The aging neuromuscular system and motor performance

changes in the basic functional unit of the neuromuscular system, the motor unit, and its neural inputs have a profound effect on motor function, especially among the expanding number of old (older than (cid:2) 60 yr) and very old (older than (cid:2) 80 yr) adults. This review presents evidence that age-related changes in motor unit morphology and properties lead to impaired motor performance that includes 1 ) reduced maximal strength and power, slower contractile velocity, and increased fatigability; and 2 ) increased variability during and between motor tasks, including decreased force steadiness and increased variability of contraction velocity and torque over repeat contractions. The age-related increase in variability of motor performance with aging appears to involve reduced and more variable synaptic inputs that drive motor neuron activation, fewer and larger motor units, less stable neuromuscular junc- tions, lower and more variable motor unit action potential discharge rates, and smaller and slower skeletal muscle ﬁbers that coexpress different myosin heavy chain isoforms in the muscle of older adults. Physical activity may modify motor unit properties and function in old men and women, although the effects on variability of motor performance are largely unknown. Many studies are of cross-sectional design, so there is a tremendous opportunity to perform high-impact and longitudinal studies along the continuum of aging that determine 1 ) the inﬂuence and cause of the increased variability with aging on functional perfor- mance tasks, and 2 ) whether lifestyle factors such as physical exercise can minimize this age-related variability in motor performance in the rapidly expanding numbers of very old adults. aging;

The large age-related variability among and within older adults, however, can be masked when the average or best performances of the age groups or individuals are compared, so that subtle age-related changes in physiology and function are not always detected. Contributing factors that interact with biological aging to increase variability between and within old adults includes physical activity, genetics, nutritional status, hormonal status, and inflammatory status among other mediators that will modify the motor unit and motor performance between people as they age and into very old age (27). Some of this large variability may also be exacerbated (or dampened) by the sampling bias associated with cross-sectional design studies that also average the motor performance data of subjects over a large age range (60 -90 yr).
In this review, we focus on the age-related changes in the motor unit and the inputs it receives from spinal and supraspinal sources, and then highlight several areas of motor performance that are directly influenced by these changes, including maximal strength and power, velocity of contraction, fatigability, and steadiness during submaximal contractions.

AGING AND THE MOTOR UNIT
The motor neuron is the final common pathway through which all synaptic inputs are translated to motor function by the musculoskeletal system. Aging is accompanied by a net loss of motor units (16,145), changes to the morphology and properties of existing motor units (55), and altered inputs from peripheral, spinal, and supraspinal centers [e.g., see Seidler et al. (132) and Tracy et al. (147)]. Ultimately, motor performance is impaired and its variability is greater with advanced age.
Sarcopenia is the age-related loss of muscle mass and associated reductions in maximal muscle strength, power, and physical function (10,42,54). It is predictive of disability, loss of independence, and mortality (108) and has a strong neurogenic component (86). One of the primary causes of sarcopenia is the loss of muscle fiber innervation by alpha motor neurons caused by spinal motor neuron apoptosis and distal axon retraction. Spinal motor neuron apoptosis can begin gradually, early in life, and accelerate after the age of 60 yr and beyond (16,145) with more rapid declines in human muscle mass and function from approximately 75 to 80 yr of age [e.g., see Andersen (2), Justice et al. (70), and McNeil et al. (104)]. In addition, age-related morphological changes in the alpha motor neuron occur during normal aging is in part due to oxidative stress and inflammation (115), and the neuromuscular junction becomes less stable (31,55). These events alter the normal cycle of denervation and reinnervation of muscle fibers that occurs over the life span in skeletal muscle. The damaged motor axon withdraws from the muscle fibers it innervates, and many but not all of these fibers are then reinnervated by adjacent axons through collateral sprouting (55). This process of denervation and reinnervation accelerates with advanced age and leads to fewer but larger surviving motor units (16,43), ultimately resulting in lasting functional consequences for men and women starting as early as age 50 -60 yr that can be magnified in very old adults (Ͻ80 yr) [e.g., see Reid and Fielding (127)]. Regular physical activity, however, may modify age-related changes in motor unit structure and function across the life span, with examples provided throughout this review. Figure 1 summarizes some of these age-related changes in structure at the different levels of the motor unit and the accompanying physiological and functional changes that ultimately contribute to declines in motor performance in old and very old adults. It is important to note that although the role of physical activity is highlighted in this review, other processes such as genetics, nutrition (decreased total caloric and protein intake), altered hormonal status, inflammatory mediators, and factors resulting in altered protein synthesis and sarcopenia (35) may also be powerful modifiers to the aging neuromuscular system, possibly contributing to the large variability in motor performance between old adults and very old adults (27). The contribution of biological aging, physical activity, and these other modulators on the motor unit and the subsequent neuromuscular system is not well understood because many of the studies in humans are of cross-sectional design and not able to control for many of these potential modifiers. In this context, the following section will broadly highlight some of the known age-related changes in motor unit at the level of the 1) motor neuron, 2) neuromuscular junction, and 3) muscle fiber, and are summarized in the model presented in Figure 1.

Motor Neuron
Skeletal muscle contraction is the end result of a series of events that is initiated with the transmission of synaptic inputs to motor neurons. Thus, age-associated changes in the synaptic input can change the output (i.e., force and movement) and the age effect determined by the gain of the system and its sensitivity to different inputs. "Gain" is the ratio of the output to the input, and motor neurons in the spinal cord are able to alter the gain (e.g., excitability) of the system, thereby providing a change in the output for a given input. For example, monoaminergic projections from the brainstem have a powerful role in influencing the gain of the spinal motor neurons (e.g., by either increasing or decreasing excitability) and thus motor function (58,69). The synaptic input onto motor neurons that originate from multiple sites within the neuromuscular system change with age, although the current review will highlight those age-related changes from supraspinal and peripheral reflex pathways. In addition, the importance of agerelated changes in both independent and shared common synaptic inputs to motor neurons within and across muscles that govern motor control are highlighted. These common synaptic inputs play an integral role in determining the input-output relationships (i.e., gain) of the motor neuron pool, and these inputs coordinate motor unit activity.
Cortical inputs to motor neurons. Significant age-associated changes in the structure and function of cortical neurons contribute to the dysfunction of the aging neuromuscular system. Neuronal atrophy is ubiquitous with advancing age throughout the central nervous system (156), with ϳ40% reduction in the volume/size of motor cortical cell bodies by 80 yr of age (51), and general reductions in both gray matter (103,130) and white matter (98) volume throughout the cortex. In contrast to spinal motor neurons, there is limited evidence that the number of neurons change substantially in the cortex with with reduced motor performance in old adults during motor tasks (76,136).
In addition to age-related structural changes, the function of cortical neurons is marked by a significant decrease in inhibition and increased activity in many cortical areas in old adults. The decreased inhibition with advancing age has been assessed predominantly with transcranial magnetic stimulation (TMS) and quantified as 1) decreased interhemispheric inhibition between cortices (100, 156); 2) shorter silent period in electromyographic (EMG) recordings in response to TMS before and after fatiguing exercise (63,113,131); and 3) reduced short-interval intracortical inhibition (53,99). In addition, there is greater activation during motor tasks throughout the aging cortex in the contralateral and ipsilateral hemispheres (100,156). An inability to inhibit appropriate cortical areas during focused and targeted motor tasks can lead to inappropriate activation of motor units (132). Accordingly, greater muscle activity is frequently observed when older adults perform fine motor tasks (74,135). It is unknown whether these age-related changes are the result of dysfunction or are compensatory [i.e., increased cortical activity to compensate for decreased cortical matter (132) or increased activity to stiffen the joint for postural support (74)], but ultimately, they influence corticospinal inputs leading to altered synaptic inputs at the motor neuron.
Reflex responses. Afferent feedback from sensory receptors to spinal motor neurons evokes rapid responses in muscle that are reduced with advancing age. One hallmark of aging is a slowing of responses due to a reduction in peripheral efferent and afferent axon action potential conduction velocity (36,129) due at least in part, to declines in the density of unmyelinated and myelinated neurons (64). Furthermore, the Hoffman (H) reflex, which predominantly assesses the efficacy of Ia afferents to activate spinal motor neurons, is often reduced in old adults compared with young adults, although these effects are highly variable and dependent on the task being performed (6,8). Moreover, the function of peripheral reflexes changes with advanced age and can be accompanied by corresponding compensatory responses in the cortex; for example, the lower efficacy of Ia afferents to discharge spinal motor neurons is accompanied by greater corticospinal excitability in old adults compared with young adults (9).
Independent and common inputs to motor neurons. Spinal motor neurons receive tens of thousands of synaptic inputs from a multitude of different sources, with many synaptic inputs shared across motor neurons within and across populations of motor neurons (52). In addition to age-associated changes in the timing and magnitude of synaptic inputs onto motor neurons, changes in the common synaptic inputs across many motor neurons also play an important role in coordinat- ing muscle activation in old adults. One frequently used method to assess coordination of motor neuron activity is to quantify the common modulation of synaptic inputs from central and peripheral sources onto spinal motor neurons, which influences motor function (29,149). This common synaptic input is received by a population of motor neurons and drives the discharge rates of these motor neurons at a common low frequency that affects the force exerted by the muscle (112). Common modulation of the motor unit activity is typically quantified by performing a coherence analysis between two physiological signals, including motor unit action potential discharge times, and EMG and electroencephalographic (EEG) recordings. Coherence is a measure of the correlation between two signals in the frequency domain and is quantified between values of zero to 1. Coherent motor unit activity may be critical during the performance of fine motor skills, including that by older adults. However, the association between oscillatory activity (identified by coherence analysis) and motor function is unclear. Although it has been suggested that the oscillatory activity may improve the recruitment of motor units and facilitate coordination of motor unit activity across multiple muscles, the coherent activity is more commonly viewed to be maladaptive; for example, in the generation of essential tremor and Parkinsonian tremor (48,124). Fine motor skills may require the separate and independent activation of multiple muscles, which would require decoupling from this coherent activity that occurs throughout the neuromuscular system to produce precise forces and movements (4). Consistent with this view, a number of studies have reported increased coherence between different physiological signals in old compared with young adults. First, old adults demonstrate greater coherence than young adults between the discharge times of pairs of concurrently active motor units in the first dorsal interosseus from 2 to 20 Hz during the performance of an index finger abduction task (133). Second, EEG-EMG and magnetoencephalographic (MEG)-EMG coherence was increased in old adults across a wide frequency range  during the performance of hand tasks (68,73). Third, interhemispheric EEG-EEG coherence between 1 to 30 Hz was increased in old adults compared with young adults during attention-demanding tasks (102,142). These studies taken together point to an inability of old adults to appropriately regulate and coordinate the common synaptic inputs onto spinal motor neurons. Because the common oscillatory synaptic input likely represents the effective neural drive to muscle (112), greater fluctuations in this common oscillatory input in old adults likely explain in large part the increased variability in motor unit action potential discharge rate (47,67,147) and impaired motor performance in old compared with young adults (133).

Neuromuscular Junction
The neuromuscular junction is the synapse between an alpha motor neuron and a skeletal muscle fiber (55). Aging is associated with remodeling of the neuromuscular junction and impaired neuromuscular transmission (31,55) that may increase the variability in motor unit activation among old adults. Typically, the amount of the neurotransmitter released in the postsynaptic membrane of the neuromuscular junction is in excess of what is necessary, so there is a safety factor (157) to ensure efficient transmission of the potential even in high-demand events such as a fatiguing contraction. The neuromuscular junction typically allows a 1:1 transmission of the action potential from the motor neuron to the muscle fiber, but this reliability declines with aging (157), and it probably reflects age-related changes and remodeling of the neuromuscular junction both morphologically and physiologically (55).
Evidence for age-related changes in the neuromuscular junction exists mostly in reports from animal studies with indirect evidence in humans. Morphological changes observed in aged animals include increased presynaptic branching, larger and more dispersed postsynaptic endplate areas, and a reduction in the coupling of the presynaptic vesicles and postsynaptic receptors [e.g., see reviews by Deschenes (31) and Hepple and Rice (55)]. The presynaptic nerve terminal branches become greater in number and length and are accompanied by a larger area of vesicles (31). However, as neuromuscular branching increases, vesicle density decreases along with the acetylcholine stores and release (31). Primary factors underlying these changes and instability of the neuromuscular junction probably involve greater mitochondrial dysfunction, oxidative stress, inflammation, and neurodegeneration, all of which increase with aging (144). The functional consequences of these agerelated changes in the neuromuscular junction include impaired neuromuscular transmission, which may result in complete blocking of the action potential (impulse blocking) and inefficient and more variable activation of the muscle. Ultimately, such changes may lead to sarcopenia. For example, in mice, a reduction in fiber size was shown to be preceded by neuromuscular junction disruption (32); however, whether sarcopenia is secondary to this disruption at the neuromuscular junction in old adults or vice versa is not understood (144).
In humans, insight into age-related changes of the neuromuscular junction is observed with increased "jitter" (12) and "jiggle," which are both detected with EMG recordings of single motor unit potentials (56,137). Jitter is the variability in the time interval of two muscle fiber action potentials from different fibers of the same motor unit and is larger with greater collateral sprouting. Jiggle is the variability in the shape of a single motor unit across consecutive discharges (56) and is related to decreased transmission stability due to increased jitter and impulse blocking (55). Cross-sectional studies show that jiggle is greater in very old adults (age, ϳ80 yr) compared with young adults (age, Ͻ30 yr) (56), and in very old elite masters athletes (age, ϳ80 yr) jiggle is reduced relative to age-matched nonathletes (119). In contrast, endurance training can alter the morphology of the neuromuscular junction and neuromuscular transmission in several muscles of young mice, but this effect is dampened or does not occur with training in old mice (31,33). Thus, physical activity and training may not always preserve the neuromuscular junctions in old and very old muscle, and other factors may preserve their function in very-old-age lifetime exercisers (119). Calorie restriction, for example, can reduce oxidative stress and preserve function of the neuromuscular junction with aging in old mice (115). Longitudinal studies, however, would provide greater insight into the role that exercise and other factors such as nutrition and genetics play in the change in neuromuscular junction stability that occurs with advanced age.

985
The Aging Neuromuscular System • Hunter SK et al.

Muscle Fiber
Advanced age is accompanied by a net loss of innervated muscle fibers and a reduction in the fiber size of the surviving motor units (93). Fiber loss results from motor unit remodeling (a greater rate of fiber denervation relative to reinnervation) (55) and also may include increased oxidative stress and apoptosis, which reduces satellite cell regeneration of muscle fibers (11,111). Old adults typically have a larger innervation ratio than that of young adults so that there are more muscle fibers within each motor unit [e.g., see Fling et al. (43)], and the fibers of a single motor unit are more likely to be clustered within a specific region of the muscle (fiber-type grouping) (91).
Fibers from the existing motor units are generally smaller in old and very old adults compared with young adults, especially in lower limb muscles [e.g., see Hunter et al. (62), Lexell et al. (93), and Venturelli et al. (152)]. The age-related reduction in fiber cross-sectional area usually occurs across all fibers, although there is some evidence that this atrophy occurs to a greater extent in fibers expressing myosin heavy chain (MHC) II isoforms in old adults younger than ϳ80 yr (62,93,123). In very old adults (older than ϳ80 yr) atrophy can be quite marked in all fibers (123). These age-related reductions in fiber size, however, are not reported by all studies and may differ between muscles and the sexes. For example, compared with young adults, there were no differences in the size of vastus lateralis muscle fibers in old men (age, ϳ69 yr) (109) and very old men and women (age, 80 -88 yr) (122,148,152), nor in the fibers of the biceps brachii of very old women (age, ϳ88 yr) (152). Aside from the possible influence of lifestyle modifiers such as physical activity or nutrition, the disparities of findings may be influenced by sampling bias of cross-sectional designed studies, the variability associated with muscle biopsies, and small numbers of subjects in some studies.
Smaller fibers in old adults compared with young adults are associated with lower protein synthesis and less satellite cells in type II fibers, which are needed for skeletal muscle growth and repair (71,153,154). Some fibers of older adults are also quite small and angular in shape (92), possibly representing denervation (123). Furthermore, with aging, muscle fibers of animals and humans are more likely to coexpress multiple MHC isoforms in single fibers, suggesting age-related changes in denervation-innervation of muscle fibers (3,109,122). Myosin ATPase histochemical staining techniques, however, are not able to easily identify coexpression of the different MHCs (123), so caution is required in categorizing fibers in older adults on the basis of type I and type II categories alone without considering the MHC isoform composition.
In general, the smaller muscle of old adults (age, approximately 70 -73 yr) has been shown to have a smaller proportional area that expresses MHC IIa isoforms compared with that of young adults [e.g., D'Antona et al. (26) and Lamboley et al. (87)]. This is consistent with studies that show increased atrophy of type II fibers (identified with myosin ATPase histochemical staining) compared with type I fibers in old adults (age, up to ϳ83 yr) (62,93). Furthermore, in old and very old adults there is evidence of a shift to a proportional increase in the expression of the MHC IIx isoform (26,109,122,152), with less coexpression in very old master athletes compared with age-matched nonathletic controls (122).
Although the size of muscle fibers decreases with advanced age, their specific tension (force per unit area of muscle) may be preserved with aging especially when young and old adults are matched for physical activity levels (148,152). This finding, however, is not universal because other studies report lower specific tension across different fibers of muscle in old individuals (mean age, approximately 70 -73 yr) (26,87,90,122) and in very old master athletes (age, ϳ80 yr) (122). Lower intracellular calcium (endogenous and sarcoplasmic reticulum) in the skeletal muscle fibers of old adults (ϳ70 yr) compared with that in young adults, and a lower calcium sensitivity to activation could in part be responsible for an age-related reduction in the specific tension (87). In fact, there are several age-related mechanisms informed by animal studies that can reduce effective excitation-coupling and be responsible for lower specific tension in old skeletal muscle [see review by Delbono (28)]. In contrast, specific tension of MHC I and IIa single fibers in old women (age, ϳ68 yr) was greater than that in activity-matched young women and men, possibly due to increased myofilament lattice stiffness (109). Specific tension of all fiber types in old men and women (age, ϳ74 yr) were elevated in a 3-yr longitudinal follow-up study (128). Taken together, the age differences in the specific tension of single fibers may not be uniform across the population of motor units and people. Physical activity may in part modulate age differences in specific tension observed in old adults Ͻ75 yr, but does not seem to be able to prevent the reduction of specific tension even in very old athletes (122).
Muscle fibers of old adults also typically exhibit reduced contractile speed compared with those of young adults. For example in single fibers, lower rates of force development in active and inactive older adults (age, ϳ80 yr) (122) and lower maximal shortening velocity (26,85,90,122) are reported, although not all studies show this to be the case in active, older adults (age, ϳ78 yr) (148). The reduced velocity of the muscle fibers is associated with slower cross-bridge kinetics (109,122). Shortening velocity can be 3ϫ faster and peak power output up to 6ϫ greater in MHC IIa fibers compared with MHC I fibers (148), so that any shift toward a reduction in MHC IIa isoforms in old muscle would reduce its contractile speed (90).
Consistent with the slowing of contractile speed, the rate of muscle relaxation of whole muscle slows with aging [e.g., see Callahan and Kent-Braun (14), Hunter et al. (62,63), and Molenaar et al. (110)], probably due to slower cross-bridge mechanics and slower rates of Ca 2ϩ uptake into the sarcoplasmic reticulum. For example, in old women (age, ϳ71 yr), the rate of relaxation of the quadriceps muscles was associated with a lower proportional area of type IIa fibers (% area of sample), and lower sarcoplasmic reticulum maximal rate of Ca 2ϩ uptake and Ca 2ϩ ATPase activity analyzed from biopsy samples of the vastus lateralis muscle (62). In the same study, 12 wk of strength training of the quadriceps muscles decreased the age difference in sarcoplasmic reticulum Ca 2ϩ uptake and Ca 2ϩ ATPase activity in the old women, but the Ca 2ϩ uptake and Ca 2ϩ ATPase activity in the young women did not change (62). Thus, the initial age difference in calcium kinetics was due in part to age differences in physical activity (62). Thus, the slower and small muscle fibers of old adults can be influenced by sedentary behavior, which can also increase the variability of fiber characteristics among old adults.

AGE-RELATED CHANGES IN MOTOR PERFORMANCE
This section will highlight the consequences of altered motor unit morphology and properties on motor function, with reference to distinct areas of motor performance including strength, power, contraction velocity, fatigability, and force steadiness.

Strength and Power
Strength. The age-related reduction in maximal isometric strength largely parallels the loss of muscle mass (44,107). Lower specific force (force per unit cross-sectional area) of limb muscles in old adults is explained in part by greater infiltration of fat and connective tissue that may be ameliorated with physical activity (46). Based on cross-sectional studies of lower limb muscles, strength is usually reduced ϳ10% per decade with reductions starting at approximately 40 -50 yr of age, with evidence of accelerated declines in very old age so that the average strength of an 80-yr-old can be ϳ40% that of a same-sex 20-to 30-yr-old (61,94). The findings from cross-sectional studies, however, may underestimate the actual age-related reductions of strength in old adults [e.g., see  (106) and with chronic disease in old adults (126). Furthermore, there can also be variability in strength between muscles with advanced age because the age-related reductions in isometric strength are typically greater in muscles of the lower limb such as the knee extensors than for the upper limb such as the finger flexor and elbow flexor muscles (44,61,125,152). This muscle group difference may be explained in part by the variability in the rate of sarcopenia in response to differing use of muscles groups between the limbs (27,61,66,151).
Maximal torque during isokinetic dynamic contractions is also reduced with advanced age in both men and women [e.g., see Lindle et al. (94)]. At fast speeds of shortening (concentric) contractions, the age-related reduction in maximal torque for some muscles such as the knee extensor muscles is larger than for slower speeds (89), and this difference can vary across muscle groups (125) (see Power). The age difference in maximal torque, however, differs with contraction mode with less of an age reduction in torque for lengthening (eccentric) than shortening (concentric) contractions (81,94,120,121,125). For example, for lengthening contractions performed between 15 to 360 deg/s with the ankle dorsiflexor muscles, old men (age, ϳ76 yr) had a larger eccentric:isometric torque ratio than young adults (121). The greater preservation of torque during lengthening contractions may be larger in old women than old men (81,94). The mechanisms for the age and sex-related differences in lengthening contraction torque is not due to cocontraction (121) but may involve elastic, structural, and cross-bridge properties of the muscle. Specifically, the contributing mechanisms likely involve slower cross-bridge kinetics, an increased proportion of weakly bound compared with strongly bound cross-bridges, and a stiffer musculotendinous complex in the muscle of old adults compared with that of young adults (81, 90, 120, 121).
Power. Age-related reductions in maximal power are greater in magnitude than for maximal isometric strength (105,125,127,134,143). Power is also more strongly associated with functional performance tasks such as stair climbing, ambulation, and rising time from a chair, than age-associated reductions in isometric strength (127). Importantly, the measure of leg power during shortening contractions is predictive of functional tasks and disability (127).
The age difference in power widens as the velocity demands increase, particularly for the knee extensor muscles, with some older adults unable to reach velocities of Ͼ270 deg/s (14,89,127). The greater age reductions in power of the lower limb at fast velocities are associated with the reductions in maximal shortening velocity of single fibers, especially MHC IIa fibers (85,90), reduced muscle mass, and altered muscle architecture (127,143). Furthermore, the role of inadequate activation of the motor units during fast contractions performed by older adults is discussed below (80,127), and its effect on agerelated reduction in power probably underestimated. Increasing physical activity, in particular with resistance and power training of the lower limbs, can have large effects on increasing power and decreasing risk of disability in old and very old adults who are close to a critical threshold of functional independence (127).  148)]. There is also evidence of lower rates of force development of the lower limb muscles in old adults (age, ϳ76 yr) compared with young adults (80). Because in old adults the rapid force development of isometric contractions with ankle dorsiflexor muscles is greater during voluntary than electrically evoked contractions (by ϳ10%), inadequate activation of the motor units likely contribute to the slower force development in old adults (80) (Fig. 2A).
Voluntary activation. During a maximal-effort isometric contraction and with adequate practice, old adults can achieve high levels of activation, although this may vary with the muscle group involved, velocity of contraction, physical activity level, and age of the old adult (82). relaxation times were thought to compensate for the lower maximal discharge rates (22,72), but these associations are not found in all muscles (23) and are probably not causally related (45,55).
Voluntary activation of motor units can also be assessed by stimulating the muscle or motor cortex during maximal efforts and is most easily performed during isometric contractions (141). Age differences in voluntary activation can depend on contraction velocity, the muscle groups involved, whether practice of maximal efforts is provided, and physical activity levels (82). Some studies show that old adults consistently have lower or more variable activation levels during maximal isometric contractions than young adults [e.g., De Serres and Enoka ]. Young adults do not need as much practice to achieve consistent levels of activation, but practice can improve activation levels in old and very old adults in one session or even within a session (63,65). Furthermore, less regular physical activity among old adults may exacerbate age differences in voluntary activation (49). Accordingly, old men and women are more variable in their activation levels across trials and among themselves (63,158) (Fig. 2D). Implications from these findings are that the activation may be inadequate during tasks that require/allow only one attempt as routinely observed in real-world situations (e.g., recovering from a disturbance in balance to prevent a fall). Although maximal voluntary contractions that are practiced in the laboratory allow comparison of the best attempts of young and old adults, the larger variability in voluntary activation within older adults potentially exacerbates the age-related declines in strength and power required during daily tasks.
During slow-to-moderate velocity-dynamic contractions, young and old adults appear to have similar levels of voluntary activation for upper limb muscles (V. Rozand, J. Senefeld, and S. K. Hunter, unpublished findings) and lower limb muscles (81) when assessed with the interpolated stimulus technique. This technique, however, is difficult to accomplish during fast contractions at which age differences in power become large (89). Insight has been gained from single motor unit recordings showing lower rates of motor unit discharge and torque development in old adults immediately before contractions requiring rapid force development (80) (Fig. 2, A and B). Such differences in neural activation of the muscle may also occur with aging during fast-velocity contractions (155), but further exploration is needed in older women and across many muscle groups in both sexes.

Fatigability
Performance fatigability is the reduction in force or power of a muscle in response to exercise (40). The consequences of greater fatigability with aging are significant because fatigability further exacerbates the age-related loss of strength and power in old adults observed before exercise. In addition, fatigue-induced variability of force or power, which is greater with aging (78,134), may further impair performance of daily activities.
Age-related changes in the inputs to the motor neuron and motor unit morphology and physiology as described earlier, can alter the rates at which the motor unit is stressed in young, old, and very old adults during a fatiguing task, resulting in an age difference in fatigability. Furthermore, when the demands of the task are altered, such as the intensity, velocity and mode of contraction (dynamic vs. isometric tasks), the muscle group involved, and age of the older adult, the magnitude of age difference in fatigability also changes [e.g., Christie et al. For maximal and submaximal isometric fatiguing contractions, healthy old adults (age, 60 -75 yr) are typically less fatigable than young adults (77) even when matched for strength (60) for both men and women (59). The greater fatigue resistance of old adults is associated with slower contractile properties, a shift to a lower proportional area of fibers expressing MHC IIa isoforms, and a lower reliance on glycolytic metabolism in old adults compared with young adults (15,62,63,77). Consequently, during isometric fatiguing contractions, the skeletal muscles of old adults have a lower accumulation of inorganic phosphate and hydrogen ions that potentially interfere with force production (88). The fatigue resistance of sustained isometric contractions may be relevant only in old adults aged Ͻ75 yr; adults aged Ͼ75 yr exhibited greater fatigability in the lower limb rather than the fatigue resistance observed in the those Ͻ75 yr (70). More studies are needed to determine the mechanisms for the shift in fatigability from the old to very old age groups across many muscle groups and determine the effect on physical function and disability.
Cortical inputs, however, are often increased during daily fatiguing tasks (e.g., increased attention), and these inputs can influence performance of fatiguing contractions differently for young and old adults. This was observed by imposing a difficult cognitive challenge (mental math) while a participant sustained a submaximal isometric contraction with the elbow flexors muscles (75,116,160) and ankle dorsiflexor muscles (150). Young women, for example, exhibited greater fatigability (decreased time to task failure) only when a high cognitive challenge was imposed during elbow flexion contraction compared with the control task (75,160), whereas the time to task failure for old women was reduced when either a low-or high-level cognitive challenge was imposed (116). The mechanisms for the greater fatigability with superimposition of a cognitive challenge are not understood. However, these studies are important because imposition of a cognitive challenge more closely resembles real-life activities, which commonly require dual tasks. Furthermore, the variability between the task durations (coefficient of variation of the time to failure across the tasks with and without a cognitive challenge) can be much larger for old adults than for young adults (150). This was observed in the ankle dorsiflexor muscles despite no differences in the mean time to failure with and without the cognitive challenge imposed (Fig. 3, A and B) (150). These data suggest that a lack of age difference when group averages of best performances are compared may not effectively detect the poor motor performance among old adults.
The age difference in fatigability also depends on the contraction velocity. For example, the age difference in fatigability observed for isometric contractions was diminished for a dynamic task with slow-to-moderate velocity contractions (Ͻ180 deg/s) for both the elbow flexor and knee extensor muscles (13,24,161). During high-velocity concentric contractions, old and very old adults exhibit greater fatigability than young adults when contractions are performed with the lower limb muscles (14,25,105,155); for example, for contractions Ͼ180 deg/s with the knee extensor muscles, old adults were more fatigable (greater reductions in power) than young adults (14,24). The mechanism for the velocity-dependent age difference in fatigability appears to involve processes that are associated with contractile speed, a slower shortening velocity that occurs with age (7,14,161), and also a reduced ability to rapidly activate the agonist muscle during plantar flexion for old men aged ϳ78 yr (155). The extent that inadequate activation during repeated fast contractions contributes to the age differences in fatigability in women, is not known. It also appears that the fatigability of fast contractions with advanced age is even larger in very old men (age, Ͼ80 yr) than men aged 60 -70 yr (105). There are gaps in the literature in understanding the mechanisms for the greater fatigability with aging, particularly in women, and in general the effect of fatigability in predicting functional performance, frailty and disability.
An important but understudied aspect of performance fatigability that is altered with age is the variability of force or power during the fatiguing task. For example, variability of the maximal torque obtained during repeated dynamic contractions of the knee extensor muscles (120 deg/s) was greater in old women compared with young women and old women who were not mobility impaired (78). Similarly, maximal velocity during a dynamic fatiguing task with the knee extensor muscles was more variable [a larger coefficient of variation (CV)] for old adults compared with young adults (134) (Fig. 3D). However, there were no age differences in variability for the same task conducted with the elbow flexor muscles (134). Although the mechanism was not identified, these data are consistent with increased variability in voluntary activation and motor performance with aging and suggests that neural mechanisms may contribute to increased variability of motor output in older adults across a fatiguing task.

Force Steadiness
Age-related changes in motor unit behavior affect the control of force among old adults. Typically, old adults are less steady than young adults (39), especially during light-load tasks that are often required during activities of daily living (79). One metric of force steadiness is the amplitude of force fluctuations around a target and is quantified as the standard deviation (SD) of the force (39). In young and old adults, the SD of isometric force is associated with the mean target force so that the SD is larger in high-intensity than low-intensity contractions, commonly referred to as signal-dependent noise

989
The Aging Neuromuscular System • Hunter SK et al. (50,138). To allow comparisons between contractions and among groups of individuals who have large strength differences (e.g., young vs. old), the SD is normalized to the mean force and expressed as the CV (39). Motor unit behavior and the inputs to the motor unit are largely responsible for the magnitude of force fluctuations (41). Single motor unit recordings and decomposition of surface EMG from multiple-array electrodes demonstrate that the variability of the motor unit discharge rate at low frequencies (1-2 Hz) can explain up to ϳ70% of the force steadiness in young adults (41). Supraspinal centers are also involved in the fluctuations of force in young adults and include both cortical and subcortical areas in the motor area and frontal lobes such as the putamen, insula, and contralateral superior frontal gyrus (162). Although previous research has found increased coherence in higher frequency bands (Ͼ2 Hz) in old compared with young adults (68,73,102,133,142), whether age-related differences exist in coherence at low frequencies (Յ2 Hz) is not known. Changes within these cortical areas with advancing age might be expected to similarly influence low-frequency inputs onto the motor neuron pool.
The larger CV of force during isometric contractions (decreased steadiness) with advanced age occurs in both the upper-and lower-extremity muscles, primarily at lower con-traction intensities [2-10% maximum voluntary contraction (MVC)] (117,150). At these low forces, greater variability in the discharge rates of motor units in old adults are primarily responsible for the larger CV of force that occurs with aging (47,147) and not the larger size of the motor units as was once thought (39,140). The source of the age-related variability in discharge rates could arise from many different sources, including the synaptic input, the intrinsic properties of the motor neuron, and a more variable after-hyperpolarization, which is known to impair the time course of a subsequent action potential (101). All these alterations in the aging motor neuron potentially can induce greater oscillations in the neural signal and synaptic noise (101) that ultimately will increase the variability of discharge rate of motor units in older adults. Simulation studies show that synaptic noise primarily accounts for the force fluctuations at very low contraction intensities (34), and therefore may explain the age differences.
Age-related differences in steadiness are further exacerbated when old adults are required to perform a dual task that involves simultaneously maintaining a force steadiness task and cognitive challenge (116,117,150). For example, force steadiness of the elbow flexor muscles was reduced (greater CV of force) when a cognitive challenge (mental math) was performed during isometric contractions in young and old adults (117) (Fig. 4, A and B). The greatest differences occurred during a very-low-intensity contraction (e.g., 5% MVC) for old adults and women (117) compared with young adults and men (Fig. 4, C and D). When the cognitive challenge was imposed, the increased CV of force during contractions at 5% MVC with the ankle dorsiflexor muscles was variable between old adults and women (Fig. 4, E and F) (150), although the mechanisms of this are not known.
The motor output variability in young and old adults is associated with functional performance tasks. For example, larger variability during sinusoidal tracking with the ankle dorsiflexor muscles was associated with poor reactive driving performance in old adults (95), and low force steadiness of the upper limb muscles was associated with pegboard tests of dexterity (1,97). Furthermore, force steadiness was worse in old adults who had a history of falling compared with those who did not have a fall history (age, Ͼ70 yr) (17). Interventions to improve steadiness (reduced CV of force) among old adults have usually involved practice and strength training with the aim to reduce discharge rate variability [e.g., Griffin et al. (47) and Kornatz et al. (84)]. Large improvements in force steadiness were obtained after short-term (Ͻ4 wk), light-load strength training and practice, highlighting the improvements that arise from neural sources (84,96,146). Modulation of motor unit spikes between 13 and 30 Hz indicate that cortical centers may be responsible for the reduction in the variability of discharge rate with practice in old adults (114). There is limited information on how such improvements in steadiness and discharge rate variability in response to physical training ultimately affect functional performance tasks.

CONCLUSIONS
Motor performance declines with advanced age and is accelerated in very old age (Ͼ80 yr), involving muscles that are weaker, slower, less powerful, less steady, and more fatigable during high-velocity dynamic tasks. The reduced maximal power of lower-extremity muscles is predictive of functional performance and disability, with less understanding of the predictive strength of other aspects of motor performance, including steadiness and fatigability, thus providing opportunities for high-impact studies.
An important but often overlooked aspect of age-related declines in motor performance is the large within-and betweensubject variability of many aspects of motor performance (e.g., force, velocity, and fatigability) that increase with advanced aging, and lead to low predictability of an already compromised neuromuscular system in old and very old adults. Imposing a cognitive challenge, which often accompanies reallife motor tasks, further exacerbates variability in motor performance among older adults. The mechanisms for the greater variability between trials with advanced age are not well understood, but they may have a neural origin as evidenced by increased variability in voluntary activation and slower and more variable motor unit discharge rates in old adults. Practice and enhanced physical activity appear to ameliorate some of  representative force signals of the elbow flexor muscles in a young and an old woman performed at 5% of MVC for a control and high-cognitive demand trial (subtracting by 13). C and D: CV of force for the elbow flexor muscles during contractions at 5, 7, 10, and 20% of MVC. Values are means Ϯ SE for men (filled symbols) and women (open symbols) during the control session (circles) and high-cognitive demand session (subtracting by 13) (High-CD, squares). [Data for A and B are from Pereira et al. (116), and for the 7-20% MVC from Pereira HM and Hunter SK, unpublished.] E and F: CV of force for the ankle dorsiflexor muscles during contractions at 5% MVC performed by men and women for a control (filled circles), lowcognitive demand (Low-CD, subtracting by 1; open triangles) and high-cognitive demand (High-CD, subtracting by 13; filled squares) trials. The mean is shown for each session for young (E) and old (F) adults. E and F show that old adults, especially women, exhibited greater between trial variability in CV of force with imposition of the cognitive demand. [E and F were created using data presented in Vanden Noven et al. (150).] differences in variability of performance and function of the motor unit, which are often attributed to aging. Many studies describing the change in motor performance and variability across the continuum of ages are cross-sectional in design. Thus, the future challenge is to perform adequately powered longitudinal studies to determine the involved mechanisms of impaired motor performance and the increased variability, and the effect on functional performance and disability in the old and very old men and women.

GRANTS
This work was in part supported by National Institute of Aging Grant R21-AG-045766 to S. K. Hunter.

DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the authors.