Training‐induced improvements in knee extensor force accuracy are associated with reduced vastus lateralis motor unit firing variability

New Findings What is the central question of this study? Can bilateral knee extensor force accuracy be improved following 4 weeks of unilateral force accuracy training and are there any subsequent alterations to central and/or peripheral motor unit features? What is the main finding and its importance? In the trained limb only, knee extensor force tracking accuracy improved with reduced motor unit firing rate variability in the vastus lateralis, and there was no change to neuromuscular junction transmission instability. Interventional strategies to improve force accuracy may be directed to older/clinical populations where such improvements may aid performance of daily living activities. Abstract Muscle force output during sustained submaximal isometric contractions fluctuates around an average value and is partly influenced by variation in motor unit (MU) firing rates. MU firing rate (FR) variability seemingly reduces following exercise training interventions; however, much less is known with respect to peripheral MU properties. We therefore investigated whether targeted force accuracy training could lead to improved muscle functional capacity and control, in addition to determining any alterations of individual MU features. Ten healthy participants (seven females, three males, 27 ± 6 years, 170 ± 8 cm, 69 ± 16 kg) underwent a 4‐week supervised, unilateral knee extensor force accuracy training intervention. The coefficient of variation for force (FORCECoV) and sinusoidal wave force tracking accuracy (FORCESinu) were determined at 25% maximal voluntary contraction (MVC) pre‐ and post‐training. Intramuscular electromyography was utilised to record individual MU potentials from the vastus lateralis (VL) muscles at 25% MVC during sustained contractions, pre‐ and post‐training. Knee extensor muscle strength remained unchanged following training, with no improvements in unilateral leg‐balance. FORCECoV and FORCESinu significantly improved in only the trained knee extensors by ∼13% (P = 0.01) and ∼30% (P < 0.0001), respectively. MU FR variability significantly reduced in the trained VL by ∼16% (n = 8; P = 0.001), with no further alterations to MU FR or neuromuscular junction transmission instability. Our results suggest muscle force control and tracking accuracy is a trainable characteristic in the knee extensors, which is likely explained by the reduction in MU FR variability which was apparent in the trained limb only.

trainable characteristic in the knee extensors, which is likely explained by the reduction in MU FR variability which was apparent in the trained limb only.

K E Y W O R D S
electromyography, firing rate variability, motor unit, muscle force accuracy, neuromuscular function INTRODUCTION The human motor unit (MU) is the final component of the neuromuscular system and is fundamental to muscle force generation (Heckman & Enoka, 2012). Each MU comprises a single somatic motor neuron, including its axon, distal axonal branches, neuromuscular junctions (NMJ) and associated innervated skeletal muscle fibres. Motor output is governed by supraspinal commands and spinal reflex pathways which collectively control MU firing rate (FR). Thus, modulation of MU FR contributes to the increase and decrease of muscle force generating capacity.
During muscle contraction, the desired force output fluctuates around an average value rather than being at a constant level (Enoka & Farina, 2021;Pethick & Piasecki, 2022). Variation in MU FR has been identified as a critical determinant influencing the control of muscle force (Enoka & Farina, 2021;Vila-Cha & Falla, 2016), with associations between muscle force and MU FR dependent on single MU force, the input-output function of motor neurons and the frequency response of the muscle to transform an activation signal into force (Enoka & Farina, 2021). Using computational models allowing manipulation of key MU parameters (MU FR and MU FR variability), increasing index finger force resulted in MU FR variability of the flexor dorsal interosseous reducing exponentially, corresponding with improved simulated force fluctuations (Moritz et al., 2005). These simulated data are consistent with experimental (Laidlaw et al., 2000) and other simulated observations  supporting evidence that MU FR variability (i.e., the variability of inter-discharge intervals across consecutive MU firings) is a, if not the, key physiological parameter influencing the ability to maintain steady muscle contractions (Vila-Cha & Falla, 2016). Compared to central MU function, much less is known with respect to peripheral MU features (i.e., NMJ transmission instability) and the influence these may have on muscle force control.
Peripheral factors such as the release of acetylcholine at the NMJ, sodium/potassium pump activity, or modification to sodium and/or potassium intracellular and/or extracellular concentrations may alter muscle fibre action potential transmission (Allen et al., 2008). Although this alteration may subsequently impact muscle contraction and thus levels of force control, this has not yet been explored in a longitudinal manner.
The coefficient of variation for force (FORCE CoV ) has been identified as a significant explanatory variable for multiple performance tasks including balance (Zech et al., 2010), walking (Davis, Alenazy et al., 2020), manual dexterity (Keogh et al., 2019;Kornatz et al., 2005), levels of tremor (Kavanagh et al., 2016;Keogh et al., 2019) and the risk of falling in older adults (Carville et al., 2007;Enoka & Farina, 2021). The use of exercise training strategies (e.g., resistance exercise training (RET)) to improve muscle force control is, therefore, of interest for multiple diverse groups of individuals, including athletes, older adults, and those who are clinically vulnerable. It should be noted that findings from such diverging ranges of populations may not be directly comparable; for example, muscle tremor may present differently in varying physiological states (i.e., influence of aged muscle vs. exercise induced fatigue). Although RET is known to improve muscle strength, the effects of such training programmes on muscle force control/accuracy and MU firing properties remain unclear (Elgueta-Cancino et al., 2022).
While improvements in knee extensor maximal voluntary contraction (MVC) force were observed following 8 weeks RET (∼80% 1-repetition maximum (1RM)) in young individuals, neither FORCE CoV nor common drive was altered (Beck et al., 2011). Conversely, performing light-load (30% 1RM) training led to improvements in both knee extensor strength and both knee extensor and elbow flexor muscle force control (Kobayashi et al., 2014), with the greatest RET-induced improvements in isometric FORCE CoV occurring in the least steady subjects (Tracy & Enoka, 2006). Offering a potential explanation for improvements in FORCE CoV with RET, 4 weeks of isometric strength training which significantly increased muscle strength also increased MU FR (+3 ± 2.5 pps) during the plateau phase of submaximal muscle contractions and decreased in the MU recruitment threshold (Del Vecchio et al., 2019). Similarly, an increase in MU FR during the plateau phase of trapezoidal dorsiflexor contractions was observed following strength training (Kim et al., 2019), with conduction of ballistic muscle contractions leading to earlier activation of MUs and increased MU FR in the dorsiflexor muscles post-training (Van Cutsem et al., 1998). Despite RET proving to be mostly an effective training mechanism to improve FORCE CoV , RET may not be accessible to all individuals due to its higher intensity, which may pose physical limitations for older individuals (Barry & Carson, 2004) and those who are injured or present with disability. Resultingly, alternative training modalities, with a focus on light-load/task specific training, still need to be established to circumvent these limitations.
The aim of the current study was, therefore, to investigate the effect of a 4-week low intensity force accuracy training strategy on levels of knee extensor muscle force control/accuracy and any subsequent alterations to central and peripheral MU function in the vastus lateralis (VL) muscle. We hypothesised that muscle FORCE CoV and sinusoidal wave tracking accuracy (FORCE Sinu ), but not muscle strength, would improve with this training strategy alongside reduced MU FR variability, and would be observed in the trained limb only.  (Guo et al., 2022), both sexes were included in the current study.

New Findings
• What is the central question of this study?
Can bilateral knee extensor force accuracy be improved following 4 weeks of unilateral force accuracy training and are there any subsequent alterations to central and/or peripheral motor unit features?
• What is the main finding and its importance?
In the trained limb only, knee extensor force tracking accuracy improved with reduced motor unit firing rate variability in the vastus lateralis, and there was no change to neuromuscular junction transmission instability. Interventional strategies to improve force accuracy may be directed to older/clinical populations where such improvements may aid performance of daily living activities.
To avoid corrective actions when reaching the target line, the first two passes were excluded from the calculation. From these six contractions, the mean FORCE CoV was subsequently calculated as (standard deviation/mean) × 100, from the plateau phase of the contraction.
Participants next completed a series of sinusoidal wave force tracking tasks (using OTBioLab software, OT Bioelettronica, Turin, Italy) at 25% MVC to assess levels of FORCE Sinu . A familiarisation contraction was performed prior to the assessment contraction.
Contractions consisted of eight oscillations at a set amplitude (±4%) lasting for 30 s. A 10-s ramp preceded and followed each oscillating section of the contraction to allow force to steadily increase and decrease to and from the desired contraction intensity ( Figure 1b).
Contractions were exported and analysed in Spike2 (version 9) software, where a virtual channel was created (by subtracting the performed path from the requested path, and rectifying) and the area under the curve (N s) of this channel was representative of the level of deviation from the target line, reflecting muscle force tracking accuracy.
Participants also completed physical function tests to assess unilateral balance of both legs pre-and post-training. All balance tests were performed using a Footscan plate (Footscan, 200 Hz, RSscan International, Paal, Belgium) allowing measurement of centre of pressure, and the displacement of this, during static one-legged standing. Participants were asked to visually focus on a fixed point in front of them for the duration of the test (30 s). A 5-s countdown was given before instruction to lift one leg, 2 s before the recording period began. Distance travelled (mm), the displacement of centre of pressure, was recorded for further analysis (Figure 1c).

Sampling of single motor units during voluntary contractions
Intramuscular electromyography (iEMG) recordings were obtained using disposable concentric needle electrodes with a recording area of 0.07 mm 2 (model N53153, Teca, Hawthorne, NY, USA), with a grounding electrode on the patella. Participants were asked to relax their muscles to enable insertion of the needle electrode into the VL muscle to enable sampling of MUs during the series of voluntary iso-metric contractions used to assess FORCE CoV (as described above).
Following each contraction, the needle electrode was withdrawn 5-10 mm and the bevel rotated 180 • , recording from a total of four to six contractions from spatially distinct areas (Jones et al., 2021). iEMG signals were sampled at 50 kHz and bandpass filtered at 10 Hz to 10 kHz. Signals were digitised with a CED Micro 1401 data acquisition unit (Cambridge Electronic Design). All iEMG and force signals were recorded and displayed in real-time via Spike2 software (version 9).

Force accuracy training
All participants were required to perform force accuracy training

Statistical analysis
Data are presented as mean ± SD unless stated otherwise. As there were no bilateral leg differences at baseline in any parameter,

Muscle force tracking accuracy
Although the interaction effect was not significant (

DISCUSSION
The aim of the current study was to investigate the effects of a In line with our hypothesis, MVC force remained unchanged in the knee extensors across both legs. However, these results are not in direct agreement with others. For example, in older adults, low intensity RET (30% 1RM) increased knee extensor MVC after both 8 (Kobayashi et al., 2014) and 16 (Tracy & Enoka, 2006) weeks of training, with numerous studies reporting that low intensity RET (i.e., ∼20-40% 1RM) can induce increases in muscle strength (Hortobagyi et al., 2001;Keen et al., 1994;Kobayashi et al., 2014;Tracy & Enoka, 2006). This notion does, however, remain inconclusive with studies such as that by Moore et al. (2004)  MVC (Oshita & Yano, 2010), and another for contraction intensities ≤5% MVC (determined via displacement of centre of pressure) (Kouzaki & Shinohara, 2010). FORCE CoV of the hip abductors and dorsiflexors has been shown to be the most significant explanatory variable in sway-area rate during light load contractions, although most of the variance across conditions was unexplained,suggesting other physiological mechanisms important for postural control likely influence unilateral balance (Davis, Allen et al., 2020). The effects of force accuracy training on functional outcomes such as unilateral balance remains, therefore, to be further examined in other muscle groups and populations such as older individuals, who display deterioration of unilateral balance with advancing age (Maki et al., 1990;Izquierdo et al., 1999;Hess & Woollacott, 2005).
The current study utilised two different force tracking tasks, varying in difficulty, to assess levels of FORCE CoV and force tracking accuracy. Our results demonstrate improvements in both FORCE CoV and FORCE Sinu following training. During isometric contractions of fluctuating force (i.e., sinusoidal contractions), the recruitment and subsequent de-recruitment of MUs needs to be aligned to match the desired trajectory to increase and decrease force (Duchateau & Enoka, 2008) and as such, they require different control strategies.
Resultingly, the fluctuation in force during the sinusoidal contractions may contribute to positive alterations of force tracking and control strategies.
Commonly, FORCE CoV has been highlighted as a critical explanatory variable with respect to muscular performance of tasks including walking (Davis, Alenazy et al., 2020), tremor (Kavanagh et al., 2016;Keogh et al., 2019) and risk of falls in older adults (Carville et al., 2007). FORCE CoV is also associated with impaired functional ability in multiple sclerosis (Davis, Alenazy et al., 2020), and progressively deteriorates from middle to older age in highly active males and females (Piasecki, Inns et al., 2021). Therefore, the impact of force  Vila-Cha & Falla, 2016), and here we demonstrate improvements in force accuracy and MU FR variability in the trained limb only, which may be a result of reduced antagonist muscle activity and/or inhibitory afferent feedback (Enoka & Farina, 2021). Further, common inputs of descending and sensory signals induce a correlation between low-frequency oscillations in FR of motor neurons, known as common drive, also significantly influence fluctuations in muscle force output (Negro et al., 2009). Interestingly, low-frequency oscillatory components of MU FR were strongly associated to explain most variation in muscle force output during submaximal contractions, with MU FR variability being poorly correlated with FORCE CoV (Negro et al., 2009). Therefore, although not assessed in the current study, the influence and potential alterations to common drive should additionally be considered as a mechanism to aid improvements in FORCE CoV . There is potential for ionic changes (e.g., modification to sodium and/or potassium ion intracellular and/or extracellular concentrations, subsequently altering muscle fibre action potential transmission; Allen et al., 2008), release of acetylcholine at the NMJ, and/or the type/intensity of muscle contraction Carville et al., 2007;Enoka & Farina, 2021) to also influence levels of muscle force control; however, we observed no alterations to NMJ transmission instability, assessed via NF MUP jiggle.

Strengths and limitations
As the training period was only 4 weeks, it offers translational relevance and application to pre/rehabilitation scenarios (Durrand et al., 2019) with, for example, pre-operative colorectal patients in the UK having a 31-day target time frame between decision to treat and operation (Boereboom et al., 2019). The short duration of each training session (∼20 min) also counters one of the most commonly cited barriers to exercise interventions: 'lack of time' (Trost et al., 2002). Secondly, the tasks constituting force accuracy training are arguably more applicable to daily movements (e.g., rising from a chair)

Conclusion
To summarise, we highlight that a 4-week period of targeted force accuracy leads to improved muscle force control and accuracy in young healthy participants, which is associated with reduced MU FR variability. Importantly, these adaptations and possible mechanisms were evident in the trained limb only. These findings may influence interventional strategies to improve force accuracy, including in older and clinical populations where such improvements may help with independence maintenance via improved performance of activities of daily living.