Sinusoidal Stretch Research Articles

Responses of isolated Golgi tendon organs of cat to sinusoidal stretch.

1. Receptor potential and tension have been recorded from isolated Golgi tendon organs in response to sinusoidal stretch. Responses depended on amplitude and frequency of stretch and on the initial (resting) tension of the preparation. 2. Both tension and receptor potential behaved as power functions of stretch amplitude over most of the range corresponding to physiological tendon strains. However, for very small stretch amplitudes (less than 8 microns), a more linear response was seen. Those characteristics of responses that depended on stretch amplitude behaved similarly at all frequencies examined. 3. Frequency dependence of tension was slight. Its character, a gradual monotonic increase in response with increasing stretch frequency and a constant phase lead of a few degrees, did not change over the examined frequency range from 0.12 to 80 Hz. In contrast, receptor potential displayed a marked frequency dependence, increasing rapidly with increasing frequency of stretch in the range from approximately 1 to 20 Hz, then slowly declining as frequency was further increased. 4. Changes in initial tension of the preparation produced marked parallel changes in the amplitude dependence of tension and receptor potential. Frequency response was not significantly affected. 5. By comparing tension and receptor potential responses, the relative contributions of mechanical and electrical properties of the receptor to the sensory transduction process was examined. The present results suggest that in tendon organs the observed nonlinear dependence on amplitude of stretch originates primarily in the mechanical stage of transduction. Dynamic sensitivity, however, seems largely attributable to ionic processes within the sensory terminal membranes.

Observations on the effects on spindle primary endings of the stimulation at low frequency of dynamic β-axons

Stimulation of single dynamic beta-axons at presumed physiological frequencies (10-30/s) significantly increases the sensitivity of spindle primary endings of cat peroneus brevis muscle to ramp and sinusoidal stretches. This action becomes greater as the amplitude of stretch and the initial muscle length is increased. Stimulation at low frequency of two dynamic beta-axons supplying the same spindle greatly enhances the effect elicited by each axon individually.

Impulse patterns of Ia afferents and Ia-activated DSCT neurons during sinusoidal muscle stretch in cats

The cat gastrocnemius muscles of one hind leg were stretched sinusoidally with amplitudes between 10 micrometer and 2.5 mm and frequencies between 1 and 30 Hz. The stretch response of deefferented muscle spindle afferents and of Ia-activated cells within Clarke's column were investigated by means of extra-cellular recordings of action potentials. The responses to 20-50 cycles were displayed in impulse patterns (raster diagrams) of the responding action potentials. The impulse patterns of Ia afferents exhibited a high degree of phase-locking (regularity) on the stretch cycle of amplitudes of about 50 micrometers at 3 Hz and all higher amplitudes or frequencies. At comparable stretch parameters the regularity in Ia afferents was 4-6 times larger than in Ia-activated DSCT neurons. The regularity in the DSCT patterns increased with an increase in stretch frequency. The impulse patterns of DSCT cells exhibited a high negative correlation between successive interspike intervals (-0.4 to -0.6) at low stretch frequencies (less than 3 Hz), which decreased with an increase in stretch frequency.

Effect of ethanol on stretch response of muscle receptors in cats

The response of muscle spindles and tendon organs to steady and sinusoidal muscle stretch was investigated at different blood alcohol concentrations (BAC). After initial anesthesia (pentobarbital), cats were spinalized, the lumbar ventral roots cut and the gastrocnemius muscles of one hindleg prepared for controlled stretching. The animals were paralyzed and artificial respiration was applied. Action potentials from isolated Ia/Ib/II afferent fibers could be recorded. Under steady stretch conditions, all fibers responded to an increasing BAC with an increase in firing rate. This could be observed already at 0.8 mg/ml BAC. The increase in discharge rate reached at the most 80 imp/s. During intoxication the regularity of firing was higher than in the no-alcohol situation. At blood alcohol concentrations higher than 5 mg/ml, the neuronal activity suddenly dropped to zero, exhibiting an irregular impulse pattern. The increase in discharge rate at steady stretch is regarded to be of minor significance in the explanation of the impairment of motor performance under ethanol. When sinusoidal stretch was applied, the increase in the mean discharge rate was smaller than at steady stretch conditions. Up to about 10 mg/ml BAC the periodical modulation of firing rate during sinusoidal stretch of a large amplitude remained mainly unchanged. After the discharge rate had dropped to zero for the steady stretch condition at high BAC, elicitation of action potentials was always possible using dynamic stretch.

Responses of cat dorsal spino-cerebellar tract neurons to sinusoidal stretching of the gastrocnemius muscle.

A defined class of cells within the nucleus dorsalis (Clarke's column) receives excitatory input from Ia afferents of mainly one muscle. Action potentials were recorded from axons of these cells (DSCT neurons) which are excited by Ia afferents of the gastrocnemius muscles. We investigated the response to sinusoidal muscle stretch over a wide range of amplitudes (10 microns-4 mm) and frequencies (0.1-130 Hz) in the deefferented preparation. The dynamic stretch was superimposed on a moderate static muscle stretch to ensure that the muscle was not slack during the phase of release. The response up to 10 Hz was displayed as PST histograms (cycle histograms) and a sinewave of stretch frequency was fitted to the PST histograms to define amplitude and phase of a response sinewave. At a constant frequency of about 3 Hz, the relation between stretch amplitude and response amplitude could well be described by decelerating intensity functions: the hyperbolic or tanh log function and a modified power function (exponent 0.48 +/- 0.12). The phase lead of the response sinewave increased with increasing stretch amplitudes of up to 0.5 mm and then decreased. At constant stretch amplitudes of 0.5-2.0 mm the frequency response was investigated. In relation to stretch frequencies between 0.1 and 1 Hz an increase in the response amplitude of 4.4 dB was observed and an increase for 13 dB/decade between 3 and 10 Hz. At 0.1 Hz the phase of the response sinewave was 48 degrees in advance and increased to a maximum lead of 89 degrees at 6-8 Hz. Above 10 Hz the positions of the responding action potentials with respect to the stretch cycle were used to define a phase, which was in advance up to 60 Hz but decreased and changed to a phase lag at higher frequencies. If in PST histograms no periods of silence occurred during the phase of stretch release, the mean discharge rate was found to be independent of the sinusoidal stretching. If the pauses were present the mean rate increased with increasing stretch frequencies or amplitudes.

The proprioceptive function of a complex chordotonal organ associated with the mesothoracic coxa in locusts

A system of chordotonal organs in the locust mesothorax consists of four subunits one of which connects to the coxa. Proprioceptive afferents from the scoloparia record the rotatory movements of the coxa. Mechanical stimulation of the sensory system by sinusoidal stretch or movements mimicking stretch as in natural walking of the locust elicits reflex activation of coxal motoneurones. Both assistance and resistance reflexes to imposed movements occur, but their intensity can vary from periods of suppression below firing threshold in a motoneurone to recruitment of additional motoneurones to the same muscle. It is concluded that some of these reflexes recorded in isolated preparations can also occur in freely walking animals where they should contribute to the muscular coordination of transitions between antagonistic movements.

After-effects of fusimotor stimulation on spindle la afferents' dynamic sensitivity, revealed during slow movements

After-effects of fusimotor action on dynamic sensitivity of spindle primary afferents from soleus resulted from low rate stimulation of both static and dynamic γ-axons. Sensitivity to small sinusoidal stretches superimposed on triangles was enhanced during triangular lengthening but not during shortening. These after-effects were elicited by fusimotor stimulation during preceding triangular shortening, but not during lengthening. Given these characteristics, they may be important functionally during alternating rhythmic movements.

Nigral output neurons are engaged in regulation of static fusimotor action onto flexors in cat.

Picrotoxin and muscimol were unilaterally injected into the postero-lateral part of the reticular zone of substantia nigra (plSNR) through chronically implanted guide cannulae in ketamine-anesthetized cats. Afferent activity of pretibial flexor muscle spindle primary and secondary endings was recorded before and after drug administration, and spindle sensitivity monitored during both sinusoidal and ramp stretch of the receptor-bearing muscle. From changes in spindle sensitivity after drug injection it is deduced that unilateral block by picrotoxin of the action of GABA on postsynaptic receptors in plSNR removes tonic static fusimotor action from flexor muscle spindle primary endings. Secondary endings seemed largely unaffected. The effect on primary endings is reversed by a subsequent injection of muscimol. It is concluded that the central nervous system, through GABA-modulated nigral output neurons, can control static fusimotor action onto flexor muscle spindle primary and secondary endings separately to some extent.

Action of vibration on the response of cat muscle spindle Ia afferents to low frequency sinusoidal stretching.

1. A study has been made of the effect of continuous vibration, at 150 Hz, upon the response of muscle spindle afferents to low frequency sinusoidal stretching (1 and 8 Hz). Using the soleus muscle of the anaesthetized cat, with severed ventral roots, recordings were made of single Ia afferents and of the massed Ia afferent discharges in the main bulk of the cut L7 dorsal root. 2. When the amplitude of vibration was large (50 micrometers, short pulses) and that of the sinusoidal stretching was not too great (50-100 micrometers, peak-to-peak) the discharge of the afferents was largely locked 1:1 to the vibration and the response to the sinusoidal stretching was abolished. 3. When the amplitude of the vibration was reduced to below that eliciting continuous afferent driving, then the response to sinusoidal stretching of any amplitude was often markedly increased. This arose through the vibration having a much more powerful excitatory action during the rising phase of the sinusoidal stretch than it did during the falling phase. 4. Averaged over a full cycle, the phase of the response to the sinusoidal stretching tended to be delayed during the vibration in comparison with the normal. This was largely dependent upon the afferents continuing to respond maximally to the vibration around the peak of the sinusoidal stretch, at which stage their unvibrated response is declining, rather than to a phase lag of the whole pattern of response. 5. The results are discussed in relation to the effects of vibration on tremor and the human stretch reflex, and on the determination of the frequency-response of spindle afferents.

An analysis of muscle spindle behavior using randomly applied stretches

The behavior of mammalian muscle spindles to relatively large amplitude randomly applied stretches is compared to the linear behavior elicited by small amplitude sinusoidal stretches. The results disclose two apparently independent sources of nonlinear behavior. One is a static nonlinearity affecting both primary and secondary endings in which the sensitivity decreases with stretch amplitude up to about 1% stretch. The other is a ‘dynamic’ nonlinearity affecting only primary endings that is manifest as a charge in the ratio of rate to proportional sensitivity depending on the duration of an applied stretch.

The effects of cerebellar stimulation on alpha motoneuron excitability and the stretch reflex in the cat

The effects of cerebellar surface stimulation on electrically evoked monosynaptic and polysynaptic reflexes and the stretch reflex were investigated in cats anesthetized with alpha-chloralose. The electrically evoked monosynaptic and polysynaptic reflexes were recorded from the ventral root following stimulation of nerves innervating the anterior tibial and gastrocnemius-soleus muscles. These responses were averaged at various times after the initiation of cerebellar stimulation and compared with the amplitude of the control response. The effects of cerebellar surface stimulation on the stretch reflex were assessed while recording the multiunit EMG activity evoked by a passive sinusoidal stretch of the gastrocnemius-soleus muscle. Cycle histograms of this EMG activity were constructed during control periods and during cerebellar stimulation. The stimuli were applied to the lateral vermis of the anterior lobe and crus I and crus II of the posterior lobe. The effects of stimulating at different frequencies and charge densities were evaluated in each animal. The magnitude of the effects produced by cerebellar stimulation on these reflexes was highly dependent on the charge density and intensity of the cerebellar stimulus. In general, higher intensities and frequencies of surface stimulation produced greater reductions in the excitability of alpha motoneurons and the amplitude of the stretch reflex. Occasionally low frequency stimulation produced an increase in alpha-motoneuron excitability, particularly when the cerebellar stimuli were applied to the hemisphere. The results of this study indicate that the maximal reduction in motoneuron excitability and the amplitude of the stretch reflex is produced at different stimulus parameters in different animals. However, in each animal the effects evoked by a single set of stimulus parameters were consistent throughout the experiment. These results demonstrate that in an animal with an intact neuraxis cerebellar stimulation applied at relatively safe charge densities can produce dramatic reductions in the amplitudes of segmental reflexes. The inter-animal variability observed emphasizes the importance of establishing the optimal set of stimulus parameters in order to maximize the reduction in motoneuron excitability and stretch reflexes in each animal.

Evidence that the human jaw stretch reflex increases the resistance of the mandible to small displacements.

1. Small 'step' or sinusoidal displacements were imposed on the mandible while human subjects maintained an average biting force of 10 N. Phase-related changes in the force resisting sinusoidal displacement were used to determine the mechanical stiffness of the human mandibular system as a function of the frequency of stretching. 2. Jaw-muscle electromyographic (e.m.g.) responses to 'step' stretches were of 8 msec latency and generated a very substantial force response. Jaw-muscle e.m.g. responses having longer latency were not observed. 3. The mechanical stiffness of the human mandible was relatively constant as a function of the frequency of stretching, having a typical magnitude of about 15 N/mm (+/- 200 micrometers stretch) or 10 N/mm (+/- 1500 micrometers stretch) at mean biting forces of 10 N. The force resisting displacement was phase-advanced at all frequencies. 4. Modulation of jaw muscle electrical activity evoked by sinusoidal stretches increased in amplitude as a function of increasing stretch frequency. E.m.g. modulation was 60--100 degrees advanced at frequencies of 1--10 Hz, but the phase decreased at higher frequencies, becoming negative (lagging stretch) at frequencies of 30 Hz and above. These characteristics are consistent with the idea that the jaw stretch reflex is dependent on jaw muscle spindle afferent fibres exciting jaw-closing motoneurones by relatively direct (but not necessarily monosynaptic) connexions. 5. The relationship between jaw-muscle activity and voluntary fluctuations of isometric biting force suggests that human jaw muscles can be modelled as a second-order linear 'filter'. The corner frequency for human jaw muscle is about 3 Hz; thus it would appear to be considerably slower than jaw muscle of monkeys. 6. The reflex stiffness of the human mandible, estimated quantitatively on the assumption that human jaw muscle stiffness is similar to the intrinsic stiffness of the gastrocnemius of the cat, ranges between 5 and 9 N/mm at frequencies between 1 and 8 Hz. Since this reflex stiffness is about the same as muscle stiffness in this frequency range, we conclude that the stretch reflex of the human mandible contributes functionally to its postural stability. 7. Reflex stiffness appears to be greater in the monkey mandible relative to muscle stiffness than in the human mandible. The difference is argued to be a manifestation of the difference in jaw muscle contraction speed between the two species. 8. The fact that the mandibular stretch reflex appears to be stronger than the stretch reflex of the limbs of intact animals and humans is discussed in terms of the special anatomical and functional features of the mandible.

Effect of catechol on the discharge of muscle spindle afferents from the hind limb of the rat

Previous work had shown that some of the effects of catechol could be via the fusimotor system. In order to determine the extent of fusimotor involvement, recordings have been made from muscle spindle afferents in split dorsal root filaments of anaesthetised rats. Catechol failed to excite de-efferented muscle spindles therefore eliminating many possible non-fusimotor effects. Over 80% of spindle afferents with intact efferents showed increased discharge frequency 1 min after injection, this increase often following a biphasic pattern with a pronounced pre-myoclonic burst and decline followed by a more sustained period of activity during the myoclonic phase. Analysis of spontaneous twitches or twitches evoked by ipsilateral auditory stimulation showed in addition a phasic increase in discharge suggesting α-γ co-activation. Both primary and secondary afferents from many muscle groups around ankle and toes had their discharge frequencies elevated. Elevation of discharge frequencies of secondaries implies increased γ-s activity confirmed by a decrease in dynamic index. Both γ-d and γ-s involvement in catechol action on primaries is suggested by dynamic index measurements. Perhaps a more continuous form of testing such as sinusoidal stretches would reveal any rapid switching between activities in the two systems.

An analysis of receptor potential and tension of isolated cat muscle spindles in response to sinusoidal stretch.

In isolated cat muscle spindles the receptor potential responses of primary and secondary endings as well as tension responses to sinusoidal length changes in the steady state have been analysed. 1. At a given stimulus frequency, receptor potential per unit length change (receptor potential gain) in both primary and secondary endings is constant when displacement is less than about 10 micrometer. With larger stretches, receptor potential gain decreases approximately as a power function of displacement, the gain of primary endings decreasing more rapidly with increasing displacement than that of secondary endings. Tension per unit length change (tension gain) shows a similar constant range above which it also decreases as a power function of displacement. 2. In spite of the large reduction in gain at high displacement amplitudes, response wave forms remained essentially sinusoidal. The gain reduction results principally from a displacement-dependent non-linearity which has a rapid onset and slow decay. 3. Receptor potential and tension responses to small amplitude sinusoidal stretch depend, in a parallel manner, on the initial length of the preparation. 4. Both receptor potential and tension responses are highly dependent on frequency of sinusoidal stretch. In primary endings receptor potential gain increased as a power function of frequency over the range 0 . 01 to about 40 Hz, above which frequency the gain decreased; phase advance remained relatively constant up to 10 Hz then decreased to become a phase lag at higher frequency. In secondary endings receptor potential gain remained fairly constant between 0 . 01 and 1 Hz then rose as a power function of frequency but less steeply than in primary endings. 3. The possible mechanisms underlying these findings are discussed.

A method for the interpretation of signals in afferent nerves from muscle spindles.

A method is proposed for the interpretation of the signals in sensory nerve fibres. It is applicable to systems in which the transfer function between the input to a sense organ and the action potential firing frequency is known. In the present case, the chelonian muscle spindle is considered since its output to ramp-hold-release and sinusoidal stretches can be rather accurately simulated.

Static and dynamic fusimotor action on the response of spindle primary endings to sinusoidal stretches in the cat

In anaesthetized cats, the discharge of single spindle primary endings of flexor and extensor muscles was studied during sinusoidal stretches of the muscle (3–4 mm peak-to-peak, 1–5 Hz). In the response to each stretch, two main components could be distinguished; they were related respectively to the speed and to the amplitude of lengthening. By analogy with the usual definitions of the ‘dynamic index’ (d) and ‘position sensitivity’ (s) during ramp stretches of the same amplitude at a velocity of 35–60 mm/sec, a ‘sinusoidal dynamic index’ (D) and a ‘sinusoidal position sensitivity’ (S) were defined. Stimulation of dynamic fusimotor fibres increased D and d, S and s, while stimulation of static fusimotor fibres decreased D and d and had a less consistent effect on S and s. The total frequency variation during a sinusoidal stretch was always increased by a dynamic fusimotor action while it could be either decreased or increased by a static fusimotor action, depending on the stretch frequency and on the rate of fusimotor stimulation. These results were obtained in various muscles, without any significant difference related to their flexor or extensor function. It is concluded that in order to identify a fusimotor effect as static or dynamic during sinusoidal muscle stretch, it is necessary to use, as here, stretch parameters which elicit a dynamic component in the primary afferent response.

The influence of the gamma system on cross-correlated activity of Ia muscle spindles and its relation to information transmission

In the decerebrate cat during static or slowly varying (<0.3 Hz) muscle stretches, the activity of muscle spindle (MS) pairs was poorly correlated and such correlation was not changed after ventral root sectioning. With slightly faster sinusoidal stretches (0.3–20 Hz) activity in pairs of MS was also poorly correlated. However, sectioning of the ventral roots produced preferential firing frequencies in the same muscle spindle pairs and increased their degree of correlation. The increase in correlation between MS activity detected after suppressing the gamma bias appeared to arise from extrinsic 60 Hz power line vibration in the range of micrometers. Nevertheless, activation of the gamma system could suppress such phase locking. When frequencies above 20 Hz were used, the gamma system could not decorrelate the MS pair activity since the two units became locked to their common input signal. It is suggested that decorrelation of MS activity by gamma influence may improve the fidelity of the information transmitted by the Ia MS ensemble by filtering distortion harmonics, as well as damping tremor oscillations in the stretch reflex loop.

The strength of the reflex response to sinusoidal stretch of monkey jaw closing muscles during voluntary contraction.

1. Rhesus monkeys were trained to exert steady biting forces of 3--60 N for 1--2 sec. This behaviour was well maintained while sinusoidal or step opening and closing movements were imposed on the jaw. 2. The amplitude of the force modulation during sinusoidal stretching was divided by the amplitude of movement to obtain the magnitude of stiffness. This estimate was made at frequencies from 2 to 50 Hz at amplitudes of 100 and 500 micrometer (half the peak-to-peak movement at the incisors). 3. Peak magnitudes of stiffness were seen with frequencies of 8--15 Hz when the amplitude of movement was small; there was a great deal of variation between individual animals. This variation was most striking with mean forces of 25--35 N. The stiffness was greatest in animals that showed considerable spontaneous tremor, and the highest levels of stiffness were often recorded with frequencies near which tremor amplitude was large. A marked phase lag in the force response was often seen during small amplitude stretching at 8--30 Hz. 4. Estimates of stiffness for larger amplitude (500 micrometer) stretching showed less variation; the magnitude of stiffness showed maximum values below 10 Hz and a minimum at 15--30 Hz. Force always showed a phase lead on position although this lead became small in the frequency range where with smaller movement there had been phase lags. The magnitude of stiffness increased with increasing mean force. 5. Bilateral electrolytic lesions were made in the brain stems of three animals; they reduced by over 95% the expected number of cells in the mesencephalic nucleus of the fifth cranial nerve on either side. These lesions interrupted the afferent pathway for the stretch reflex and so abolished excitatory electromyogram (e.m.g.) responses to step stretches of the jaw closing muscles. 6. Such reflex responses as persisted after the lesions were small and inhibitory. E.m.g. silences followed both step stretch and release; the response to release was a 'load compensation' that could not be attributed to spindle afferents. 7. After the lesions the responses to movements of 100 micrometer showed neither negative values for the phase nor marked peaks in the stiffness magnitude at low frequencies; these features therefore take origin in the action of the stretch reflex. The stiffness that was measured after the lesions may be attributed to the non-reflex components resisting stretch, particularly to the properties of the contracting muscles. Thus, the phase of the force response was markedly advanced at all frequencies and the stiffness seen for 100 micrometer was similar to that for 500 micrometer. Stiffness increased with increasing mean force, as before surgery. 8. Vector subtraction of the stiffness seen at each frequency after interrupting the stretch reflex from that seen before doing so gave a quantitative estimate of the strength of the stretch reflex. The reflex activity calculated in this way showed attenuation and progressive phase lag as the frequency increased above 10 Hz...

Small-signal analysis of response of mammalian muscle spindles with fusimotor stimulation and a comparison with large-signal responses.

1. Response dynamics of primary and secondary muscle spindle endings to small-amplitude sinusoidal stretches were found to be unaltered by tonic repetitive stimulation of fusistatic or fusidynamic fibers. 2. Overall sensitivity of these receptors is decreased by fusistatic stimulation and either unchanged, increased, or decreased by fusidynamic stimulation at rates of 75/s or greater. 3. In the case of primary endings, the results obtained with small-amplitude sinusoidal stretches are not compatible with the response of these receptors to large-amplitude ramp stretches. The difference is explained by dependence of receptor dynamics on stretch amplitude. Fusistatic stimulation tends to prevent those changes in dynamics, whereas fusidynamic stimulation tends to enhance them. 4. In the case of secondary endings, the results obtained with small- and large-amplitude stretches appear to be compatible with a linear model for this receptor (i.e., one with dynamics independent of input parameters). 5. By modulating the frequency of stimulation applied to fusimotor fibers and comparing the resulting afferent response to the receptor response to stretch dynamic characteristics of intrafusal muscle contraction can be deduced. The results suggest that the dynamics of fusiastatic and fusidynamic contraction are the same and, furthermore, that they are the same as those of extrafusal muscle. We note that the result is incompatible with measurements of the time course of twitch and tetanus development and suggest, therefore, that muscle dynamics are a function of contractile state.

Responses of primary and secondary endings of isolated mammalian muscle spindles to sinusoidal length changes.

1. Responses of primary and secondary endings of isolated cat spindles to sinusoidal length changes have been recorded before and after block of impulse activity by tetrodotoxin. 2. Primary endings may discharge with each cycle of sinusoidal stretch at 25-50 Hz, with stretch amplitudes applied to the spindle poles as small as 1 micron. Thresholds are higher at lower frequencies. 3. In primary endings, amplitude of the receptor potential varies with frequency and magnitude of sinusoidal stretch. At a given stretch amplitude, the receptor-potential response increases markedly between 1 and 10 Hz. At a fixed frequency, for example, at Hz, the response to graded amplitude of sinusoidal stretch is highly nonlinear, sensitivity decreasing with large amplitudes. 4. Secondary endings show a much higher threshold than primary endings to sinusoidal stretch. Thus, at 25 Hz, secondary endings required stretch amplitudes of 50-100 micron to evoke discharge. Relatively large amplitudes of stretch were also required to evoked detectable receptor potentials. Over the range studied, the receptor potential varied more linearly with stretch amplitude in secondary than in primary endings.