmeter connected to the dynamometer was provided. After the subject felt familiar with the required submaximal level of contraction, the visual feedback was gradually suppressed and replaced by oral information. The test session started only after the grip performance had reached a steady state level and varied less than 4%.
For each trial, the subject started to exert one of three different grip forces , namely, 0, 10, or 20% of the MVC, while viewing the voltmeter. (The 0% MVC corresponded to a resting situation in which the fingers are wrapped around the handle without exerting a significant grip.) Once the proper level was reached, the visual feedback was suppressed, and oral feedback was given by the experimenter until the percentage of the MVC stabilized and the subject felt ready for the trial to begin . Force stabilization was reached within 5 to 10 s. No external feedback was provided during the trial.
Data were collected after force stabilization while the subject maintained the submaximallevel of contraction for 60 s, consisting of a 15-s control period followed by a 45-s period during which vibration was applied continuously. The vibration frequency (40,80, 100, 120, 150, or 200 Hz) was varied randomly across the contraction levels for constant peak-topeak vibration displacement amplitudes of 200 and 300 /lm.
For each subject, the experiment was carr ied out in 2 d to reduce any possible effect of boredom and fatigue . Data were collected for a total of 36 trials for each subject (three levels of contraction x six vibration frequencie s x two vibration displacement amplitud es). The subjec ts served as their own controls. Because of the practical limitations imposed by the setting of the vibration displa cement amplitud e, the trials were randomized only across the vibration frequen cies and muscle contraction levels. A 2-min rest period separated each consecutive trial.

Data recording and analysis
The raw electromyographic (EMG) data, the rms EMG data, and the grip force signal were continuously monitored on two oscilloscopes and simultaneously digitized(1000 Hz) and stored by a computer.
The rms EMG data and the grip force were timeaveraged before ( 15 s) and durin g (40 s) the vibration periods. Changes in these values were analyzed as a function of the vibration parame ters. The first 5 s of data of the vibration periods were not included in the analysis to eliminate the transient effec ts.
For each experimental condition, the grip force and the rms value of the vibration-induced increa se in muscle contraction (rms VIC), defined as rms VIC = ave(rms EMGvib)-ave (rms EMG,) . the time averages (ave) having been obtained during and before vibration respectively , were computed and averaged over the subjects. A repeated-measures analysi s of variance (ANOVA) treating the subjects as a random blocking factor was performed on the average rms VIC value and the grip force to test whether they were affected by the initial muscle contraction level, vibration displacement amplitude, or vibration frequency. The Tukey method of multiple comparisons with the 95% family comparison coef-EMG Scand J Work Environ Health 1993, vol 19, no I ficient was applied to evaluate the effect of the factor levels.

Muscle activity
As an example, the time recordin gs reprodu ced in figure 2 illustrate the effects of the application of the vibration (120 Hz, 300 urn) on the EMG activity of the finger flexor muscle and the grip force exertion. Vibration induced an increase in muscle contraction, as clearly shown by the rms EMG trace. This effect primarily resulted from the expression of the TVR.
For all of the subjects, there was a significant increase in the time-averaged rms EMG data during vibration (P = 0.01) for both the finger flexor and wrist flexor muscles. However, the rms VIC varied with the initial muscl e contraction level, vibration displacement amplitude, and vibration frequency.
Initial level of muscle contraction. For both muscles, the initial contraction level had significant effects on the rms VIC (P = 0.0001 ), as illustrated in figure 3. For the finger flexor muscle, the rms VIC was maximum at 10% MVC and was less pronounced at 20% MVC. For the wrist flexor muscle, however, the increase in the rms VIC was similar at 10 and 20% MVC.
Vibration displacement amplitude. For the finger flexor muscle, there was no significant difference in T=15s T=20s     more obvious for the finger flexor muscle than for the wrist flexor muscle , except at 20% MVC. An example is presented in figure 7.

Grip f orce
Although the EMG activity of the flexor muscles increased as the vibration incre ased, a decrease in the time-averaged grip force was observed during the vibration exposure (P =0.05). This apparent paradox was observed for nine out of the ten subj ects. A re-the rms VIC resulting from the vibration displacement amplitude of 200 J..Lm and 300 J..Lm (P = 0.43 ). However, the increase in the rms VIC with the vibration displacement amplitude was significant (P =0.0 15) for the wrist flexor muscle . This change is illustrated in figure 4. The rms VIC was summed over the vibration frequencies and initial muscle contracti on levels.
Vibration frequ ency. For both the finger flexor and wrist flexor muscles, the vibration frequency had significant effects on the rms VIC (P = 0.000 I) . Figures  5 and 6 show changes in the rms VIC for the finger fle xor muscle and wrist flexor muscle , respecti vely, represented as a function of vibration frequ ency at constant vibration displacement amplitudes for three different initial muscle contraction levels (top to bottom) . For the finger flexor muscle, the rms VIC increased with the vibration frequency up to 100 Hz; it the n incre ased at a slower rate up to 150 Hz, whereafter it decreased . This trend was consi stent for all of the experimental conditions with the exception of those for 20% MVC. For the wrist flexor muscle, the trend was slightly different. The rms VIC increa sed with the vibration frequency up to 100 Hz and then decreased. Least-square regres sion s were performed to determine whether second-order concave curvilinear associations between the rms VIC and vibration frequency existed. Table I     peated-measures ANOV A was performed on the differential grip force(DGF = GFyib-GFref) to test whether the changes in grip force correlated with the vibration displacement amplitude, frequency, and initial muscle contraction level. Neither significant main effects nor interaction effects were found (P =0.05).

Discussion
Differences in the response of the muscles to the various vibratory situations were observed. They appeared to be dependent on the specific muscle, its initial level of activity, and the parameters of the vibratory stimulus. Nevertheless, the increase in muscle contraction resulting from the application of the vibration was primarily related to the development of the TVR. This conclusion is supported by the observation of the classical attributes of the TVR, such as latency of onset, progressive build-up, and variability. Furthermore, the EMG activity increased concomitantly with a grip force decrease. The versatility of the TVR (12,17,22) certainly contributed to the large inter-and intraindividual variability.

Sensitivity of the tonic vibration reflex to initial muscle contraction level
The contribution of the TVR increased with initial voluntary muscle contractions of moderate level (10% MVC). This effect may have resulted from a combination of the following factors affecting the muscle spindles: (i) an increase in the responsiveness of the primary spindle endings to the vibratory stimulus as a result of the increased fusimotor drive (5, 31); (ii) a facilitation of the accessibility of alpha motoneurons by Ia-afferents as induced by the descending voluntary command; (iii) an increase in vibration transmissibility produced by the stiffening of the tissues accompanying muscle contraction (Martin, unpublished results). However, a higher level of voluntary contraction (20% MVC) failed to increase the TVR. A saturation or even a smaller increase was observed for the wrist flexor muscle and finger flexor muscle, respectively. Such a result suggests that the driving of alpha motoneurons by the vibration-induced activity of the la-afferents is close to its maximum at moderate levels of voluntary contraction.

Sensitivity of the tonic vibration reflex to vibration amplitude
The increase in vibration amplitude resulted in either an increase or an nonsignificant change in the TVR evoked in the wrist flexor and finger flexor muscles, respectively. These differential effects may have been due to the spatial arrangement of the muscle tendons, which are not in the same plane, and the biomechanical properties of the tissues. It seems that an increase in vibration displacement amplitude fails to produce a larger stretch of the tendon of the flex-40 or digitorum profundus (the measured finger flexor muscle). However, the apparent higher sensitivity of the primary spindle endings of the flexor digitorum profundus, as described later, may lead to an early saturation of the number of Ia-afferents recruited in the muscle by the vibratory stimulus in the 200-300 urn displacement amplitude range. As indicated by a microneurographic study (27), almost all of the laafferents of the pretibial muscles are activated in synchrony by low-amplitude tendon vibration below 100 Hz. Therefore, we suggest that this sensitivity contributes to the maximal strength of the TVR at a moderate vibration intensity for some muscles.

Sensitivity of the tonic vibration reflex to vibration frequency
Several studies have investigated the frequency dependency of the TVR at constant acceleration (32,33). These studies have shown a monotonous decrease in muscular activity as the frequency increases. As suggested previously (9,10), the decrease in displacement amplitude as the frequency increases at constant acceleration yields a spatial derecruitment of the firing of la-fibers. Therefore this process results in a decrease in the TVR component. In the present study, since the displacement amplitude was constant regardless of the frequency, we assume that a quasi identical recruitment took place at all of the frequencies. The positive slope of the TVR up to 100-150 Hz probably resulted from an increase in the depolarization of motor neurons with the firing frequency of the la-afferent fibers, which increases the number of responding motoneurons. The negative slope exhibited beyond correlated highly with the "frequency response" of the primary spindle endings. Indeed, these receptors can respond in I: I synchrony up to about 100-150 Hz (5,26,27); beyond this "cutoff' frequency, driving is less secure and most of the receptors start to misbehave and respond at subharmonic frequencies or at random. Thus a derecruitment process affecting the motoneurons is likely to occur. Since it has been shown that the firing frequency of the motor units is not significantly modified by the vibration frequency in the 30-95 Hz range (23), the contribution of a temporal summation to TVR changes can be ruled out. The slight difference in the "cutoff' frequency of the TVR elicited from the finger flexor and wrist flexor muscles seems to indicate a higher stretch sensitivity of the spindle endings in the measured finger flexor muscle. It may result from the functional difference of the two muscles. Indeed, precise regulation of muscle activity, and thus accuracy of proprioceptive feedback, is conceivably more critical for the fingers than the wrist.
Some of our results that were related to vibration displacement amplitude and frequency effects differed from those obtained earlier by Eklund & Hagbarth (17). However it is important to note that Ek-lund & Hagbarth applied vibration using an "eccentric vibrator" strapped to the limb over the muscle tendons. The vibration amplitude was two to ten times higher than the levels used in our experiment. Thus we assume that under their experimental conditions the spread of vibration to the limb and the activation of the skin area covering the vibrated muscles, which is known to have a facilitatory effect (34), contributed to the overwhelming strength of the TVR.

Grip force
The mean results over the 10 subjects indicate a decrease of grip force while the activity of the flexor muscles increases. However a detailed analysis of the individual result s in each situation showed either an increase or a decrease in the time-averaged grip force , the latter being predominant. In addition, an initial increase in the force at vibration onset was often observed, followed by a decrease, as illustrated in figure 2. These results are not inconsistent. They rather confirm the plasticity of the organization of a motor response as a function of the context, as suggested by many authors (7 ,II,35,36). It is worth mentioning that, in the present case, the subjects were required to maintain a selected force level while relying only on proprioceptive feedback.
During vibration, the subjects generally reported the sensation of an increase in grip force. Such a perception probably led to a reorganization of the central command in an attempt to match the previbration sensation. This reorganization seems indicated by force changes over time . It may be that vibrationinduced activity of Golgi tendon organs is interpreted by the central nervous system as an increase in muscle contraction. This latter phenomenon leads to a decrease in the grip force through a complex interaction of the flexor and extensor components, since flexor muscle activity does not decrease. This hypothesis parallels the classical interpretation of movement amplitude undershoot resulting from vibrationinduced activation of agonist muscle spindles (37)(38)(39).
The increase in grip force observed for some of the subjects probably resulted more simply from a lack of voluntary control of the grip force, and this lack of control allowed an opportunity for the full expression of the common effects of the TVR. This result contrasts with the general observation of force increase provoked by "whole-limb" vibration (3,25). However, it is not a paradox if we assume that force control, similarly to position or velocity control, of slow movements is based on the differential imbalance between agonist and antagonist proprioceptive feedback (4,(39)(40)(41). Thus, in that case, vibrationsuperimposed activity of "antagonistic" proprioceptive afferents decreases the relative differences of the respective signals and leads to an increase in movement amplitude and velocity (4,39,41). Further-Scand J Work Environ Health 1993. vol 19. no 1 more, in the experiment carried out by Radwin et al (25) , the task consisted of holding various loads attached to a vibrating handle. Thus maintaining a constant grip force while holding the load was not specifically required.
To conclude, our results indicate, first, that the contribution of the TVR to muscle activity may already reach a maximum at low vibration intensities or at moderate levels of voluntary muscle contraction comparable with the average exertion usually required to hold pneumatic screwdrivers (42) . Second , the general increase in muscle activity resulting from vibration exposure contributes to muscle stress and fatigue. Finally, high -frequency vibration (> ISO Hz) tend s to induce a relatively weaker TVR and thus superimposes less muscle and tendon stress. Such remarks are of particular importance for the design of handheld vibrating tools . From a neurological point of view, they may provide justification for increasing tool speeds when possible. Obviously, maintaining a constant vibration displacement amplitude while increasing the frequency increases the acceleration considerably. Such an effect may increase tissue insults and is not recommended.