When newborn infants are supported in an upright position with their feet contacting a flat surface they will make co-ordinated stepping movements that are very similar in character to those of an older child or adult (Andre-Thomas and Autgaerden, 1966; Peiper, 1963; Zelazo, et al., 1972). The activity bears such striking similarity to the mature form of independent walking that it is generally viewed as the appropriate starting point for studying the development of upright locomotion. That starting point may be even earlier, given that human fetuses as young as 13–14 weeks gestational age will produce alternating steps while somersaulting in the uterus (de Vries et al., 1982). These seemingly precocious stepping behaviours highlight two important questions for the development of independent locomotion: (1) how does mature locomotion emerge from early stepping; and (2) how important is it to maintain and promote early stepping to improve later locomotion?

From a clinical standpoint, the challenge is to determine whether and how the earlier-appearing stepping patterns can be progressively guided into functional patterns of locomotion when the integrity of the musculoskeletal and nervous systems is compromised by disease or trauma. It is somewhat surprising that although locomotion is one of the most extensively studied skills in motor development (Bernstein, 1967; Dominici et al., 2011; Forssberg, 1985; McGraw, 1940; Shirley, 1931; Sutherland et al., 1980; Thelen and Smith, 1994), experimental discoveries on infant stepping are only now being integrated into clinical practice (Teulier et al., 2009; Ulrich et al., 2001). We argue that there is a pressing need to implement these discoveries at a much faster rate so that the odds of acquiring independent locomotion are improved for the millions of children worldwide who suffer from some form of locomotor disability.

The goal of the current chapter is to examine whether treadmill-induced stepping might be used to promote functional mobility in infants with cerebral palsy (CP). While no evidence currently exists to support the efficacy of treadmill stepping for infants with CP, an impressive body of evidence gathered on typically developing infants and infants with other neurological disorders shows that treadmill stepping practice can hasten the onset of independent walking and improve the quality of gait. We will review this evidence and discuss how researchers are currently augmenting and varying perceptual information during training to further enhance the efficacy of the treadmill stepping paradigm. These modifications to the paradigm are providing opportunities to initiate training interventions at earlier and earlier ages and in more ecological settings. Based on the scientific community’s increasing understanding of the early plasticity in neural and behavioural systems, as well as its appreciation of the widespread nature of critical periods in perceptual and motor development, we argue that early interventions may be essential to ensuring the infant with CP reaches his or her motoric potential.

The birth of the infant treadmill stepping paradigm

The infant treadmill stepping paradigm emerged in the context of a theoretical debate about the most appropriate explanation for the disappearance of newborn stepping. One of the most curious features of early stepping is its disappearance at around 2–3 months of age followed by its reappearance shortly before the onset of independent walking (McGraw, 1932, 1940, 1945). The ontogeny of stepping is a classic example of the U-shaped character of several developmental phenomena (Strauss, 1982). Because the stepping pattern has traditionally been viewed as a primitive reflex, controlled by simple circuits in the brainstem and spinal cord, its disappearance and reappearance was attributed to maturation of cortical brain areas that suppressed the reflex temporarily before bringing it back under voluntary control (Fiorentino, 1981; Forssberg, 1985; McGraw, 1945; Peiper, 1963). This belief was seriously challenged by the fascinating discovery that practicing stepping not only prevented it from disappearing but also increased the quantity of stepping and led to an earlier onset of independent walking (Andre-Thomas and Autgaerden, 1966; Peiper, 1963; Zelazo et al., 1972).

A series of ingenious experiments conducted by Esther Thelen and her colleagues ultimately debunked the notion that stepping’s disappearance was attributable to cortical maturation. The series of experiments began with a simple observation—that despite the impressive similarity between the kinematic features and muscle activation patterns underlying supine kicking and stepping, supine kicking did not disappear, whereas stepping did (Thelen et al., 1981). Assuming that stepping and kicking were basically the same behaviours expressed in different postural contexts, Thelen wondered why the brain would inhibit the behaviour in one context but not the other. The rapid accumulation of leg fat mass relative to muscle mass at the time stepping disappeared proved to be the pivotal discovery in overturning the cortical maturation hypothesis (Thelen and Fisher, 1982). It became obvious that the legs simply became too heavy to lift when the infant was supported in an upright position. The most compelling evidence to support this argument was that stepping could be inhibited if small weights were added to the infant’s legs and that infants who had stopped stepping could be made to step again if their heavy legs were buoyed in a tank of water (Thelen et al., 1984).

The treadmill stepping paradigm was born in the context of the experiments described above. Thelen and colleagues discovered that they could also induce stepping to reappear by supporting the infant on a moving treadmill belt (Thelen, 1986; Thelen and Ulrich, 1991; Thelen et al., 1987). Moreover, and of great importance for clinicians, the steps induced in typically developing infants on the treadmill were more mature than those seen when the infant was supported with their feet touching a stationary surface. Notably, a greater degree of hip extension is observed on the treadmill and often infants show the heel-to-toe progression characteristic of adult walking (Thelen and Smith, 1994; Thelen and Ulrich, 1991; Jensen et al.,1994). These discoveries all emerged from experiments that were originally designed to test Thelen’s explanation for the disappearance of the stepping pattern.

Plasticity and adaptability of the stepping pattern

The treadmill stepping paradigm has revealed a high degree of adaptability in the infant stepping pattern. Stepping shows a level of sophistication and contextually adapted modulation when leg movements are mechanically driven by the treadmill that belies its original reputation as a rigidly stereotyped spinal reflex. For example, the legs can maintain a stable phase relation even when each leg is placed on separate treadmills running at different speeds from each other (Thelen and Smith, 1994; Thelen et al., 1987; Yang et al., 2005). In addition, the pattern has been shown to adapt to repeated perturbations (Pang et al., 2003) and to scale to the speed of the treadmill (Lamb and Yang, 2000). Furthermore, sideways and backward stepping have been elicited, with both scaled to treadmill speed (Lamb and Yang, 2000), and the individual legs have been driven to step in different directions when two treadmill belts are moved in opposite directions (Yang et al., 2004).

Plasticity in the stepping pattern is the feature that makes stepping an ideal target for clinical intervention. Moreover, the degree of responsiveness in the pattern to variations in treadmill parameters opens exciting new avenues for using the treadmill to train early infant locomotion.

Continuity from stepping to walking

The evidence described in the preceding paragraph provides some insight into the stepping pattern’s plasticity. However, the primary evidence for plasticity comes from the early experiments in which infants were given stepping practice for extended periods of time. Those experiments demonstrated clearly that stepping would not disappear if it was practiced and, more importantly, that practice led to an earlier onset of independent walking (Andre-Thomas and Autgaerden, 1966; Peiper, 1963; Zelazo et al., 1972). The latter finding provides particularly strong support for continuity in locomotor development from the earliest stepping movements to the emergence of independent walking. More recent support has been provided by researchers who have used sophisticated neural modelling to argue that the basic patterns of lumbosacral motor neuron activity seen in neonatal stepping are retained in adult walking, even though new patterns are also evident (Dominici et al., 2011). Despite the obvious similarities between stepping and walking, however, it is abundantly clear that the two are not isomorphic.

The research to date suggests strongly that stepping is an important, and probably necessary, precursor to walking, but two caveats have to be kept in mind. First, the underlying electromyographic patterns and movement kinematics are much more variable in early stepping than mature walking. Healthy infants produced a wide variety of muscle activation combinations and timings when stepping on a treadmill, even though decreases in co-contraction suggest an increase in muscular co-ordination over time (Teulier et al., 2012). This finding implies that the earliest locomotor patterns do not simply reflect the output of a stereotyped central pattern generator (CPG). Second, walking is far more demanding than stepping. For example, supported stepping places no demands on balance and steering, no demands on movement planning and initiation, no demands on monitoring the external layout, and only minimal (though adjustable) demands on strength. Nevertheless, as a precursor to independent walking, the stepping pattern is highly amenable to clinical intervention.

To understand what can be accomplished with treadmill stepping interventions, as well as their limitations, it is helpful to further examine the demands on independent walking and further delineate the differences between stepping and walking. The analysis will also help to clarify the range of different variables that need to be addressed in interventions designed to promote motor development. The analysis begins with a brief summary of Newell’s model of constraints on behaviour (Newell, 1984, 1986).

Constraints on the development of walking

Drawing on the seminal works of Kugler et al. (1980, 1982) as well as Higgins (1977), Newell (1986) argued that all motor behaviour emerges from the interaction of constraints from three broad sources: the individual’s body; the tasks in which individuals participate; and the environment. The myriad of constraints channel movement dynamics and specify what sources of information are likely to be most useful in the control of action.

Individual constraints

With respect to walking, the capacity to generate alternating limb movements, sufficient strength to support the body and generate propulsive force, and balance have been identified as important individual constraints, though it is important to realize that a new skill like walking will be influenced by a range of perceptual, affective, attentional, motivational, postural, anatomical, and physiological variables interacting in a particular context (Thelen and Smith, 1994; Thelen and Ulrich, 1991).

The constraints can be viewed as important substrates of skillful walking that, if weak or missing, might be the target of interventions to promote walking. Sufficient strength and balance, in particular, are often considered key contributors to independent walking (Thelen and Smith, 1994). Adolph et al. (1998) also identified strength as an important contributor to the development of hands-and-knees crawling. Infants who had engaged in several weeks of belly crawling prior to the onset of hands-and-knees crawling mastered hands-and-knees crawling much faster than infants who had skipped belly crawling altogether. Belly crawling presumably ‘shored up’ the prerequisites, e.g., strength, that were necessary for proficient hands-and-knees crawling.

Task constraints

Task constraints and environmental constraints are often given far less attention in the organization of behaviour than individual constraints. However, they are just as important. Task constraints in walking would largely involve the locomotor goals that the child sets itself. These goals change of course as the child becomes a more competent walker. Choices of places to go, for example, increase exponentially as walking skill improves. Task constraints also involve the rules associated with tasks. The young child need not conform to any formal rules when first learning to walk, though the older child must, particularly when navigating city streets, where a number of rules are in place to protect pedestrian traffic.

Environmental constraints

Environmental constraints represent the final class of constraints from which behaviours emerge. These constraints include the field of external forces (such as gravity) in which all movements are made as well as the social and cultural forces that shape the activities we choose to participate in as well as the manner in which we express those choices (Gentile 1987). All movements represent a blend of internal forces that are generated through muscular contraction and external forces that surround the individual. Gravity is the most pervasive external force that influences the character of a movement—witness the dramatic changes in gait that occur when astronauts ‘walk’ on the surface of the moon—though inertial forces, frictional forces, and various reaction forces will also heavily influence the character of a movement. The most obvious environmental constraints, however, include the physical features of the environment because movements are organized relative to these features.

Gentile (1987, 2000) has provided a particularly clear description of the environmental constraints that impose themselves on movement dynamics. She has coined the term regulatory conditions to describe the physical features of the environment to which movements must conform if they are to be successful. In the case of walking, important regulatory conditions include the size of the surface on which walking occurs, its slope, its solidity, and its texture. People walk very differently on a slippery surface like ice, for example, than on a textured surface like concrete. Any variations in the regulatory conditions force some form of adaptation or reorganization of the movement pattern. This point is crucially important because the regulatory conditions will vary in an infinite number of ways. Consequently, in a task like walking, the motor system is continually challenged to come up with adaptations that accommodate to changing regulatory conditions. This notion highlights how important the processing of sensory information is during functional locomotion. We will return to this point later in the chapter when we discuss ways to enhance the efficacy of treadmill training.

The point of this primer on constraints has been to reinforce the complexity associated with learning and performing an ostensibly simple task like walking. That complexity needs to be given careful consideration in the development of interventions designed to promote independent walking. While it is clear from the earlier discussion that treadmill stepping is remarkably adaptable to contextual variations in the way the treadmill belt moves, functional locomotion is far more demanding. Moreover, adaptations in gait require continuous monitoring of the environment and one’s relation to it to ensure that the locomotor goal is achieved and that it is achieved efficiently. As such, the demands on attention are far greater during independent walking than they are during supported stepping. With these points in mind, it is possible to design interventions that progressively challenge the infant to acquire the prerequisite skills and capacities to engage in functional locomotion. The primary goal of interventions should be to develop the substrates of independent locomotion, provide opportunities to integrate these substrates into functional patterns, and systematically expose the patterns to contextual variations to promote the development of flexible locomotor strategies. In addition, where possible, the infant should be exposed to the sources of perceptual information that will be available during independent locomotion.

When should interventions start?

Having established the goals for treadmill training interventions, we now turn to the important question of when interventions should be started. Discussions about the most appropriate times to initiate motor skill learning have historically been couched relative to the critical periods and readiness concepts (Anderson et al., 2012). The critical period concept originated in embryology, where the embryo’s ultimate form was shown to be exquisitely sensitive to the timing of developmental disruptions (Spemann, 1938; Stockard, 1921). Critical periods are now well-established phenomena in embryological development and nearly every parent is familiar with the notion that teratogens (external agents) will have widely different effects depending on the timing of exposure. The period of most rapid growth or differentiation of an organ or system is generally considered the most susceptible time (Moore and Persaud, 1998).

The most widely cited example of a critical period in mammalian development concerns the visual system. Hubel and Wiesel (1970) demonstrated that surgical closure of one eye during a brief period after birth causes a severe visual impairment in species such as cats and monkeys when the eye is later reopened. Critical periods have been demonstrated for auditory development, tactile development, and motor and neuromuscular development in the rat (Jamon and Serradj, 2009). A special issue of Developmental Psychobiology maintains that critical periods in human sensory development are pervasive (Maurer, 2005).

Though critical periods in human development have most often been discussed relative to neurological development, particularly as it pertains to the perceptual systems, an experiment by Walton et al. (1992) suggests that they are equally present in other physiological systems. Using tail suspension to simulate weightlessness, they discovered a critical period in rat locomotor development during which unloading of the weight-bearing limbs caused permanent disruptions to swimming and walking. Jamon and Serradj (2009) have reviewed the effects of exposure to hypo- and hypergravity and confirmed that muscular development is highly sensitive to disruptions in the normal forces on the limbs during early postnatal development. For example, rats reared in hypergravity showed marked changes in the contractile and morphological properties of their muscles. Their review also showed that altered gravity has significant effects on the development of the vestibular system during a critical period in development. Given the number of different components from which independent locomotion is assembled, each with its own developmental trajectory, it is likely that multiple critical periods exist in the development of locomotion.

This brief discussion of critical periods suggests that interventions should be initiated as early as possible to enable infants with disabilities to achieve their locomotor potential. When combined with our increasing understanding of the plasticity in early neurological systems, the case for early interventions is even more compelling (Ulrich, 2010). In the remaining sections of this chapter, we will highlight how the treadmill paradigm has been used to promote the development of independent walking. In addition, we will describe recent discoveries about the sophistication of early perception–action coupling that can be used to enhance the effectiveness of treadmill training and potentially allow interventions to begin even earlier.

Only gold members can continue reading. Log In or Register to continue