Soleus H-reflex modulation during a double-legged drop landing task

Abstract

Muscle spindle afferent feedback is modulated during different phases of locomotor tasks in a way that facilitates task goals. However, only a few studies have studied H-reflex modulation during landing. This study aimed to characterize soleus (SOL) H-reflex modulation during the flight and early landing period of drop landings. Because landing presumably involves a massive increase in spindle afferent firing due to rapid SOL muscle stretching, we hypothesized H-reflex size would decrease near landing reflecting neural modulation to prevent excessive motoneuron excitation. The soleus H-reflex was recorded during drop landings from a 30 cm height in 9 healthy adults. Electromyography (SOL, tibialis anterior (TA), medial gastrocnemius, and vastus lateralis), ankle and knee joint motion and ground reaction force were recorded during landings. Tibial nerve stimulation was timed to elicit H-reflexes during the flight and early ground contact period (five 30 ms Bins from 90 ms before to 60 ms after landing). The H-reflexes recorded after landing (0–30 and 30–60 ms) were significantly smaller (21–36% less) than that recorded during the flight periods (90–0 ms before ground contact; P≤ 0.004). The decrease in H-reflex size not occurring until after ground contact indicates a time-critical modulation of reflex gain during the last 30 ms of flight (i.e. time of tibial nerve stimulation). H-reflex size reduction after ground contact supports a probable neural strategy to prevent excessive reflex mediated muscle activation and thereby facilitates appropriate musculotendon and joint stiffness.

INTRODUCTION

Muscle spindle afferents provide excitatory feedback to a muscle’s motoneuron pool in response to muscle stretch. The excitatory stretch sensitive afferent feedback, mainly from primary Ia muscle spindle afferents, functions to increase muscle stiffness and by association joint stiffness (Nichols and Houk 1976; Abelew et al. 2000; Mirbagheri et al. 2001). In addition to responding to overt muscle length perturbations (e.g. tendon tap, unexpected joint rotation), evidence indicates muscle spindle afferent feedback influences motor activity on an ongoing basis during postural and locomotor tasks (Yang et al. 1991; Abelew et al. 2000; Mazzaro et al. 2005; Mayer et al. 2018). Indeed, cats lacking stretch sensitive excitatory Ia feedback from the triceps surae muscles exhibit significant ankle yielding (i.e. limp) during downslope walking (Abelew et al. 2000; Maas et al. 2007). In humans walking on a level surface, triceps surae activity increases when they unknowingly step onto a slight incline that imposes additional ankle dorsiflexion, but this response is absent in persons lacking muscle spindle feedback (af Klint et al. 2008). These observations indicate that muscle spindle afferent feedback contributes to muscle activation to meet task demands.

The functional role of Ia feedback has been studied during several different locomotor tasks in humans, most commonly using the soleus Hoffmann reflex or H-reflex. Despite some limitations such as bypassing the muscle spindle receptor, the H-reflex, often viewed as an electrical analog of the spinal stretch reflex, is a convenient laboratory tool that evaluates Ia excitation of homonymous motoneurons (Zehr 2002). Soleus H-reflex size, reflecting the excitability of the H-reflex pathway, is not fixed but rather changes as a function of posture and task. For example, the H-reflex size progressively decreases from supine to sitting to standing (Hayashi et al. 1992; Angulo-Kinzler et al. 1998; Chalmers and Knutzen 2002). In addition, the soleus H-reflex size is modulated during different phases of cyclic motion such as walking (Capaday and Stein 1986; Stein and Capaday 1988; Yang and Whelan 1993; Brooke et al. 1995). The size of the H-reflex is typically very small or absent during the swing phase of walking when soleus activation could result in toe drag and tripping, whereas the H-reflex typically increases during the stance phase when the excitatory stretch sensitive feedback can contribute to soleus activation and ankle propulsion (Stein and Capaday 1988; Yang et al. 1991; Mazzaro et al. 2005). The general idea is that the excitability of a reflex pathway appears to be task and phase dependently modulated such that it functions appropriately to contribute to task execution (Zehr and Stein 1999; Thompson and Wolpaw 2015). Characterizing the phase-dependent reflex modulation during a specific task can, therefore, help to determine if and how the reflex normally functions during that task and how impaired modulation may manifest as motor impairments (Zehr and Stein 1999). Such knowledge has motivated targeted neurorehabilitation training regimens such as the use of operant conditioning to correct altered reflex behavior contributing to motor impairments (Manella et al. 2013; Thompson et al. 2013; Thompson and Wolpaw 2021).

During dynamic tasks such as hopping and landing from a jump, modulation of muscle spindle afferent feedback would seem to be critical. These tasks involve rapid changes in angular velocity at the ankle and knee joints after ground contact, which could result in a high proportion of muscle spindle afferents rapidly firing (Prochazka et al. 1977b). Indeed, prior studies strongly suggest that spindle afferent feedback contributes to triceps surae muscle activation during hopping (Zuur et al. 2010) and landing (Duncan and McDonagh 2000). However, excitatory reflexes that are too strong (i.e. inappropriately modulated) could result in excessive reflex-mediated muscle contraction and resulting stiffness that could damage the musculotendon complex (Leukel et al. 2008a). Several studies have reported modulation of H-reflex size during landing tasks at the time of the stretch-mediated reflex (i.e. approximately 50 ms after ground contact) indicating support for neural modulation to prevent excessive reflex activation (Dyhre-Poulsen et al. 1991; Leukel et al. 2008a; Leukel et al. 2008b). However, the phase-dependent modulation across the flight and transition to the landing period has only been descriptively reported (Dyhre-Poulsen et al. 1991). Thus, the timing and extent of H-reflex modulation during the different phases of landing remains unclear.

The purpose of this study was to characterize soleus H-reflex amplitude modulation during a double-legged drop landing task. Specifically, we aimed to determine whether a systematic reflex modulation could be identified as the task transitions from flight to landing. Based on a descriptive study of H-reflex modulation during landing (Dyhre-Poulsen et al. 1991), we anticipated that the H-reflex size will decrease starting during the flight phase nearing ground contact, and further decrease during the early landing period (Leukel et al. 2008a), in preparation for the massive increase in muscle spindle firing due to rapid ankle dorsiflexion. Through characterization of soleus H-reflex modulation during landing, this study aims to advance understanding of the neural control of landing.

METHODS

Participants

Nine healthy participants (3 female; age 19–45 yrs (mean 32); weight 52–100 kg (mean 73)) with no neurological or musculoskeletal injuries that would impair their ability to complete jumping and landing tasks participated in this study. Prior to participation, study details were described, and participants provided written informed consent as approved by the Medical University of South Carolina Institutional Review Board.

General Procedures

Participants attended a single experimental session. After cleaning the skin and placing EMG electrodes over the soleus (SOL), tibialis anterior (TA), medial gastrocnemius (MG), and vastus lateralis (VL) (see EMG recording and electrical stimulation), a SOL H-reflex – M-wave recruitment curve was obtained by stimulating the tibial nerve while the participant maintained upright standing posture. When the participant had maintained the SOL EMG activity at their natural standing level (typically about 20 μV) for 2 s and at least 4 s had passed since the last stimulus, a 1-ms square pulse was delivered. Stimulation intensity was increased from below H-reflex threshold to just above the level that was required to elicit the maximum M-wave (Mmax) in steps of 1.25–2.5 mA so as to determine the maximum H-reflex (Hmax) and Mmax. The recruitment curve data were used to determine stimulation intensity during landing trials.

Then, the participant was asked to perform between 105–150 drop landings (Figure 1A). The landing task involved dropping down from a 30 cm box leading with the test limb, which was self-selected by each participant as the most comfortable leading leg. The participant was instructed to perform the landing maneuver as consistently as possible with both feet contacting the ground at the same time on separate force plates embedded in a stationary split-belt treadmill (Bertec Corporation, Columbus, OH). Through a couple of practice landing trials, the participant was familiarized with the task and instrumentation. After familiarization, five to ten landing trials were recorded without tibial nerve stimulation to identify the landing flight time. Flight time was defined as the time from the non-test leg foot leaving the box as determined by a force sensitive resistor (Model 406 force sensing resistor, Interlink Electronics, Westlake Village, CA) to ground contact by the test leg detected by the force plate (Figure 1B). Ground reaction forces were recorded from the force plates and sagittal plane ankle and knee joint angles were recorded with electrogoniometers (Biometrics Ltd., Ladysmith, VA). The electrogoniometers were placed on the lateral surface of the leg or foot crossing the joint rotation axes. For the knee, the two bases of the goniometer were aligned over the lateral femur and lower leg crossing the lateral femoral epicondyle. For the ankle, the two bases were aligned along the fibula and fifth metatarsal crossing the lateral malleolus.

Figure 1.

A) Participant completing the drop landing task with lower case letters (a-d) identifying the different phases of the landing task. B) A typical example of raw data collected during an unstimulated trial. During stimulation trials (not shown here), the H-reflex size was examined during flight and the early landing periods (b-d).

To elicit the H-reflexes during and after landing, tibial nerve stimulation was triggered by the non-test leg’s step-off signal (detected by a force sensitive resister, see Fig. 1A) with various delays, aiming to deliver the stimuli so that the H-reflex response occurred from 90 ms before to 60 ms after ground contact. For each participant, the H-reflex latency and flight time estimated during familiarization trials were used to adjust the timing of tibial nerve stimulation. For example, if flight time was estimated to be 200 ms in duration, a stimulus trigger delay of 150 ms would result in delivering tibial nerve stimulation 50 ms before ground contact with the H-reflex occurring just before ground contact (i.e. typically 20–10 ms before contact). Generally, a block of 15–20 landing trials with H-reflex stimulations were followed by 2–5 landing trials without stimulation. Rest was granted as often as needed.

Throughout the landing H-reflex data collection, the soleus M-wave amplitude was monitored and the stimulus current was adjusted as needed to obtain H-reflexes that were accompanied by M-waves just above threshold (Capaday and Stein 1986; Makihara et al. 2012; Thompson and Wolpaw 2021). In addition, in 6 of 9 participants, several tibial nerve stimuli at a supramaximal intensity elicited the Mmax during the landing periods of interest (i.e., from 120 ms before to 30 ms after ground contact). This allowed normalization of landing H-reflexes to landing Mmax values of the corresponding landing phases in those individuals, in addition to normalizing landing H-reflexes to the standing Mmax values.

EMG recording and electrical stimulation

After the skin was cleaned with isopropyl alcohol wipes, self-adhesive Ag-AgCl electrodes (2.2 × 3.5 cm; Vermed, Nissha Medical Technologies, Buffalo, NY) with center-to-center 3 cm apart were placed over the SOL, TA, MG, and VL. For the soleus, a pair of electrodes were placed longitudinally just below the gastrocnemii. For MG and TA, the electrodes were placed over the center of the muscle bellies. For the VL, the electrodes were placed on the distal third of the muscle in the direction of muscle fibers. The reference electrode was placed over the patella or proximal medial tibia. Surface EMG signals were differentially amplified and band-pass filtered at 10–1000 Hz (AMT-8, Bortec Biomedical, Calgary, AB, Canada), and digitized and stored at 4 kHz (Axon Digidata 1440A, Molecular Devices, San Jose, CA). The electrodes and EMG pre-amplifiers were secured to the skin with tape and a tubular bandage to minimize movement artifacts.

To elicit H-reflexes and M-waves, the tibial nerve was stimulated in the popliteal fossa with cathode (2.2 × 2.2 cm; Vermed) and proximal anode (2.2 × 3.5 cm, Vermed) oriented along the course of the tibial nerve. Stimulation location was chosen to minimize the SOL H-reflex threshold, maximize the SOL Mmax size, and minimize the TA excitation. During standing or during landing, single 1 ms square pulses were delivered to the tibial nerve using a Grass S48 stimulator with a CCU1 constant current unit and SIU5 stimulation isolation unit (Natus Neurology; Grass Instruments, West Warwick, RI, USA).

Data Analysis

Several kinetic and kinematic variables were calculated during landing. The variables included sagittal plane ankle and knee joint angles at initial contact with the ground, the change in ankle and knee joint angle during the last 30 ms before ground contact, the peak ankle and knee joint angular velocities after landing, and the peak ground reaction force. These variables were included to descriptively characterize landing behavior and aid in interpretation of H-reflex modulation. For example, ankle joint angle can influence H-reflex size (Frigon et al. 2007; Dutt-Mazumder et al. 2020) and peak ankle angular velocity after landing provides context for why modulation of muscle spindle feedback may be requisite. Each of the landing variables were averaged for each participant across landing trials without stimulation.

In order to characterize EMG activity during landing, SOL, MG, TA, and VL signals from non-stimulated trials were rectified, aligned to the moment of ground contact, and averaged for each participant. Then, the mean EMG value was calculated for each of the five 30 ms analysis bins: Bin 1: −90 to −60 ms; Bin 2: −60 to −30 ms; Bin 3: −30 to 0 ms; Bin 4: 0 to 30 ms; Bin 5: 30 to 60 ms post ground contact.

H-reflex and M-wave sizes were calculated as the peak-to-peak amplitude during their respective time windows after stimulation (typically, 6–23 ms for the M-wave and 30–48 ms for H-reflex). The H-reflex size during landing was expressed relative to the standing Mmax. To evaluate the extent of H-reflex size modulation during landing, only the landing H-reflexes with comparable M-wave sizes (Voigt et al. 1998; Leukel et al. 2008b) were included in the analysis. In the present study, the landing H-reflexes that were accompanied by M-waves of 3–25% standing Mmax were averaged within five 30 ms timing bins (i.e., Bin 1: −90 to −60 ms; Bin 2: −60 to −30 ms; Bin 3: −30 to 0 ms; Bin 4: 0 to 30 ms; Bin 5: 30 to 60 ms post ground contact) for each participant. The bin-averaged H-reflex sizes were normalized to the Mmax during standing (Capaday and Stein 1986; Makihara et al. 2012; Thompson and Wolpaw 2021). In 6 of 9 participants in whom landing Mmax values were obtained during each of the 5 timing bins, the corresponding H-reflex sizes were also expressed as % landing Mmax.

Statistical Analyses

All statistical analyses were conducted with SPSS software (v.27, IBM, Armonk, NY) using a significance level of P ≤ 0.05. Normality was tested with the Shapiro-Wilk’s test. Descriptive statistics were used to characterize standing Hmax amplitude as well as kinematic and kinetic landing variables for non-stimulated landing trials. Repeated measures Analysis of Variance (ANOVA) tests were used to determine whether SOL, TA, MG and VL EMG amplitudes were different across the 5 time bins spanning flight and early landing periods. For repeated measures analyses, homogeneity of variances were tested with Mauchly’s test of sphericity and if sphericity could not be assumed, the Greenhouse-Geisser correction was used.

To determine the effect of stimulus timing on H-reflex size during landing, a linear mixed model was used. Because SOL background EMG amplitude and TA EMG (e.g. due to reciprocal inhibition) could influence H-reflex size (Zehr 2002; Knikou 2008) and modulatory pathways from VL and MG activity could also influence SOL H-reflex (Pierrot-Deseilligny et al. 1979; Meunier et al. 1990), mean SOL, TA, VL and MG EMG within each time bin from the non-stimulated landing trials were included as time-varying covariates in the model. The linear mixed model included stimulation timing (Bins 1 to 5) and the EMG covariates (Bins 1 to 5) as fixed factors and used a first-order autoregressive covariance structure. A Q-Q plot indicated the normality assumption was met. Bonferroni adjustment for multiple pairwise comparisons was used for all post hoc analyses. The same linear mixed model approach was used to evaluate the H-reflex size normalized to the landing Mmax values (n=6) across time bins.

RESULTS

All data are reported as mean ± SD. The participants standing Hmax was 49 ± 19% Mmax (Hmax: 6.0 ± 3.0 mV; Mmax: 12.3 ± 5.2 mV; lower and upper 95% CI: 35 and 65%). Participants completed 122 ± 16 drop landings (range: 104–155), of which 22 landings were completed without stimulation, on average. On average, 8–16 H-reflex trials were averaged across participants for each of the 5 landing bins and the M-wave amplitudes were similar across the 5 bins (5.5 ± 2, 6.5 ± 2, 8 ± 2, 10 ± 3, 9 ± 4 % Mmax, respectively).

Joint motion and EMG during drop-landing

Participants contacted the ground with the ankle plantarflexed and knee flexed with negligible change in joint angle during the last 30 ms before ground contact (Table 1, Figure 2). Upon contact with the ground, the ankle and knee rapidly flexed (peak angular velocity range: 501–835°∙s⁻¹ and 370–500°∙s⁻¹, Table 1). The peak ground reaction forces were close to 2 times body weight (Table 1).

Table 1.

Summary of landing kinematics and kinetics obtained from non-stimulated landing trials

		95% CI
	Mean ± SD	Lower	Upper
Ankle Angle at foot contact (°)	30.4 ± 7.1	24.9	35.8
Ankle Angle Change −30 ms to GC (°)	1.9 ± 1.3	0.84	2.9
Knee Angle at foot contact (°)	20.2 ± 7.5	14.5	25.9
Knee Angle change −30 ms to GC (°)	7.7 ± 2.7	5.65	9.8
Peak ankle Angular velocity (°·s−1)	708.8 ± 124.2	613.3	804.2
Peak knee angular velocity (°·s−1)	429.1 ± 47.7	392.4	465.8
Peak ground reaction force (× body weight)	1.9 ± 0.63	1.5	2.4
Peak ground reaction force time (ms)	63.6 ± 18.6	49.3	77.9

GC, ground contact

Figure 2.

Ensemble average (mean ± SD) of participants’ ground reaction force (GRF), ankle, and knee joint angle profiles from 90 ms before to 60 ms after ground contact during the landing task. The vertical dashed line represents the time of ground contact.

EMG during the landing task is shown in Figure 3. The SOL, TA, and VL EMG gradually increased during the flight phase and early landing period (F_{1.8, 14} = 4.23, P = 0.039; F_{1.6, 13} = 12.3, P = 0.001; F_{4, 32} = 49.7, P < 0.001). Pairwise comparisons for SOL revealed that Bin 3 was significantly greater than Bin 1 (P = 0.05). Pairwise comparisons for TA revealed that Bin 4 was greater than Bins 1–2 (P < 0.024), Bin 3 was greater than Bin 1 (P < 0.021), and Bin 2 was greater than Bins 1 (P < 0.03). Pairwise comparisons for VL revealed Bin 5 was significantly greater than Bins 1–4 (P < 0.04), Bin 4 was significantly greater than Bins 1–3 (P < 0.017), Bin 3 was significantly greater than Bins 1–2 (P < 0.04), and Bin 2 was significantly greater than Bin 1 (P < 0.03). The MG EMG significantly decreased during the early landing period (F_{1.7, 13} = 13.7, P = 0.01). The pairwise comparisons revealed that Bins 1 and 2 were significantly greater than Bins 4 and 5 (P < 0.046) and Bin 3 was significantly greater than Bin 4 (P = 0.035).

Figure 3.

EMG activity during landing without tibial nerve stimulation (Mean ± SD). These data reflect average rectified EMG within each of the bin times across participants (shown as numbers). Contact with the ground is at time 0, thus Bins 1 through 3 reflect flight and Bins 4 and 5 the period just after landing. Please refer to text for description of statistical evaluation of modulation of EMG across bins.

H-reflex modulation during landing

The H-reflex size decreased from the flight (Bin 1: 59 ± 15%, 6.9 ± 2.7 mV) to early landing period (Bin 4: 32 ± 19%, 3.5 ± 2.2 mV) and further decreased during Bin 5 (18 ± 13%, 1.9 ± 1.1 mV) (Figure 4). There was a statistically significant main effect of stimulus timing on the H-reflex size (F_{4, 36} = 13.5, P<0.001) with a significant influence from background SOL (F_{1, 34.5} = 10.1, P=0.003) and nonsignificant influence from TA, VL or MG amplitudes (all P>0.26). Pairwise comparisons with Bonferroni correction revealed that the H-reflex sizes during the landing periods (Bins 4 and 5) were significantly lower than the flight periods (Bins 1, 2 and 3; mean differences range: −19 to −41%, all P ≤ 0.008). In addition, the H-reflex size recorded during landing Bin 5 was significantly lower than Bin 4 landing period (Bin 4; mean difference: −18%, P < 0.001).

Figure 4.

A) Soleus EMG sweeps from a representative individual (participant #2) illustrating reduction in H-reflex size during late flight and early landing periods (Bins 3, 4, 5). B) Modulation of H-reflex size expressed as % standing Mmax and C) expressed as % landing Mmax across participants (shown by numbers) when the tibial nerve was stimulated during the respective 30 ms time bins (Bin 1: −90 to −60 ms, Bin 5: 30–60 ms; time 0 is ground contact). Please refer to text for description of statistical evaluation of modulation of H-reflex across bins.

In the 6 of 9 participants in whom the Mmax was measured during landing, the same pattern of H-reflex modulation was observed with slight reductions in the relative H-reflex size due to slightly larger Mmax values recorded during landing. The H-reflex size decreased from the flight phase (Bin 1: 51 ± 11%) to early landing (Bin 4: 31 ± 12%) and further decreased during the landing period (Bin 5: 19 ± 10%) (Figure 4C). There was a statistically significant main effect of stimulus timing on the H-reflex size (F_{4, 23.8} = 9.9, P<0.001) with a nonsignificant influence from background SOL, TA, VL and MG EMG (all P>0.05). Pairwise comparisons with Bonferroni correction revealed that the H-reflex sizes recorded during landing period Bin 4 was significant smaller than flight period Bin 3 (mean difference: −12%, P=0.04) and landing Bin 5 was significantly smaller than flight Bins 1–3 (mean differences: −26 to −32%, all P<0.02) and landing Bin 4 (mean difference: −14%, P<0.001).

DISCUSSION

The purpose of this study was to characterize soleus H-reflex amplitude modulation during the flight and early landing period of a double-legged drop landing task. We found that the H-reflex size was significantly decreased during the early landing period (Bins 4–5), compared to the flight phase (Bins 1–3). This supports the hypothesis that the H-reflex amplitude is phase-dependently modulated during drop landing. Below, we discuss how the reflex modulation during landing differs from walking and hopping, how our results support a central origin of H-reflex modulation during landing rather than a peripheral origin (i.e., arising from movement related factors) and the potential functional implications of H-reflex modulation during landing.

Muscle activation during landing

During landing, lower limb muscles are activated before and after ground contact presumably to achieve musculotendon stiffnesses that is adequate for absorbing impact and controlling center of mass deceleration without damaging muscles or joints. In this study, the activity of ankle and knee joint muscles increased during the flight phase (Fig. 1 & 3) similar to prior studies (e.g. Santello 2005). This prelanding muscle activation attributed to preprogrammed supraspinal control (Santello 2005; Taube et al. 2012) provides the initial braking forces during landing and influences musculotendon mechanics (i.e. enable energy storage in tendon, influence relative muscle fascicle strain and lengthening velocities) (Sousa et al. 2007; Konow et al. 2012; Sawicki et al. 2015; Gordon et al. 2020).

Muscle activity after landing can be generated supra-spinally in a preprogrammed manner and in response to sensory afferent firing caused by ground contact (Santello 2005; Taube et al. 2012). The contribution of sensory afferent feedback to muscle activation during the early landing period has been well documented (Prochazka et al. 1977a; Duncan and McDonagh 2000; Zuur et al. 2010). Excitation of muscle spindle afferents Ia and II (Prochazka et al. 1977b) from rapid stretching of the triceps surae likely contributes to triceps surae activation and thereby control of ankle joint stiffness (Prochazka et al. 1977b; Duncan and McDonagh 2000; Taube et al. 2008; Zuur et al. 2010). In addition to excitation of spindle afferents, we acknowledge that the rapidly increasing ground reaction force (i.e., reaching its peak value around 60 ms) and muscle forces would excite load-sensing afferents such as Ib afferents from Golgi tendon organs and cutaneous afferents (Duysens et al. 2000; Nichols 2018). While the functional role of load-sensing afferents during landing is yet to be determined, the load-sensitive afferent firing could independently influence triceps surae activation or influence H- or stretch reflex mediated triceps activation during landing Bins 4 and 5 (Prochazka et al. 1977a; Lyle and Nichols 2018).

Phase specific modulation of the H-reflex during landing

In this study, we evaluated soleus H-reflex size during the flight and early landing period as a proxy for modulation of muscle spindle afferent reflex gain. A key finding was a consistent reduction in H-reflex size across participants when the tibial nerve stimuli were delivered starting around 30 ms before ground contact, such that the H-reflex occurred just after foot contact and prior to when dorsiflexion induced spinal stretch reflexes could occur. This implies that a finely timed modulation of the reflex gain occurs in anticipation of foot contact with the ground, and is similar in principal to the consistent onset timings of lower limb muscles coupled to contact (Santello 2005). The smaller H-reflexes observed starting after ground contact are in general agreement with the observation by Dyhre-Poulsen et al. (1991) and Leukel et al. (2008a) who observed smaller H-reflexes during the time corresponding with spinal stretch reflex mediated excitation compared to the time of ground contact.

The modulation of H-reflex size observed during landing, a discrete task, contrasts with the modulation pattern from rhythmic tasks such as walking and hopping. In the present study, the H-reflexes were significantly greater during the flight phase (Bins 1–3) than during the landing period (Bins 4–5). The latter corresponds with H-reflexes occurring just after ground contact (Bin 4, 0–30 ms) and when stretch reflexes from rapid dorsiflexion are typically observed (Bin 5, 30–60 ms). The pattern of phase modulation during landing observed in the present study is generally opposite to that observed during hopping and walking in which the H-reflex size is smallest during the flight or swing phase and greatest during the ground contact period (Capaday and Stein 1986; Dyhre-Poulsen et al. 1991; Voigt et al. 1998; Thompson and Wolpaw 2021). Note that phase-dependent modulation during hopping has only been studied in 9 participants across two studies (Dyhre-Poulsen et al. 1991, n=3; Voigt et al. 1998, n=6) with high variability in H-reflex modulation across participants (Voigt et al. 1998), and thus, the modulation pattern is not fully established. The different modulation patterns observed during rhythmic task such as hopping and walking could be due, at least in part, to the engagement of central pattern generator (CPG) circuits; presumably, no CPG would be involved during the discrete task of landing. Nonetheless, the different phase-dependent modulation highlights the importance of characterizing the normal pattern of modulation during typical anticipated conditions in order to understand the potential role of muscle spindle afferent feedback specific to that task.

Findings from this study suggest that it is unlikely that H-reflex modulation during landing is significantly influenced by movement related afferent firing (e.g. changes in joint postures) or ongoing muscle activity. First, it is known that the soleus H-reflex is larger when the ankle is in a plantarflexed position (e.g., 15°) when compared to a neutral or dorsiflexed position during standing (Dutt-Mazumder et al. 2020) or sitting (Frigon et al. 2007). In the present study, the ankle was in a plantarflexed position with <3° change in angle during the 30 ms period prior to ground contact (see Figs 1 and 2). Importantly, tibial nerve stimulation during this exact timing period resulted in smaller H-reflexes immediately after ground contact (i.e., Bin 4). This suggests that ankle posture and afferent activity associated with ankle joint motion at the time of stimulation was not responsible for the H-reflex decrease observed during this early phase of landing. The imposed dorsiflexion due to ground contact could contribute to the decreased H-reflex size based solely on ankle posture (Frigon et al. 2007; Dutt-Mazumder et al. 2020); however, the excitatory muscle spindle feedback resulting from the rapid ankle dorsiflexion would be expected to increase the size of the H-reflex during Bin 5 (i.e. 30+ ms post ground contact) though the H-reflex was the smallest at this time. Second, H-reflex size is often affected by ongoing soleus and TA activity. H-reflex size generally increases with increasing soleus background EMG up to moderate levels (Capaday and Stein 1986; Frigon et al. 2007; Makihara et al. 2012), whereas the soleus H-reflex decreases when the antagonist TA is active (Crone et al. 1987). Heteronymous recurrent and Ib inhibitory pathways, indirectly reflected by VL and MG EMG activity in this study, could have also contributed to decreased H-reflex size (Pierrot-Deseilligny et al. 1979; Meunier et al. 1990). Thus, dynamic changes in EMG activity that occurred during landing (e.g., Fig. 2) could potentially influence the H-reflex modulation observed. However, the statistical analyses that included soleus, TA, VL and MG activity as covariates indicated that the ongoing EMG activity did not significantly influence H-reflex size. Nonetheless, we cannot completely rule out the possibility that the increased TA activity during landing contributed to the reduced soleus motoneuron excitability through reciprocal inhibition. However, we believe this is unlikely since soleus activity was also increasing in the late flight period and tonic activation of soleus (Crone et al. 1987) and soleus-TA co-contraction (Nielsen and Kagamihara 1992) has been shown to reduce reciprocal inhibition of soleus from TA. Thus, the soleus H-reflex reduction during the late flight to early landing is unlikely caused by the soleus and TA EMG activity. Instead, we speculate that (1) an increase in presynaptic inhibition, regulated by descending supraspinal mechanisms, beginning around 30 ms prior to foot contact resulted in reduced H-reflex size during the landing period, and (2) during the early landing period, in addition to presynaptic inhibition, sensory feedback from load receptors (Duysens et al. 2000; Nichols 2018) and the dorsiflexed ankle joint posture (Frigon et al. 2007; Dutt-Mazumder et al. 2020) might have acted to further reduce H-reflex size.

Functional importance of reflex gain reduction during early landing

Identifying the typical pattern of H-reflex modulation during tasks provides critical context for interpreting the functional role of muscle spindle feedback. For example, identifying altered soleus H-reflex modulation during walking in persons after spinal cord injury has led to operant conditioning interventions to correct the altered modulation and improve walking (Thompson et al. 2013; Thompson and Wolpaw 2021).

A prevailing belief is that muscle activation before landing and triceps surae reflex responses after ground contact are both scaled so as to achieve a level of musculotendon stiffness appropriate for task demands (Dyhre-Poulsen et al. 1991; Santello 2005; Leukel et al. 2008a; Taube et al. 2008; Taube et al. 2012; Gordon et al. 2020). Modulation of reflex size occurring just after foot contact likely reflects a pre-programmed control for dealing with the impending rapid lengthening of the soleus (Sousa et al. 2007; Konow et al. 2012) and resulting large muscle spindle firing (Prochazka et al. 1977b). If the gain of the reflex remains too high, the robust firing of Ia afferents could result in excessive muscle contraction and increased musculotendon stiffness. Thus, a reasonable neural strategy to avoid overload injury is to reduce the reflex gain during the early landing. Consistent with this expectation, the present results and several prior studies indicate a reduced H-reflex size corresponding with the timing of the stretch mediated reflex response, i.e. Bin 5 in this study (Dyhre-Poulsen et al. 1991; Leukel et al. 2008a; Leukel et al. 2008b). Leukel et al. (2008a) also discovered that the H-reflex size at the time of the stretch evoked reflex progressively decreased when participants landed from increasing heights (i.e., loading intensity effect: 30, 50 and 70 cm); the authors found H-reflex size was not modulated across heights when timed to occur at ground contact. The present study has further identified that the reflex modulation begins just after ground contact (0–30 ms after contact, Bin 4) and thus modulation of the reflex at the spinal cord begins within 30 ms of landing. Taken together, the reduction in H-reflex size after ground contact during drop landing supports a purposeful reduction in the relative stretch reflex gain so as to achieve an adequate level of musculotendon stiffness.

Alterations to this time-critical phase-dependent reflex modulation may increase injury risk. For example, reduced joint flexions and higher ground reaction forces during landing, in which sports injuries often occur, have frequently been found in female athletes and thought to increase risk of injury (Schmitz et al. 2007; Lyle et al. 2014). Several studies found that female athletes absorb energy using primarily the ankle and knee during landing whereas male athletes absorbed more relative energy at the knee and hip (Decker et al. 2003; Schmitz et al. 2007). The greater ankle joint stiffness in female athletes raises the possibility that the female athletes, in addition to increased ankle muscle coactivation before landing (Lyle et al. 2014), may have a higher triceps surae reflex gain (i.e. larger reflexes) which could result in excessive triceps activation during ground contact. The extent to which higher (or lower) reflex gain during landing influences landing kinetics needs to be further examined in the future.

Methodological considerations

Several methodological factors that do not influence the results but warrant additional consideration for interpretation are discussed here. First, in this study, the H-reflex was used to characterize reflex modulation rather than the stretch reflex. As illustrated in Figures 1 and 2 and summarized in Table 1, extremely rapid ankle dorsiflexion occurs during landing (i.e., over 700°∙s⁻¹). Thus, attempting to impose additional dorsiflexion perturbations on top of such fast dorsiflexion rotations would be meaningless since spindle afferents would already be firing at a high rate, or worse might cause injuries. Note that the previous studies that examined stretch reflexes during walking used dorsiflexion perturbation speeds of 250–300°∙s⁻¹ on top of an ongoing ankle rotation of 30 – 40°∙s⁻¹ (Andersen and Sinkjaer 1999; Mazzaro et al. 2005; Thompson et al. 2019). For these reasons, we elected to study reflex modulation during landing with the H-reflex. While overall phase-dependent modulation is similar between the H- and spinal stretch reflexes during walking, there are times where their modulation patterns do not necessarily match (Andersen and Sinkjaer 1999; Thompson et al. 2019); the H-reflex is more sensitive to presynaptic inhibition than stretch reflexes (Morita et al. 1998) while other factors that could influence muscle spindle afferent sensitivity (e.g. alpha-gamma coactivation) would not affect the H-reflex modulation. Thus, when interpreting the results of this and other studies using H-reflex as a proxy for understanding the potential functional role of stretch reflexes, it should be noted that H-reflex behaviors do not represent how muscle spindle Ia afferents (and Ib and II) are firing naturally during a given task nor does this method account for alpha-gamma coactivation. Rather, the H-reflex indicates how a synchronous synaptic input from Ia afferents may (or may not) result in muscle contraction.

Second, to estimate H-reflex size modulation across the flight and early landing periods, each participant was asked to perform over 100 drop landings. As such, participants in this study were all recreationally active and had prior experience with jumping and landing. To maintain and monitor the experimental condition (including possible fatigue effects), we randomized the stimulation timing and collected 2–5 non-stimulated landing trials after every 15–20 trials. The landing variables assessed (e.g. ankle and knee joint motion, ground reaction force) were the same throughout the experiment in each participant. Thus, it is unlikely that there was a systematic bias in H-reflex modulation within and across participants.

Finally, in the primary analyses of H-reflex modulation, the H-reflex sizes were normalized to the Mmax sizes measured during standing. In the secondary analyses, in 6 of 9 participants, the H-reflexes were normalized to the Mmax obtained during each of the landing period bins; we found that the Mmax was slightly smaller during standing than during landing. As a result, the primary analyses’ H-reflex normalization to the standing Mmax resulted in a systematically larger H-reflex (in %Mmax). However, this issue of normalization method does not affect the reflex modulation pattern nor influence the conclusions.

Conclusions

This study characterized the soleus H-reflex modulation during the flight and early landing period of drop landing. The H-reflex size after landing (0–30 and 30–60 ms) was significantly smaller (21–36% less) than that recorded during the flight periods (90–0 ms before ground contact). The time-critical decrease in H-reflex size starting about 30 ms prior to ground contact likely reflects the CNS control of reflex gain to facilitate an appropriate level of musculotendon stiffness in anticipation of rapid lengthening imposed upon landing.