Effects of aging on transmission from Ia afferents to α-motoneurons and presynaptic mechanisms across muscle lengths and contraction types

Abstract

Presynaptic inhibition of Ia afferents, particularly via primary afferent depolarization (PAD), plays a crucial role in modulating transmission from Ia afferents to α-motoneurons according to task demands. In young adults, PAD is enhanced during eccentric contractions and at longer muscle lengths, supporting adaptable motor control. However, aging is associated with reduced muscle spindle sensitivity, which may impair the regulation of presynaptic inhibition via PAD. This study investigated whether aging alters the modulation of PAD according to contraction type and muscle length. Nineteen older adults (69 ± 4 years) and 16 young adults (23 ± 2 years) participated in three experimental sessions. The maximal H-reflex (H_max) and maximal M-wave (M_max) were recorded in the soleus muscle to assess Ia-alpha motoneuron transmission (H_max/M_max). PAD was evaluated using the D1 inhibition method (H_D1/H_test), while ongoing presynaptic inhibition was assessed via heteronymous facilitation (H_fac/H_test). Measurements were performed during isometric and eccentric contractions at intermediate (0°) and long (− 15° dorsiflexion) muscle lengths. Young adults showed significant modulation of both H_D1/H_test and H_fac/H_test across conditions, while older adults did not (p > 0.05). Similarly, H_max/M_max was reduced during eccentric contractions at intermediate muscle length in young but not in older adults. These results indicate that aging impairs regulation of PAD according to muscle mechanics, potentially contributing to deficits in sensorimotor control.

Introduction

Regulation in neural transmission from Ia afferent fibres to alpha motoneurons (Ia-motoneuron transmission) is essential for fine motor control, balance, posture and coordinated movement [1–4]. This regulation is influenced by various motor task constraints and characteristics, including muscle length and contraction type, i.e. isometric, concentric or eccentric contraction [5–7]. Aging is often associated with an impaired capacity to regulate Ia-motoneuron transmission in response to motor tasks [8–11], potentially leading to a decline in motor function [12, 13]. A deeper understanding of the mechanisms underlying age-related impairments in the regulation of Ia-motoneuron transmission according to contraction type and muscle length is essential for advancing our knowledge of motor dysfunction in older adults.

Ia afferent fibres are fast-conducting sensory axons from muscle spindles that detect changes in muscle stretch rate [14]. Their discharge rate increases during muscle lengthening and sustained static muscle length [15]. Within the spinal cord, Ia afferents form excitatory synaptic connections with alpha motoneurons [16]. The spinal regulation of Ia afferent input involves presynaptic inhibition through primary afferent depolarization (PAD), mediated by GABA_ergic interneurons. These interneurons, modulated by supraspinal and spinal inputs, enable fine-tuned control of Ia afferent transmission. In healthy young adults, PAD is modulated by various motor task constraints and is particularly high during eccentric contractions and at longer muscle lengths, where the firing rate of Ia afferent fibres increases significantly [5, 17–20]. This modulation leads to reduced Ia-motoneuron transmission [5–7, 19].

Aging impacts muscle spindles through various morphological and functional changes, including capsule thickening [21] and a reduction in the number of intrafusal fibres [22]. Muscle spindle sensitivity also declines, due to diminished responsiveness to stretching and changes in muscle length [23, 24]. As PAD regulation adapts to changes in Ia afferent discharge, it is plausible that aging may attenuate the modulation of PAD regulation during muscle lengthening, especially during eccentric contractions with changes in muscle length. Some evidence suggests that aging impairs the capacity to modulate PAD according to motor tasks, which are assumed to have different Ia afferent activities. For example, while young adults adjust PAD based on contraction intensity, postural control or specific motor tasks, older adults often exhibit diminished PAD regulation [8, 9, 11, 25]. However, further direct evidence is needed to clarify how aging influences PAD regulation across different contraction types and muscle lengths, and how these changes ultimately impact Ia-motoneuron transmission.

This study investigated whether aging alters PAD regulation based on contraction type and muscle length. Using the Hoffmann reflex, we assessed PAD and Ia-motoneuron transmission in the soleus muscle of young and older adults. Measurements were done at two muscle lengths (intermediate and long) and during two contraction types (isometric and eccentric). We hypothesized that, unlike young adults who modulate PAD based on contraction type and muscle length, older adults exhibit diminished or no PAD modulation, which is associated with impaired presynaptic regulation of Ia-motoneuron transmission.

Methods

Participants

A total of 35 participants participated in the study:16 young adults (8 males and 8 females; age, 23 ± 2 years; height, 175 ± 4 cm; weight, 70 ± 13 kg; BMI, 23.0 kg·m⁻²) and 19 older adults (8 males and 11 females; age, 69 ± 4 years; height, 168 ± 7 cm; weight, 64 ± 10 kg; BMI, 22.8 kg·m⁻²). Individuals were classified as part of the “young” group if they were between 18 and 35 years old, and as part of the “older” group if they were between 60 and 80 years old [26]. Exclusion criteria included any neurological or neuromuscular disorders, musculoskeletal injuries that could impair torque production and the use of medications known to affect the nervous system. During the medical screening, we ensured that participants were (1) physically active according to WHO guidelines (i.e., at least 150 min per week of moderate-intensity aerobic activity or 75 min per week of vigorous-intensity aerobic activity) and (2) not engaged in any resistance training. These inclusion criteria were applied equally to both cohorts to minimize potential bias from resistance training that may influence neuromuscular function. In addition, all older participants underwent a medical screening with a geriatrician, including the WHO ICOPE (Integrated Care for Older People) assessment, to ensure that only healthy older adults were included. Participants were instructed to refrain from engaging in strenuous exercise for 48 h prior to the testing sessions. Ethical approval for the project was obtained from the Ethics Committee on Non-Interventional Research of Nantes University (No. 21022024–3). All procedures conducted in this study adhered to the requirements of the Declaration of Helsinki (last revised in 2011). Data from the young adult group were previously reported in an earlier study [18].

Study design

All 35 participants were engaged in two experimental sessions. All younger participants and 11 older participants (among the 19) also participated in a third session. Each experiment was separated by a minimum of 24 h. During each experiment, assessments of contraction type (isometric and eccentric contractions) and muscle length (intermediate and long muscle lengths) were randomly assigned to avoid potential biases stemming from the order of testing procedures.

In the first experimental session (Experiment A; duration, 2 h 30), we assessed soleus fascicle length and the recruitment curves for the soleus H-reflex and M-wave for each contraction type and soleus length. The second experimental session (Experiment B; duration, 2 h) examined post-activation depression by PAD using the D1 method. In the third experimental session (Experiment C; duration, 2 h), heteronymous Ia facilitation was assessed. The order of Experiments B and C was randomized across participants. All data were collected from the right leg of the participants.

At the end of each experimental session, two maximal voluntary concentric and isometric dorsiflexions were performed to assess maximal dorsiflexor activity (3 s of duration, interspaced by 1-min rest).

Experimental setup

Mechanical data

The participants were seated with the trunk inclined at 30°, the hip flexed at 85° (0° representing the anatomical position), and the knee joint fully extended at 0° (representing full extension). Torque values were collected using an isokinetic dynamometer (Biodex 3, Shirley, NY, USA) (Fig. 1). A Biodex ergometer enabled the instantaneous recording of muscle torque at a constant velocity of 20°·s⁻1. The output signal from the dynamometer was collected at 2 kHz using a commercial acquisition system (CED Power 1401-3A, Cambridge Electronic Design; Cambridge, UK) and was displayed and stored with Signal 7 software (Cambridge Electronic Design; Cambridge, UK). The right foot was strapped to the footplate, while the trunk and thighs were secured to the seat of the isokinetic dynamometer. The ankle angle was modulated according to experimental conditions. During isometric contractions, the ankle angle was set at 0° (neutral position) and − 15° (dorsiflexion) for intermediate and long muscle lengths, respectively. During eccentric contractions, the ankle range of motion was set at 30°. The ankle angle moved from 15° to − 15° and from 0° to − 30° for intermediate and long muscle lengths, respectively. Throughout the three experiments, the dynamometer movement cycles lasted 5 s, including 1.5 s for performing the movement, 1.5 s to return to the initial position and 1 s of rest before the next cycle. Within the latter period, participants performed a 1-s submaximal isometric plantarflexion as a pre-activation contraction to standardize thixotropic effects across conditions [27].

Fig. 1.

Participant positioning on the dynamometer for plantar flexor testing. Trunk reclined at 30°, hip flexed at 85°, knee fully extended (0° = full extension). Torque recorded with an isokinetic dynamometer

Electromyography (EMG)

EMGs were recorded from the soleus, tibialis anterior and vastus lateralis muscles using pairs of self-adhesive surface electrodes (Kendall™ H124SG; Covidien, Dublin, Ireland) configured in a bipolar arrangement with a 30-mm inter-electrode distance. The soleus electrodes were placed 2 cm below the muscle–tendon junction of the gastrocnemii. For the tibialis anterior, electrodes were placed on the muscle belly parallel to the longitudinal axis of the muscle, located one-third of the distance between the head of the fibula and the tip of the medial malleolus. The vastus lateralis electrodes were positioned two-thirds along the line from the anterior superior iliac spine to the lateral side of the patella. Prior to electrode application, the skin was shaved, gently abraded and cleaned with alcohol. The signals were sampled at 5 kHz, amplified 1000 ×, and bandpass filtered from 5 to 500 Hz using a bio-amplifier (g.BSamp 0201a; Guger Technologies, Shieldberg, Austria).

Percutaneous electrical nerve stimulation

The electrophysiological responses (i.e., H reflex and M wave) of the soleus were evoked by percutaneous electrical stimulation of the posterior tibial nerve using a 1-ms single rectangular pulse at high voltage (400 V), delivered by a stimulator (Digitimer, model DS8R Biphasic Constant Current Stimulator, Hertfordshire, UK). A self-adhesive electrode (Kendall™ H124SG; Covidien, Dublin, Ireland) was used as a cathode and was placed in the popliteal fossa. The anode electrode (5 × 10 cm, Medicompex SA, Ecublens, Switzerland) was positioned on the anterior surface of the knee, below the patella. Once positioning was confirmed, the cathode electrode was firmly secured to this site with a strap and tape, as recommended [28]. To investigate PAD (D1 method), the common peroneal nerve was stimulated using a DS7R Current Stimulator (Digitimer, model DS7R Biphasic Constant Current Stimulator, Hertfordshire, UK), with the cathode electrode placed close to the head of the fibula and the anode electrode near the medial part of the tibial head. For assessment of heteronymous Ia facilitation, the femoral nerve was stimulated with the cathode positioned over the nerve in the femoral triangle and the anode placed over the greater trochanter. The location of the stimulation electrode was identified to obtain an M wave in the vastus lateralis, which is associated with an upward movement of the patella. The motor threshold for stimulation of the common peroneal and femoral nerves was determined as the lowest intensity that evoked at least three M waves out of five stimulations. To ensure that the electrophysiological recordings reflected neurophysiological modulations rather than methodological artefacts, several controls were implemented. Stimulation was automatically triggered at fixed joint angles (0° and − 15°) and verified by ultrasound to correspond to distinct fascicle lengths. The maximal M-wave (M_max) remained stable across sessions and conditions (p > 0.05), and all responses were normalized to M_max. In addition, M_atHmax/M_max and H_test/M_max ratios confirmed consistent recruitment of the α-motoneuron pool across conditions. These procedures ensured that the observed effects were not influenced by electrode displacement, limb movement or differences in muscle architecture.

Ultrasound recording

Soleus fascicle length was measured using a Mach30 ultrasound scanner (Hologic-Supersonic Imagine, Aix-en-Provence, France) with B-mode image acquisition and a linear probe (SL18-5, 7.5 MHz central frequency, Hologic-Supersonic Imagine, Aix-en-Provence, France). Participants were seated in the isokinetic dynamometer with their trunk inclined 20° backward from the vertical and their knee fully extended. The lateral malleolus was aligned with the axis of rotation of the dynamometer. Images were captured from + 30° (plantarflexion) to − 30° (dorsiflexion) during eccentric submaximal contractions, with 0° representing the anatomical position. Fascicle length changes were evaluated during submaximal eccentric contractions (50% of maximal EMG_RMS) at a constant angular velocity of 20 deg·s⁻1.

Proximal and distal insertions of the soleus muscle were identified, positioning the probe just distal to the myotendinous junction of the medial gastrocnemius muscle to ensure clear visibility of the soleus fascicles in this region. Despite the multipennate structure of the soleus muscle, a study on its three-dimensional architecture demonstrated no significant differences in fascicle length across its four compartments: medial-anterior, lateral-anterior, medial-posterior and lateral-posterior [29]. Therefore, the chosen probe position provides a fascicle length measurement representative of the entire soleus muscle. The probe was secured in a foam holder and firmly attached to the leg with tape. To condition the muscle, five ankle angle rotations ranging from 30° (plantarflexion) to − 30° (dorsiflexion) were performed back and forth [30]. B-mode ultrasound images were continuously recorded through three acquisitions during eccentric contractions at 50% of maximal EMG_RMS, covering the full range of motion to assess the architecture of the soleus muscle (Fig. 2). The acquired ultrasound images were sampled at 50 Hz and processed using a customized MATLAB script (UltraTrack software; 5–2-2, Digitizing Services, UK) [31].

Fig. 2.

Examples of two ultrasound images in the sagittal plane during eccentric contraction at two different ankle angles in the soleus of one elderly participant. This image was used to calculate soleus fascicle length (blue dashed line) from the visible insertion of the fibres between the deep and superficial aponeuroses (yellow lines)

Experimental protocol

Experiment A

Experiment A was conducted in two parts. In the first part, three maximal voluntary isometric contractions (MVCs) at 0° were performed to determine the maximal root mean square (RMS) value of soleus EMG. The maximum soleus EMG_RMS activity recorded during isometric MVCs was used to establish the activation level targeted, i.e., 50% of maximal EMG_RMS provided to the participants during the study. This submaximal activation level was selected to avoid inducing muscle damage and/or neuromuscular fatigue during and between experimental sessions. Moreover, it is functionally relevant because it is quite close to the loads required in activities of daily living [32–34]. In this phase, participants were instructed to contract their plantar flexors to move and maintain their soleus EMG_RMS biofeedback at the target corresponding to 50% of the maximal EMG_RMS. The biofeedback represented the RMS EMG signal, which was computed in real time by a digital computing channel. This channel instantaneously calculated the RMS EMG signal with an integration time of 500 ms. This method ensured consistent muscle activation levels across all conditions [5, 6, 18]. In addition, to ensure that the modulation of our neurophysiological variables reflected neural activity rather than peripheral or recording factors, we measured M_max in each condition. M_max remained stable across all conditions (p > 0.05). Furthermore, all trials were visually inspected to confirm the absence of artefacts, and any trials showing obvious movement contamination were excluded from analysis. Subsequently, soleus fascicle length was measured.

The second part of the experiment aimed to obtain maximal H-reflex (H_max) and maximal M wave (M_max) for both contraction types (eccentric and isometric) at each muscle length (intermediate and long) in soleus. Accordingly, we determined the soleus H-reflex and M wave recruitment curves during plantarflexion at 50% of maximal EMG_RMS for each contraction type and soleus length. This resulted in a total of four recruitment curves per participant (Fig. 3). For each recruitment curve, the stimulation intensity was gradually increased in 2-mA and 3-mA increments for young and older individuals, respectively. The H-reflex threshold to the stimulation intensity was where no further increase in M wave amplitudes was observed between two stimulations (i.e., M wave plateau). We chose a greater increment in older adults compared to younger adults because the relationship between the amplitude of the evoked potentials and stimulation intensity is less steep in older adults. Even with the different increments used between young and older subjects, the M_atHmax/H_max ratio remained consistent across conditions and groups, allowing for a comparison of neural mechanisms rather than recording and stimulation conditions [35]. Finally, supramaximal stimulations were delivered at 120% of this stimulation intensity to ensure the recording of soleus M_max. Five electrical nerve stimulations were delivered at each stimulation intensity with an interstimulus interval of 5 s [4].

Fig. 3.

Recruitment curves of the soleus H-reflex (blue) and M-wave (red) according to muscle length and contraction type in a representative participant from the younger group (A) and a participant from the older group (B)

Electrical tibial nerve stimulations were delivered at ankle angles of 0° and − 15° (representing the midpoint of each range of motion) for intermediate and long muscle lengths, respectively. Stimulation of the tibial nerve was driven by the CED acquisition system. During eccentric contractions, stimulations were automatically triggered when the ankle angle passed 0° (intermediate) or − 15° (long).

Experiment B

To investigate presynaptic modulation by PAD, we employed the D1 method as described by Mizuno et al. [36]. This technique involves eliciting an electrical volley on the nerve supplying the α-motoneuron pool of the antagonist muscle, thereby activating the Ia terminals and resulting in depression of the H-reflex [36]. The H-reflex of the soleus was conditioned using a stimulus applied to the fibular nerve, activating GABA_ergic interneurons responsible for the PAD of Ia afferents from the soleus muscle [37]. To evoke the conditioned H-reflex of the soleus (H_D1), a train of three 1-ms stimulations at 300 Hz was applied to the common peroneal nerve, with an intensity set at 120% of the stimulation intensity at the motor threshold of the tibialis anterior. The interval between the first stimulus of the train and the soleus H-reflex test (H_test) was 21 ms [38, 39]. A total of 20 H_D1 and 20 H_test responses were randomly evoked across all conditions (eccentric and isometric contractions at intermediate and long muscle lengths) [5].

Experiment C

A limitation of the D1 method is that changes in PAD activity may be contaminated by potential occlusion or reflex gain phenomena in the reflex loop. To avoid this misinterpretation, it is recommended to compare the results obtained with the D1 method to those obtained with an independent method that assesses ongoing post-activation depression with heteronymous Ia facilitation [5, 16]. To investigate heteronymous Ia facilitation, we employed the method of Hultborn et al. [40], which evaluates the monosynaptic facilitatory effect of femoral nerve stimulation on the amplitude of the soleus H-reflex. To minimize contamination from polysynaptic inputs, femoral nerve stimulation was delivered following tibial nerve stimulation due to the shorter neural pathway of the heteronymous Ia afferent pathway compared to the homonymous pathway [40]. Considering the average duration of a synaptic event (0.5 to 1 ms), we systematically assessed the onset of facilitation in 1-ms increments to mitigate the influence of polysynaptic excitatory inputs [41]. For each participant, the onset of H-reflex facilitation in response to femoral nerve stimulation was determined by adjusting the delay between the test (tibial nerve) and conditioning (femoral nerve) stimuli in 1-ms increments, ranging from 9 ms (test stimulus preceding conditioning stimulus by 9 ms) to 1 ms (test stimulus preceding conditioning stimulus by 1 ms). The conditioning stimulation intensity was set at 130% of the motor threshold, and at least five trials were averaged for each delay [42]. A total of 20 H_fac and 20 H_test responses were randomly evoked in all conditions (eccentric and isometric contractions at intermediate and long muscle lengths).

To determine non-conditioned H-reflexes, the H_test was based on the smallest H_max amplitude measured across the four experimental conditions (Fig. 4). This approach ensures that each condition is subjected to the same H_test size, allowing for accurate comparisons between the different experimental conditions. Before testing each condition, soleus H_test was adjusted to ensure a consistent amplitude across conditions, as the number and/or origin of the motoneurons recruited may vary with stimulation intensity [43]. Each condition was evaluated in a randomized order, with a 5-min rest period between each tested condition.

Fig. 4.

Sigmoidal fittings of the soleus recruitment curve for the H-reflex (blue circles) and M-wave (red squares) to determine optimal amplitudes and stimulation intensities for post-activation depression (PAD) assessment, as well as H_slope (blue line) and M_slope (red line) evaluation for a representative participant. H_max: maximal H-reflex; H_test: test H-reflex for PAD assessment; M_max: maximal M-wave; M_atHmax: M-wave corresponding to H_max; H_slope: slope of the ascending portion of the H-reflex curve; M_slope: slope of the M-wave curve

Data analysis

Evoked potentials

For each combination of contraction type and muscle length, the mean peak-to-peak amplitude of the following soleus evoked potentials were calculated over all trials: H_max, M_max, M_atHmax, H_D1, H_fac and H_test. All ratios of evoked potentials were calculated for each experimental condition.

To estimate the effectiveness of Ia afferents in activating α-motoneurons, we calculated the H_max/M_max ratio. H_max/M_max is commonly used as an indicator of the maximal proportion of α-motoneurons recruited via the Ia afferent pathway. To control the stability of stimulation conditions in each condition, the M wave elicited concomitantly with H_max (M_atHmax)—a small fraction of M_max (M_atHmax/M_max)—was measured and analysed [35]. In addition, the slope of the ascending limb of the recruitment curve reflects the rate of motoneuron recruitment as a function of increasing Ia input to the motoneuron pool and is generally used to examine H-reflex gain [1, 44]. This reflex gain serves as a complementary measure, more closely linked to transmission efficiency between Ia afferents and α-motoneurons. By employing this approach, the influence of peripheral factors on the results is minimized, ensuring the accuracy and reliability of the findings [45]. The sigmoidal slopes for both H-reflex (H_slope) and M wave (M_slope) were determined by fitting the relationship between the amplitude of H-reflex in the ascending part of the recruitment curve or M wave in entire recruitment curve by stimulation intensity [46]. The H_slope/M_slope ratio was then calculated.

To examine presynaptic modulation by PAD using the D1 method, we calculated the H_D1/H_test ratio. A higher H_D1/H_test ratio is associated with lower post-activation depression by PAD [36].

To examine heteronymous Ia facilitation, we calculated the H_Fac/H_test ratio. A higher H_Fac/H_test ratio is associated with greater heteronymous Ia facilitation [40].

The H_test/M_max ratios obtained in the study of post-activation by PAD and heteronymous Ia facilitation were used to ensure that the amplitude of the non-conditioned H-reflex remained consistent across the four experimental conditions. Hence, we controlled such that approximatively the same amplitude of α-motoneuron pool was investigated through experimental conditions.

EMG activity during voluntary contractions

The RMS of the soleus EMG signals over a 500-ms period preceding stimulation was normalized to the corresponding M-wave amplitude (M_max) (EMG_RMS/M_max) for all experimental conditions. This normalization procedure was implemented to ensure that the EMG activity of the soleus (agonist) remained consistent across conditions. Similarly, the RMS of the tibialis anterior EMG signal over the same time period preceding stimulation was calculated to assess coactivation across all conditions. To quantify the level of coactivation in isometric contractions, tibialis anterior EMG_RMS during submaximal isometric contractions was divided by the peak tibialis anterior EMG_RMS obtained during the dorsiflexors MVC. Tibialis anterior RMS during eccentric plantar flexions was normalized to the maximal concentric dorsiflexion. Tibialis anterior EMG_RMS was calculated across all trials within each experimental condition.

Changes in fascicle length

Ultrasound images were compiled to track the length of the soleus fascicles using a custom MATLAB script that enabled the tracking of fascicle behaviour during eccentric submaximal contractions. Fascicle length was measured at two distinct angles: 0° (intermediate muscle length), and − 15° (long muscle length). A single fascicle was tracked at each angle. In cases where fascicles were not fully visible, fascicle length and the deep aponeurosis were extrapolated linearly to estimate the complete fascicle length.

Statistical analysis

All descriptive statistics presented in the text, tables and figures are expressed as mean values ± standard deviation. The significance level for all analyses was set at p < 0.05. Normality of the data and homogeneity of variances were verified and validated using the Shapiro-Wilks W test and Levene’s test, respectively. Two-factor ANOVAs [muscle length (intermediate vs. long) × age (young vs. old)] were conducted to assess changes in fascicle length at different ankle angles (0° and − 15°) during eccentric submaximal contractions. Three-factor ANOVAs [contraction type (eccentric vs. isometric) × muscle length (intermediate vs. long) × age (young vs. old)] with repeated measures on contraction type and muscle length were performed to compare soleus M_atHmax/M_max, H_max/M_max, H_slope/M_slope, H_D1/H_test and H_Fac/H_test across conditions. Three-factor ANOVAs [contraction type (eccentric vs. isometric) × muscle length (intermediate vs. long) × experiment (experiment B vs. experiment C)] with repeated measures on contraction type, muscle length and experiment were conducted to compare H_test/M_max. Similarly, four-factor ANOVAs [contraction type (eccentric vs. isometric) × muscle length (intermediate vs. long) × experiment (experiment A vs. experiment B vs. experiment C) × age (young vs. old)] with repeated measures were used to compare soleus M_max and EMG_RMS/M_max for soleus, as well as tibialis anterior coactivation. When significant interactions were found, a post hoc Holm test was performed. Statistical analyses were performed using JASP (Version 0.17.2.1, Amsterdam, NED) and GraphPad Prism software (Version 10.2.1; GraphPad Software Inc., San Diego, CA, USA).

Results

Changes in fascicle length with ankle angle

Changes in fascicle length were calculated to ensure that the muscle lengths associated with the angle of electrical nerve stimulation were distinct. As shown in Fig. 5, muscle length at − 15° of ankle angle was 15.1% greater than at 0° (p < 0.001), with no significant difference between young and older participants (p = 0.718).

Fig. 5.

Changes in soleus muscle fascicle length during 50% submaximal eccentric contraction across ankle angles. Data are presented as mean ± standard deviation with individual values shown. Young adults are represented in blue with filled circles, while older adults are represented in orange with open circles. *Significant difference for ankle angle effect at p < 0.001

EMG activity

We compared soleus EMG_RMS/M_max between eccentric and isometric contractions at intermediate and long muscle lengths to ensure consistent agonist muscle activation across experimental conditions. No significant effect was found (all p values > 0.05) (Table 1). However, an age effect was observed (p < 0.001), with young participants displaying a 33% higher soleus 50% EMG_RMS/M_max ratio than older participants, regardless of muscle length or contraction type.

Table 1.

Effect of muscle length, contraction type and age on soleus and tibialis anterior EMG-RMS activity

		Young		Old
		Intermediate	Long	Intermediate	Long
Eccentric
	SOL EMGRMS/Mmax	0.025 ± 0.006	0.026 ± 0.006	0.016 ± 0.006	0.017 ± 0.006
	Tibialis anterior Coactivation (%)	10.269 ± 2.571	10.124 ± 1.510	8.818 ± 0.937	8.850 ± 1.071
Isometric
	SOL EMGRMS/Mmax	0.024 ± 0.005	0.025 ± 0.008	0.018 ± 0.007	0.018 ± 0.006
	Tibialis anterior Coactivation (%)	10.139 ± 1.492	9.891 ± 1.909	8.782 ± 1.058	8.838 ± 1.510

Tibialis anterior coactivation during plantar flexion was assessed to determine if antagonist activation remained consistent across all experimental conditions. No main effects or interactions were found for tibialis anterior coactivation levels (all p values > 0.05) (Table 1). However, an age effect was observed (p < 0.001), with older participants showing approximately 1.18% lower coactivation than younger participants. The mean tibialis anterior coactivation was around 10.10% in younger participants and 8.83% in older participants during all plantar flexor conditions.

Data are pooled across the four experimental conditions and presented as mean ± standard deviation. EMG electromyography, SOL soleus, M_max maximal M wave, RMS root mean square.

Presynaptic modulation by PAD (D1 method)

SOL H_D1/H_test was assessed to estimate the level of PAD. Figure 6A shows raw traces of the conditioned (H_D1) and non-conditioned (H_test) H-reflexes measured during isometric and eccentric contractions at intermediate and long muscle lengths in a representative older participant.

Fig. 6.

Modulation of post-activation depression via primary afferent depolarization in relation to muscle length, contraction type and age. A Representative traces of the test H-reflex (H_test) and conditioned soleus H-reflex (H_D1) evoked by fibular nerve stimulation preceding posterior tibial nerve stimulation at a 21-ms conditioning-test interval during isometric and eccentric contractions at intermediate and long muscle lengths. B Mean absolute data and individual values are shown as mean ± standard deviation. H_D1 and H_test are presented as a ratio during isometric and eccentric contractions at intermediate and long muscle lengths. Young adults are represented in blue with filled circles, while older adults are represented in orange with open circles. The electrical intensity for H_test was normalized across all contractions for each muscle length. Abbreviations: H_D1, conditioned soleus H-reflex; H_test, non-conditioned soleus H-reflex; Ecc, eccentric; Iso, isometric. *Significantly different from eccentric contraction type. ^#Significantly different by age

A significant three-way interaction was observed between contraction type, muscle length and age (F(1, 33) = 4.314, p = 0.046, ηp² = 0.116), suggesting differential modulation of H_D1/H_test across age groups. In older participants, H_D1/H_test was not significantly influenced by muscle length or contraction type (p > 0.05), indicating an impaired capacity for modulation of PAD in response to these factors. This lack of adaptation contrasts with younger participants, where both muscle length and contraction type had a significant impact.

In younger participants, H_D1/H_test decreased by 13.4% during isometric contractions at long compared to intermediate muscle lengths (p < 0.001). This reduction was more pronounced during eccentric contractions (29.3%) (p < 0.001). Additionally, contraction type influenced PAD, with a 25.3% lower H_D1/H_test during eccentric compared to isometric contractions at long muscle lengths (p < 0.001). Age-related differences were particularly evident during eccentric contractions at long muscle lengths, where H_D1/H_test was 28.9% higher in older compared to younger participants.

Heteronymous Ia facilitation (HF)

SOL H_Fac/H_test was assessed to estimate the level of heteronymous Ia facilitation. Figure 7A illustrates raw traces showing the conditioned (H_Fac) and non-conditioned (H_test) H-reflex measured during isometric and eccentric contractions at intermediate and long muscle lengths in a representative older participant.

Fig. 7.

Effects of age, muscle length and contraction type on heteronymous Ia facilitation. A Representative traces show the non-conditioned soleus H-reflex (H_test) and the conditioned H-reflex (H_Fac) in a representative older participant. H_test and H_Fac were evoked by posterior tibial nerve stimulation, with H_Fac conditioned by femoral nerve stimulation applied after posterior tibial nerve stimulation at conditioning intervals between 9 and 1 ms. Traces are shown for both isometric and eccentric contractions at intermediate and long muscle lengths. B Absolute and individual data are presented as mean ± standard deviation. H_Fac and H_test ratios are shown for isometric and eccentric contractions at intermediate and long muscle lengths. The electrical intensity to evoke H_test was normalized across all contractions at each muscle length. Young adults are represented in blue with filled circles, while older adults are represented in orangey with open circles. Abbreviations: H_Fac, conditioned soleus H-reflex; H_test, non-conditioned soleus H-reflex; Ecc, eccentric; Con, concentric; Iso, isometric. * indicates significant difference from eccentric contraction type. # indicates significant difference from intermediate muscle length. α indicates significant difference by age within the same experimental condition

A significant three-way interaction was found between age, contraction type and muscle length for H_fac/H_test (F(1, 21) = 18.557, p < 0.001, ηp² = 0.469), indicating differential modulation of heteronymous facilitation across age groups. In older participants, H_fac/H_test remained relatively stable across contraction types and muscle lengths, with no significant differences observed between eccentric and isometric contractions or between intermediate and long muscle lengths. However, age-related differences emerged when comparing younger and older participants. During isometric contractions at intermediate muscle lengths, younger participants exhibited a 27.0% higher H_fac/H_test than older adults (p < 0.001), while this difference remained significant but slightly lower at long muscle lengths (+ 18.7% in younger adults). Similarly, during eccentric contractions at intermediate lengths, H_fac/H_test was 24.0% higher in younger compared to older participants (p < 0.001). However, at long muscle lengths during eccentric contractions, no significant age-related differences were observed (p > 0.05).

In younger participants, both muscle length and contraction type significantly influenced H_fac/H_test. During isometric contractions at long muscle lengths, H_fac/H_test decreased by 11.2% compared to isometric contractions at intermediate length. A similar but more pronounced reduction (19.5%) was observed during eccentric contractions at long muscle lengths compared to intermediate lengths. Furthermore, eccentric contractions led to a 5.8% lower H_fac/H_test than isometric contractions at intermediate muscle lengths, with this difference increasing to 14.6% at long muscle lengths.

Effectiveness of activated Ia afferents to discharge α-motoneurons and reflex gain

In this study, SOL H_max/M_max and H_slope/M_slope were evaluated to estimate the effectiveness of activated Ia afferents in stimulating α-motoneurons and the reflex gain, respectively. Figure 8A presents raw traces of the SOL H-reflexes and M-waves elicited during eccentric and isometric contractions at intermediate and long muscle lengths in representative young and older participants.

Fig. 8.

Changes in maximal soleus H-reflex amplitude normalized to the maximal M-wave amplitude across muscle length, contraction type and age. A Representative traces of the maximal soleus H-reflex evoked by posterior tibial nerve stimulation at intermediate and long muscle lengths during isometric and eccentric contractions in a representative young (red line) and older (black line) participant. B Absolute and individual data for elderly (n = 19) and young participants (n = 16) are expressed as mean ± standard deviation, showing changes in the maximal H-reflex normalized to the maximal M-wave at intermediate and long muscle lengths during isometric and eccentric contractions. Young adults are represented in blue with filled lines, while older adults are represented in orange with open lines. Abbreviations: H_max, maximal soleus H-reflex; M_max, maximal M-wave; Ecc, eccentric; Iso, isometric. * Significantly different from other conditions (p < 0.05)

A significant three-way interaction was observed between contraction type, muscle length and age for H_max/M_max (F(1, 33) = 31.716, p < 0.001, ηp² = 0.490). Regardless of age, muscle length and contraction type influenced H_max/M_max, but the effects differed between young and older participants. During isometric contractions, a significant decrease in H_max/M_max at long muscle length compared to intermediate length was observed in young adults (− 24%, p < 0.001), while this effect was not significant in older adults (p = 0.064). In contrast, during eccentric contractions, the opposite pattern was found: older adults exhibited a significant 17.8% decrease in H_max/M_max at long compared to intermediate muscle length (p < 0.001), whereas no difference was observed in young adults (p > 0.05).

Contraction type also had a differential impact depending on age. At an intermediate muscle length, young adults showed an 18.6% lower H_max/M_max during eccentric compared to isometric contractions (p < 0.001), while no difference was observed in older adults (p = 0.271). Conversely, at a long muscle length, the effect was reversed: older adults exhibited a 17.2% lower H_max/M_max during eccentric compared to isometric contractions (p < 0.001), while no significant difference was found in young adults (p > 0.05).

Figure 9 shows the soleus H-reflex slope to M-wave slope ratio according to muscle length and age. For the H_slope/M_slope, a significant main effect of contraction type was observed (F(1,29) = 6.853, p < 0.05, ηp² = 0.191), indicating that this ratio differs between isometric and eccentric contractions, regardless of age and muscle length. Post hoc analysis revealed an 8.1% reduction in H_slope/M_slope during eccentric compared to isometric contractions (p = 0.014).

Fig. 9.

Soleus H-reflex slope to M-wave slope ratio according to muscle length and age. Data are presented as mean ± standard deviation, with individual values shown. Young adults are represented in blue with filled circles, while older adults are represented in orange with open circles. H_slope, slope of the soleus H-reflex; M_slope, slope of the M-wave; Inter, intermediate muscle length; Long, long muscle length. * Significantly different from other conditions (p < 0.05)

Additionally, a significant interaction between age and muscle length was found (F(1,29) = 26.161, p < 0.001, ηp² = 0.474), suggesting differential modulation of H_slope/M_slope by muscle length in young versus older participants. Specifically, age influenced H_slope/M_slope, with a 61.5% reduction at intermediate muscle lengths in older adults compared to younger adults (p < 0.001), and a 47.8% reduction at long muscle lengths (p < 0.001). Muscle length also affected H_slope/M_slope. In young participants, the ratio decreased by 14.6% at long muscle lengths compared to intermediate lengths (p < 0.001); however, this modulation of the H_slope/M_slope across muscle lengths was not observed in older adults (p > 0.05) (Fig. 9).

We measured SOL M_max to ensure stability of EMG recording conditions across experimental conditions. For M_max, no interaction effect between conditions was observed (all p values > 0.117, Table 2).

Table 2.

Effect of contraction type, muscle length and age on soleus evoked potential

		Young		Old
		Intermediate	Long	Intermediate	Long
Eccentric
	Mmax (mV)	5.968 ± 1.158	5.747 ± 1.060	4.399 ± 1.777	4.306 ± 1.732
	Hmax/Mmax (u.a)	0.549 ± 0.122	0.417 ± 0.101	0.459 ± 0.136	0.413 ± 0.148
	MatHmax/Mmax (a.u)	0.155 ± 0.069	0.141 ± 0.019	0.171 ± 0.113	0.206 ± 0.117
	Htest/Mmax (a.u)	0.398 ± 0.111	0.392 ± 0.102	0.347 ± 0.093	0.327 ± 0.099
Isometric
	Mmax (mV)	5.883 ± 1.019	5.992 ± 1.058	4.459 ± 1.822	4.286 ± 1.698
	Hmax/Mmax (u.a)	0.447 ± 0.117	0.427 ± 0.101	0.416 ± 0.136	0.342 ± 0.134
	MatHmax/Mmax (a.u)	0.158 ± 0.050	0.164 ± 0.102	0.170 ± 0.075	0.178 ± 0.082
	Htest/Mmax (a.u)	0.399 ± 0.119	0.396 ± 0.109	0.339 ± 0.092	0.340 ± 0.095

To verify that similar numbers of efferent and potentially afferent nerve fibres were stimulated in each condition, we measured SOL M_atHmax. Results show that both M_max and M_atHmax amplitudes remained stable across all conditions, indicating that equivalent proportions of α-motoneurons were activated (Table 2). This consistency suggests that H_max/M_max modulations are attributable to neural mechanisms rather than recording conditions [35].

Absolute data are presented as mean ± standard deviation. M_max, maximal M wave; H_max, maximal soleus H-reflex; M_atHmax, M wave corresponding to maximal soleus H-reflex; H_test, non-conditioned soleus H-reflex corresponding to its test value for D1 and heteronymous Ia facilitation methods.

Discussion

This study investigated the impact of aging on presynaptic modulation by PAD with respect to contraction type and muscle length. Our findings revealed that in older adults, PAD remains unchanged across muscle lengths or contraction types. In contrast, younger adults exhibited greater PAD at longer muscle lengths compared to shorter and during eccentric contractions compared to isometric contractions. These results highlight an age-related impairment in the presynaptic control of Ia-motoneuron transmission in response to muscle constraints that alter Ia afferent discharge.

Age-related modulation of presynaptic inhibition by PAD

Muscle lengthening and muscle length significantly influence PAD regulation in healthy young adults [5, 18, 19]. Specifically, PAD is greater during eccentric contractions compared to isometric ones and at longer muscle lengths compared to shorter ones. This modulation likely results from increased Ia afferent discharge associated with muscle lengthening and long muscle length [18]. In contrast to younger participants, older individuals did not exhibit presynaptic modulation by PAD based on contraction type or muscle length. Furthermore, to evaluate presynaptic mechanisms, complementary approaches are necessary, such as assessing ongoing Ia presynaptic inhibition by examining monosynaptic facilitation of the soleus H-reflex via femoral nerve stimulation [40]. Interestingly, older adults exhibited limited facilitation (cf. methodological considerations), with no variation across experimental conditions, in contrast to younger participants. This supports the findings of an absence of presynaptic mechanism modulation in older adults based on contraction type and muscle length.

The age-related decline in presynaptic modulation by PAD in response to muscle length and contraction type may result from a combination of peripheral and central changes, leading to differential motor control. Peripheral changes may be attributed to morphological changes in muscle spindles [21, 22] and decreased sensitivity with aging [23, 24], which diminish responsiveness to stretch and changes in muscle length, thereby affecting the Ia afferent discharge rate. These peripheral changes are accompanied by central alterations at both spinal and supraspinal levels. At the spinal level, Terao et al. (1996) demonstrated a significant reduction in small neurons, i.e., mainly interneurons, distributed within the intermediate zone of the ventral horn of the spinal cord. Furthermore, their findings further indicate that among the spinal ventral horn cells, these small neurons in the intermediate zone, rather than large α-motoneurons and medium-sized γ-motoneurons, are more profoundly affected by aging [47]. Consequently, these age-related morphological changes in the spinal cord could impair the integration of afferent inputs, thereby also altering the regulation of PAD.

At the supra-spinal level, aging induces structural and functional changes in brain regions involved in motor control [13, 48]. Aging affects sensorimotor structures, potentially leading to functional impairments and reduced motor performance. This decline may necessitate compensatory mechanisms to preserve movement execution [49]. Indeed, age-related reductions in proprioceptive input from muscle spindles may impair the efficiency of automatic control processes, requiring a greater reliance on controlled processes and consequently increasing cognitive load [12]. Neuroimaging and electrophysiological studies show age-related reductions in cortical inhibition and increased activation across multiple cortical areas during motor tasks. These findings suggest a shift toward greater reliance on voluntary, supraspinal control in older adults, potentially due to reduced efficiency of automatic spinal mechanisms [13]. Such changes in the supraspinal control of movement may affect the processing and integration of afferent feedback, thereby influencing the regulation of presynaptic inhibition of Ia afferents by PAD, and ultimately modifying its contribution to motor output.

Taken together, these findings suggest that aging significantly impacts PAD regulation and alters motor control strategies. Our results are in line with previous studies examining PAD modulation in different contexts. For instance, Baudry et al. [8] found that older adults exhibited no variation in PAD between position and force control tasks, relying instead on increased antagonist coactivation for limb stabilization. In contrast, younger adults demonstrated task-specific PAD modulation, with PAD decreasing during position control compared to force control [8]. Similarly, Butchart et al. [9] showed that older participants lacked consistent PAD modulation across torque levels, whereas younger adults exhibited torque-dependent PAD modulation [9]. Earles et al. (2001) also observed reduced PAD modulation in older individuals during submaximal efforts, with greater reliance on direct motor activation to maintain motor output [25]. Notably, PAD modulation in older adults may depend on task specificity. For example, Baudry and Duchateau [43] found that older adults showed heightened PAD and greater co-contraction under challenging balance conditions (e.g. with eyes closed or on compliant surfaces) compared to younger adults [43]. Differences between these findings and our results may stem from variations in the motor tasks performed.

Although presynaptic mechanisms play a crucial role in regulating synaptic transmission between Ia afferent fibres and motoneurons, the findings on PAD do not consistently align with those on the efficacy of Ia-motoneuron transmission, especially in older adults.

Age-related efficiency of Ia-motoneuron transmission

In young adults, the capacity of Ia afferents to activate alpha motoneurons is reduced during eccentric contractions at intermediate muscle lengths [6, 7], consistent with our findings showing lower H_max/M_max and H_slope/M_slope ratios during eccentric compared to isometric contractions at intermediate lengths. This inhibition likely results from recurrent inhibition via Renshaw cells [18, 19, 50, 51] and post-activation depression by PAD [5, 18, 19]. However, despite greater PAD during eccentric contractions at long muscle lengths, H_max/M_max did not differ from isometric contractions at the same length, likely due to differences in recurrent inhibition.

In older adults, H_max/M_max was modulated by muscle length and contraction type, but differently compared to young individuals. Specifically, H_max/M_max was similar between isometric and eccentric contractions at intermediate muscle lengths. This highlights a distinct behaviour of the H_max/M_max in older adults compared to young adults, where it is typically reduced during eccentric contractions at intermediate muscle lengths [6, 7, 18]. Furthermore, H_max/M_max was reduced during eccentric contractions at long muscle lengths in older adults, whereas this was not the case in young adults. These results are surprising, as presynaptic inhibition by PAD does not explain this decrease. It is possible that the loss of neuromuscular spindle sensitivity [23, 24] does not generate sufficient afferent discharge at an intermediate muscle length to regulate H_max/M_max, whereas this regulation may still occur at a longer muscle length. It is therefore plausible that aging also alters other spinal and/or supraspinal mechanisms contributing to the modulation of Ia–alpha motoneuron transmission according to muscle length and lengthening. At the spinal level, there are several contributing age-related changes such as altered recurrent inhibition via Renshaw cells, potential reductions in Group II afferent facilitation [52] and modified Group Ib afferent input due to muscle–tendon adaptations [53]. For example, regarding recurrent inhibition, our research team previously demonstrated that recurrent inhibition was more pronounced during eccentric contractions than during isometric contractions at intermediate muscle length. However, at long muscle length, recurrent inhibition was similar between these two contraction types. Similarly, H_max/M_max was comparable across different contraction types at long muscle length [18]. This was not the case in the current study for older participants, suggesting that recurrent inhibition could be a key candidate explaining the observed reduction in Ia-motoneuron alpha transmission in older subjects.

Similar findings were reported by Chen et al. (2014), who observed that older adults exhibited a significantly lower H_max/M_max during isometric contractions in the dorsiflexion position (− 20°) compared to the neutral (0°) and plantar flexion (20°) positions. However, these results should be interpreted with caution due to differences in joint angle (− 15° in our study) and the absence of M_max measurements for each active position in their study, which complicates the interpretation of H_max [54]. In our study, the amplitude of H_max/M_max was not significantly different between young and older adults, except during eccentric at long muscle length. This contrasts with studies conducted under static and isometric conditions [55, 56] and in dynamic conditions [57], which indicated lower spinal excitability in older adults. For example, Skarabot et al. (2019) demonstrated that the amplitudes of the H-reflex, normalized to RMS, were lower in older adults regardless of contraction type (i.e. eccentric and concentric). However, the differences observed may be related to the muscle group investigated in their study (i.e., tibialis anterior). Finally, our results showed that the H_slope/M_slope ratio was significantly lower in older adults, regardless of contraction type or muscle length. This indicates a reduced reflex gain [45] during muscle contractions in this population, suggesting that the recruitment rate of motoneurons in response to increased Ia afferent input is reduced with aging. This reduction may be attributed to morphological changes in muscle spindles [21, 22] decreased spindle sensitivity [23, 58], and alterations in the regulation of presynaptic mechanisms [9, 25, 59].

Beyond the neurophysiological mechanisms, these findings may have functional relevance in older adults. The absence of PAD modulation with respect to contraction type and muscle length may compromise the capacity to fine-tune reflex pathways during daily motor activities that demand rapid adjustments, such as maintaining balance during gait or responding to unexpected perturbations. Such reduced adaptability could contribute to deficits in postural control, diminished locomotor efficiency and ultimately a greater risk of falls and fall-related hospitalizations, which represent major causes of morbidity and loss of independence in older adults [60]. While our study was not designed to directly assess clinical outcomes, it highlights a potential mechanistic link between altered spinal reflex regulation and age-related functional decline, thereby emphasizing the need to explore whether targeted interventions (e.g. resistance or balance training) can mitigate these impairments.

Methodological considerations

While this study provides valuable insights into motor control and aging, certain limitations should be acknowledged. One limitation relates to the measurement of heteronymous Ia facilitation. Accurately evaluating presynaptic mechanisms requires combining two complementary methods. The first assesses PAD activity via changes in the H-reflex amplitude conditioned by fibular nerve stimulation, reflecting its responsiveness to conditioning inputs [36]. The second evaluates the monosynaptic facilitation of the soleus H-reflex by femoral nerve stimulation, which indicates the level in ongoing Ia presynaptic inhibition [40]. This dual approach helps mitigate potential biases related to reflex recruitment gain [42]. However, in older adults, we observed minimal facilitation (< 10% of H_test), which is consistent with previous findings of age-related declines in heteronymous Ia facilitation [59]. This limited facilitation complicates interpretation of PAD amplitude and raises the possibility of occlusion effects, even though findings were consistent across conditions and unaffected by muscle length or contraction type. We contend that these limitations do not affect our main results. Thus, this study offers a solid foundation for understanding age-related changes in motor control and opens several avenues for future research.

Future directions

In this study, presynaptic modulation via PAD differs between older adults and young participants in this study. However, future investigations should explore other spinal mechanisms that may also contribute to altered Ia-alpha motoneurons transmission. At the spinal level, age-related changes such as altered recurrent inhibition via Renshaw cells, potential reductions in Group II afferent facilitation [52] and modified Group Ib afferent input due to muscle–tendon adaptations [53] may all play a role. Additionally, it would be valuable to link the modification of these mechanisms’ behaviour with aging to a potential decrease in the sensitivity of muscle spindles. Such age-related reductions in proprioceptive input from muscle spindles may impair the efficiency of automatic control processes, necessitating a greater reliance on controlled processes and consequently increasing cognitive load [12]. Therefore, it would be interesting to investigate the contribution of supraspinal changes to the regulation and modulation of PAD. It would be also valuable to examine inter-individual effects to determine which conditions (e.g., sex, level of physical activity or age) impact PAD regulation. Finally, although PAD appears to be altered with aging, it remains unclear whether older adults can modulate PAD in response to physical interventions such as balance training, resistance training or aerobic exercise, even though these modalities have been shown to induce other neuromuscular adaptations [44, 61]. To date, only one study has investigated whether physical intervention is able to modulate PAD [62]. This study showed that 8–9 weeks of eccentric resistance training reduce PAD amplitude in young adults.

Conclusion

This study demonstrates that aging impairs the regulation of presynaptic mechanisms involved in Ia-motoneuron transmission in response to muscle length and contraction type. Unlike young adults, older adults do not adjust PAD according to mechanical constraints, indicating altered spinal reflex control. This dysregulation likely results from age-related peripheral and central changes affecting sensorimotor integration. These findings highlight the need to investigate potential compensatory spinal or supraspinal mechanisms and to determine whether physical interventions can help preserve the adaptability of presynaptic regulation in response to motor task demands and constraints.

Author contribution

Yohan Betus, Julian Colard, Thomas Cattagni and Marc Jubeau: conception or design of the work.

Yohan Betus, Julian Colard, Thomas Cattagni and Marc Jubeau: acquisition, analysis or interpretation of data for the work.

Anne-Sophie Boureau, Guillaume Chapelet and Christophe Cornu: participant recruitment, notably through the coordination and conduct of medical consultations.

All authors drafted and revised the manuscript, approved the final version and agree to be accountable for all aspects of the work, ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This study was supported by the French National Research Agency (ANR) under grant number ANR-22-CE17-0002.

Agence Nationale de la Recherche,ANR-22-CE17-0002,Christophe Cornu

Data availability

The analyzed data are available at the following address: https://doi.org/10.6084/m9.figshare.30439091

Declarations

Conflict of interest

The authors declare no competing interests.