Development and aging of human spinal cord circuitries

Abstract

The neural motor circuitries in the spinal cord receive information from our senses and the rest of the nervous system and translate it into purposeful movements, which allow us to interact with the rest of the world. In this review, we discuss how these circuitries are established during early development and the extent to which they are shaped according to the demands of the body that they control and the environment with which the body has to interact. We also discuss how aging processes and physiological changes in our body are reflected in adaptations of activity in the spinal cord motor circuitries. The complex, multifaceted connectivity of the spinal cord motor circuitries allows them to generate vastly different movements and to adapt their activity to meet new challenges imposed by bodily changes or a changing environment. There are thus plenty of possibilities for adaptive changes in the spinal motor circuitries both early and late in life.

our ability to perform purposeful movements and interact with the environment depends on the integration of descending motor commands and sensory feedback signals in the neural motor circuitries of the spinal cord (Burke and Pierrot-Deseilligny 2012; Jankowska 1992; Nielsen and Sinkjaer 2002). Any single motor neuron or interneuron in the spinal cord receives convergent input from a number of descending motor tracts and sensory afferent pathways. It is the integration of the activity from all of these inputs that determines the discharge of the neuron and eventually the muscle activity when we move. This is what Charles Sherrington referred to as “integration in the final common path” (Sherrington 1906). The neural circuitries of the spinal cord ensure that information about the interaction of our body with the environment is integrated into the motor commands that reach our muscles and that all our movements are optimally adjusted to the environment. In this way, sensory feedback may also directly contribute to muscle activity through servo feedback mechanisms, thereby reducing the necessity of descending drive (af Klint et al. 2009; Grey et al. 2004, 2007; Sinkjaer et al. 2000). Importantly, sensory feedback through spinal neural circuitries may also act as error signals that can be used to optimize the performance of future movements (Choi et al. 2016). Transmission in spinal neural circuitries has been shown to adapt to changes in the environment (Fortin et al. 2009; Rossignol and Bouyer 2004) and undergo plastic changes in relation to motor learning (Wolpaw 2007).

This adaptability of the neural circuitries is also the basis of adjustments of their properties according to changes in motor demands during development and when we grow old and our bodies begin to deteriorate. Once the spinal circuitry is established early during development of the nervous system, it thus undergoes changes that are necessary to ensure that our interaction with the environment remains optimal despite significant changes in the anatomical and biomechanical properties of our various body parts. It is the purpose of this review to provide an overview of the available knowledge of these age-related adaptations in human spinal cord circuitries and attempt a critical discussion of their functional significance.

In the Beginning—Establishment of Spinal Circuitries and Our First Movements

Although human newborn babies appear helpless and with limited movement abilities compared with other animals, they have already developed the capacity to move their head, trunk, and extremities for several months before they are born, as most mothers will be able to testify. The nervous system begins to develop at day 17 after gestation, whereas differentiation of myotomes, which develop into muscles, begins in week 5 (Musumeci et al. 2015). At this time, sensory and motor connections develop and the first elementary reflex circuitries are established (Chen et al. 2003). Very soon after this (in week 7), the first movements of the fetus may be detected in ultrasound examinations (de Vries and Fong 2006). During the third trimester these movements increase in complexity and may for instance involve opening of the mouth before movement of the hand toward the mouth as evidence of coordinated activity of neural circuits that involve both the spinal cord and brain stem (de Vries and Fong 2006). Development of the human cerebral cortex begins already in the fifth postgestational week, but it is very protracted compared with other animals and does not acquire its characteristic folding into gyri and sulci until the last 2 mo before birth (Meyer 2001). The subplate, which is assumed to be important for guiding growth of axons in and out of the cortex, acquires its maximal size around the middle of the third trimester and may be seen until 6 mo postpartum (Meyer 2001; Wang et al. 2010). Consistent with this, the available evidence suggests that the corticospinal tract has reached all levels of the spinal cord well before birth and has established functional contacts with motor neurons and interneurons in the ventral horn and intermediate nucleus at the time of birth (Eyre 2003; Eyre et al. 1991, 2000). Myelination of the corticospinal and other descending motor tracts begins in the first few months of life but is not fully developed and mature until well into the teen years (Yeo et al. 2014). This optimization of conduction in descending pathways is likely to have an impact also on the spinal cord circuitries on which they impinge.

The synaptic contacts from the corticospinal tract on motor neurons and interneurons in the spinal cord have been shown to undergo considerable activity-dependent reorganization during a critical period within the first weeks following birth in rats (Gibson et al. 2000) and cats (Martin 2005; Martin et al. 2004, 2007). If the animal is prevented from using a limb or if a lesion of one hemisphere is induced during this critical period, corticospinal projections to the relevant motor neurons retract and the animal fails to develop skilled movement abilities involving the affected limb (Martin et al. 2007). There is evidence to suggest that a similar activity-dependent reorganization of corticospinal contacts in the spinal cord takes place in human subjects, but it is unclear to what extent a distinct critical period of development exists. Some evidence suggests that the period in which so-called fidgety movements (small, high-frequency movements of distal limb segments) are observed (~9–25 wk) may delineate a period in which the corticospinal tract undergoes activity-dependent reorganization (Ritterband-Rosenbaum et al. 2017). If so, this would be assumed also to affect transmission in spinal cord circuitries, and this also seems to be the case (see below).

Stretch reflexes are elicited at birth and may show early developmental changes.

Stretch reflexes are readily activated in babies born at term (Fig. 1; Leonard et al. 1991, 1995; Myklebust and Gottlieb 1993; O’Sullivan et al. 1991). Several studies have also documented the presence of stretch reflexes and their electrical equivalent, the Hoffmann reflex (H reflex), in preterm babies born as early as week 25 (Allen and Capute 1990; Hakamada et al. 1988). In adults, stretch reflexes are usually only elicited in the muscle from which the Ia afferents originate (Schieppati 1987), although collaterals of Ia afferents are widely distributed to motor neurons of synergist muscles throughout a limb (Mendell and Henneman 1968; Meunier et al. 1993; Scott and Mendell 1976). Apparently, the synaptic strength of the Ia afferent input to these synergistic motor neurons is insufficient to make them discharge at rest in adults. This may be because there are too few synaptic contacts to the individual motor neurons and/or because the synapses are functionally suppressed by presynaptic inhibition (Rudomin and Schmidt 1999). There is good evidence to support that the latter possibility may be the case (Hultborn et al. 1987; Rudomin and Schmidt 1999).

Fig. 1.

Stretch reflexes and reciprocal inhibition in newborns, adults, and elderly subjects. Descending pathways, such as the corticospinal tract (green lines), have only just innervated spinal interneurons (blue) and motor neurons (gray) in newborns (A) and continue maturation throughout childhood and adolescence to reach an adult maturation level in the late teens (B). In elderly subjects (C), reduced transmission in the corticospinal tract has been found. Muscle spindle afferents (red lines) are fully developed at birth and cause excitation of not only the muscle from which the afferents originate but also heteronymous and antagonist muscles. In adults, stretch reflexes are only elicited in the muscle from which the Ia afferents originate, although subthreshold excitation may also be seen in heteronymous motor neurons. This restriction of excitation may be explained by increased presynaptic inhibition of Ia afferents or reduced excitability of motor neurons during development. In elderly subjects muscle spindles, sensory afferents, and motor neurons degenerate, leading to further reduction of stretch reflexes (dotted red and green lines in C). Interneurons mediating reciprocal inhibition in adults (blue) are excitatory in newborns because of the lack of the potassium-chloride cotransporter KCC2, and activation of muscle spindle afferents or corticospinal pathways may therefore cause reciprocal excitation and coactivation of antagonists. In elderly subjects, reciprocal inhibition is reduced, presumably because of degeneration of muscle spindle afferents and descending fibers (dotted red and green lines in C). Note that this is a schematic drawing, which for the sake of clarity only shows some of the possible explanations for the changes that occur in stretch reflexes and reciprocal inhibition from birth to maturation and aging.

Several studies have suggested that stretch reflexes are more easily elicited in newborns than in adults and that they may also be elicited in synergist heteronymous muscles (Leonard et al. 1995; Myklebust and Gottlieb 1993; O’Sullivan et al. 1991, 1998). These studies indicate that reflexes become gradually weaker up until ages 2–4 yr, when stretch reflexes are only elicited in homonymous muscles, similar to adults (Leonard et al. 1995; Myklebust and Gottlieb 1993; O’Sullivan et al. 1991). It is noteworthy that responses in antagonist muscles may also be seen in infants until ∼9–12 mo of age (Mc Donough et al. 2001; Myklebust and Gottlieb 1993) or maybe even longer (Leonard et al. 1995). These studies should be interpreted with some caution, since all of them are based on mechanical stimuli applied to muscle tendons. Therefore, it cannot be fully excluded that the observed irradiation of reflexes is caused by propagation of the mechanical stimulus and activation of (sensitive) muscle spindles in heteronymous and antagonist muscles. The restriction of the irradiation during development may thus be explained simply by body growth, causing less favorable conditions for propagation of the mechanical stimulus to other muscles. It would be interesting if other electrophysiological or biochemical techniques could confirm that the early irradiation of reflexes and the subsequent restriction in the first years after birth reflect genuine changes in the central connections of the sensory afferents.

If we put this concern aside, the observed irradiation of reflexes to heteronymous muscles may be considered unsurprising given the known widespread distribution of Ia afferent collaterals to heteronymous motor neurons in both the cat and human spinal cord (Meunier et al. 1993; Nelson and Mendell 1978; Scott and Mendell 1976). It is not unlikely that the synaptic effects of sensory afferents are larger and that motor neurons are more excitable in newborns than in older children and adults. This could be explained by numerous factors such as intrinsic differences in the properties of sensory afferent terminals and motor neurons as well as differences in presynaptic and postsynaptic inhibitory and excitatory influences on the stretch reflex circuitry (Tahayori and Koceja 2012).

Developmental switch from excitation to inhibition in spinal pathways between antagonists.

The developmental change from reciprocal excitation to inhibition around 9–12 mo of age is especially noteworthy. Reciprocal inhibition was described originally by Charles Sherrington in the cat spinal cord and later shown by John Eccles to be mediated by spinal glycinergic inhibitory interneurons that project to antagonist motor neurons and receive collaterals from Ia afferents originating in agonist muscles (Eccles et al. 1956; Sherrington 1906, 1932). Subsequent animal and human studies have documented that the interneurons receive convergent input from supraspinal descending fibers, including the corticospinal tract, and are activated as part of the descending command for movements that involve selective extension or flexion (Fig. 1; Crone et al. 1987; Crone and Nielsen 1989; Hultborn and Lundberg 1972; Hultborn and Udo 1972; Jankowska et al. 1976; Nielsen et al. 1993). This pathway is therefore fundamental for the regulation of limb movements, and it may therefore be surprising that it appears to start out by being excitatory early in development. However, there is convincing experimental evidence and a profound mechanistic understanding to support this.

Inhibitory neurotransmitters such as glycine and GABA have been shown to be dominantly excitatory during early development (Kilb 2012; Lynch 2004; Plotkin et al. 1997; Vinay and Jean-Xavier 2008). In the rat spinal cord, inhibition becomes dominating around the end of the first postnatal week, which, as mentioned above, is considered a critical period during motor development (Clowry et al. 2006; Gibson et al. 2000), and is associated with upregulation of the potassium-chloride cotransporter KCC2 (Plotkin et al. 1997; Rivera et al. 2005). KCC2 maintains a low intracellular chloride concentration. Opening of GABA- or glycine-regulated chloride channels therefore leads to influx of ions and hyperpolarization in the presence of adult levels of KCC2. During early development the absence of KCC2 leads to efflux of anions and depolarization when GABA and glycine are released at synaptic sites. The excitation by GABA and glycine early during development is essential for establishment of long-lasting synaptic contacts (Kilb 2012; Vinay and Jean-Xavier 2008). This is an activity-dependent process, which depends highly on oscillatory release of neurotransmitters to bind presynaptic and postsynaptic neurons together (Graziadio et al. 2010; Minlebaev et al. 2011). This is also consistent with the idea that the first postnatal week in the rat is a critical period in which considerable activity-dependent reorganization of the corticospinal and spinal networks takes place (Clowry et al. 2006; Gibson et al. 2000). If a similar switch from GABAergic and glycinergic excitation into inhibition takes place during early postnatal human development, this may explain the observed changes in stretch reflex irradiation and especially the change from excitation to inhibition of antagonists within the first year of development. Analysis of expression of the genes for subunits of glycine receptors in human infants suggests that a switch may take place ∼6–9 mo after birth (Bar-Shira et al. 2015). This coincides quite well with the change from reciprocal excitation into inhibition in healthy infants around 9 mo of age (Mc Donough et al. 2001; Myklebust and Gottlieb 1993; O’Sullivan et al. 1998), and it also follows the period of fidgety movements in healthy human infants in which considerable corticospinal and spinal reorganization is assumed to take place (Eyre et al. 2001; Hadders-Algra 1993; Ritterband-Rosenbaum et al. 2017). It is interesting in this relation that infants younger than 6–9 mo have been shown to exert considerable coactivation of antagonist muscles compared with older infants, who show a clearer reciprocal activation pattern (Thelen 1985). It is an exciting possibility that this is a behavioral correlate of a change from excitation into inhibition in the reciprocal Ia pathway linking antagonist muscles (Fig. 1).

The observation that infants may show alternating rhythmic activity already at birth (spinal stepping) and appear to have a relatively well-developed spinal locomotor circuitry (Lam et al. 2003; Pang et al. 2003; Pang and Yang 2001; Yang et al. 1998, 2004) does not contradict the above line of thought. Alternating locomotor activity has been shown to occur independently of reciprocal inhibition in the cat spinal cord (Pratt and Jordan 1987).

Integration of reflexes into voluntary movements during development.

As infants grow older and begin to move around to explore their environment, reflexes are integrated into the central motor commands, which makes a continuous adaptation of muscle activity according to the environmental conditions and changes in body proportions and properties possible (Tahayori and Koceja 2012). This is a gradual maturation process, which takes years to complete. One example is the modulation of the stretch reflex circuitry in relation to locomotion (Hodapp et al. 2007; Willerslev-Olsen et al. 2013). Stretch reflexes as well as the electrical surrogate of the stretch reflex, the H reflex, are modulated with the excitability of the spinal motor neurons and the amount of transmitter released from Ia afferent terminals, which in turn depends on the level of presynaptic inhibition (Fig. 2; Capaday and Stein 1986; Ethier et al. 2003; Simonsen and Dyhre-Poulsen 1999; Sinkjaer et al. 1996). In both adults and children, the reflexes are therefore large in the stance phase and small in the swing phase during walking (Hodapp et al. 2007; Willerslev-Olsen et al. 2013). However, the suppression of the reflexes in the swing phase is more efficient in adults than in children, presumably because of more efficient activity in inhibitory pathways mediating reciprocal inhibition and presynaptic inhibition (Faist et al. 1996; Petersen et al. 1999; Willerslev-Olsen et al. 2013). Reflexes are also smaller in the stance phase in adults, presumably because of larger presynaptic inhibition of the Ia afferent terminals (Capaday and Stein 1986; Ethier et al. 2003; Faist et al. 1996; Hodapp et al. 2007). Reflexes are reduced gradually over several years during late childhood, so that reflexes elicited in the stance phase have reached the adult level at the age of ∼12–14 yr (see Fig. 2 and Hodapp et al. 2007; Willerslev-Olsen et al. 2013). It has been suggested that this may reflect a gradual increase in presynaptic inhibition of Ia afferents as children optimize their walking pattern during late childhood (Hodapp et al. 2007; Willerslev-Olsen et al. 2013). At least the changes in the reflexes parallel changes in a number of gait parameters that reflect the efficiency and stability of gait (Norlin 1981). Estimates of corticospinal tract maturation and functional corticospinal drive during gait also reach adult values around the same time (Nezu et al. 1997; Petersen et al. 2010; Yeo et al. 2014).

Fig. 2.

Age-dependent changes in stretch reflex activity during gait in children. A and B: when a fast (300°/s) dorsiflexion perturbation of the ankle joint is imposed late in the stance phase during treadmill walking (A), a series of stretch reflex bursts may be seen in the soleus EMG (B). The red line in B is the average of the soleus EMG during 30 steps with perturbation, whereas the black line is the average of 30 steps without perturbation. The time of the perturbation is marked by an arrow. The first reflex burst, M1, is seen at a latency of 40 ms, whereas the subsequent M2 reflex has a latency of 60 ms. In both cases the latency is so short that transmission in spinal networks has to be involved. The M1 reflex has been shown to be mediated by the monosynaptic Ia reflex pathway, whereas the M2 reflex is likely caused by transmission in at least disynaptic, slower conducting group II pathways. M3, on the other hand, has a sufficiently long latency (90 ms) to be caused by transmission in a supraspinal pathway, and evidence has been presented that supports that the motor cortex and corticospinal tract are involved in mediating this reflex. C–E show the size of the 3 reflex responses in relation to age in 44 healthy children aged 5–14 yr. Each circle marks the size of the respective reflexes in an individual child. As can be seen, M1 (C) and M2 (D) decreased with age, whereas there was no change in M3 (E).

It is notable that long-latency (M3) reflexes appear not to show an age-dependent decline similar to the short-latency (M1) and medium-latency (M2) stretch reflexes (Fig. 2). There is good evidence to suggest that the M3 reflex is mediated by a transcortical pathway, and the increasing relative importance of this reflex compared with the spinal M1 and M2 reflexes may therefore be related to the maturation of the corticospinal tract (Nezu et al. 1997; Petersen et al. 2010; Yeo et al. 2014). Integration of error signals from the ankle muscles with visual and vestibular information in older children makes a context-dependent regulation of ankle joint stability in the stance phase of gait possible. This may be related to a larger versatility in gait behavior as the child grows older and obtains an optimal ability to navigate and avoid obstacles in the environment.

In the End—Changes in Spinal Circuitries with Aging

While the spinal circuitries seem to be continuously modified to meet the changing motor demands during development, the question is how these reflex networks adapt when our bodies begin to deteriorate with age and how this affects our motor control.

Age-related decrease in stretch reflexes.

Numerous studies have reported an age-related reduction in the proportion of elderly subjects in whom the Achilles tendon reflex can be elicited by a tendon tap (Bryndum and Marquardsen 1964; Milne and Williamson 1972). Carel et al. (1979) measured the half-relaxation time of the Achilles tendon reflex torque response in 1,837 women and 7,937 men from 10 to 69 yr and found a gradual slowing of the reflex with advancing age (Carel et al. 1979). Furthermore, Koceja (1993) found that the force of the reflex response was reduced and the latency increased in old compared with young subjects (Koceja 1993). These age-related changes in the stretch reflex could be due to changes in the sensitivity of muscle spindles, the transmission in the Ia afferent pathway, spinal reflex circuitries, or descending pathways affecting this transmission, or the excitability of the motor neuron and/or changes in the motor unit (for a detailed review see Mynark and Koceja 2001).

Swash and Fox (1972) investigated 22 postmortem subjects from newborn to 81 yr old and found an increase in the capsular thickness of the muscle spindle with age due to an increase in laminar collagen. This thickening might impair the muscle spindle's ability to deform, and thereby reduce its sensitivity to stretch. Furthermore, they observed a slight reduction in the number of intrafusal muscle fibers with advanced age, which could also reduce muscle spindle sensitivity (Swash and Fox 1972). Liu et al. (2005) also found an age-related reduction in the number of intrafusal fibers per muscle spindle in the human biceps brachii muscle and showed that it was mainly due to a reduction in nuclear chain fibers. Besides changes in the muscle spindle, the transmission along the Ia afferent might also be affected by aging (Liu et al. 2005).

The H reflex activates the axons of the Ia afferents directly and is therefore independent of potential age-related changes in the muscle spindle. It has now been reported in several studies that both the maximal H reflex (H_max) and the maximal M wave (M_max), which result from direct stimulation of motor axons, decrease with aging (deVries et al. 1985; Kido et al. 2004; Raffalt et al. 2015). deVries et al. (1985) found a significant 33% reduction in H_max in healthy active elderly people aged 61–74 yr compared with a young control group (18–30 yr). Increased H-reflex threshold and latency are also frequently reported with aging (deVries et al. 1985; Sabbahi and Sedgwick 1982).

It is not surprising that both H_max and M_max are reduced with aging, as it is well documented that both the diameter of muscle fibers (Larsson and Ansved 1995; Lexell et al. 1988) and the number of motor neurons (Kawamura et al. 1977; Kawamura and Dyck 1977; Tomlinson and Irving 1977; Wright and Spink 1959) are reduced with aging. Muscular atrophy has been reported to begin already at 25 yr of age, with a 10% decrease in fiber area at the age of 50 after which the atrophy accelerates to reach ~50% at the age of 80. Especially type II muscle fibers are affected (Lexell et al. 1988), which partly explains the age-related reduction in muscular strength. On the basis of linear regression analyses of the number of large-diameter myelinated fibers in postmortem human ventral lumbar spinal roots, the loss of lumbar motor neurons was estimated to be ~5% per decade (Kawamura et al. 1977). However, Tomlinson and Irving (1977) reported no loss of motor neurons until after 60 yr of age, after which they observed a loss of 8.8% from 61 to 70 yr and 21.2% from 71 to 80 yr. Whereas changes in H_max and M_max may both reflect muscular changes, age-related changes in the H_max-to-M_max ratio has been investigated in several studies as a measure of spinal reflex excitability (deVries et al. 1985; Sabbahi and Sedgwick 1982; Scaglioni et al. 2003). However, whereas some studies show an age-related reduction in the ratio (deVries et al. 1985), others have not (Sabbahi and Sedgwick 1982; Scaglioni et al. 2003). Since H-reflex size also depends on the transmission between the Ia afferent fibers and the alpha motor neurons, the efficiency of spinal reflex pathways making inhibitory contact pre- and postsynaptically must also be considered.

Reduced efficiency of spinal reflex pathways.

Several studies have found reduced efficiency and impaired modulation of spinal reflex pathways with aging. These include the pathways responsible for inhibition of the Ia afferent before its synapse with the motor neuron (presynaptic inhibition) and for postsynaptic inhibition of the motor neuron through Ia inhibitory interneurons activated from the antagonist (reciprocal inhibition). Age-related changes in presynaptic inhibition have been investigated by conditioning the soleus H reflex with prior vibration of the antagonistic muscle (Butchart et al. 1993) or by brief electrical stimulation of the Ia afferents from the antagonist (Earles et al. 2001) or a heteronymous muscle (Koceja and Mynark 2000; Morita et al. 1995). These studies have shown conflicting results. Butchart et al. (1993) and Earles et al. (2001) found evidence of reduced presynaptic inhibition in resting older compared with young subjects, whereas Morita et al. (1995) found the opposite. It seems likely that this discrepancy is explained by the different ways in which presynaptic inhibition was evaluated in the two studies, and a clarification of whether changes in presynaptic inhibition contribute to the age-dependent decline in stretch reflexes requires further investigation. Reciprocal inhibition also appears to be reduced with aging. Short-latency reciprocal inhibition, measured as the amount of inhibition in the ongoing voluntary EMG activity during isometric plantar and dorsiflexion contraction, has been found to be reduced with aging (Crone et al. 2007; Kido et al. 2004). It is possible that some of the age-related changes in spinal reflex pathways are adaptive and that they occur in response to the functional impairments caused by other central or peripheral changes with aging. The reductions in presynaptic and/or reciprocal inhibition observed at rest could potentially be an adaptation to an age-related decrease in stretch reflex size. The ability to modify the efficiency of these networks during movement is much more important from a functional perspective.

Butchart et al. (1993) found that during voluntary plantar flexion contraction presynaptic inhibition was gradually decreased with increased torque production in the young group but only slightly decreased in the older group. This suggests that the functional modulation of presynaptic inhibition is also less efficient in older subjects (Butchart et al. 1993). Koceja and Mynark (2000) also found impaired modulation of presynaptic inhibition during standing compared with a supine position in elderly subjects. These functional impairments in elderly subjects could be partly due to supraspinal changes. An age-related increase in the threshold for eliciting motor evoked potentials after transcranial magnetic stimulation and reduced motor evoked potential size have been reported (Bhandari et al. 2016; Oliviero et al. 2006; Rossini et al. 1992), suggesting that changes in the corticospinal tract occur with aging, which could affect the functional modulation of spinal reflex networks.

Is task-related modulation of spinal reflex pathways affected by aging?

The modulation of stretch reflexes and H reflexes in the gait cycle described above in children appears to be preserved in middle-aged (Raffalt et al. 2015) and old (Chalmers and Knutzen 2002) adults, although a minor reduction in the H-reflex amplitude during stance (Chalmers and Knutzen 2002) and swing (Raffalt et al. 2015) compared with young adults has been observed.

Age-related changes in spinal reflexes during postural tasks have been investigated to explore potential links to postural stability and balance. Mynark and Koceja (2002) investigated the ability to maintain postural stability in young and elderly subjects while perturbing the balance with evoked H reflexes (Mynark and Koceja 2002). They showed that elderly subjects were as good as young subjects at downtraining their H-reflex responses to maintain postural stability and that training resulted in a 10% decrease in postural sway area in the elderly subjects on day 2 of the training (Mynark and Koceja 2002).

More studies are needed to determine the functional consequences of age-related changes in spinal reflex pathways. Also, the potential for using focused training paradigms to preserve or even increase the functional modulation of these pathways (Geertsen et al. 2008; Mynark and Koceja 2002) should be explored.

Concluding Remarks

Spinal neural circuitries may provide the most direct and simple translation of sensory information into movement in the form of relatively stereotyped reflexes, but even our most simple reflex, the monosynaptic stretch reflex, shows adaptive changes as we grow older. These adaptations help to ensure that our movements are optimally adjusted to the limits and possibilities of a changing body in a constantly changing environment. Since spinal neural circuitries make an integrated contribution to all our movements, they also need to adapt to the changing descending commands from the motor cortex when we become better at what we are doing throughout our lifetime. These concerted plastic changes at cortical and spinal levels are key to understanding the age-dependent changes in our motor abilities. There is much more research to do to obtain a full understanding of this complex relationship.

GRANTS

This work was funded by the Elsass Foundation.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

S.S.G., M.W.-O., J.L., and J.B.N. drafted manuscript; S.S.G., M.W.-O., J.L., and J.B.N. edited and revised manuscript; S.S.G., M.W.-O., J.L., and J.B.N. approved final version of manuscript.