The Stepping Reflex Disappearance Study (Water immersion) – Esther Thelen

The trajectory of developmental science throughout the twentieth century was largely dictated by a powerful, reductionist metaphor: the brain as a hardwired, hierarchical master computer whose genetic and neuroanatomical maturation unilaterally directs the emergence of physical behaviors. Within this prevailing paradigm, infant motor milestones—from rolling over and grasping to crawling and unassisted bipedal locomotion—were mapped as the direct, deterministic readouts of a ripening central nervous system. Nowhere was this conceptual framework more firmly entrenched than in the clinical and theoretical consensus surrounding primitive neonatal automatisms. Among these, the neonatal stepping reflex stood as an archetypal textbook phenomenon: a vigorous, coordinated alternation of the lower limbs elicited by holding a newborn upright, an ancient motor program that reliably and mysteriously vanished between the second and third months of human life, only to re-emerge months later in the form of intentional, voluntary stepping.

For more than four decades, the universal disappearance of the stepping reflex was brandished as triumphant proof of cortical encephalization. Classic neuroembryological and pediatric dogma, formulated by luminaries such as Arnold Gesell and Myrtle McGraw, maintained that the developing cerebral cortex gradually extended descending inhibitory axons into the subcortical and spinal motor centers, suppressing archaic subcortical reflexes to prepare the neural architecture for voluntary, cortical control. The sudden cessation of stepping was not viewed as an ecological or biomechanical event, but as a silent, invisible triumph of cerebral myelination and cortical dominance over lower spinal automatisms. This interpretation reigned unchallenged in developmental textbooks, pediatric neurology clinics, and medical curricula worldwide, reinforcing a fundamentally disembodied conception of human ontogeny.

However, in 1984, the developmental psychologist and biomechanist Esther Thelen and her colleagues published a series of deceptively simple yet conceptually earth-shattering experiments that shattered this classical hegemony. By applying the non-linear principles of Dynamic Systems Theory to infant morphology and physical dynamics, Thelen interrogated a glaring, previously ignored contradiction: why would an executive cortical command selectively extinguish stepping while infants were held upright, yet allow virtually identical coordinated leg movements to persist unabated when those same infants kicked vigorously while lying on their backs? Suspecting that the primary catalyst of reflex disappearance was physical rather than purely neuroanatomical, Thelen placed non-stepping two-month-old infants waist-deep in water and observed an extraordinary phenomenon: the supposedly inhibited neural reflex instantly, fluidly, and robustly returned. This landmark water immersion study permanently displaced the brain-centric orthodoxy of human motor development, demonstrating that behavioral change is not dictated by an isolated central nervous system, but emerges from the dynamic, reciprocal self-organization of neural circuits, bodily morphology, gravitational fields, and environmental task constraints.

1. Historical Context of Infant Motor Development and Neuromaturational Theories

1.1 Pre-1980s Dominance of Cortical Inhibition Models

To understand the revolutionary nature of Esther Thelen’s empirical contributions, one must first appreciate the intellectual monolith that dominated mid-twentieth-century developmental psychology and pediatric neurology: the neuromaturational perspective. Championed in the foundational treatises of Arnold Gesell and Myrtle McGraw, this framework posited that human behavioral ontogeny was the direct, unmediated expression of an unfolding genetic blueprint instantiated through the structural maturation of the central nervous system (CNS). Gesell’s meticulous cinematographic inventories of infant growth codified motor development into rigid normative sequences, conceptualizing milestone achievement as a strictly chronological, biologically determined process. McGraw’s elegant neurobehavioral investigations further solidified the conviction that the emergence and dissolution of early motor patterns mirrored the progressive structural differentiation of neural tissue, specifically the ascending myelination of descending motor pathways.

Central to this neuromaturational paradigm was the doctrine of strict neural hierarchy and cephalocaudal (head-to-tail) and proximodistal (inward-to-outward) control gradients. Development was envisioned as an imperial conquest executed by the cerebral cortex over phylogenetically older, primitive subcortical and spinal structures. In neonates, the lack of extensive cerebral myelination, synaptogenesis, and dendritic arborization was believed to leave the infant under the autocratic dominion of brainstem and spinal cord reflex loops. As the neocortex matured, it was presumed to actively suppress, override, and replace these primitive automatisms through descending, inhibitory corticospinal projections. Thus, any behavioral pattern observed in early infancy was classified either as a subcortical reflex awaiting cortical suppression or as a higher-order voluntary action signaling successful corticalization.

Crucially, this structural-determinist paradigm maintained an almost total theoretical neglect of the physical body and the ambient physical environment. Biomechanical variables—such as gravitational load, soft-tissue mass, inertial moments of the limbs, connective tissue elasticity, and muscle torque capacities—were treated as incidental, passive epiphenomena. The infant was treated essentially as an isolated, disembodied brain wired to an arbitrary biological chassis. Motor competence was equated with neuroanatomical maturity; if a behavior failed to manifest, the default theoretical conclusion was that the requisite neural architecture had not yet developed or that an inhibitory neural switch had been engaged. Consequently, the dynamic real-world interface between somatic growth, physical laws, and active movement was rendered invisible within the prevailing scientific discourse.

1.2 The Canonical View of the Neonatal Stepping Reflex

Among the catalog of primitive reflexes documented by early clinicians, the neonatal stepping reflex—frequently termed the primary walking reflex or the automatic stepping response—held a place of preeminent scientific and clinical fascination. When a healthy human neonate, just hours or days old, is held vertically under the axillae with their bare soles allowed to touch a flat horizontal surface, and the torso is tilted slightly forward, the infant will systematically execute rhythmic, well-coordinated, alternating walking steps. This behavior is characterized by distinct swing and stance phases, complete with hip and knee flexion followed by terminal extension and heel-to-toe or flat-footed contact, bearing an uncanny resemblance to mature bipedal locomotion.

Yet, the developmental timeline of this dramatic behavior exhibited a striking, highly reproducible anomaly. While present and vigorous at birth, the automatic stepping response begins to deteriorate significantly over subsequent weeks, typically undergoing complete cessation or systematic disappearance between the sixth and twelfth weeks of postnatal life (approximately two to three months of age). Following this protracted period of absence—during which infants held upright in an identical manner exhibit passive limb collapse, rudimentary flexion without rhythmic progression, or completely static, flaccid postures—rhythmic, intentional stepping emerges de novo around eight to twelve months of age, culminating in independent walking.

For more than half a century, the standard neurological interpretation of this developmental trajectory was settled law: the disappearance of the stepping reflex was heralded as an unassailable clinical behavioral index of normal, ascending cortical inhibition. Classic pediatric neurologists like Albrecht Peiper and André-Thomas argued that the archaic spinal central pattern generator responsible for primitive stepping was brought under the inhibitory subjugation of the rapidly maturing frontal lobes and precentral motor cortex. Because this functional suppression was deemed essential for the eventual emergence of voluntary, goal-directed, cortically mediated locomotion, the disappearance of the stepping reflex was formally integrated into standard neonatal neurological batteries, including the Prechtl Neurological Examination, the Dubowitz assessment, and the Brazelton Neonatal Behavioral Assessment Scale. An infant who continued to step vigorously at four or five months was frequently pathologized, viewed with clinical suspicion as potentially harboring delayed corticalization, spastic diplegia, or upper motor neuron lesions.

1.3 Theoretical Limitations of Hierarchical Neurological Determinism

Despite its clinical utility and internal aesthetic coherence, the hierarchical neuromaturational model suffered from profound theoretical vulnerabilities, chief among them being its inability to explain contextual variability in infant motor output. In clinical and laboratory observation, the stepping reflex proved notoriously sensitive to environmental conditions, maternal holding techniques, behavioral arousal states, and localized physical contexts. Infants classified as lacking the reflex in one clinical examination would frequently display robust stepping if held at a slightly different postural tilt or stimulated during a transient state of motor excitation. A strictly hardwired cortical-inhibition model could not easily reconcile how descending inhibitory tracts could be instantaneously neutralized by mild changes in holding mechanics or situational arousal.

This empirical fragility exposed a deeper philosophical flaw within the discipline: what philosophers of mind and cognitive scientists have identified as the “ghost in the machine” or homunculus fallacy. By locating the sole agency of behavioral development within the central nervous system, neuromaturational theory treated the brain as an omnipotent, prescriptive executive controller that held internally pre-programmed representations of every milestone. If a motor pattern was expressed, the brain had instructed it; if it disappeared, the brain had inhibited it. This circular logic treated the biological body as a transparent, passive transmission device that seamlessly executed central commands without introducing physical constraints, nonlinear biomechanical noise, or self-organizing dynamic properties into the behavioral equation.

Traditional developmental equations completely omitted physical-dynamic, anatomical, and morphological parameters. No formal accounts were taken of limb segment mass, changes in subcutaneous fat deposition, skeletal muscle cross-sectional area, or the gravitational torque required to accelerate an infant leg through a spatial trajectory. The explanatory failure to incorporate these physical realities revealed an acute epistemological void. By the late 1970s, it became increasingly apparent to a vanguard of cross-disciplinary researchers that explaining developmental transitions required abandoning the disembodied, top-down neurological determinism of the past in favor of an embodied, ecological framework capable of treating the infant as an integrated biological system embedded within a demanding physical world.

2. Esther Thelen’s Paradigm Shift: Dynamic Systems Theory in Motor Control

2.1 Core Principles of Dynamic Systems Theory (DST) Applied to Infancy

Into this theoretical vacuum stepped Esther Thelen, whose revolutionary work fundamentally re-engineered developmental science by importing the conceptual and mathematical framework of Dynamic Systems Theory (DST). Drawing heavily upon the physical biology of Nikolai Bernstein, as well as the non-equilibrium thermodynamics of Ilya Prigogine and the synergetics of Hermann Haken, Thelen posited that motor behavior is not the deterministic printout of a prescriptive central code, but an emergent property born of the real-time, non-linear interaction of heterogeneous, decentralized subsystems. Under DST, the human infant is conceptualized as a multi-component system comprising neural assemblies, anatomical structures, biomechanical forces, perceptual systems, motivational states, and environmental physics.

Within this systems framework, no single component holds executive hegemony over the others. Motor output is structured by the confluence of three distinct categories of constraints: organismic constraints (e.g., neural connectivity, muscle strength, body mass, segment lengths, skeletal flexibility), environmental constraints (e.g., ambient medium, friction, surface stability, the constant downward pull of the gravitational field), and task-specific constraints (e.g., navigating a barrier, maintaining an upright visual field, reaching for a salient object). Development is radically reconceptualized: rather than being a linear, maturationally paced progression dictated by a genetic clock, it is a process of soft assembly. Motor patterns are not hardwired programs stored in cerebral engrams; they are dynamically assembled, flexible coordinative structures configured on the fly to meet the demands of a specific task within a specific physical context.

By defining movement as soft-assembled, Thelen radically challenged the notion of static developmental milestones. A motor milestone is not an all-or-nothing neurological switch that turns on permanently on a predetermined postnatal day; it is a temporary, dynamically stable behavioral solution that crystallizes when all requisite subsystems have achieved a functional threshold of confluence. If even one necessary subsystem lags behind—whether it be the contractile force of an extensor muscle or the visual-vestibular calibration of balance—the overt behavioral pattern cannot manifest, even if the underlying neural network is fully developed and structurally intact.

2.2 Self-Organization, Attractor States, and Nonlinear Development

A foundational tenet of Dynamic Systems Theory is the phenomenon of spontaneous self-organization, a concept derived from non-equilibrium physics. Complex physical systems, when subjected to energetic flows and environmental constraints, can spontaneously generate order and complex spatial-temporal patterns without the intervention of an internal or external organizer. Just as thermal convection currents in a heated fluid spontaneously organize into the geometric symmetry of Bénard cells, or meteorological systems organize into coherent cyclones, the human neuromuscular system self-organizes into stable, coordinated movements through the spontaneous, physical-mathematical cooperation of its constituent anatomical parts.

Within this dynamic landscape, behavioral coordination patterns are mathematically modeled as attractor states residing within an epigenetic potential landscape. An attractor represents a preferred, highly stable state of coordination toward which the biological system naturally gravitates because it minimizes energetic expenditure and maximizes biomechanical efficiency. Highly stable attractors (represented as deep energetic wells) resist perturbation, producing stereotyped, highly reproducible movements. Conversely, shallow attractor wells represent unstable, highly variable coordination states that are easily disrupted or disassembled. During developmental transitions, existing attractor states become destabilized—often through small, continuous changes in a single underlying component—causing the system to undergo a qualitative, non-linear reorganization known as a phase shift or bifurcation.

This dynamic perspective provided an elegant mathematical explanation for the notorious non-linearity of infant motor development. In traditional neuromaturational theory, structural neural growth proceeds along a relatively smooth, continuous curve, making the sudden, discontinuous appearance and disappearance of behaviors an explanatory puzzle. In contrast, Dynamic Systems Theory shows how continuous, quantitative scalar shifts in a single physical parameter (such as body weight, limb length, or contractile muscle force) can drive a complex system past a critical tipping point, precipitating sudden, qualitative reconfigurations of behavioral output. The emergence or loss of a motor pattern does not require the sudden construction or destruction of a neural program; it simply reflects the non-linear shifting of attractor dynamics within a self-organizing physical system.

2.3 Challenging the Brain-as-Dictator Metaphor in Developmental Psychology

The core of Esther Thelen’s intellectual insurgency was an uncompromising assault on the brain-as-dictator metaphor that had long held developmental psychology captive to a Cartesian dualism. For generations, cognitive science and developmental psychology had operated under an explicit encephalization bias, exalting the brain as an omniscient, detached computational machine while treating the peripheral body as merely a passive, generic actuator. Thelen argued that divorcing motor output from anatomical and physical realities was a fatal scientific error, one that fundamentally misconstrued the evolutionary and biological nature of the human organism.

In Thelen’s radical paradigm, agency is decentered from the cerebral cortex and relocated to the holistic, real-time coupling between the brain, the somatic body, and the physical environment. The central nervous system is demoted from an all-powerful puppet master to an active, co-equal partner in a decentralized collaborative network. Neural commands do not dictate physical motion in a vacuum; they must navigate the physical properties of biological tissues—the visco-elasticity of tendons, the force-velocity profiles of muscle fibers, and the complex cross-joint inertial torques generated whenever multiple skeletal segments accelerate through space. The brain does not calculate every complex differential equation of multi-joint physics; it relies on the self-organizing physical properties of the periphery to absorb, guide, and simplify the motor problem.

By establishing that biomechanics, physical mass, and morphological geometry are constitutive components of motor coordination, Thelen laid the empirical and philosophical groundwork for a dramatic re-evaluation of early neonatal automatisms. If motor behavior is fundamentally embodied, then the enigmatic disappearance of the neonatal stepping reflex could no longer be uncritically accepted as a purely neuro-cortical phenomenon. It was time to look beyond the cranial vault and systematically examine the changing physical body of the growing human infant.

3. The Paradox of the Disappearing Reflex: Formulating the Research Hypothesis

3.1 The Developmental Discrepancy Between Supine Kicking and Upright Stepping

The flashpoint of Esther Thelen’s historic inquiry emerged from a striking behavioral paradox that had been hidden in plain sight for decades. In the late 1970s and early 1980s, Thelen conducted exhaustive, fine-grained observational and cinematographic studies of spontaneous motor behavior in healthy human infants. During these investigations, she documented that while the neonatal stepping reflex systematically deteriorated and vanished between the second and fourth months of life when infants were suspended upright, these very same two-, three-, and four-month-old infants continued to perform vigorous, rhythmic, alternating lower-extremity kicking when placed lying on their backs in a supine position.

When Thelen meticulously examined the spatio-temporal structure and joint kinematics of these supine kicks, she uncovered a profound morphological revelation: the kinematic profile of the infant supine kick was virtually identical to the kinematic profile of the neonatal stepping reflex. Both behaviors exhibited the same canonical, highly stereotyped sequence of intra-limb joint coordination—a rapid, coupled flexion of the hip, knee, and ankle joints, followed by a coordinated, terminal extension phase, accompanied by a precise temporal alternation between the left and right lower extremities. The topological morphology of the movement was invariant; the motor system was executing the exact same coordinated synergy in both postures.

This empirical observation presented a catastrophic logical contradiction to the traditional neuromaturational doctrine of cortical inhibition. If the lumbosacral spinal central pattern generators or subcortical motor circuits responsible for alternating leg coordination had been functionally suppressed, overridden, or cortically inhibited by ascending neocortical pathways, how could that identical neuro-pattern generator fire with unbridled vigor and rhythmic precision the moment the infant was placed in a supine posture? To claim that the cerebral cortex possessed a magical, hyper-selective inhibitory mechanism that suppressed a neural circuit exclusively when the infant was held vertically, while completely disinhibiting the identical circuit the moment the infant reclined horizontally, was biologically implausible. The contradiction demanded an entirely new hypothesis: posture and physics, rather than cortical encephalization, were dictating the phenotypic expression of the neuromuscular pattern.

3.2 Biomechanical Dynamics: Muscle Mass Versus Subcutaneous Adiposity

To resolve this paradox, Thelen turned her scientific attention to the fundamental, dramatic anthropometric transformations that characterize the early postnatal period of human life. Human infants experience a growth trajectory that is biologically unique among primates: during the first two to three months after birth, they undergo a monumental surge in body mass driven predominantly by the rapid, disproportionate accumulation of subcutaneous adipose tissue. In evolutionary terms, this neonatal fat accumulation serves critical metabolic and thermoregulatory functions, providing an energy reserve for rapid cerebral expansion and insulating a vulnerable organism against hypothermia.

However, from a physical and biomechanical standpoint, this surge in adiposity introduces a profound mechanical challenge. While the infant’s total body mass and the physical volume of the lower limbs increase rapidly between four and twelve weeks of life, the development of functional skeletal muscle mass does not keep pace. Muscle tissue matures along a significantly slower biochemical and structural trajectory; myofibrillar protein accretion, cross-sectional muscle area expansion, and voluntary motor unit recruitment undergo a protracted, gradual maturation. The immediate physiological consequence of this developmental divergence is a precipitous drop in the infant’s muscle-strength-to-body-mass ratio. In simple terms: over the first two to three months of life, the infant’s legs become substantially heavier, thicker, and more massive without a commensurate increase in the muscular strength required to move them.

Synthesizing these anthropometric facts with mechanical physics, Thelen formulated her revolutionary counter-hypothesis: the neonatal stepping reflex does not disappear because of active neuro-cortical suppression; it disappears because the infant’s lower extremities become too heavy for their underdeveloped muscle tissue to lift against the unforgiving downward pull of Earth’s gravity. When an infant is held upright, initiating a step requires the hip flexor muscles (primarily the iliopsoas) to generate sufficient dynamic torque to lift the substantial mass of the thigh, shank, and foot upward against gravity into the swing phase. In contrast, when the same infant is supine, the leg is kicked across a horizontal plane where the gravitational vector acts perpendicularly to the plane of motion, or the limb is supported by the flat surface of the crib, radically altering the mechanical torque required to produce alternating flexion and extension. The reflex, Thelen reasoned, had not been silenced by the brain; it was mechanically suppressed by somatic weight.

3.3 Operationalizing the Dynamic Hypothesis for Empirical Testing

Transforming this theoretical insight into rigorous, falsifiable experimental science required operationalizing the dynamic hypothesis through precision empirical manipulation. Thelen recognized that if her biomechanical model was correct, the overt presence or absence of stepping behavior should be directly manipulable by altering the physical, gravitational constraints acting upon the infant’s lower limbs, completely independent of their underlying neurological maturation.

Thelen and her team formulated two reciprocal, highly daring experimental predictions designed to test the limits of physical-dynamic causality in infant behavior:

• The Offloading Prediction (Buoyancy/Water Immersion): If older, two- to three-month-old infants—who had already experienced the natural “disappearance” of the stepping reflex and demonstrated zero stepping in standard clinical conditions—were placed in an environment that physically counteracted or minimized the gravitational load on their lower limbs without altering their central nervous systems, their underlying neuromuscular competence would instantly resurface. Specifically, submerging them in a buoyant medium (water) would offset their excess limb mass, decrease the mechanical torque required for hip flexion, and immediately restore the coordinated stepping reflex.
• The Mass-Augmentation Prediction (Artificial Loading): Reciprocally, if younger, two- to four-week-old neonates—who still possessed a robust, highly active stepping reflex—had their limbs artificially weighted with mass equivalent to the adipose tissue they would naturally accumulate over the subsequent two months, the stepping reflex would instantly vanish. The added mechanical inertia would mimic future somatic growth, exceeding the neonate’s immediate force-generation capacity and extinguishing the reflex weeks ahead of its natural developmental schedule.

These two experimental conditions established an unprecedented evidentiary threshold. If changing the effective mass of an infant’s legs could instantaneously toggle a classic neurological reflex on and off at will—in direct defiance of the infant’s chronological age and structural cortical status—the classical neuromaturational model of cortical inhibition would be definitively falsified. The stage was set for one of the most elegant, consequential experimental designs in the history of developmental science.

4. Experimental Methodology of the Landmark 1984 Thelen Studies

4.1 Cohort Selection, Demographics, and Ethical Rigor

The empirical architecture designed to test these dynamic hypotheses culminated in the landmark 1984 study published by Esther Thelen, Donna M. Fisher, and Ridley-Johnson in the journal Child Development, entitled “The Relationship Between Physical Growth and a Newborn Reflex”. The experimental paradigm was executed with extraordinary methodological rigor, combining longitudinal and cross-sectional cohorts of healthy, full-term human infants to track the exact developmental inflection points of physical growth and reflex expression.

The research design sampled infants systematically across their critical first three months of life, stratifying cohorts to capture subjects before, during, and well after the conventional normative window of reflex disappearance. The primary experimental cohorts consisted of normal, healthy, full-term infants: a younger group (tested around 2 to 4 weeks of age) who exhibited vigorous, baseline neonatal stepping, and an older group (tested between 4 and 12 weeks of age) who had reached the developmental threshold where upright stepping had dropped to zero or near-zero levels. Stringent screening criteria ensured that all enrolled infants had normal birth weights, unremarkable APGAR scores, no history of perinatal distress, and completely intact, age-appropriate neurological profiles.

Conducting complex biomechanical and hydro-immersion experiments on fragile human infants required exacting ethical and procedural protocols. The laboratory environment was meticulously thermally regulated; ambient air temperatures were maintained at high comfort levels to prevent cold-induced motor bracing, and all aquatic immersion testing utilized precisely calibrated, skin-neutral water temperatures to completely avert thermal shock or autonomic distress. Furthermore, recognizing that infant motor output is notoriously coupled to neurobehavioral state, Thelen and her colleagues standardized all experimental trials to occur strictly during periods of alert, contented, non-crying behavioral equilibrium (Prechtl States 3 and 4). If an infant became drowsy, fussed, or transitioned into active crying, testing was immediately paused until equilibrium was restored, ensuring that alterations in stepping frequency were purely a function of biomechanical manipulation rather than systemic emotional distress or autonomic agitation.

4.2 Kinematic Recording and Observational Protocols

To capture the transient, multi-dimensional physical movements of the infants with mathematical precision, Thelen abandoned crude, subjective clinical tallies in favor of quantitative, microgenetic kinematic instrumentation. The laboratory was equipped with high-speed cinematographic and synchronized video recording systems, spatially configured to capture orthogonal, multi-planar views of the infant’s lower extremities with high spatial and temporal fidelity.

High-contrast, non-toxic optical reference markers were affixed directly over specific anatomical landmarks of the infant’s pelvis and lower limbs: the greater trochanter of the hip, the lateral condyle of the femur (knee joint axis), the lateral malleolus of the fibula (ankle joint axis), and the distal head of the fifth metatarsal. Spatial calibration grids were recorded within the camera fields before and after each experimental session, allowing the two-dimensional and three-dimensional spatial coordinates of these markers to be digitized frame-by-frame. This permitted the continuous, high-resolution mathematical reconstruction of joint angular trajectories, angular velocities, and intra-limb phase dynamics across every phase of movement.

Step-cycle metrics were governed by rigorous, objective behavioral coding criteria. A fully coordinated step was operationalized not as an arbitrary leg jerk, but as a complete, synchronized cycle: an active initiation phase characterized by concurrent hip and knee flexion, a swing phase where the foot cleared the supporting surface and moved anteriorly relative to the trunk, a terminal extension phase where the knee uncoupled and the foot descended, and a definitive stance phase where the foot contacted the support surface and sustained weight-bearing alternation with the contralateral limb. Inter-observer reliability was rigorously maintained, with multiple independent coders blinded to the specific dynamic hypotheses analyzing the video records, achieving inter-rater reliability coefficients consistently exceeding r = .90.

4.3 Electromyographic (EMG) Instrumentation and Data Acquisition

A crucial dimension of the 1984 experimental paradigm was demonstrating that the physical manipulations of limb weight did not disrupt, fundamentally alter, or fabricate novel central motor commands, but acted directly on the biomechanical peripheral execution of existing neural signals. To directly observe the underlying neuromuscular drive, Thelen and her colleagues integrated surface electromyography (EMG) into their experimental recording apparatus.

Miniature, highly sensitive surface electrodes were affixed over the bellies of major antagonistic muscle pairings across the infant’s lower extremities: the rectus femoris and vastus lateralis (quadriceps/knee extensors) paired against the biceps femoris (hamstrings/knee flexors), and the tibialis anterior (ankle dorsiflexor) paired against the gastrocnemius-soleus complex (ankle plantarflexors). Inter-electrode distances were carefully calibrated to minimize electrical cross-talk between the small, closely spaced muscle groups of the infant leg, and the skin was prepared with rigorous cleansing to keep electrical impedances below acceptable clinical thresholds.

The raw electromyographic signals were amplified, bandpass filtered, and simultaneously recorded alongside the kinematic video frames using automated temporal pulse synchronization. This integration allowed the researchers to capture the exact burst durations, amplitude envelopes, and reciprocal firing latencies of the antagonistic muscle groups relative to kinematic movement onsets. Baseline EMG profiles were systematically established across resting states, spontaneous supine kicking, and held-upright postures. This electromyographic apparatus provided an unassailable window into the infant’s nervous system: it allowed Thelen to verify whether descending neural bursts were actively firing into the peripheral musculature even when the limb itself was physically incapable of translating that neural energy into overt, visible spatial movement.

5. The Water Immersion Experiment: Buoyancy as an Environmental Perturbation

5.1 Physical Principles of Hydrostatic Buoyancy and Gravitational Offloading

The core empirical breakthrough of Thelen’s research program rested upon an ingenious environmental perturbation: utilizing the physical properties of water to selectively cancel out the downward force of gravity. The physical mechanism governing this experimental manipulation is dictated by Archimedes’ principle, which states that any body completely or partially submerged in a fluid experiences an upward buoyant force equal to the weight of the fluid displaced by that body. Mathematically, the net vertical force acting upon an infant’s submerged lower extremity can be expressed as:

F_net = F_gravity – F_buoyant = (m_limb × g) – (ρ_fluid × V_submerged × g)

Where m_limb is the somatic mass of the leg, g is the acceleration due to gravity, ρ_fluid is the density of the water, and V_submerged is the physical volume of the submerged limb. Because the biological tissues comprising the human leg (specifically adipose tissue, which has a specific gravity less than 1.0, and muscle/bone, which have specific gravities slightly greater than 1.0) possess an average mass density that closely approximates the density of freshwater, immersing the limb dramatically reduces its effective underwater weight to a mere fraction of its dry-land value.

The biomechanical consequences of this hydrostatic offloading are profound. When an infant is submerged to the waist, the upward buoyant force directly counteracts the downward gravitational vector, radically diminishing the net gravitational torque that the immature proximal hip flexors (primarily the iliopsoas and rectus femoris) must generate to elevate the thigh. Crucially, this dramatic reduction in gravitational load is achieved without any mechanical encumbrance, external splinting, or restrictive mechanical rigging; the infant retains 360 degrees of ambient joint freedom. While the water introduces a minor component of viscous fluid drag that damps high-velocity movements, at the slow speeds of infant stepping, this hydro-resistive force is negligible compared to the massive, liberating reduction in gravitational mass. Furthermore, the hydrostatic pressure of the water provides a subtle, uniform, stabilizing surround that supports the trunk and lower pelvis, functionally simplifying the complex biomechanical control problem facing the infant.

5.2 The Submersion Protocol and Experimental Controls

The physical execution of the water immersion experiment was as elegant as it was conceptually radical. Thelen and her co-investigators constructed a specialized, optically transparent testing vessel that allowed completely unimpeded, multi-angle cinematographic tracking of the submerged lower extremities. The water level within the chamber was carefully calibrated to the individual anthropometric proportions of each infant, consistently reaching the level of the xiphoid process or mid-trunk, thereby ensuring that the entirety of both lower extremities was fully submerged while the head, neck, and upper chest remained comfortably above the surface.

Thermal neutrality was maintained with obsessive precision. The water was continuously monitored and kept between 34°C and 36°C (93.2°F to 96.8°F), precisely matching the normal cutaneous surface temperature of an infant. This vital control eliminated any possibility that thermal shock, cold-induced shivering, or temperature-related autonomic arousal could trigger non-specific neuromuscular activation. The experiment utilized a rigorous within-subject, repeated-measures control design. Each infant was tested across tightly sequenced, alternating baseline and experimental blocks:

• Condition A (Air Baseline): The infant was held upright under the axillae by an experienced experimenter, suspended vertically over a dry, solid platform with the soles lightly touching the surface, and tilted slightly forward in the classic posture used to elicit the stepping reflex. Behavioral stepping frequency was quantified over standardized time intervals.
• Condition B (Water Immersion): The infant was immediately transferred into the transparent aquatic vessel, held in an identical upright, forward-tilted posture under the axillae, with their feet lightly contacting the bottom floor of the tank, and their lower body fully submerged up to the waist. Identical observational and kinematic recording protocols were executed.
• Condition C (Return to Air): The infant was lifted out of the water, dried, and immediately re-tested in the dry-air baseline condition to assess post-immersion rebound, fatigue, or carry-over effects.

The primary experimental focus was targeted precisely at older infants (around 4 to 8 weeks of age and beyond) who, during the pre-test air baseline, exhibited the canonical developmental disappearance of the reflex: zero steps, completely unresponsive lower extremities, or passive, flaccid dragging of the limbs across the support surface.

5.3 Empirical Findings: Instantaneous Restoration of Stepping Behavior

The empirical results of the water immersion trials were instantaneous, dramatic, and unequivocal. The moment these older infants—who had demonstrated absolute zero stepping in the air baseline—were lowered into the buoyant warm water, their lower extremities spontaneously sprang into action. The infants began executing robust, rhythmic, highly coordinated alternating steps. Submerged to the waist, their legs moved through fluid, sweeping swing and stance cycles, with the left and right legs alternating in a classic reciprocal locomotion pattern that was visually and statistically indistinguishable from the stepping reflex of a newborn infant.

Quantitative analysis revealed a statistically massive, highly significant increase in stepping frequency under water immersion compared to air. Infants who had produced an average of zero to one rudimentary step over multiple trials in air produced dozens of coordinated steps in water within identical timeframes. Kinematic reconstruction of the joint trajectories demonstrated that the submerged steps were not unorganized, spastic, or thrashing leg movements; they displayed the complete, mature kinematic architecture of true stepping: rapid hip and knee flexion driving the limb upward through the water, followed by forward excursion and smooth extension of the knee, culminating in downward stance contact.

Crucially, this behavioral resurgence occurred immediately upon submersion, without any preceding practice trials, motor conditioning, reinforcement schedules, or priming periods. The infant did not need to “learn” how to walk in water; the coordinated motor behavior was unlocked the microsecond the physical constraint of gravitational load was lifted from the system. This single empirical observation delivered a devastating blow to the classical neuromaturational paradigm. If the neural circuitry governing the stepping reflex had been systematically inhibited, destroyed, or suppressed by a maturing cerebral cortex, it would be physically impossible for that identical neural pattern to reappear instantaneously simply because the legs were wet. The neural program had never disappeared; it had merely been masked by the physics of somatic growth.

6. The Counter-Condition: The Weighted Leg Experiment

6.1 Rationale for Artificially Modifying Limb Inertia in Neonates

Scientific rigor demands that if a dynamic hypothesis claims a specific physical parameter is the primary causal agent governing a developmental transition, the researcher must demonstrate bidirectional control over the phenomenon. Esther Thelen recognized that demonstrating the restoration of stepping via gravitational offloading in older infants was only half of the empirical equation. To conclusively prove that somatic mass accumulation is the true rate-limiting constraint responsible for reflex disappearance, she had to execute the reciprocal experiment: artificially extinguishing the stepping reflex in young neonates who still naturally possessed it.

Thelen formulated an ingenious experimental counter-condition based on the logic of simulated ontogeny. In normal human development, an infant accumulates massive amounts of subcutaneous adipose tissue over their first ten weeks of life, progressively increasing the inertial mass and physical weight of the lower limbs. Thelen reasoned that if this natural accretion of mass is what causes the stepping reflex to disappear, one should be able to compress two months of biological fat accumulation into a five-minute laboratory manipulation. By attaching artificial, calibrated weights to the lower legs of young neonates (2 to 4 weeks old) who were actively and vigorously stepping, the researchers could experimentally recreate the high-mass/low-strength ratio that characterizes an older infant.

The precise mechanical objective was to artificially increase the limb’s downward gravitational force vector to a point where it exceeded the maximum isometric and dynamic torque capacity of the neonate’s hip flexor musculature. If the neuromaturational theory were correct, and the presence of stepping in young neonates is an uninhibited, mandatory subcortical reflex hardwired to fire upon tactile and proprioceptive upright stimulation, the infant’s CNS should continue to fire, and the legs should continue to step regardless of mild inertial variations. If, however, Thelen’s dynamic systems hypothesis were correct, artificially augmenting the limb’s mass would instantly extinguish the reflex, simulating the natural developmental “disappearance” weeks ahead of its normative chronological schedule.

6.2 Methodology of Precision Weight Loading

To execute this artificial loading without causing mechanical injury, soft-tissue bruising, or localized joint immobilization, Thelen and her team engineered custom-fabricated, precision-calibrated cuff weights. These wearable weights were constructed from thin, flexible fabric channels embedded with fine, distributed lead strips, designed to wrap smoothly and securely around the infant’s lower extremities. The mass of the weights was precisely calculated for each individual subject, scaled proportionately to the infant’s specific anthropometric measurements (body mass, thigh circumference, and shank length) to mirror the exact percentage increase in lower-limb mass that an infant typically experiences between four and eight weeks of postnatal life (typically an addition of only a few ounces per limb, yet mechanically monumental relative to the infant’s scale).

The attachment locations were strategically selected: the lead-strip cuffs were affixed circumferentially around the lower shank, immediately superior to the ankle joint malleoli. This distal placement was biomechanically deliberate: by placing the additional mass at the distal end of the limb segment, the researchers maximally increased the limb’s moment of inertia and rotational torque demands around the proximal hip joint axis without mechanically splinting, restricting, or encumbering the natural range of motion of the knee or ankle joints. The infants retained full, unhindered mechanical freedom to flex, extend, abduct, and rotate their joints through their complete anatomical arcs.

The experimental protocol utilized a tightly controlled, within-subject ABA design consisting of three continuous phases:

• Phase 1 (Baseline Unweighted): The young neonate was held upright in standard fashion, and baseline stepping frequency, step amplitudes, and kinematic joint trajectories were recorded in the unweighted state.
• Phase 2 (Mass-Loaded Condition): The calibrated lead cuffs were secured to the infant’s lower legs. The infant was held in the exact same posture, and stepping behavior was recorded across identical temporal durations.
• Phase 3 (Wash-out / Unweighted Recovery): The lead weights were immediately unstrapped and removed from the legs, and the infant was instantly re-tested in the unweighted state to evaluate immediate behavioral recovery and rule out systemic physiological fatigue.

Throughout all phases, continuous behavioral monitoring guaranteed that the infant remained in an alert, calm, non-distressed behavioral state, confirming that any change in motor output was directly attributable to mechanical loading rather than emotional distress or behavioral shutdown.

6.3 Observed Outcomes: Immediate Elimination of the Stepping Reflex

The empirical outcome of the mass-loading experiment provided the mirror-image proof of the water immersion study. When the calibrated weights were affixed to the legs of young neonates who had just seconds prior been stepping with rhythmic, vigorous abandon, the stepping reflex experienced immediate, catastrophic suppression. The stepping frequency dropped precipitously, with the majority of infants displaying a total, complete cessation of coordinated stepping movements.

When held upright with the weights attached, the infants attempted to move, but their motor output was profoundly altered. Kinematic analysis revealed that while the infants still attempted to initiate movements, they could no longer successfully complete the swing phase of the step. The legs remained held against the support surface, or the infants produced small, abortive, low-amplitude twitches of the hip that failed to overcome the augmented gravitational inertia of the limb. The physical weight of the cuffs had rendered the limbs mechanically unliftable for the infant’s developing muscular system.

The proof that this cessation was purely mechanical—and not the result of sudden behavioral apathy, generalized fatigue, or neural exhaustion—was delivered during Phase 3. The exact instant the lead cuffs were unstrapped and removed from the infants’ ankles, their stepping behavior immediately surged back with full, unabated vigor. The infants instantly resumed their rhythmic, alternating stepping at baseline frequencies and amplitudes. By artificially manipulating a single physical variable—limb mass—Thelen had successfully induced the “disappearance” of a classic neurological reflex in two-week-old infants, and then instantly resurrected it seconds later. The classic dogma that reflex disappearance is a permanent, biologically fixed milestone of cortical maturation was empirically pulverized.

7. Kinematic and Neuromuscular Parallels: Kicking Versus Stepping

7.1 Comparative Kinematic Analysis of Supine Kicking and Upright Stepping

To establish an unassailable empirical foundation for her dynamic systems model, Thelen recognized that she needed to mathematically prove that the motor coordination patterns driving upright stepping and supine kicking were fundamentally the same behavioral animal dressed in different gravitational clothing. To achieve this, she performed exhaustive kinematic comparative analyses, utilizing phase-plane trajectories, cyclograms (angle-angle plots), and continuous angular velocity profiles across the hip, knee, and ankle joints during both motor activities.

When the spatial and temporal data were plotted, the kinematic similarities were extraordinary. In both supine kicking and upright stepping, the lower extremity operates through a tightly coupled, topologically equivalent kinematic synergy. During the flexion phase of both movements, the hip, knee, and ankle do not move independently; they exhibit strong, synchronized intra-limb coupling. The hip flexes at a fixed phase relationship with the knee, while the ankle dorsiflexes in tight lockstep. As the limb reaches peak excursion, the system transitions into an extension phase where the knee and hip extend concurrently toward the resting boundary. When Thelen plotted knee angle against hip angle in phase-plane portraits, the resulting orbital loops tracing the trajectory of the limb through state space were virtually identical in both postures.

The only fundamental difference between the two behaviors lay in the external mechanical work required to negotiate gravity. In the upright position, the infant’s body axis is parallel to the downward gravitational vector (g). To execute a step, the hip flexor musculature must lift the entire mass of the lower limb directly upward against this vertical vector, performing continuous positive mechanical work against gravity. In the supine posture, the infant’s longitudinal body axis is perpendicular to the gravitational vector. When a supine infant kicks, the leg flexes upward toward the abdomen, but the physical support surface of the mattress or crib absorbs the baseline gravitational load of the pelvis, and the limb segments oscillate across a mechanical plane where the gravitational resistance against hip flexion is drastically altered. The infant is executing the exact same motor synergy, but the physics of the supine posture grant a massive mechanical advantage that masks the underlying muscle weakness.

7.2 Electromyographic Equivalence of Muscle Activation Profiles

Kinematic trajectories describe the geometry of movement through space, but electromyography (EMG) reveals the physiological instructions issuing from the central nervous system to the peripheral muscles. To verify whether the neural driving engines behind supine kicking and upright stepping were identical, Thelen analyzed the fine-grained temporal architecture of muscle activation using surface EMG across the rectus femoris, hamstrings, tibialis anterior, and gastrocnemius.

The electromyographic findings demonstrated a stunning neurophysiological equivalence between the two behaviors. Both upright stepping and supine kicking were driven by the exact same underlying temporal muscle burst morphology:

• Agonist Burst Timing: Movement initiation in both behaviors was preceded by a sharp, high-amplitude burst of electrical activity in the primary flexors (iliopsoas, rectus femoris, and tibialis anterior).
• Co-activation and Reciprocity: Early infant motor output is characterized by a high degree of antagonistic co-activation, where flexors and extensors contract concurrently to stiffen the compliant infant joints. Thelen’s EMG records revealed the identical, distinctive signature of co-activation—with reciprocal relaxation emerging only at specific phase transitions—in both supine kicks and upright steps.
• Inter-limb Phasing: Electromyographic recordings across both legs simultaneously demonstrated that the alternating, reciprocal temporal pattern (where the firing of flexor bursts in the left leg alternates precisely with the firing of flexor bursts in the right leg with an approximate 50% phase lag) was completely identical whether the infant was suspended upright or lying flat on their back.

This electromyographic congruence provided definitive biological evidence that the descending and spinal motor commands issuing from the infant’s nervous system are entirely functional, active, and persistent across both postures. The spinal pattern generator does not turn off when the infant is rotated from horizontal to vertical. The nervous system broadcasts the identical rhythmic motor program into the peripheral musculature in both postures; however, in the upright posture, the peripheral physical chassis simply lacks the structural force to translate that electrical signal into spatial joint displacement against the gravitational load.

7.3 Gravity as a Contextual Modulator Rather Than a Constant

These kinematic and electromyographic discoveries forced a radical philosophical reappraisal of the role of gravity in biological movement. In traditional, computational, and neuromaturational models of motor control, gravity was treated essentially as an irrelevant constant—a passive, background environmental parameter that the executive brain effortlessly accounts for through internal representational algorithms. Thelen turned this assumption on its head, demonstrating that gravity is an active, dynamic, context-defining constraint that directly co-authors the final phenotypic form of biological movement.

Because the biological body is an interconnected, multi-segment physical pendulum, gravitational forces generate massive, non-linear inter-segmental torques whenever a limb segment accelerates. A change in the infant’s orientation in space instantly and completely transforms the coordinate frame and the mechanical equations governing movement. The exact same neural motor command, generated by the exact same collection of spinal motor neurons, will produce radically different behavioral outputs depending entirely on how the limb is positioned relative to the gravitational field. If the torque demands imposed by gravity exceed the contractile force of the muscles, the movement mechanically collapses into zero displacement, creating the optical illusion of a non-existent or “inhibited” neural program.

By establishing that posture and gravitational orientation dynamically modulate the manifestation of motor competence, Thelen demonstrated that motor development cannot be measured or understood through cerebral metrics alone. Motor competence is not an abstract, internal neural property possessed exclusively within the brain tissue of the infant; it is an emergent, real-time relational phenomenon that exists only within the functional coupling between the infant’s physical body, their neuromuscular activation patterns, and the external physical field in which they are embedded.

8. Physical Growth and Anthropometrics as Rate-Limiting Controllers

8.1 The Concept of the ‘Rate Limiter’ in Motor Competence

To formalize these biomechanical insights within the mathematical framework of Dynamic Systems Theory, Esther Thelen and her lifelong collaborator Linda B. Smith introduced the critical concept of the rate limiter (or control parameter) into developmental psychology. Borrowed from chemical kinetics and non-linear dynamics, a rate limiter is defined as the slowest-developing component or subsystem within a complex, multi-component system that constrains, limits, or dictates the overall behavioral output of the entire system.

A complex behavior like bipedal locomotion requires the simultaneous, harmonious integration of dozens of distinct physiological and physical subsystems: central pattern generators, myelinated descending corticospinal tracts, sensory afference from muscle spindles and cutaneous receptors, vestibular balance mechanisms, postural righting reflexes, joint range of motion, extensor and flexor muscle strength, bone density, and an optimal body-mass-to-strength ratio. Within this multi-causal network, the system cannot advance to a new behavioral attractor state (such as independent upright stepping) simply because one or two components (such as cortical myelination or spinal CPGs) are mature. The entire system is held hostage by its slowest-developing component. Until that specific rate-limiting subsystem crosses a critical, functional threshold, the overt behavior remains latent and unexpressed.

In the case of the disappearing stepping reflex, Thelen identified that the absolute rate limiter was not neural readiness, but the ratio of functional muscular strength to somatic adipose mass. The underlying neural competence for alternating bipedal stepping is present and functional at birth, and persists continuously throughout infancy. However, the child’s overt performance is constrained by the lagging developmental curve of skeletal muscle hypertrophy relative to explosive adipose accumulation. This distinction between latent neural competence and observable somatic performance dismantled decades of diagnostic dogma, providing an elegant explanation for why infant development is characterized by prolonged behavioral plateaus followed by sudden, non-linear transitions—all without requiring the sudden appearance of novel, prescriptive neural programs.

8.2 The Postnatal Growth Curve: Fat Mass Versus Muscle Mass

The biological engine driving this rate-limiting mechanism lies in the unique, highly asynchronous anthropometric growth curves of the human infant during the first six months postpartum. Pediatric growth studies demonstrate that human neonates experience an astonishing growth spurt immediately following birth. Over the first twelve weeks of life, a healthy infant typically gains between 20% and 30% of their total birth weight. Crucially, body composition analyses reveal that this rapid weight gain is profoundly skewed toward the deposition of fat mass.

Subcutaneous adipose tissue expands at an exponential velocity during this early postnatal window, serving as an evolutionary metabolic safety net. In stark contrast, functional skeletal muscle tissue expands at a significantly slower, more linear rate. Myofibrillar protein synthesis, the maturation of actin and myosin cross-bridge cycling, and the proportional cross-sectional area of skeletal muscle fibers undergo a steady but protracted developmental timeline. As a consequence, during the exact window when the stepping reflex classically disappears (between 6 and 10 weeks of age), the infant experiences an anthropometric nadir in their strength-to-weight ratio. Their limbs become physical anchors: dense with viscous, non-contractile fat mass that adds mechanical inertia and weight without contributing a single dyne of contractile force.

This anthropometric imbalance imposes severe mechanical consequences on the infant’s musculoskeletal geometry. Because the mass of the infant’s leg is distributed unevenly—with substantial adipose accumulation across the distal thighs and shanks—the limb’s radius of gyration and moment of inertia around the hip joint axis increase dramatically. When an infant is held upright, the rotational torque required to accelerate the leg upward into a step can be mathematically modeled as:

τ_required = I × α + m_limb × g × r_cm × sin(θ)

Where I is the moment of inertia, α is the angular acceleration, r_cm is the distance from the hip axis to the limb’s center of mass, and θ is the joint angle relative to the vertical gravity vector. As m_limb and r_cm balloon due to fat deposition, the required torque (τ_required) expands exponentially. The immature, cross-sectionally narrow iliopsoas and rectus femoris muscles simply cannot generate the necessary isometric or isotonic contractile force to overcome this massive rotational torque demand. The limb remains tethered to the ground by simple Newtonian physics.

8.3 The Shift from Continuous Growth to Discontinuous Behavioral Phenomena

One of the most profound theoretical achievements of Thelen’s 1984 work was resolving a long-standing epistemological dilemma in developmental biology: how can continuous, gradual physical growth produce sudden, discontinuous, all-or-nothing behavioral changes? In classical developmental theory, whenever a behavior disappeared or appeared abruptly, scientists assumed that an abrupt, qualitative neural event—such as the sudden closure of a synapse, the sudden onset of myelination, or the flipping of a genetic switch—must have orchestrated the transition.

Using the mathematical principles of Dynamic Systems Theory and catastrophe theory, Thelen demonstrated that smooth, completely continuous, quantitative changes in a single physical parameter can naturally drive a complex system across a critical threshold, precipitating a sudden, non-linear catastrophe or bifurcation. Consider the infant’s growth: body weight, limb mass, and muscle fiber diameter all increase along smooth, continuous, unbroken trajectories. There are no sudden leaps or breaks in the infant’s physical growth charts. Yet, because movement requires muscle force to exceed the mechanical threshold of gravitational torque, the interaction between these two continuous curves generates a sharp, discontinuous behavioral cliff:

• At Birth (0 to 2 weeks): Muscle strength is marginally sufficient to overcome the relatively light, lean mass of the newborn limb. The infant steps.
• At the Tipping Point (6 to 8 weeks): Subcutaneous fat deposition outpaces muscle hypertrophy, pushing the limb’s gravitational inertia past the maximum force-generation threshold of the hip flexors. In a single week, the behavior drops from frequent stepping to absolute zero. The reflex appears to have “vanished.”
• At Re-emergence (8 to 12 months): Gradual, continuous muscle hypertrophy, skeletal elongation, and neural recruitment eventually cross the threshold in the opposite direction. Muscle force finally catches up to and exceeds somatic mass, restoring the capacity to lift and swing the leg in an upright posture. Voluntary stepping emerges.

This dynamic model effortlessly explained historical anomalies and individual differences that had bedeviled traditional pediatric milestone charts for decades. Thelen observed that leaner, longer infants—who naturally possessed lower fat-to-muscle ratios and longer mechanical lever arms—retained the neonatal stepping reflex significantly longer into their second and third months than plump, heavy infants who experienced explosive fat deposition. Under a neuromaturational model, one would have to make the absurd claim that plumper infants experience faster cortical myelination and brain maturation than lean infants. Under Thelen’s dynamic systems model, the variation was revealed for what it truly was: a direct, predictable consequence of morphological variance and Newtonian mechanics.

9. Deconstructing the ‘Disappearance’: Masking vs. Cortical Inhibition

9.1 Refuting the Classical Cortical Suppression Hypothesis

The empirical results of Esther Thelen’s 1984 experiments systematically dismantled the foundational assumptions of classical pediatric neurology regarding the stepping reflex. For nearly a century, the authoritative treatises of Myrtle McGraw, Albrecht Peiper, and their intellectual successors had asserted that the loss of early reflexes was a direct manifestation of structural encephalization: as the cerebral cortex matured, its expanding descending pathways (specifically the corticospinal tracts) established active, tonic inhibitory control over primitive spinal and subcortical pattern generators. Reflex suppression was championed as the premier behavioral biomarker of a healthy, ripening cerebrum.

Thelen exposed the fatal neurophysiological flaw at the heart of this dogma. A structural, neuroanatomical inhibitory mechanism cannot be posture-specific or density-dependent. If the cerebral cortex had extended functional inhibitory synaptology into the spinal cord to suppress the alternating locomotor pattern generator, that pattern generator would be rendered structurally dormant across all behavioral domains. It is a biological and logical impossibility for cortical inhibition to actively suppress alternating leg coordination when an infant is held upright in air, yet instantaneously permit that exact same alternating coordination to fire the moment the infant’s body is rotated ninety degrees into a supine crib posture, or the moment the infant’s waist is lowered into a tub of warm water.

Thelen proved that the classical neuromaturationalists had fallen victim to a profound category error: they had confused the absence of physical performance with the absence of biological competence. They looked at a baby whose legs were not moving in an upright posture, inferred that the brain had actively silenced the motor program, and built an entire theoretical edifice around an unobserved neural ghost. The stepping reflex never undergoes structural neurological disappearance. By demonstrating that the reflex could be instantly resurrected via buoyancy or prematurely extinguished via lead cuffs, Thelen proved that the motor program was alive, functional, and intact within the nervous system, entirely uninhibited by the cerebral cortex.

9.2 The ‘Masking’ Hypothesis: Latent Motor Capacities

In place of the cortical suppression dogma, Thelen introduced the masking hypothesis. Under this framework, the developmental hiatus of the stepping reflex is conceptualized not as a neural disappearance, but as a temporary physical masking effect driven by normal somatic ontogeny. The underlying neuromuscular coordination machinery remains continuous, robust, and developmentally active beneath the surface, but its overt mechanical manifestation is physically masked by the transient mismatch between somatic mass and muscular strength.

This insight re-established a profound, unbroken continuum across human locomotor development. Rather than viewing human motor ontogeny as a fractured sequence of disjointed phases—where primitive reflexes abruptly emerge, get destroyed by cortical inhibition, and are later replaced by completely new, cortically programmed voluntary actions—Thelen demonstrated deep developmental continuity. Modern ultrasonography has shown that human fetuses begin executing coordinated, alternating stepping and kicking movements in the amniotic fluid as early as 10 to 12 weeks of gestational age. That ancient, self-organizing locomotor synergy persists continuously across the fetal-to-neonatal transition, manifests as the neonatal stepping reflex, continues unabated as infant supine kicking, re-emerges during water immersion, and eventually transforms into mature, voluntary bipedal walking once somatic strength, postural balance, and dynamic equilibrium cross their developmental thresholds.

Development was thus radically redefined. It is not the linear accretion of rigid, hierarchical neural switches; it is the ongoing, fluid management of changing somatic and environmental constraints. The infant’s motor journey is a process of learning to negotiate their ever-changing physical body within an unforgiving physical world. The latent motor capacity is always there, waiting for the somatic chassis and the environmental context to align in a configuration that allows the movement to soft-assemble into reality.

9.3 Broader Implications for Pediatric Neurological Assessment

The deconstruction of the cortical inhibition hypothesis had massive, reverberating consequences for clinical pediatrics and neonatal neurological assessment. For generations, standardized infant neurological examinations had utilized the presence, absence, and temporal disappearance of the stepping reflex as a critical diagnostic metric to evaluate central nervous system integrity, screen for cerebral palsy, and detect developmental delays. An infant who failed to step at two weeks, or an infant who continued to step vigorously at four months, was frequently flagged for suspected neuropathology.

Thelen’s findings revealed the extreme peril of drawing definitive neurological conclusions from raw, uncorrected behavioral milestone assessments without accounting for the infant’s physical morphology, weight-to-length ratio, and body composition. A heavy, rapidly fattening, perfectly healthy infant might show an “advanced” early disappearance of the stepping reflex simply because their legs had become exceptionally heavy, leading a clinician to falsely assume accelerated cortical maturation. Conversely, a low-birth-weight, malnourished, or exceptionally lean infant might continue to step vigorously at five months simply because their low limb mass never exceeded their muscle strength, prompting an unwarranted clinical suspicion of delayed cortical inhibition or spastic diplegia.

Thelen’s work catalyzed a paradigm shift in pediatric clinical thinking, demanding that infant evaluations move away from rigid, milestone-tallying rubrics toward holistic, somatic-ecological assessment frameworks. Clinicians began to recognize the necessity of introducing gravity-eliminated, buoyant, or mechanically supported testing postures when assessing latent neurological capacity. Evaluating an infant’s neuromuscular competence required stripping away the confounding physical mask of gravitational load. By demonstrating that morphology and biomechanics are inseparable from behavioral output, Thelen forced pediatric neurology to acknowledge that the body must be evaluated just as rigorously as the brain.

10. Methodological and Conceptual Legacy in Developmental Science

10.1 Methodological Innovations: Perturbation Analysis in Infancy

The impact of Esther Thelen’s 1984 study reached far beyond the stepping reflex; it revolutionized the entire methodological architecture of developmental psychology. Prior to Thelen, developmental psychology was overwhelmingly an observational, correlational, and descriptive science. Researchers cataloged milestones, mapped chronological ages onto developmental charts, and generated observational inventories. Experimental interventions were largely confined to cognitive and perceptual domains (e.g., preferential looking tasks or habituation paradigms), while physical motor development was viewed as an unalterable biological maturation process that simply had to be observed as it unfolded.

Thelen shattered this observational passivity by introducing perturbation analysis into infant motor research. Borrowing methodologies from non-linear physics, biomechanics, and engineering, Thelen demonstrated that the true, deep architecture of a developmental system can only be exposed by actively perturbing it—pushing it away from its equilibrium state and observing how it self-organizes, adapts, and reorganizes. By applying precise physical perturbations—submerging infants in warm water, strapping precision weights to their ankles, placing them on motorized treadmills, and altering their postural coordinate frames—Thelen turned the nursery and the developmental laboratory into spaces for dynamic, mechanical experimentation.

Furthermore, Thelen pioneered the integration of high-speed microgenetic kinematic analysis, dynamic electromyography, and phase-plane modeling into developmental science. She showed that gross milestone tallies (e.g., recording whether an infant “can” or “cannot” walk) were fundamentally inadequate for scientific inquiry. Understanding motor ontogeny required continuous, high-resolution kinematic tracking capable of capturing intra-limb joint coupling, acceleration profiles, and the underlying muscle activation synergies. This methodological revolution inspired a new generation of developmental scientists to treat infant movement not as a trivial precursor to language and thought, but as a rich, complex, mathematically rigorous laboratory of biological self-organization.

10.2 The Rise of Embodied Cognition and Ecological Psychology

The conceptual shockwaves of Thelen’s stepping reflex study served as one of the primary founding catalysts for the modern paradigm of embodied cognition. Throughout the mid-to-late twentieth century, cognitive science had operated under an extreme computational, Cartesian dualism: the mind was an abstract software program running on the hardware of the brain, processing arbitrary mental symbols that bore no intrinsic relationship to the physical flesh of the body or the physical structure of the world. Cognition was disembodied, cerebral, and isolated.

Thelen’s work, synthesizing Dynamic Systems Theory with the ecological psychology of James J. Gibson, shattered this computational isolationism. In her landmark 1994 book with Linda B. Smith, A Dynamic Systems Approach to the Development of Cognition and Action, Thelen argued that there is no clean, Cartesian dividing line between the body, the brain, and the world. Perception, action, and cognition do not occur sequentially as isolated, centralized computational steps; they are continuously co-constructed through the dynamic, real-time coupling of an embodied agent navigating an environment rich in physical affordances.

By showing that a foundational behavioral change—the disappearance and emergence of walking—could be caused by non-neural, physical-dynamic properties like body fat and water buoyancy, Thelen demonstrated that the mind is fundamentally grounded in somatic realia. Thought is not an abstract computational process residing exclusively within the neocortex; it emerges from our history of physical, bodily exploratory actions within a gravitational field. Thelen dismantled Cartesian dualism within developmental science, establishing that to understand the human mind, one must understand the movements of the biological body that houses it.

10.3 Transforming Pediatric Physical Therapy and Early Intervention

The theoretical paradigm shift engineered by Thelen had immediate, transformative ramifications for clinical physical therapy, neuro-rehabilitation, and early pediatric intervention. For decades, pediatric physical therapy had been dominated by traditional neuromaturational techniques—such as the classic Bobath Neurodevelopmental Treatment (NDT) approach—which operated under the conviction that motor rehabilitation must strictly follow a fixed, linear, hierarchical neurological sequence. Therapists were trained to passively inhibit “abnormal” primitive reflexes and facilitate isolated, maturationally sequenced motor patterns, believing that therapy had to repair the central nervous system’s internal neural program before functional movement could occur.

Thelen’s dynamic systems framework swept away this passive, neuromaturational dogma, replacing it with dynamic, task-oriented, and context-rich physical therapy. Clinicians recognized that they did not need to wait passively for the damaged or delayed central nervous system to mature; nor did they need to spend valuable clinical time trying to “inhibit” primitive reflexes. Because motor behavior is soft-assembled through the interaction of multiple subsystems, physical therapists could actively intervene by identifying and manipulating specific, non-neural rate limiters:

• Targeting Rate Limiters: Instead of focusing vaguely on “brain damage,” therapists began systematically analyzing the biomechanical bottlenecks preventing functional movement—such as trunk extensor weakness, ankle stiffness, or unfavorable mass-to-strength ratios—and designing targeted, physical interventions to cross those specific thresholds.
• Aquatic Therapy (Hydrotherapy): Thelen’s water immersion paradigm was directly imported into pediatric clinics worldwide. Aquatic therapy became a cornerstone intervention for infants and children with cerebral palsy, spinal muscular atrophy, and developmental delays. Buoyancy was exploited to offload gravity, eliminate the crushing mass-to-strength penalty, and allow children with severe motor impairments to actively self-organize coordinated, reciprocal locomotor patterns that were mechanically impossible on dry land.
• Environmental and Task Affordance Engineering: Therapists transformed clinic spaces into dynamic physical environments, using supportive harnesses, body-weight-supported treadmills, elastic resistance bands, and customized seating systems to artificially manipulate physical constraints, thereby unlocking latent motor competence and driving neuroplastic functional reorganization.

11. Contemporary Replications, Extensions, and Neurophysiological Insights

11.1 The Motorized Treadmill Experiments: Expanding Dynamic Systems

Building upon the triumph of the water immersion study, Esther Thelen and Beverly D. Ulrich executed another monumental series of experiments in the late 1980s and early 1990s that extended the dynamic systems model even further: the motorized infant treadmill studies. If the water immersion study demonstrated that buoyant offloading could resurrect stepping, the treadmill studies demonstrated that dynamic, environmental task energy could literally drive the self-organization of walking in infants who could not step on a static surface.

In these landmark studies, Thelen and Ulrich placed infants ranging from 1 to 7 months of age—infants well into the “disappeared” phase of the stepping reflex who exhibited absolute zero steps when held over a stationary floor—upright over a small, motorized treadmill belt. When the treadmill belt was activated and began moving backward beneath the infants’ feet, an astonishing behavioral transition occurred: the moving belt immediately elicited continuous, highly organized, alternating bipedal stepping. The infants stepped with beautiful, rhythmic alternating patterns, complete with coordinated swing and stance phases, matching their stepping frequencies precisely to the variable speed of the moving belt.

The dynamic explanation was breathtaking in its physical elegance: the backward motion of the belt mechanically pulled the infant’s stance leg into extension, passively stretching the anterior hip flexor musculature (specifically the iliopsoas). This dynamic stretch loaded the elastic tendons of the hip and triggered the intrinsic proprioceptive stretch reflexes of the muscle spindles, injecting external mechanical energy into the musculoskeletal system. When the foot reached the end of the belt, this stored elastic recoil and proprioceptive afference spontaneously launched the limb into an active swing phase. The external environment—the moving belt—provided the dynamic mechanical work that the infant’s underdeveloped musculature could not generate on a static floor. Thelen subsequently extended this work to infants with Down syndrome, showing that early treadmill training acted as a profound physical catalyst, significantly accelerating the onset of independent walking by systematically strengthening specific subsystem rate limiters.

11.2 Contemporary Neurophysiological Validation of Spinal Central Pattern Generators

In the decades following Thelen’s original 1984 experiments, modern neurophysiology, optogenetics, and advanced neuroimaging have overwhelmingly confirmed the biological realities that her dynamic systems model predicted. Modern neurobiology has definitively proven that the mammalian lumbosacral spinal cord houses highly sophisticated, self-organizing neural networks known as Central Pattern Generators (CPGs) for locomotion. These spinal CPGs are capable of generating rhythmic, alternating, reciprocal locomotor output completely autonomously, without requiring any continuous, rhythmic timing signals from the brain, and even in the complete absence of descending supraspinal inputs.

Contemporary neurophysiological research has validated Thelen’s core assertion: spinal CPGs do not operate as rigid, immutable programs that fire in a biological vacuum; they are profoundly coupled to, and modulated by, peripheral sensory afference. Research by neurophysiologists such as Serge Rossignol, Sten Grillner, and V. Reggie Edgerton has shown that sensory feedback from muscle spindles (Group Ia and II afferents) and Golgi tendon organs (Group Ib load receptors)—which directly signal the mechanical state of muscle stretch, limb loading, and gravitational force—continuously gates and reconfigures the spinal CPG circuitry in real time.

When an infant is submerged in water, the sudden unloading of the cutaneous plantar receptors and the reduction of force on the Golgi tendon organs during the stance phase alters this sensory afferent loop, lowering the activation threshold required for the spinal CPG to trigger the subsequent swing phase. What Thelen observed in her transparent water tank was the direct, macroscopic behavioral manifestation of this dynamic sensorimotor coupling: the physical offloading of gravitational mass altered the peripheral sensory feedback stream, allowing the spinal central pattern generators to fire rhythmically without encountering the mechanical blockage of excess somatic weight.

11.3 Critiques, Nuances, and Integrative Perspectives

While Esther Thelen’s dynamic systems revolution permanently overturned the simplistic, unilateral cortical inhibition models of the past, the passage of decades has allowed developmental neuroscience to construct a more nuanced, integrative synthesis. Contemporary researchers have offered constructive critiques, cautioning against swinging the pendulum too far toward an extreme biomechanical reductionism that would treat the central nervous system as merely a passive, incidental conduit for physical forces.

Modern developmental neuroscientists emphasize that descending supraspinal pathways—specifically the maturation of the reticulospinal, vestibulospinal, and corticospinal tracts—do play a critical, highly sophisticated role in the ontogeny of locomotion. However, their function is not to act as autocratic, prescriptive dictators that unilaterally program or suppress movements. Rather, descending cortical and brainstem inputs operate as dynamic modulators, stabilizers, and adaptive tuners of the underlying spinal and biomechanical machinery. Cortical maturation is essential for the anticipatory postural adjustments, dynamic balance calibration, visuomotor steering, and flexible goal-directed navigation required for real-world unassisted walking.

The contemporary scientific consensus has crystallized into a balanced, multi-causal, integrative model. The emergence of motor skills is neither purely cortical nor purely biomechanical; it is the non-linear synthesis of both. Biomechanics, somatic mass, and physical laws define the operational possibility space and impose absolute physical rate limiters on behavioral expression; meanwhile, the central nervous system continuously learns to navigate, exploit, and master these physical dynamics through ongoing neuroplastic self-organization. Esther Thelen’s work did not eliminate the brain from developmental psychology; it properly embodied it within the physical reality of the biological organism.

12. Epistemological Implications: Rethinking the Nature-Nurture Dichotomy in Motor Skills

12.1 Rejection of Genetic Preformationism and Prescriptive Epigenesis

At its deepest epistemological level, Esther Thelen’s water immersion study dealt a fatal blow to the persistent doctrines of genetic preformationism and prescriptive epigenesis that had long haunted the biological and psychological sciences. Throughout the twentieth century, human development was endlessly partitioned across the sterile, Cartesian battleground of the “nature versus nurture” dichotomy. Motor development, in particular, was routinely surrendered to the “nature” side of the ledger, conceptualized as a hardwired biological program written into the genetic code, mechanically unspooling along an inevitable, pre-scripted chronological timeline.

Thelen’s dynamic systems framework exposed the profound biological poverty of this view. Development is not the passive unfolding of a pre-existing internal program; it is generative, indeterminate, historical, and continuously constructed in real time through the ongoing, reciprocal dialogue between the organism and its context. There is no central executive code—neither in the DNA nor in the cerebral cortex—that contains the explicit instructions for walking. The instructions do not exist beforehand; they are soft-assembled and created in the moment through the dynamic interaction of physical mass, gravity, muscle contraction, and fluid dynamics.

This insight aligned Thelen’s work with the broader revolution in evolutionary developmental biology (evo-devo) and probabilistic epigenesis, championed by developmental biologists such as Gilbert Gottlieb. Causality in development is not linear, deterministic, or top-down (where genes create brains, brains create behaviors, and behaviors hit the environment). Causality is circular, reciprocal, and emergent. The physical environment (e.g., a tub of water) can instantaneously alter biomechanical forces, which in turn alters peripheral neuromuscular firing, which reshapes the recruitment of spinal circuits, which ultimately feeds back into structural brain development and gene expression. The nature-nurture dichotomy evaporates into an integrated, self-organizing physical system.

12.2 The Organism-Environment-Task Triad as the Unit of Explanation

The permanent conceptual legacy of Esther Thelen’s paradigm shift is the redefinition of the fundamental unit of scientific explanation in developmental psychology. Prior to Thelen, the unit of analysis was the isolated, individual child—or more specifically, the isolated child’s internal cognitive or neuroanatomical structures. One studied the child’s brain to explain the child’s behavior.

Thelen shattered this localized boundary, establishing that the indivisible, fundamental unit of developmental analysis must always be the Organism-Environment-Task Triad:

• The Organism: Encompassing the totality of the child’s biological and physical reality—their neural networks, neuromuscular physiology, skeletal architecture, soft-tissue mass, adipose distribution, and momentary metabolic state.
• The Environment: Encompassing the complete physical field in which the organism is embedded—the ambient medium (air vs. water), the support surface (rigid floor vs. compliant mattress vs. moving treadmill), and the immutable, ubiquitous downward vector of the gravitational field.
• The Task: The immediate, functional goal or physical constraint governing the moment—whether it be maintaining vertical stability, lifting a foot over an obstacle, reaching for a caregiver’s hand, or simply oscillating the limbs in exploratory play.

Under this triad, behavior is not located “inside” the child, waiting to be expressed. It exists only at the intersection of these three dynamic domains. Change the task, and the behavior changes; change the environment (from air to water), and an “inhibited” behavior instantly emerges; change the organism’s physical mass (adding lead weights), and a robust behavior instantly vanishes. The developing child is no longer an isolated, self-contained biological machine, but an active, self-regulating biological system embedded within a continuous physical, gravitational, and perceptual field. Early motor development is revealed to be the foundation of all human intelligence: a process of exploratory physical problem-solving wherein the child learns to discover the dynamic affordances of their body and their world.

12.3 The Enduring Significance of Esther Thelen’s Water Immersion Study

Few scientific investigations in the history of psychology achieve the rare, poetic distinction of overturning an entire scientific paradigm through an experimental design of such pristine, transparent simplicity. Esther Thelen did not require multi-million-dollar neuroimaging scanners or complex gene-sequencing arrays to overthrow decades of entrenched medical and psychological dogma. She needed only a transparent tank of warm water, a few strips of lead weights, and the profound, transformative courage to look at a human infant through the lens of physical reality.

The stepping reflex disappearance study remains the canonical, premier pedagogical masterpiece of Dynamic Systems Theory in action. It stands as a timeless masterclass in scientific epistemology, demonstrating how a rigid, unquestioned theoretical orthodoxy—the cortical inhibition hypothesis—can blind generations of brilliant clinicians and researchers to an obvious physical truth: that baby legs simply get heavy. By lowering non-stepping infants into that simple water tank and watching their legs dance in coordinated, buoyant rhythms, Thelen permanently liberated developmental science from the disembodied confines of cerebral reductionism.

Esther Thelen’s enduring vision was an integrated, non-dualistic science of mind, body, and movement across the entire human lifespan. Her legacy lives on across modern robotics, where engineers build self-stabilizing, passively dynamic walking robots that exploit physical mechanics rather than complex central algorithms; across contemporary pediatric physical therapy clinics, where children are placed into buoyant pools and dynamic harnesses to unlock latent motor potential; and across developmental cognitive neuroscience, which now recognizes that abstract thought is forever rooted in the physical choreography of our bodies. In the final analysis, Thelen taught us that human development is not a pre-scripted, genetic march dictated from the heavens of the neocortex, but a magnificent, self-organizing dance between biological life and the physical laws of the universe—a dance that begins the very moment a newborn child enters the gravitational field of Earth.

Conclusion

The stepping reflex disappearance study executed by Esther Thelen, Donna M. Fisher, and Ridley-Johnson in 1984 represents one of the most consequential conceptual bifurcations in the history of developmental science. For nearly a century, the disappearance of the neonatal stepping reflex was universally brandished as the definitive behavioral proof of cortical encephalization—an unshakeable article of faith asserting that a maturing cerebral cortex progressively and permanently suppresses primitive subcortical spinal automatisms. This disembodied, brain-centric paradigm dominated pediatric neurology, clinical milestones, and psychological theory, reducing the human body to an arbitrary, passive vessel governed by an imperial central nervous system.

Through the revolutionary application of Dynamic Systems Theory, Thelen exposed the profound logical and physical flaws of this classical model. By observing that infants continue to kick with identical kinematic and electromyographic synergies while supine, and by demonstrating that the stepping reflex could be instantaneously resurrected in older infants via the buoyant offloading of water immersion—and reciprocally extinguished in young neonates via artificial weight loading—Thelen proved that the reflex never structurally disappears from the central nervous system. The developmental hiatus of stepping is an optical and mechanical illusion: a temporary physical masking effect caused by the rapid, natural accretion of subcutaneous adipose tissue that temporarily outpaces skeletal muscle force generation within the relentless gravitational field of Earth.

In proving that a classic neurological milestone is governed by biomechanics, anthropometrics, and physical laws, Esther Thelen fundamentally dismantled the Cartesian dichotomy between mind and body, nature and nurture, and competence and performance. Her water immersion experiment revealed that human motor development is a decentralized, self-organizing, soft-assembled process born of the continuous, dynamic dialogue within the Organism-Environment-Task triad. The historical, methodological, and clinical reverberations of this single study continue to shape embodied cognition, ecological psychology, robotics, and pediatric physical therapy to this day. Ultimately, Thelen demonstrated that to understand the developing human mind, we must never lose sight of the physical body—and that sometimes, the most profound secrets of human biology can be unlocked simply by placing a baby in a tank of warm water and watching them step.

References

• Bernstein, N. A. (1967). The co-ordination and regulation of movements. Pergamon Press.
• Gesell, A. (1928). Infancy and human growth. Macmillan.
• Gibson, J. J. (1979). The ecological approach to visual perception. Houghton Mifflin.
• Gottlieb, G. (1991). Experiential canalization of behavioral development: Theory. Developmental Psychology, 27(1), 4–13. https://doi.org/10.1037/0012-1649.27.1.4
• Grillner, S. (2003). The neurobiology of the central pattern generator for locomotion: A review. Current Opinion in Neurobiology, 13(6), 678–685. https://doi.org/10.1016/j.conb.2003.10.003
• McGraw, M. B. (1943). The neuromuscular maturation of the human infant. Columbia University Press.
• Peiper, A. (1963). Cerebral function in infancy and childhood. Consultants Bureau.
• Prechtl, H. F. (1977). The neurological examination of the full-term newborn infant (2nd ed.). Spastics International Medical Publications / Heinemann Medical.
• Rossignol, S., Dubuc, R., & Gossard, J. P. (2006). Dynamic sensorimotor interactions in locomotion. Physiological Reviews, 86(1), 89–154. https://doi.org/10.1152/physrev.00028.2005
• Thelen, E. (1984). Learning to walk: Ecological demands and phylogenetic constraints. Advances in Infancy Research, 3, 213–260.
• Thelen, E. (1995). Motor development: A new synthesis. American Psychologist, 50(2), 79–95. https://doi.org/10.1037/0003-066X.50.2.79
• Thelen, E., Fisher, D. M., & Ridley-Johnson, R. (1984). The relationship between physical growth and a newborn reflex. Infant Behavior and Development, 7(4), 479–493. https://doi.org/10.1016/S0163-6383(84)80007-7
• Thelen, E., & Smith, L. B. (1994). A dynamic systems approach to the development of cognition and action. MIT Press.
• Thelen, E., & Ulrich, B. D. (1991). Hidden skills: A dynamic systems analysis of treadmill stepping during the first year. Monographs of the Society for Research in Child Development, 56(1), 1–103. https://doi.org/10.2307/1166099
• Ulrich, B. D., Ulrich, D. A., Angulo-Kinzler, R. M., & Yun, J. (2001). Treadmill training of infants with Down syndrome: Evidence-based developmental outcomes. Pediatrics, 108(5), e84. https://doi.org/10.1542/peds.108.5.e84
• Varela, F. J., Thompson, E., & Rosch, E. (1991). The embodied mind: Cognitive science and human experience. MIT Press.