Influence of sex on the age‐related adaptations of neuromuscular function and motor unit properties in elite masters athletes

Masters athletes maintain high levels of activity into older age and allow an examination of the effects of aging dissociated from the effects of increased sedentary behaviour. Evidence suggests masters athletes are more successful at motor unit remodelling, the reinnervation of denervated fibres acting to preserve muscle fibre number, but little data are available in females. Here we used intramuscular electromyography to demonstrate that motor units sampled from the tibialis anterior show indications of remodelling from middle into older age and which does not differ between males and females. The age‐related trajectory of motor unit discharge characteristic differs according to sex, with female athletes progressing to a slower firing pattern that was not observed in males. Our findings indicate motor unit remodelling from middle to older age occurs to a similar extent in male and female athletes, with discharge rates progressively slowing in females only.


Introduction
Master athletes (MAs) provide an attractive model within aging research, offering examination of the effects of aging independent of the commonly observed associated sedentary behaviour (Lazarus & Harridge, 2017). Regular exercise from middle to older age may minimise the musculoskeletal decline reported to begin from around the age of 40 years (Mitchell et al. 2012), with a wealth of literature clearly demonstrating the detrimental effects of the loss of muscle mass and function into older age (Larsson et al. 2019). Encompassing recent definitions of sarcopenia (Cruz-Jentoft et al. 2019) this potentially avoidable condition is explicable by both the atrophy and loss of individual muscle fibres, and associated motor unit (MU) remodelling (Piasecki et al. 2016b;Wilkinson et al. 2018), with a multitude of underlying factors. This general musculoskeletal decline with age may contribute to the increased incidence of falls (Yeung et al. 2019) and associated co-morbidities (Pacifico et al. 2020).
Studies involving MAs have demonstrated the benefits of lifelong exercise on minimising loss of strength (McKendry et al. 2018), improved immunological function (Duggal et al. 2018) and greater aerobic capacity when compared to age-matched controls (McKendry et al. 2020). Furthermore, regular exercise that is maintained from middle into older age helps to maintain lower body fat percentage, and possibly enhanced lean mass and skeletal muscle strength into later life (Piasecki et al. 2019a;Crossland et al. 2020). Understandably, MAs are not entirely resistant to the effects of age and notable declines have been observed within older competitive athletes in overall performance (Lazarus & Harridge, 2017;Bagley et al. 2019). Data at the single fibre level are less clear, suggesting older athletes have preserved contractile properties in the vastus lateralis, yet of type II fibres only (Power et al. 2016b;Gries et al. 2019). Moreover, as individual fibre power normalised to size appears to improve with age (Grosicki et al. 2016), the neural input to muscle warrants greater research interest.
Evidence of preservation of MU number in older athletes is equivocal; it was shown to be preserved in the tibialis anterior (TA) of one study (Power et al. 2010) but not another (Piasecki et al. 2016a). Nor is it preserved in the biceps (Power et al. 2012) or vastus lateralis of lifelong runners (Piasecki et al. 2019b), the latter including both power and endurance disciplines. Expansion of the MU is believed to partly compensate for MU loss, acting to reinnervate locally denervated fibres in aging muscle (Hepple & Rice, 2016). There is mounting evidence to suggest that life-long exercise exerts beneficial effects within the peripheral motor system, attenuating the denervation of muscle fibres and/or improving rates of reinnervation, as demonstrated by fewer markers of denervation in older athletes (Sonjak et al. 2019), larger MU potentials (MUPs) recorded from indwelling needle electrodes (Piasecki et al. 2019b), less neuromuscular junction (NMJ) transmission instability (Power et al. 2016a) and increased fibre-type grouping compared to that expected during normal aging (Mosole et al. 2014). However, the last of these has been disputed (Messa et al. 2020) and may highlight methodological limitations of histology when assessing MU remodelling.
Contraction of skeletal muscle relies upon successful synaptic input from descending cortical motor pathways to lower motor neurons and their associated NMJs, and the rate at which they discharge (the firing rate) is dictated by this descending drive in addition to neuromodulatory afferent feedback (Heckman & Enoka, 2012). MU firing rate (FR) is a key component of muscle strength (Del Vecchio et al. 2019) and decreases observed with age (Watanabe et al. 2016;Piasecki et al. 2016c) undoubtedly contribute to associated strength reductions. Although complete mechanisms for a general reduction in FR with older age are unclear, they are probably multifactorial and encompass hormonal fluctuations and environmental factors such as activity level (Hunter et al. 2016).
The majority of data relating neural adaptations to age and/or activity are based entirely on males, which is largely generalisable across human physiological research (O'Halloran, 2020). The lack of female data represents a significant shortfall of translational research when considering the male-female disability paradox; women tend to live longer than men but with greater disability (Kingston et al. 2014). However, recent histological findings from the vastus lateralis (VL) showed female MAs (ß81 years) also display greater reinnervation capacity than their inactive age-matched counterparts, based on morphological markers from muscle biopsies (Sonjak et al. 2019). Furthermore, few sex differences were noted in MU parameters of the TA in older athletes (Power et al. 2016a), indicating peripheral MU adaptations may not be sex-specific. Further knowledge gaps exist in the ages of the populations studied, with a focus on comparisons between the young (<40 years) and old (>65 years) populations, often omitting the middle ages where functionality begins to decline (Mitchell et al. 2012).
The aims of the present study were to investigate the age-related trajectory of neuromuscular function and MU adaptations in male and female elite MAs, from middle to older age. It was hypothesised that functional performance and all MU characteristics would decline with increasing age and do so to a similar extent in both sexes.

Ethical approval
The study conformed to the standards set by the Declaration of Helsinki, except for registration in a database. The study was approved by the local ethics committee at Nottingham Trent University (ethics number 619), and all participants provided written informed consent. Participant recruitment and data collection took place at the British Masters Athletics federation National track and field championships, 10-11 August 2019, Birmingham, UK.

Participant recruitment
A total of 30 MAs (16 males) took part in the study (age range 44-83 years). Of those tested, five competed at 1500-5000 m (two male), 22 competed at 800 m and below (11 male), and three took part in field events only (jumping events; two male). Participants had been training and competing for an average of 32 ± 18.1 years, training between 5 and 7 h per week for the majority of their training years and at the time of recruitment. The age-graded performance (AGP) of an athlete enables a direct comparison to the current world record, within the athletes specific age group and discipline, and is expressed as a percentage of that world record. The mean AGP for this cohort was 83.8 ± 7.37%, indicating a high level of performance relative to respective age group records and favoured event. For example, a marathon of 3 h and 30 min as a 70-year-old male gives an AGP of 80%.

Strength and force steadiness assessments
TA strength of the right leg was assessed with participants seated with the foot secured to an isometric force dynamometer (purpose-built calibrated strain gauge, RS Components Ltd, Corby, UK) with the knee and ankle bent at approximately 90 degrees. Participants were familiarised to dorsal flexion by performing three isometric contractions lasting 2-3 s at around 80% of maximal effort. Next, isometric maximum voluntary contraction (MVC) of the TA was assessed three times with 60 s rest between efforts. The highest value was accepted as MVC. Force was recorded at 100 Hz and displayed in real-time using Spike2 software (v8.01). Force steadiness was quantified as the coefficient of variation [CoV; (SD/mean) × 100] of force set at target lines of 10, 25 and 40% MVC. Participants were given a single familiarisation trial at each contraction intensity, before performing six contractions at 10%, six contractions at 25% and two contractions at 40% MVC, with each lasting approximately 12-15 s and a rest of 30 s in between contractions. The mean CoV at each contraction level was calculated.

Balance and jump mechanography
Limb dominance was assessed by asking which limb would be used to kick a ball. All participants were right-leg dominant. An RS Foot scan (Gait and Motion Technology Ltd, Bury St Edmunds, UK) pressure sensor plate was used to assess balance of all participants. The centre of pressure (COP), or postural sway, of the vertical plane was measured throughout assessment and is expressed as total distanced moved in millimetres (mm). This was assessed for 30 s with participants standing in the centre of the sensor plate, on the right leg only. Following this, a G-walk (a small inertial sensor placed at the base of the spine, secured with a Velcro waist belt; Gait and Motion Technology Ltd) was used to assess power (kW) from a series of counter movement jumps. Participants were instructed to jump as high as possible, with hands J Physiol 599.1 remaining on their waist with a trained assistant present and in reach of the participants in case of a fall or falter . Each participant repeated the jump sequence three times, with approximately 30 s rest between jumps, and the highest value was recorded.

Intramuscular EMG
After establishing the MVC, a concentric needle electrode (Model N53153; Teca, Hawthorne, NY, USA) was inserted at the muscle belly of the TA, to a depth of 0.5-1.5 cm. The intramuscular electromyography (iEMG) signals were bandpass filtered from 10 to 10 kHz and sampled at 50 kHz. iEMG signals were displayed in real-time using Spike2 software (v8.01) and data were stored for off-line analysis.

Sampling of individual MUs during voluntary contractions
Each participant performed a voluntary isometric contraction at 25% MVC. Needle position was initially adjusted, where needed, to obtain intramuscular MUPs with peak second derivative values >5 kV/s 2 , to ensure the recording needle electrode was close to depolarizing fibres (Stashuk, 1999a). Each voluntary contraction lasted 12-15 s, keeping as close as possible to a force target line shown on the computer monitor that was set at 25% MVC with real-time visual feedback. After each contraction the needle electrode was repositioned by combinations of rotating the bevel 180 degrees and withdrawing by approximately 5-10 mm. Participants had 30 s rest between each contraction. Needle re-positioning, voluntary contraction and signal recording was repeated until between four and six recordings from varying depths and perspectives had been obtained. iEMG signal analysis iEMG data are available for 16 males and 12 females. The mean (SD) number of individual MUs sampled was 26 (10) for males and 29 (8) for females. iEMG signals were analysed as previously described (Piasecki et al. 2016a) using decomposition-based quantitative electromyography (DQEMG) (Stashuk, 1999b). Extracted motor unit potential trains (MUPTs) with fewer than 40 MUPs were excluded. All MUP templates were visually inspected and their markers adjusted, where required, to correspond to the onset, end, and positive and negative peaks of the waveforms. The MUP area is the integral of absolute values of MUP between the onset and end and is expressed as μV.ms. MUP duration (ms) is measured from the onset to the end of the MUP. The number of phases and turns are measures of MUP complexity and are classified as the number of components above or below the baseline (phases) and a change in waveform direction of at least 25 μV (turns). A near fibre MUP (NFM) is the 'acceleration' of an MUP and is identified by applying a second-order low-pass differentiator to the MUP. This ensures only potentials from fibres close to the needle electrode significantly contribute to the recorded NFMs. NFM segment jitter (NFM SJ) is a measure of the temporal variability of individual fibre contributions to the NFMs of a MUPT. It is calculated as a weighted average of the absolute values of the temporal offsets between matched NFM segments of consecutive isolated (i.e. not contaminated by the activity of other MUs) NFMs across an MUPT expressed in microseconds (Fig. 1). NFM jiggle is a measure of the shape variability of consecutive NFMs of an MUPT expressed as a percentage of the total NFM area (Allen et al. 2015;Piasecki et al. 2016a). Firing rate was assessed as the rate of consecutive observations of the same MUP, expressed as number of observations per second (Hz). MU FR variability was measured using FR mean absolute consecutive difference (FR:MACD) across the train measured in hertz [FR:MACD = mean(abs(IFR i+1 − IFR i )]; here mean represents the mean across the MUPT and IFR i is the local 'instantaneous' FR at the time of the ith sampled MU firing. The local 'instantaneous' FR, at each sampled MU firing time, is the inverse of the Hamming-weighted mean interdischarge interval (IDI) of the previous five and following five consistent IDIs. A consistent IDI is one that is within ±3 standard deviations of the mean IDI of the MUPT.

Statistical analysis
All statistical analysis was completed using STATA (v.15). The age-graded performance of male and female athletes was compared using an unpaired t test. Neuromuscular functional parameters were assessed across age and sex using multiple linear regression, first examining interaction effects. Where none existed, interactions were removed from the model and effects of age and sex were investigated, adjusting for each other. Individual athletes have multiple values for MU parameters therefore multilevel mixed effects linear regression models were used to investigate these parameters with Age and Sex as factors. Interactions were first examined, where not present they were removed from the model, Sex and Age were explored individually, and mutually adjusted. Significance was accepted at P < 0.05.

Results
The age range of the female athletes was 45-82 years, and 44-83 years for male athletes. The AGP was 81.8 ± 6.8 for males and 85.8 ± 7.5 for females and did not differ between sexes (P > 0.1).

Discussion
To our knowledge these findings are the first to identify, using intramuscular techniques supported by functionality data, neuromuscular alterations from middle to older age in male and female competitive MAs. The data herein show there is no sex-based difference in the age-related decline of strength or control in the TA muscle, with a similar pattern in peripheral MU adaptations across age. Yet there is disparity between the sexes in MU firing rate, with females demonstrating a reduction over the lifespan that was not observed in males. Despite the MAs demonstrating exceptionally high ability as evidenced by their high AGP and regular physical activity (our cohort trained for an average of 5-7 h per week), functionality still declined into older age. These age-related decrements of strength ) and control (Castronovo et al. 2018;Mani et al. 2018) have been well described, and lifelong exercisers are no exception to this, although it is probable they have attenuated balance declines when compared to age-matched controls (Leightley et al. 2017).
Comparisons of detailed muscle function may not be logically extrapolated to athletic performance, but the current data do support previous observed reductions in competitive performance (Ganse et al. 2018), also with negligible sex differences in an/aerobic power (Bagley et al. 2019) or from world record performances (Gava & Ravara, 2019).
MUPs recorded from indwelling electrodes during voluntary contractions allow a detailed overview of individual MU structure and electrophysiological function, with NFMs consisting of significant contributions from only a few local MU fibres. We have previously shown power and endurance MAs have larger MUPs than age-matched controls (Piasecki et al.  Table 1. MVC, maximum voluntary contraction. 2019b), possibly as a result of larger MUs from greater reinnervation-based MU expansion, and in line with this we again observed a progressive increase in MUP area and duration with older age. Notably, this occurred to a similar extent in males and females, further supporting the notion that older female athletes also have a greater capacity for reinnervation-based MU expansion (Sonjak et al. 2019). MUP complexity also increased with age without a sex difference, in line with increases in MUP size here and in previous studies of the VL of non-athletes (Piasecki et al. 2016c). The complexity of an MUP is related to the temporal dispersion of its comprising individual fibre potentials, and reinnervation can cause increased MUP complexity.
The increases in NFM jiggle and NFM segment jitter are indicative of greater NMJ transmission instability, and all follow the same age-related pattern of an increase, with lack of a sex difference. Although there are no available data on NFM segment jitter, previous investigations of NFM jiggle in this muscle found no age-related differences between young (ß26 years) and old (ß71 years) non-athletic men, compared to a slightly greater instability in male MAs (ß69 years) (Piasecki et al. 2016a). Similar methods deduced very old (ß80 years) MAs had lower instability than their age-matched counterparts (Power et al. 2016a), highlighting the potential impact of prolonged exercise. The plasticity of the NMJ in response to age and exercise has been well described via histological markers (Soendenbroe et al. 2020), but with regard to in vivo electrophysiological measures such as those applied here, it is less clear what a favourable outcome would be as newly formed NMJs from axonal sprouts (i.e. 'rescued' fibres) are likely to demonstrate instability (Balice-Gordon, 1997).
The outcome of the MU discharge parameters yielded a slightly different finding; FR decreased with age for females but not for males. The firing rate of an MU is contraction-level-and contraction-type-specific (Duchateau & Baudry, 2014), and is closely related to the MU/fibre type (i.e. slow or fast). Here we assessed FR during an isometric contraction at a force normalised to individual maximum (25% MVC) to help mitigate these  contributing factors. The TA muscle is predominantly composed of type 1 fibres (ß70%) (Henriksson-Larsén et al. 1985;Nakagawa et al. 2005) and in young individuals this composition is not affected by sex (Porter et al. 2002). Therefore, it is possible the female athletes display a more prominent shift to a slower muscle phenotype with increasing age when compared to males, although muscle biopsies (particularly difficult in this cohort and muscle group) would be required to confirm this. Moreover, evidence emerging from cross-innervation and electrical stimulation studies has demonstrated fibre phenotype is governed by its innervation status (reviewed by Blaauw et al. 2013), so sex-based differences in fibre composition in these athletes would be a consequence of their differing FR, and not a cause. Alterations to MU FR have shown a high level of plasticity even in response to relatively short (2 weeks) exercise interventions (Martinez-Valdes et al. 2017), and are exercise type-specific with a reported decrease from endurance training and increase from resistance training, following a 6-week intervention (Vila-Chã et al. 2010). Therefore, the competitive discipline and training regime of an athlete could influence MU FR. However, given the even distribution across short (<800 m) and longer (>1500 m) event specialism of males and females in this current study, exercise discipline is unlikely to explain this age by sex difference. The ability to voluntarily maximally contract the TA muscle does not differ between sexes (Russ & Kent-Braun, 2003) and differs little into older age in other leg muscle groups , so, when near maximal muscle contractions are considered, there appears little sex-based influence of neural drive. The large age range of these athletes offers insight into the often-overlooked neuromuscular health trajectory from middle to older age. However, it also presents a number of challenges such as the associated change in the hormonal milieu, specifically the decrease in sex hormones known to influence neuromuscular function (Sipilä et al. 2013;Swiecicka et al. 2020) and regulate skeletal muscle signalling (Laakkonen et al. 2017). This is particularly important with regard to the transition from pre-to post-menopausal state as performance effects may be minimal in pre-menopausal (McNulty et al. 2020). Animal models have demonstrated a neuroprotective effect of oestrogens and androgens centrally (Spence & Voskuhl, 2012), and in humans some aspects of centrally mediated function alter concordantly with hormonal fluctuations (Smith et al. 2002;Ansdell et al. 2019). Therefore, the normal age-related decreases in sex hormones, of which highly active individuals are not spared, may exert a greater central influence on females than on males. Further research of the hormonal influence on neuromuscular deterioration in aging males and females is warranted.

Strengths and limitations
This is the first study, to our knowledge, to demonstrate a detailed trajectory of functionality and associated MU properties from middle to older age, within a particularly valuable cohort. Applying iEMG at a range of muscle depths ensures we have sampled from a range of active MUs, and the inclusion of NFMs allowed a detailed view of MU remodelling at the single MU level. Our findings demonstrate that despite high levels of physical activity, MAs experience declines in functionality but also demonstrate features of peripheral MU remodelling. Importantly, these changes occur to a similar extent between sexes. The MU FR reduced in females but not males, indicating the importance of the CNS and afferent feedback in aging neuromuscular function. There are a several limitations within this study that are important to acknowledge; in our assessment of functionality, we focused on postural sway taken from one footed balance, which can be improved with targeted balance training (Kiss et al. 2018), and bilateral differences between dominant and non-dominant limbs were not explored. Whilst we have identified some interesting sex-based differences, these data are related to MUs active during mid-level contractions and reveal nothing of the level of remodelling of higher recruitment threshold MUs. The hormonal milieu between sexes differs, particularly in relation to sex hormones and we are unable to describe their possible influence on MU function.

Conclusion
Highly active competitive MAs exhibit an age-related decrease in neuromuscular function from middle to older age, and this is associated with increased MU remodelling. These changes occur to a similar extent in males and females. Discharge properties of MUs differ in the age response, with firing rate decreasing in females but not in males, possibly reflecting a greater tendency towards a phenotypically slower muscle in the female athlete population.