INTRODUCTION AND PURPOSE
The first year of life is generally considered the most critical period of motor development, offering the best opportunities for introducing interventions to facilitate development of motor skills for children whose neural development has been interrupted via spinal lesion, such as myelomeningocele (MMC), a form of spina bifida.1 Spina bifida most frequently affects the lumbar and sacral spine regions,2 resulting in varying degrees of sensorimotor deficits in the legs. Infants with MMC have variations and delays in the median age for motor skill achievement,3 contributing to a cascade of effects across systems as children grow and enter adulthood.
Infants with thoracic and lumbar lesions produce spontaneous leg movements in utero that are as frequent as those of fetuses who are developing typically, potentially due to stimulation and support from the in utero fluid environment, sporadic uterine contractions, and maternal motor activity.4 , 5 However, after birth, leg activity in infants with MMC becomes depressed or disappears,4–7 possibly resulting from inadequate antigravity strength and poor sensory feedback as a result of movement efforts. By the end of the first year, infants born with MMC exhibit shorter shank length, smaller shank circumferences,8 and decreased bone density compared with infants developing typically.9 Growth impairments and bone weakness are hypothesized to result from inherently weaker muscles in the lower limbs, reduced neural capacity, and reduced time spent in weight-bearing antigravity movement. Physical activity affects development of the early nervous system for infants who are developing typically and atypically.10–12 Lee and colleagues11 tested the 1a proprioceptive pathways in infants with MMC and found that responsiveness to stimuli for these pathways was either depressed or not intact. Together, these findings highlight the critical need to increase active leg movements (eg, spontaneous activity and weight-bearing movements) to create a strong foundation of muscle strength and control in the lower body of infants with MMC.
Interventions begun earlier in development and at a higher intensity may result in greater and longer positive outcomes for infants with disabilities than for those who begin later and at lower intensities.13 Rehabilitation should begin early and intensively, emphasizing continual self-exploration and selection of movement patterns. Interventions that incorporate these qualities facilitate development of functionally relevant motor skills (eg, locomotion) and perpetuate long-term, positive effects on motor behavior across the lifespan.14 However, for many infants with MMC, therapy is often initiated only after a therapist documents delayed achievement of motor milestones with an accepted assessment instrument; thus, important opportunities for skill development are often missed. Even when prescribed promptly, therapeutic treatments tend to be passive, less aggressive, and less focused on functional skills than those for older children.15
Practice stepping on a treadmill encourages performance of more cycles of leg activity similar to the functional skill of walking. Use of step training has been shown to facilitate earlier development of continuous, well-coordinated stepping patterns in infants who are developing typically16 , 17 and acquisition of independent walking as well as motor milestones in infants with Down syndrome (DS).18 , 19 For parents, in-home, individualized treadmill stepping practice for infants with DS provided a structured environment with easy directions, fostering a perception that they were contributing directly to advancing their infants' motor development.
Based on findings from studies with infants with DS, it is possible that infants with MMC who receive early treadmill stepping practice may optimize development of their motor function, contributing to better health-related quality of life in later childhood and adulthood. Recent longitudinal research studies have demonstrated that infants with MMC can, without practice, produce a variety of step types while being supported on a moving treadmill.8 , 20 , 21 While their overall step rate was significantly lower than that in infants developing typically across the first year of life, by enhancing sensory input from the treadmill's surface (ie, visual flow: checkerboard pattern; friction: tacky surface), infants with MMC showed an increase in step frequency.20 These results provide valuable insight into environmental factors that may stimulate treadmill stepping for infants with MMC, promoting improvements in neuromuscular strength and control and, possibly, accelerating acquisition of functional developmental milestones.
The purposes of our pilot study were to determine (1) the feasibility of an early treadmill training program for infants with MMC affecting the lumbar-sacral levels; (2) the acquisition of gross motor skills; (3) the quality of steps performed; and (4) integrity of peripheral 1a pathways tested before and following treadmill step training that was initiated within the first 6 months postbirth.
Twelve infants with MMC were initially enrolled: 6 began training 1 month postbirth, 4 began after 1 month due to medical complications or late identification, and 2 were excluded from data analysis because of a move from the geographical area and lack of compliance. Therefore, data for 10 infants are presented in this article. Of the 10 infants, 8 received daily treadmill stepping practice for 12 months and 2, due to family issues, received treadmill stepping practice for only 6 months. Exclusion criteria included central nervous system or chromosomal abnormalities beyond those associated with MMC (eg, hydrocephalus and Arnold-Chiari II syndrome), lesion levels higher than the lumbar region, gestational age less than 30 weeks, or older than 6 months at the time of identification. If gestational age was less than 37 weeks, corrected age was calculated. The Table provides participant medical characteristics. Study procedures were approved by the Institutional Review Board at the University of Michigan.
For initial training sessions, parents were instructed to provide a total of 10 minutes per day of in-home training, 5 days per week that included newborn stepping, bouncing, and treadmill stepping. How and when the 10 minutes of training was split among the 3 components during early sessions were based on parental perception of infant stepping response on the treadmill.
- Newborn stepping: Parents held infants upright with bare feet in contact with a firm stationary surface, slight forward tilting of the trunk, and slow forward movement over the surface.
- Bouncing: Parents held infants upright with bare feet resting on their thighs. Parents rhythmically moved infants downward and upward, encouraging leg extension and weight acceptance.
- Treadmill stepping: Parents held infants upright in a partial, body weight–supported position with bare feet in contact with the belt of a custom-engineered pediatric treadmill (Carlin's Creations, Sturgis, Michigan; 82 × 42 × 18 cm). All participants received the same motorized treadmill model (pediatric treadmill 3; Carlin's Creations) for the duration of the study. Belt speed for the early training sessions was not specified but was based on parents' determination of “best” speed for their individual infants.
Our protocol took advantage of infants' early ability to perform newborn steps and parents' familiarity with “bouncing babies” on their laps to encourage leg extension and acceptance of weight through infants' legs. Practice performing newborn stepping helped infants learn to lift their legs and perform alternating steps, whereas bouncing encouraged leg extension and weight acceptance. To facilitate treadmill stepping at practice onset through enhancement of sensory input, the treadmill belt was made from Dycem (Dycem Ltd, Warwick, Rhode Island) with a checkerboard pattern.8 , 20
As training continued, parents were encouraged to increase the proportion of total time each day that was devoted to treadmill stepping. After 4 to 6 weeks of training, parents were instructed to discontinue newborn stepping and bouncing activities, focusing on treadmill stepping. Sessions continued to occur 5 days per week. When infants began taking continuous, alternating steps, the Dycem belt was removed and training continued with a regular treadmill belt. The practice protocol was progressive, with variations in treadmill duration and belt speed specific to each infant's changes in the stepping response rate. The progressive changes in duration and treadmill belt speed were based on parental reports and discussions with the research team regarding infant tolerance to training sessions.
Treadmill stepping practice (ie, duration, frequency, belt speed, infant response) was monitored, with daily logs completed by parents. However, parental compliance in the use of exercise logs and type of information included was erratic. Duration of treadmill use was also recorded in minutes by a gauge on the side of each treadmill. Families were visited biweekly to check gauge values, review parents' log, answer questions, and offer advice for optimizing practice sessions.
In addition to treadmill stepping practice, 8 infants received other therapy services. Of those who reported receiving other therapy services, none reported gait training activities as part of the therapeutic sessions. No additional information was provided by parents regarding dosage or specific type(s) of therapy services received during participation in the treadmill training program.
Three sets of assessments were collected for each infant: anthropometrics and motor skill acquisition, treadmill stepping behavior, and reflex testing.
Anthropometrics and Motor Skill Acquisition. Anthropometrics and motor skill acquisition were assessed biweekly for the first 3 months and then monthly for a total of 17 assessments per infant. Anthropometrics included body weight, recumbent length, leg length, and umbilicus skinfold. Acquisition of gross motor skills was assessed with the Motor subscale from the Bayley Scales of Infant Development III (BSID III).22
Treadmill Stepping Behaviors. Treadmill assessments of infant stepping were conducted 8 times during the training program. Three of these 8 assessments were formal, occurring in our research laboratory: pre-, mid- (after 6 months training), and posttraining. The other 5 assessments were informal, occurring in family homes at 2, 4, 8, 10, and 12 months of training. Two digital cameras (60 Hz; Canon ZR 960) recorded stepping behavior for coding (Figure 1A). Each treadmill assessment consisted of 4 trials, with belt speed maintained at 0.144 m/s for 60 seconds.
Integrity of Peripheral 1a Pathways. At pre- and postassessments, functioning of 3 monosynaptic reflex pathways of lower-limb tendons were examined bilaterally: gastrocnemius (GA), tibialis anterior (TA), and quadriceps (QA). As in previous works,11 , 12 the 3 monosynaptic T-reflexes were elicited by electromagnetic stimulators (Ling Dynamic V203) equipped with a smooth hammerhead. Twenty taps were delivered to the distal tendon of the stimulated muscle, with a minimum interstimuli interval of 8 seconds. This procedure was used for the 3 tendons tested. The order of muscle testing was randomized across infants. Recordings of electromyographic (EMG) activity during reflex testing for stimulated muscles were used to identify responses.
Anthropometrics. Weight-for-length percentiles were calculated on the basis of WHO standards.23
Motor Skill Acquisition. Subscales of the BSID III were used to assess gross motor milestones (75 items). The age at onset for each milestone was calculated from the actual/corrected birth date of each infant.
Treadmill Stepping Behaviors. Video recordings were behavior coded for the treadmill testing sessions using frame-by-frame analysis (60 Hz) with Peak Motus version 8 software (Peak Motus; Vicon, OMG plc, Oxford, UK). Three behavior coders were trained for at least 6 weeks prior to validity testing. Steps were categorized into 4 interlimb stepping patterns based on behavioral criteria provided by Thelen (1991) and included (from least to most complex) the following:
- Parallel: Both legs swung forward simultaneously.
- Double: During a series of alternating steps, a “stutter step(s) was performed.
- Single: The step of one leg did not overlap with a step of the contralateral leg.
- Alternating: The step of one leg overlapped with a step of the contralateral leg.
For the purpose of the questions in this pilot study, we focused on basic treadmill parameters for alternating strides performed because this interlimb stepping pattern is the most complex, requiring significant motor control and coordination.24 Step parameters reported from the behavior-coded video included step frequency, step type, and steps per second.
Interrater reliability was high between behavior coders for this study and previously validated coders on identical trials from training tapes (r > 0.85).
Integrity of Peripheral 1a Pathways. A custom-designed Labview virtual instrument12 quantified reflex responses via reading EMG data. The raw EMG data were examined to assess the peak-to-peak amplitude and latency to the first peak of muscle activation for each individual stimulus presentation. All responses occurring in the stimulated muscle were calculated in a ratio for each tendon tested: the number of responses relative to the total number of stimulations provided.
Analyses focused on determining the following: (1) feasibility of step training based on data obtained from treadmill counters; (2) changes from pre- to posttraining in overt behaviors, including developmental motor milestones and treadmill stepping behaviors—total step frequency and ratio of alternating steps; and (3) changes in anthropometrics and ratios of tendon reflex responses in the lower limbs. We used the general linear model instead of analyses of variance for all analyses to account for the repeated sessions and unequal number of sessions over training. All analyses are reported with Wald χ2. To examine effect of age at entry, infants were divided into 2 subgroups (infants who started at 1 month of age and infants older than 2 months). The effect of chronological age was controlled for during analyses of integrity of peripheral 1a pathways and anthropometric measures.
Compliance With Participation in the Training
Most infants successfully completed the training, but compliance varied across infants throughout training. Each infant's average treadmill practice was calculated by tracking each treadmill's counter cross-referenced with parental report (Figure 1B). For the first month of participation, treadmill counters were low because infants received more practice with newborn stepping and/or bouncing. Type and duration of activity(ies) (ie, newborn stepping, bouncing, treadmill stepping) included during the first month of participation were decided by each infant's parent and researchers based on an individual child's responsiveness. However, duration of treadmill usage generally increased with participation (Wald χ2 = 9.32, P = .097) but had increased variability for some infants at the midpoint of training (4-8 months). No effect for age at study entry on compliance was detected.
Weight-for-length percentiles significantly decreased over the course of practice (Wald χ2 = 146.57, P = .00) (Figure 2A). Leg length and umbilicus skinfold measurements significantly increased with stepping practice despite controlling for age (Wald χ2 = 316.66, P = .00; Wald χ2 = 117.43, P = .00, respectively) (Figures 2B and 2C).
Infants with MMC achieved most gross motor milestones within the ranges described for healthy infants (Figure 2D), including sitting alone by 6 months, crawling by 12 months, standing alone by 16.8 months, and walking independently by 19.5 months.
Treadmill Stepping Behaviors
Infants had a slow increase in the total number of steps (regardless of different stepping patterns) and alternating steps with increasing practice stepping (Wald χ2 = 44.63, P = .00; Wald χ2 = 28.83, P = .00, respectively) (Figures 1C and 1D). After 2 months of practice, infants had small increases in the overall number of steps and the number of alternating steps. This small increase was followed by a small decline between 4 and 8 months but then another gradual increase from approximately 8 to 12 months. Infants who began stepping practice at 1 month of age tended to take fewer steps and fewer alternating steps than infants who began after 2 months (Wald χ2 = 16.36, P = .02; Wald χ2 = 37.30, P = .00, respectively).
Integrity of Peripheral 1a Pathways
Integrity of monosynaptic pathways in the lower limbs had differing response ratios from pre- to post–treadmill stepping practice for QA, GA, and TA (Figure 3). Controlling for age, the ratio of T-reflexes in the QA at pretesting tended to show more responses than those for TA or GA (Wald χ2 = 18.43, P = .00), but no differences were found across muscles at posttesting (P > .05). In addition, the ratio of QA T-reflex responses showed a significant increase post–stepping practice (Wald χ2 = 9.48, P = .00), but those of TA and GA failed to reach significance (Ps > .05). Infants who began stepping practice at the age of 1 month tended to show lower responses in both QA and TA than infants who began after 2 months; however, the differences were not statistically significant after controlling for age (Ps > .05).
Results of this study provide support for the feasibility of early, home-based treadmill stepping practice for infants with MMC starting within the first 6 months postbirth and continuing for 12 months. Infants with MMC and their parents tolerated the intensive protocol. Infants demonstrated acquisition of developmental gross motor skills at an earlier age than what has been previously reported for infants with MMC who did not receive early stepping practice.8 , 20 Infants produced more alternating steps, requiring more complex patterns of coordination. Improvements in stepping performance were associated with changes in reflex activity.
This pilot study provides evidence that infants with MMC with lumbar–sacral-level lesions tolerated a parent-administered, individualized, progressive program that included treadmill stepping practice for 10 minutes, 5 days per week for 12 months. Despite variations in parental compliance over the course of training, on average, infants received 6 to 7 minutes of treadmill step training per day. Overall, parents reported enjoying the opportunities dedicated to facilitating development of their infants. At entrance into the program, parents seemed eager to provide 10 minutes of exercise a day for their infants. However, on the basis of data from treadmill counters and parental logs, compliance declined during the midpoint of participation in the program. Parents reported that provision of daily treadmill training required considerable attention and time commitment. We propose that weekly phone conversations, sharing training protocols with the child's pediatric physical therapist (if receiving treatment), and improved instrumentation for compliance measurement should be added in future programs to optimize the effects of treadmill training.
Developmental Motor Milestones
The results suggest that treadmill step training helped infants develop rhythmic, coupled movements through performance of repeated cycles of stretching the leg extensors and swinging forward, consequently increasing muscle strength and control. A high proportion of rhythmic, coupled movements—alternating steps—were associated with the age of independent walk onset in infants born preterm and full-term at 7 months' corrected age25 and infants with DS at 11 months of age.26 These results and ours indicate that the ability to perform rhythmic, coupled movements is essential to developing functional skills such as independent walking. This mechanism may have also facilitated earlier achievement for prerequisite locomotor skills (ie, rolling back to stomach, sitting alone, pulling to sit) for our infants with MMC who received step training compared with their peers who did not.3 , 8 Infants in our pilot study achieved developmental motor skills such as rolling from the back to stomach, sitting alone, and pulling to sit, on average, at least 1.5 months earlier and walking 2.2 months earlier than milestones described for infants with MMC in the extant literature.8
Treadmill Steps—Quality Versus Quantity
Treadmill stepping behaviors—total steps and alternating steps—showed improvements over the course of step training. Comparison of our results with those reported by Teulier et al8 shows that the total number of steps produced by infants with MMC was similar between studies. However, infants in our training program generally produced more alternating steps during treadmill training than previously described for infants with MMC who did not receive treadmill step training and produced a large proportion of more simplistic, parallel steps. This was a somewhat unexpected result. Alternating steps require continuous proprioceptive dialogues between the stretch of one leg and pattern in the opposite leg; simple steps (ie, single, parallel) require mainly a mechanical stretch of extensors provided by the moving treadmill belt.24 Thus, producing alternating steps requires more leg control and muscle strength than single or parallel steps.
A greater proportion of alternating steps may explain why the total number of steps performed remained similar compared with infants with MMC who did not receive treadmill step training. For infants, an increase in response frequency came first in the form of the most advanced pattern, alternating steps, requiring more muscle strength, balance control, and limb coordination. Increasing the total number of steps meant performing alternating steps. Because alternating steps required more muscle strength, control, and coordination, infants may have had difficulty performing increasing numbers of steps. This situation stands in contrast to the data of infants with MMC and infants developing typically without treadmill practice. Infants with MMC generally show an increase in step frequency using simple patterns (single, parallel) during early development.8 , 20 For infants who are developing typically, a positive correlation was observed between step frequency and alternating steps over 16 sessions of treadmill practice: as infants took more frequent steps, the more advanced pattern (alternating) emerged as the preferred pattern.27
Mechanisms Underlying Overt Motor Changes
Infants with MMC who participated in this treadmill stepping practice had changes in responsiveness of peripheral 1a pathways. Our previous research has shown that without training, the peripheral 1a pathways are depressed across the first 10 months of life in infants with MMC.11 However, in the present study, the reflex response frequency at posttraining was similar to that of age-matched infants developing typically.12 In addition, the improvement we observed from pre- to posttraining was larger than that shown by toddlers developing typically during the first 3 months of walking experience.10 Our data effectively show that step training was associated with the gain in sensitivity of 1a proprioceptive pathways. Therefore, we contend that in-home treadmill step training should be considered an integral complement to clinical interventions because it enhances infants' capacity to compensate for sensorimotor impairments.
We also observed a tendency for infants' weight gain to uniquely affect the developmental trajectories of stepping behaviors during participation in the treadmill stepping program. Two-thirds of infants showed an increase in step rate prior to rapid weight gain. However, when infants had dramatic increases in weight gain relative to growth in length, they showed less capacity to control the lower limbs by decreasing the number of steps performed on the treadmill. The asynchronous development of muscle mass and concomitant strength in the lower limbs could explain the unanticipated delay in the increased number of total steps performed. This finding aligns with previous research. Thelen and colleagues28–30 illustrated that by 2 to 3 months of age, newborn stepping behaviors were depressed because of weight gain (muscle to fat ratio of legs). Because gains in fat mass outstripped gains in muscle mass, infants' legs were not strong enough to perform stepping behaviors, contributing to the “disappearance” of stepping during early infant development. For infants in the current study, instead of stepping behaviors “disappearing,” infants with MMC took fewer steps despite daily treadmill stepping practice. Thus, monitoring and management of infant adiposity may promote a more linear increase in the number of steps performed with practice.
This study had limitations that need to be considered when interpreting our results and considering future research studies that use a similar approach for infants with neuromotor delay. One of the most relevant limitations was that the sample size and medical profiles of each participant were varied, similar to other infants with MMC.8 , 18 , 19 In addition, no control group was included for direct comparison of results. However, other research studies involving infants with MMC who had similar medical profiles have been reported in the literature, allowing us to compare the findings of our pilot study with previously reported results. Low compliance with the training protocol also may have affected results. Thus, definitive conclusions are difficult to make about efficacy of treadmill step training.
This pilot study supported that treadmill stepping practice facilitated development of rhythmic, coupled movements through performance of repeated cycles of leg flexion and extension, increasing muscle strength and limb control and encouraging acquisition of developmental motor skills. In general, the individualized practice sessions were tolerated by infants and parents. Therefore, therapeutic programs for infants with neuromotor delays, such as MMC, need to consider providing daily, intensive movement practice to facilitate development and performance of the complex motor skills necessary for functional independence.
1. Sharkey MA, Palitz ME, Reece LF, et al The effect of early referral and intervention on the developmentally disabled infant: evaluation at 18 months of age. J Am Board Fam Pract. 1990;3(3):163–170.
2. Emery JL, Lendon RG. The local cord lesion in neurospinal dysraphism (meningomyelocele). J Pathol. 1973;110(1):83–96.
3. Williams EN, Broughton NS, Menelaus MB. Age-related walking in children with spina bifida. Dev Med Child Neurol. 1999;41(7):446–449.
4. Sival D, Begeer JH, Staal-Schreinemachers AL, Vos-Niel JM, Beekhuis JR, Prechtl HF. Perinatal motor behaviour and neurological outcome in spina bifida aperta. Early Hum Dev. 1997;50(1):27–37.
5. Sival D, Brouwer OF, Bruggink JL, et al Movement analysis in neonates with spina bifida aperta. Early Hum Dev. 2006;82(4):227–234.
6. Rademacher N, Black D, Ulrich B. Early spontaneous leg movements in infants born with and without myelomeningocele
. Pediatr Phys Ther. 2008;20(2):137–145.
7. Smith BA, Teulier C, Sansom J, Stergiou N, Ulrich BD. Approximate entropy values demonstrate impaired neuromotor control of spontaneous leg activity in infants with myelomeningocele
. Pediatr Phys Ther. 2011;23(3):241–247.
8. Teulier C, Smith BA, Kubo M, et al Stepping responses of infants with myelomeningocele
when supported on a motorized treadmill. Phys Ther. 2009;89(1):60–72.
9. Lee DK, Muraszko K, Ulrich BD. Bone mineral content in infants with myelomeningocele
, with and without treadmill stepping practice
. Pediatr Phys Ther. 2016;28(1):24–32.
10. Lee DK, Ulrich BD. Functioning of peripheral Ia pathways in leg muscles of newly walking toddlers. Hum Mov Sci. 2015;40:193–210.
11. Lee DK, Teulier C, Ulrich BD, Martin B. Functioning of peripheral Ia pathways in infants with myelomeningocele
. Infant Behav Dev. 2013;36(1):147–161.
12. Teulier C, Ulrich BD, Martin B. Functioning of peripheral Ia pathways in infants with typical development: responses in antagonist muscle pairs. Exp Brain Res. 2011;208(4):581–593.
13. Ramey CT, Ramey SL. Early intervention
and early experience. Am Psychol. 1998;53(2):109–120.
14. Hadders-Algra M. Early brain damage and the development of motor behavior in children: clues for therapeutic intervention? Neural Plast. 2001;8(1/2):31–49.
15. Hinderer K, Hinderer S, Shurtleff DB. Myelodysplasia. In: Campbell SK, Palisano RJ, Vander Linden DW, eds. Physical Therapy for Children. Philadelphia, PA: Elsevier Saunders; 2006:xv, 1181.
16. Super CM. Environmental effects on motor development: the case of “African infant precocity.” Dev Med Child Neurol. 1976;18(5):561–567.
17. Zelazo PR, Zelazo NA, Kolb S. “Walking” in the newborn. Science. 1972;176(32):314–315.
18. Ulrich DA, Lloyd M, Tiernan C, Looper J, Angulo-Barroso R. Effects of intensity of treadmill training on developmental outcomes and stepping in infants with Down syndrome: a randomized trial. Phys Ther. 2008;88(1):114–122.
19. Ulrich DA, Ulrich BD, Angulo-Kinzler RM, Yun J. Treadmill training of infants with down syndrome: evidence-based developmental outcomes. Pediatrics. 2001;108(5):e84.
20. Pantall A, Teulier C, Smith B, Moerchen V, Ulrich B. Impact of enhanced sensory input on treadmill step frequency: infants born with myelomeningocele
. Pediatr Phys Ther. 2011;23(1):42–52.
21. Saavedra SL, Teulier C, Smith BA, et al Vibration-induced motor responses of infants with and without myelomeningocele
. Phys Ther. 2012;92(4):537–550.
22. Bayley N. Bayley Scales of Infant and Toddler Development (3rd Ed): Motor Scale. San Antonio, TX: Psychological Corporation; 2006.
23. WHO Multicentre Growth Reference Study Group. WHO Child Growth Standards: Length/Height-for-Age, Weight-for-Age, Weight-for-Length, Weight-for-Height and Body Mass Index-for-Age: Methods and Development. Geneva, Switzerland: World Health Organization; 2006.
24. Thelen E, Ulrich BD. Hidden Skills: A Dynamic Systems Analysis of Treadmill Stepping During the First Year. Chicago, IL: University of Chicago Press; 1991.
25. Luo HJ, Chen PS, Hsieh WS, et al Associations of supported treadmill stepping with walking attainment in preterm and full-term infants. Phys Ther. 2009;89(11):1215–1225.
26. Ulrich BD, Ulrich DA, Collier DH. Alternating stepping patterns: hidden abilities of 11-month-old infants with down syndrome. Dev Med Child Neurol. 1992;34(3):233–239.
27. Vereijken B, Thelen E. Training infant treadmill stepping: the role of individual pattern stability. Dev Psychobiol. 1997;30(2):89–102.
28. Thelen E. Learning to walk is still an “old” problem: a reply to Zelazo (1983). J Motor Behav. 1983;15(2):139–161.
29. Thelen E, Fisher DM, Ridley-Johnson R, Griffin NJ. Effects of body build and arousal on newborn infant stepping. Dev Psychobiol. 1982;15(5):447–453.
30. Thelen E, Fisher DM. Newborn stepping: an explanation for a “disappearing” reflex. Dev Psychol. 1982;18(5):760–775.