Participants

Nine healthy and right-footed volunteers (5 males and 4 females) aged 30 ± 7 yr (mean ± STD) participated in this study. All subjects gave written informed consent and all procedures were approved by the Human Ethics Committee of the Institute of Biomedical Sciences at the University of São Paulo. The experiments were conducted in accordance with the Declaration of Helsinki.

Data acquisition and stimulation

The electromyograms (EMGs) were recorded using round-shaped surface electrodes (Ag-AgCl, 0.8 cm diameter, with an inter-electrode distance of 2 cm) applied in bipolar configuration over the SO and tibialis anterior (TA) muscles. The electrodes were positioned over the SO with the most proximal contact 4 cm beneath the inferior margin of the two heads of the gastrocnemii muscles. For TA, electrodes were positioned midway over the muscle belly. A ground electrode was placed over the tibia. The EMG signals were amplified and filtered (5 Hz to 2 kHz) by a MEB-2300K system (Nihon-Kohden, Japan) and then converted by an A/D board (National Instruments, Austin, TX.) at a 5 kHz sampling rate. Data were stored in hard disk for later off-line processing.

Direct motor responses (M-waves) and H-reflexes from the SO were obtained by electrical stimulation (rectangular pulses, 1 ms duration) applied to the posterior tibial nerve (PTN, test stimuli) delivered through surface electrodes (area = 2 cm²) positioned in the popliteal fossa [23]. For D1 and D2 studies, conditioning stimuli were applied to the common peroneal nerve (CPN) at the fibular head. Conditioning stimulus intensity was set at 1.1 times the motor threshold (1.1 x MT), with MT defined as the smallest stimulus intensity required to produce an observable M-wave of the TA.

The mechanical test stimuli (to elicit the T-reflex) were applied to the Achilles tendon of the right leg by means of a LW-126-13 vibration system (Labworks, USA), consisting of a power amplifier and a shaker (cylindrical body, with diameter 10.5 cm and length 13.5 cm). The tip of the shaker (round-shaped plastic tip, 1 cm diameter) was pressed against the Achilles tendon in order to keep a steady pressure and a fixed position on the tendon [22, 24–26] (Fig. 1). A LabView system (National Instruments, Austin, TX) was utilized to generate one cycle of a sinusoid with a period of 10 ms [22, 24] which was delivered to the input of the shaker's power amplifier in order to obtain the desired parameters for the mechanical stimulation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schematic representation of the experimental setup and temporal organization of the experiment.

Upper part: subject positioned during the experiment (left panel) showing the location of both the recording and the stimulation electrodes along with the shaker. In the right panel a detail of the stimulated leg and the points of both electrical (in gray and yellow) and mechanical (in magenta) stimuli delivery. The blue and green dots indicate the recording places for the SO and TA muscles, respectively. Lower part: schematic sweeps representing the EMG recordings and stimulus patterns whose colors are associated with the respective dots at the leg.

https://doi.org/10.1371/journal.pone.0121496.g001

General Procedures

Subjects were seated on a customized chair designed for stabilizing their right leg and feet. The hip, knee and ankle of both legs were maintained at ~90°. The shaker was on the floor in front of the chair (see the scheme in Fig. 1).

At the beginning of the experiments, five maximal motor waves (Mmax) were elicited at 0.5 Hz by delivering a supramaximal electrical stimulation to the PTN. Mmax and reflex amplitudes (i.e., T and H-reflexes) were estimated online from single unrectified EMG sweeps displayed on the screen of an oscilloscope (Hewlett Packard, Colorado Springs, CO). The intensities of the test stimuli used to evoke H and T-reflexes (electrical and mechanical stimuli, respectively) were adjusted so that the amplitude of the unconditioned response (i.e., control reflex) corresponded to ~20% of the amplitude of the Mmax (defined as the maximal peak-to-peak amplitude observed among the 5 Mmax).

It has been postulated that the test H-reflex in the 20–30% range of the maximal direct muscle response (Mmax) is more responsive to conditioning regardless of its nature [27]. We hypothesized that this holds true for T-reflexes as well. Thus, a device that allows the generation of T-reflexes with amplitudes approximately in the range given above was currently employed [8, 24–26] (see Data Acquisition and Stimulation).

To evaluate the effect of D1 inhibition, H-reflexes of the SO muscle were conditioned by either a single electrical pulse (1 ms duration) applied to the CPN (1.1 x MT) at C-T interval of 15 ms or a train of 7–10 pulses with the same duration/intensity delivered at 1 Hz, with the last pulse preceding the test stimulus by 15 ms. The number of pulses in the train varied pseudo-randomly within the range of 7 to 10 pulses in order to avoid, or at least minimize, a bias in the results due to a voluntary (or even an involuntary/subconscious) prediction of the time of the test pulse by the subject [21, 28]. For the same reason, the interval between test stimuli (applied to the PTN) pseudo-randomly ranged between 11 s and 13 s. The situations in which the reflexes were conditioned by a single pulse were termed “single”. The term “train” was employed when the reflexes were conditioned by a train of stimuli. The control H-reflex (without any conditioning), the H-reflex conditioned by a single pulse and the H-reflex conditioned by the train were evoked in a random fashion. Twenty reflex responses were obtained in each condition (i.e., train, single and control). The duration of the entire trial was ~13 min.

In another trial, a 100 ms C-T interval (instead of 15 ms) was used to evaluate the H-reflex under the effect of D2 inhibition. The same protocol described above was followed for T-reflexes, except that 3 to 6 ms were added to the C-T intervals (for both D1 and D2 trials) according to the subject’s height to compensate for the distance of the test stimulus applied to the triceps surae tendon [13].

The M-wave in the EMG of the TA muscle was constantly monitored to assure a steady intensity of the conditioning stimulation throughout the experiment. For some subjects, it was not possible to measure the M-wave amplitude of the TA EMG signals for the condition using C-T interval of 15 ms due to the interference from the PTN stimulus artifact (“crosstalk”). In those cases, the “crosstalked” stimulus artifact from the SO EMG in the control condition (with no conditioning and hence no M-wave in the TA EMG) was subtracted from the TA EMG in both PSI conditions (single and train) in order to reveal a clear and measurable M-wave. For these and all other cases, no differences in the amplitude of the TA M-wave were detected among conditions.

For each subject, the order of the trials (with T or H reflexes, D1 or D2 conditioning) was randomly presented and a resting period of ~5 min was allowed between trials. Each subject completed all trials in a single experimental session, which lasted ~3h.

Traces showing the temporal organization of the experiment are depicted in Fig. 1. The traces in blue color show schematic representations of both T- (upper traces) and H-reflexes (lower traces). TA M-wave is displayed in green and the mechanical and electrical test stimuli are shown in magenta and yellow, respectively. The bottom grey traces represent the conditioning stimuli (Fig. 1).

Signal processing and data analysis

In order to evaluate the effect of conditioning (that induces either D1 or D2 inhibition) on both T and H-reflexes, the amount of inhibition (AOI) was measured for “single” and “train” conditions [21]. Therefore, conditioned H- and T-reflexes were expressed as the AOI, which consisted of the difference between control and conditioned reflex response over its control value, as follows: in which the term refers to the mean control reflex amplitudes and is the mean reflex amplitudes conditioned by either a single or a train of electrical pulses to the CPN (n = 20 for each subject). Thus, the amount of inhibition for “single” (AOI_Singl) and “train” (AOI_Train) conditions were calculated for each subject. A three-way ANOVA with repeated measures [type of conditioning (“single” vs “train”), type of reflex (H- vs T-reflex) and type of inhibition (D1 vs D2)] was used to detect main effects and interactions among experimental conditions.

In order to evaluate the relative effect of the conditioning “train” with respect to the “single” conditioning, the ratio AOI_Train /AOI_ Singl (train to single ratio, TSR) was computed for both H- and T-reflexes and for both D1 and D2 inhibitions (a similar procedure has been adopted to compare healthy and hemiplegic patients by Roche et al. [21]). Here, we were interested in quantifying how much the reflex inhibition due to a conditioning train was higher than that due to a single stimulus conditioning.

The amplitude values of T-reflex in condition “single” obtained from two subjects were very similar to the respective control values. This resulted in T_AOI_Single values close to zero (very small effect of the single pulse) or even negative due to the intrinsic variability of the conditioned T-reflex itself. These negative and near-to-zero values found in two out of nine subjects strongly affected the ratio T_AOI_ Train/T_AOI_ Single, since values close to zero in the denominator yielded quite large ratio values. As negative values of this ratio are meaningless, data associated with those two subjects were removed from the analysis.

A two-way ANOVA with repeated measures was used to detect possible differences and interactions of TSR values between type of reflex (T vs H) and type of inhibition (D1 vs D2). The same statistical procedure (two-way ANOVA) was used to investigate possible differences in the amplitude of the control (unconditioned) H- and T- reflexes (expressed in %Mmax) measured during D1 and D2 trials.

All statistical analyses were performed using the statistical package SPSS 15.0 for Windows (SPSS, Inc., Chicago, IL), with significance level set at p < 0.05.

Possible mechanisms for the enhanced PSI effect

The results obtained from both H- and T-reflexes show that the conditioning train of pulses induced an increased inhibition when compared with a single conditioning pulse. This was expected for H-reflexes with C-T interval of ~20 ms (D1 inhibition) as previously reported [21]. However, for the other conditions we examined, the results are new and hence add to the present knowledge in the literature. A brief general suggestion on possible mechanisms behind the facilitation of PSI in healthy and post-stroke subjects was presented by Roche et al. [21]. The authors suggested an imbalance between excitatory and inhibitory synaptic actions along the PSI pathway. On the other hand, Lamy et al. [29] suggested that the increased PSI levels they obtained for trains of pulses in the conditioning pathway could be due to a mechanism similar to post-tetanic potentiation which would occur at rates such as 1 Hz. But the interpretation of our results (see below) will lead to a quite different suggestion because post-tetanic potentiation has not been reported at such low rates of stimulation.

It is tempting to suggest that a cumulative facilitation is responsible for the enhanced transmission along the inhibitory presynaptic pathway in response to a train of conditioning stimuli. Nevertheless, such a cumulative effect due to a spatio-temporal summation in interneurons interposed in this pathway is highly unlikely, as the conditioning stimuli were delivered at a very low rate (1 Hz). For the same reason, post-tetanic potentiation should be discarded as a possible synaptic mechanism [11, 30].

Another possibility is a long-loop effect from different types of conditioning stimuli that lead to increased cortical excitability [31, 32]. The descending influence from both corticospinal and reticulospinal fibers can suppress the activity of first or last-order interneurons within the PSI pathway, i.e., they decrease the PSI level, instead of increasing it as currently observed [33–35].

Post-activation depression (or homosynaptic depression—HD), which is the only known synaptic phenomenon with long duration (even longer than 10 s) [36], can be a possible mechanism responsible for the observed results. This depression is presumably caused by a decrease in synaptic transmission due to depletion or decreased probability of release of neurotransmitter [37–40].

Experiments conducted in both humans and cats have shown that HD level is not the same for the whole set of Ia terminals within the spinal cord [29, 41]. It is suggested that the phenomenon of HD is not ascribed solely to the pre-synaptic neuron (Ia afferent) but possibly to an interaction between both pre and post-synaptic cell [29, 42]. Lamy et al. [29] reported that the synapse between Ia terminals and inhibitory Ia interneuron (that mediates reciprocal inhibition at lumbar level) is susceptible to HD. Hence, one may speculate that the depression in synaptic contacts with inhibitory interneurons interposed in the inhibitory pathway can, eventually, lead to a facilitation of PSI.

The architecture of the PSI neuronal network has not yet been fully disclosed, but it is hypothesized that interneurons that mediate primary afferent depolarization (PAD) from muscle afferents are located within the intermediate zone of the lumbosacral region [43]. Inhibitory interneurons in this region that receive direct or indirect contacts from peripheral and/or descending inputs [44], arguably operate as modulators of the PSI pathway. Fig. 5 summarizes the proposed mechanism: the HD in synapses on inhibitory interneurons interposed within the PSI pathway (detailed in the circled inset at the bottom of the figure) might play an important role in the enhanced effect of PSI induced by repetitive conditioning stimuli at 1 Hz. Furthermore, low threshold cutaneous afferents eventually recruited by the conditioning stimulus activate inhibitory interneurons whose action down regulates the PSI level (detailed in the inset at the upper part of Fig. 5) [34, 35]. These afferents could also be subject to HD thereby increasing the PSI level (Fig. 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Scheme showing a putative mechanism for the enhancement of PSI due to a low frequency (1 Hz) train of conditioning stimuli.

The hypothesis is the occurrence of HD in the transmission to inhibitory interneurons within the neural circuit that generates PSI (bottom circled inset). Another possibility is the occurrence of HD at synapses from low threshold cutaneous afferents that project to inhibitory interneurons whose activation lead to a decrease in the PSI level (top circled inset). The final result from the action of both processes is an increased excitability of the last order (PAD) interneuron.

https://doi.org/10.1371/journal.pone.0121496.g005

Apart from the neuronal architecture and/or the afferent fibers being stimulated, it seems reasonable to consider a mechanism based on HD to explain the observed enhanced PSI level. An additional experimental protocol that directly assessed the contribution of conditioning stimulation frequency has been devised in the present study. The preliminary results obtained from 3 subjects clearly suggested an increment in the PSI level when a double pulse was applied at 1.5 Hz as compared to 0.5 Hz (Figure A in S1 File). As HD is dependent on the stimulus frequency [45], these preliminary results reinforce the hypothesis of HD being an important mechanism mediating the observed increases in PSI level.

The inhibitory effects observed here were qualitatively similar for both intervals between conditioning and test stimuli. One possibility is that they are part of the same inhibitory pathway, and a ‘facilitatory’ period between D1 and D2 could be ascribed to contamination from a cutaneous pathway [33]. However, the similar inhibitory effects can still be conveyed by different subsets of interneurons. Specific experiments in cats, for example, are necessary to unravel possible differences between both inhibitory processes in terms of neural circuitry.

Enhanced PSI effect on T-reflexes

Results from post stimulus time histogram in humans indicated that the later part of the excitatory post-synaptic potential (EPSP) is not affected by PSI when a Ia afferent fires at high frequency (due to muscle stretch), which might be the reason for the observed lower sensitivity of T-reflexes to PSI as compared to H-reflexes in C-T paradigms [13]. Enríquez-Denton et al. [14] argued that the accumulation of Ca²⁺ in SO Ia synaptic buttons due to high frequency firing contributes to more neurotransmitter release and to lessen the effect of PSI (compared to a case when there is only one Ia spike reaching the terminal, i.e. during the generation of an H-reflex). However, as our data showed that PSI was enhanced, we propose that conditioning stimulation with trains leads to disinhibition of the pathway mediating PAD, decreasing to a larger extent (as compared to the “single” condition) the influx of Ca²⁺ thereby reducing neurotransmitter release [14].

Supposedly, the ensued higher PAD excitability would be the result of firing in pre-PAD excitatory interneurons released from the control of the inhibitory interneurons (Fig. 5). The increased activity of the PAD interneurons would thoroughly reduce the amplitude of the EPSP. As a final result, the conditioning train would be effective in enhancing the effect of PSI with the consequent reduction in the amplitude of the T-reflex.

One interesting result presented here is that, although the “single” effect was weak for the T-reflex, the effect of the “train” condition was substantial. This was reflected by the train-to-single ratio (TSR) that was more prominent for T-reflexes as compared to H-reflexes. For instance, the values shown in Fig. 4A indicate that the D1 effect for T-reflexes was nearly 5 times higher under the “train” conditioning than under the “single” conditioning, while for H-reflexes the “train” conditioning caused a relative inhibition ~2 times higher as compared to “single”. This result might be explained by the number of afferents activated by the tendon percussion as compared to electrical nerve stimulation.

H and T-reflexes have a different proportion of Ia afferents participating in the generation of responses with similar sizes. This means that a lower number of Ia afferents will produce T-reflex amplitudes similar to those of an H-reflex generated by a comparatively larger number of Ias [11, 13, 46]. As the test stimulus intensity was carefully controlled and kept constant along the experiment for both H and T-reflexes, a proportionally stronger inhibitory effect could, in principle, be expected on a T-reflex as compared to H-reflex. This was found only when the inhibition was caused by a train of stimuli, putatively due to complex interactions between the inhibitory mechanisms and those related to the burst of Ia discharges associated with the tendon tap (as previously described).

Nevertheless, it is quite reasonable to assume that the mechanical and electrical test stimuli activate somewhat different sets of Ia afferents (with diverse firing rates), representing an additional factor that accounts for the presently observed differential effect on the PSI level.

We aknowledge that the techniques used here are not refined enough to provide a definitive explanation. Further experiments (preferably using animal preparations and/or computational modeling) are necessary to disclose the mechanisms responsible for the observed results.

Functional significance

The T-reflex seems to better represent the functioning of the stretch reflex pathway than the H-reflex, since the former also depends on the muscle-spindle responsiveness and it is associated with a few asynchronous Ia discharges in each axon instead of a single synchronized discharge of all the activated Ia axons [3]. During the performance of a smooth motor task, the excitability of the stretch reflex pathway is constantly being modulated by peripheral and descending signals [6, 47] so the action of the Ia afferents will change smoothly from one level to another. Nonetheless, for brisk movements or when a limb encounters an unexpected obstacle there may occur an Ia afferent firing burst equivalent to that of a tendon tap. The functional significance of the interaction between mechanically-evoked repetitive discharge of Ia afferents and PSI has been discussed [13]. It was proposed that PSI is modulated according to Ia afferent discharge, so that when a sudden muscle stretch takes place PSI is partially suppressed, allowing an appropriate reflex reaction [13].

The present results corroborate and extend previous studies, which reported that low frequency activation of afferents from the antagonist muscle is able to increase the level of PSI onto Ia terminals of the agonist [21, 29]. We propose that these volleys from antagonist Ia afferents (as induced by a low-frequency train of stimuli) might not occur during a motor behavior, but the HD in different synapses within the polysynaptic inhibitory pathway (or any other mechanism triggered by the low frequency stimulation that changes synaptic transmission across the interneurons within this path) will be changing dynamically and hence modulating the PSI level.

Therefore, a general scenario emerges as the reduction in PSI during the high frequency afferent discharge of the agonist (homonymous) muscle (see Morita et al. [13]) might be counteracted by a low frequency volley from the antagonist muscle. It is conceivable that this reciprocal mechanism may act on the stabilization of the ankle joint during a perturbation by promoting a balanced inhibitory action, which depends on the state of the antagonist muscle. One example of such condition is the alleged increased PSI of SO Ia terminals when the subject walks on a narrow base, which requires co-contraction of both ankle flexors and extensors [48]. Alternatively, given the slow time course of the train effect, this mechanism could also play a role during either upright postural maintenance or gait, by ensuring coordinated muscle activation within the limb.

The strength of this modulation depends on the firing behavior of the afferents from the antagonist muscle. It has been stated that the presence of two independent mechanisms with different time courses (namely HD and PSI) provides flexibility for the control of Ia afferent transmission [49]. We presently suggest that this concurrent control, which assists in the modulation of the stretch reflex excitability, is not restricted to the homonymous synapse but may also act at a premotoneuronal level.

One must bear in mind, however, that most of the mechanisms behind the observed effects cannot be directly assessed with the techniques currently employed in humans and either functional or physiological interpretations must be done with reservations. Even with these clear limitations, the present investigation opens the possibility of studying the effects of PSI on the stretch reflex pathway and possible interactions with the fusimotor system [8, 49, 50]. Further, it brings new elements both to re-evaluate a previously reported lack of sensitivity of the T-reflex to PSI and to discuss the functional role of the stretch reflex pathway during a given motor task.